Close
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Perceived vulnerability to HIV infection in persons at risk for the disease: An examination of STD clinic patients
(USC Thesis Other)
Perceived vulnerability to HIV infection in persons at risk for the disease: An examination of STD clinic patients
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
INFORM ATION T O USERS This manuscript has been reproduced from the microfihn master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter 6ce, while others may be from any type o f computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 North Zed) Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PERCEIVED VULNERABILITY TO fflV INFECTION IN PERSONS AT RISK FOR THE DISEASE; AN EXAMINATION OF STD CLINIC PATIENTS by Monica Susanne Ruiz A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA in Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (PREVENTIVE MEDICINE) May 1997 Copyright 1997 Monica Susanne Ruiz Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Num ber: 9733129 UMI Microform 9733129 Copyright 1997, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeh Road Ann Arbor, MI 48103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES. CALIFORNIA 90007 This dissertation, written by Monica Susanne Ruiz under the direction of h ? .T ......... Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillment of re quirements for the degree of DOCTOR OF PHILOSOPHY P i i i i i l i T r î i i i i J « i n r *Studies Date ...... DISSERTATION COMMITTEE L Chairperson Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DEDICATION To Mom and Dad, whose faith in my abilities never wavered... not even once. 1 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGMENTS The author wishes to thank the following persons for their invaluable contributions to this research: Drs. Gary Marks and Jean Richardson, who helped me develop the idea and write the grant; Drs. Gary Richwald and Joseph Courtney of the Los Angeles County Department of Health Services, Sexually Transmitted Diseases Program, who allowed me into the Public Health Clinic system; Dr. Portia Choi and the staff of the Ruth Temple Sexually Transmitted Diseases clinic, who allowed me into their facility; The individuals attending the Ruth Temple STD clinic who participated in this study; Hyacinth R.C. Mason, whose wealth of experience with the public health clinic system was invaluable in troubleshooting logistical snafus; Leo Walker, Martha Guerrero, Raul Hernandez, Maria Sotelo, Zephaniah Dawnes, and Maureen Cairns: my indispensable "staff who helped with the subject recruitment, data collection, questionnaire coding, and data entry; Sean Mandel, who willingly sacrificed his computer so that I could do data analysis; My Committee Members, who offer inspiration through their example as well as their suggestions; and, of course... I l l Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The "Goddesses o f the Lower Chalet,” who listened to the complaints, freely gave love and support, and lightened the workload with laughter and friendship. This research was supported by National Instimte of Mental Health Minority Dissertation Grant 1 R03 MH54334-01, awarded to the author. IV Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS I. Introduction . . p.I II. Background and Significance . . . . . p.2 HTV Infection and AIDS Sexually Transmitted Diseases The STD-pnV Connection III. Theoretical Model to be Examined . . p. 17 Perceived Vulnerability and Unrealistic Optimism Behavioral Factïïrs Motivational Factors Cognitive Factors Perceived Vulnerability and Behavior IV. The Study . . . . . . . . p.33 Hypotheses Regarding the Main Effects Model Hypotheses Regarding the Moderated Effects Model V. Methods . . . . . . . . p.41 Data Collection Site Data Collection Personnel Subject Recruitment and Questionnaire Administration VI. Measures . . . . . . . . p.45 Survey Instruments Intake Measures Follow-up Measures VII. Results, Part I: Comparison o f Sample Characteristics and Major Study Variables . . . . . . p.7l Sample Characteristics Sexual Behavior at Intake and Follow-up Univariate Analyses of Time 1 Study Variables VIII. Results, Part II: Substantive Analyses . . . . p.88 Data Analytic Approaches to Testing Theoretical Models Cross-sectional Analyses: Tests of the Path Model Tests of the Mediating and Moderating Models Longitudinal Analyses Additional Analyses of Longitudinal Data Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IX. Results, Part III: Ancillary Analyses of the Study Variables p. 127 Further Exploration of Variable Interactions Examining Curvilinear Distributions of Theoretical Model Variables X. Discussion, Part 1 : Summary of Findings . . p. 144 XI. Discussion, Part 1 1 : Methodological Interpretation of Findings p. 148 Xn. Discussion, Part III: Theoretical Interpretation of Findings . p. 151 XIII. Discussion, Part IV: Directions for Future Research p. 155 References . . . . . . . p. 165 Appendix 1. Intake questionnaire, male version. p. 172 Appendix 2. Follow-up questionnaire. . . p. 195 Appendix 3. Items used in the measurement of sexual risk behavior. p.200 Appendix 4. The Repression-Sensitization scale (Epstein & Fenz, 1967). . p.202 Appendix 5. Items used to measure perceived seriousness of disease. . p.204 Appendix 6. STD knowledge items (Geringer et al., 1993). p.205 Appendix 7. HIV/AIDS knowledge items (Nyamathi et ai., 1993). . p.206 Appendix 8. Items used to measure perceived vulnerability. p.207 VI Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES 1. Means, standard deviations, ranges, and skewness for study variables . . . . . p.45 2. Varimax rotated factor patterns for perceived seriousness variables . . . . . p.59 3. Pearson correlations and Cronbach's alphas for seriousness indices pertaining to STD and AIDS concern/threat . . . . . . p.60 4. Pearson correlations and Cronbach's alphas for seriousness indices pertaining to STD and AIDS treatability and curability . . . . . p.62 5. Varimax rotated factor patterns, correlation matrices, and Cronbach's alphas for behavioral intentions variables . . . . . . . p.69 6. Demographic characteristics of the study sample (N=394). . p.72 7. Sexual behavior of sample at intake (N=306) and follow-up (N=I26) assessments . . p. 78 8. Correlations between demographic variables and major study variables . . . p.83 9. Pearson correlations among major study variables . p.85 10. Results from multiple regression analyses of Equation I A: Intentions for future condom use and reduction in number of partners regressed on theoretical variables, STD-specific models (cross- sectional analytic sample; N=306) . . . . . p.94 V ll Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11. Results from re-analyses of regression model #1 A, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample, N=306) . . . . . . p.95 12. Results from multiple regression analyses of Equation IB: Intentions for future condom use and reduction in number of partners regressed on theoretical variables, HIV-specific models (cross- sectional analytic sample; N=3 06) . . . . . p. 100 13. Results from re-analyses of regression model #1B, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample. N=306) . . . . p. 102 14. Results from multiple regression analyses of Equation 2: Perceived vulnerability to future STD infection regressed on theoretical variables. STD- specific model (cross-sectional analytic sample; N=3 06) . p. 104 15. Results from re-analyses of regression model #2, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample. N=3 06) . . . . . p. 105 16. Results from multiple regression analyses of Equation 3: Perceived vulnerability to negative consequences from future STD infection regressed on theoretical variables, STD-specific model (cross- sectional analytic sample; N=306) . . . . . p. 107 17. Results from re-analyses of regression model #3, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample, N=306) . . . p.l07 18. Results from multiple regression analyses of Equation 4: perceived vulnerability to HIV infection regressed on theoretical variables, HIV-specific model (cross-sectional analytic sample; N=306) . . . p. 109 vui Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19. Results from re-analyses of regression model #4, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample, N=306) . . . . . p.llO 20. Results from multiple regression analyses predicting sexual risk behavior at follow-up, STD-specific models (longitudinal sample; N=126). . p. 113 21. Results from multiple regression analyses predicting sexual risk behavior at follow-up, HIV-specific models (longitudinal sample; N=126).. . p.I 15 22. Regression analyses of longitudinal data in sexual risk behaviors at Time 2 are matched with corresponding intentions variables . . . p.I 19 23. Regression models in which interactions pertaining to gender x prior sexual risk behavior were significant (N=306) . . . . . p. 131 24. Regression models in which inreractions pertaining to perceived seriousness variables were significant (N=306) . . . . . p. 133 25. Regression model in which the interaction pertaining to gender X denial was significant. (N=306) p. 135 26. Regression models pertaining to Time 2 behavior, in which the age x STD history interactions were significant (N= 126) . . . . p. 137 27. Regression models in which interactions pertaining to perceived seriousness variables were significant predictors of frequency of condom use during vaginal sex with steady parmers (N= 126) p. 13 9 28. Regression model in which the age x AIDS knowledge interaction was a significant predictor of frequency of condom use during vaginal sex with steady partners (N= 126) . p. 141 ix Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES 1. Theoretical Model . 2. Moderated Effects Model p-34 p.39 3. Mediation Model . . . . . . . . p.90 4. Moderation Model . . . . . . . p.92 X Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ABSTRACT Objectives. As AIDS prevention efforts continue to focus on behavior change as the key element to curbing the spread o f HIV, researchers must better understand how psychological and sociological mechanisms influence personal beliefs about health, illness, and disease susceptibility. This study examined the degree to which behavioral, cognitive, and motivational factors influence individuals' perceptions of vulnerability to STD and HTV infection, intentions for risk reduction, and subsequent sexual risk behavior. Methods. Study participants (N=394) were recruited from the STD clinic waiting area of a public health facility serving an ethnically diverse community in Los Angeles, California. Participants completed two questionnaires; the first was self-administered at recruitment, the second was telephone-administered approximately a month following the clinic visit. Results. Prior STD history, prior sexual risk behavior, and age were consistently associated with intentions for condom use. None of the behavioral, cognitive, or motivational variables were found to be significant predictors of perceived vulnerability to STD infection or negative consequences from STDs, although prior STD history and AIDS seriousness were found to be significant predictors of perceived vulnerability to HIV. Further, stronger intentions for condom use were found to be associated with lower sexual risk behavior at follow-up. Although perceived vulnerability was not xi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. shown to be a significant mediator, several interactions between demographic and theoretical model variables were found to be significant predictors of perceived vulnerability, intentions for risk reduction, and subsequent sexual risk behavior. Conclusions. Although the results of this study do not lend support to hypotheses pertaining to perceived vulnerability as a predictor of intentions or behavior, they do lend support to several other hypotheses about the underpinnings of perceived vulnerability and the influence of other variables on personal risk assessments and subsequent behavior. Future research should continue to investigate these issues and the many other complex factors which influence sexual behavior in a maimer that is humanistic, culturally sensitive, and socially relevant. XU Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I. INTRODUCTTON As the pandemic of HIV/AIDS continues to affect the lives of men, women, and children in all areas of the world, the global community is becoming increasingly aware that prevention is the most effective way to curb the spread of this disease. To this end, researchers and educators working in the areas of disease prevention and public health are focusing their attention on the behavioral and psychosocial factors which are thought to influence the actions of individuals with an elevated risk of exposure to HTV; for example, those with a history of sexually transmitted disease (STD) infection. With an improved understanding of how psychological and sociological mechanisms— such as personal beliefs about health and illness— affect behavior, researchers and educators may be better able to help these individuals to realistically acknowledge their risk of infection and to modify their behavior accordingly. This study seeks to examine the degree to which cognitive factors (appraisal of disease seriousness and knowledge about STDs/HIV), motivational factors (denial), and behavioral factors (previous STD history and pattern of risky sexual behavior) influence individuals' perceptions of vulnerability to STD infection and HIV infection and their intentions to adopt risk-reducing behaviors (condom use and reduction in number of sex partners). Further, we examine the extent to which perceived vulnerability mediates the relationship between cognitive/motivational/behavioral factors and intentions for risk reduction behavior. Finally, because perceptions are likely to be influenced by a variety Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of factors acting in concert, we will examine the extent to which perceived vulnerability is influenced by the interaction of cognitive, behavioral, and motivational variables. We will examine these issues in a sample of adult men and women who are presenting themselves for treatment in a public STD clinic. This population is at high risk for HIV infection since genital lesions caused by STDs are known to facilitate exposure to and transmission of HTV from an infected partner to a non-infected partner. Further, the presence of sexually transmitted diseases in this sample may indicate the practice of high-risk sexual behaviors which may increase the risk of transmission of other STDs, including HIV. ri. BACKGROUND AND SIGIVIFICANCE HIV Infection and AIDS As of 1990, the World Health Organization (WHO) estimated that 8-10 million adults and one million children worldwide were infected with Human Immunodeficiency Virus (HIV), the etiologic agent which is believed to cause Acquired Immunodeficiency Syndrome (AIDS). It is also estimated that, by the year 2000,40 million people will be infected with HIV; the majority of those infected will be from developing countries in sub-Saharan Afiica, South and Southeast Asia, Latin America, and the Caribbean (CDCb, 1991). WHO further estimates that, during the present decade, more than 10 million children will have at least one parent who will have died from HIV infection/AIDS (CDCb, 1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In the United States, a comparison of the first 100,000 AIDS cases with the second 100,000 AIDS cases reflects great changes in the types of persons who are affected by HIV disease. For example, the first 100,000 cases, reported from 1981 through August 1989, largely occurred among homosexual and bisexual men and injecting drug users (IDUs); of those cases, 61% occurred among gay and bisexual men with no IDU history and 20% occurred among female or heterosexual male IDUs (CDC, 1992). On the other hand, the second 100,000 cases, which were reported to the Centers for Disease Control (CDC) in the period from September 1989 to November 1991, reflect a 44% increase in the proportion of persons with AIDS who have reported exposure to HIV through heterosexual contact; women accounted for 61% of these cases (CDC, 1992a). The expansion of the AIDS case definition in 1993 (CDC. 1992b) resulted in a variety of significant changes in the prevalence of AIDS cases. For example, the overall number of AIDS cases reported in 1993 increased substantially, resulting in a 3% increase from 1992 in the estimated number of persons diagnosed with AIDS-related opportunistic infections (CDC, 1995a). Analyses of these new diagnoses show disproportionate increases in clinically-defined AIDS for women and racial/ethnic minorities and a leveling of cases diagnosed among homosexual and bisexual men (CDC, unpublished data, 1994, ref. in CDC, 1995a). Indeed, the trends in diagnoses for 1994 indicate a continued increase in the proportion of AIDS cases accounted for by Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. women, children, and minority groups, and a decrease in the proportion of cases accounted for by gay/bisexual men (CDC, 1995a). Among all persons at risk for HIV/AIDS, women constitute the fastest-growing risk group for developing AIDS (Stein et al., 1991; CDC, 1995a). In 1993, HIV/AIDS was the fourth leading cause of mortality among American women aged 25-44 years and, in 1994, over three quarters of the cases among women occurred among African- Americans and Hispanics (CDC, 1995b). For African-American women between the ages of 15-44, for example, the death rate due to HIV disease is comparable to rates reported from West Africa (31 per 100,000) (Chu et al., 1990). Further, general AIDS prevalence rates (per 100,000 population) in 1994 were 16 times higher for African- American women and 7 times higher for Hispanic women, as compared to White women (CDC, 1995b). AIDS among women was primarily associated with injecting drug use (41%) and heterosexual contact with an at-risk partner (38%) (CDC, 1995b). A disproportionate number of AIDS cases continue to be reported among ethnic minorities. African-Americans and Hispanics represented 27% and 15% (respectively) of the first 100,000 reported AIDS cases (CDC, 1992). Of the second 100,000 AIDS cases, the proportions for both ethnic groups increased to 31% for African-Americans and 17% for Hispanics (CDC, 1992). In 1993 alone, AIDS cases diagnosed among persons of color represented 55% of the 106,949 total cases reported in the U.S. (CDC, 1994). The increase in AIDS cases among minority groups is consistent with incidence trends of past years, signifying a continuing trend of HIV infection and AIDS among Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. communities of color. Additionally, the increase in cases has significantly influenced death rates for these groups. This is especially true for young adults, as, in 1991, HIV infection was the leading cause of death for Afiican-American and Hispanic males aged 25-44 years (CDC, 1994). The fact that the first 100,000 reported AIDS cases reflect an 8-year period of time, whereas the second 100,000 cases reflect only a 2-year period, emphasizes the escalation of the HIV epidemic in the United States. Although gay and bisexual men are still highly represented in the number of AIDS cases reported to the CDC, the number and proportion of heterosexually transmitted cases of HIV infection has been increasing steadily; indeed, the time period from 1989 to 1990 alone saw a 40.3% increase in reported AIDS cases in heterosexuals (CDCb, 1991). Factors associated with an increased risk for heterosexual HTV transmission include having multiple sex parmers and the presence of other sexually transmitted diseases (STDs) (CDC, 1992; Fennema et al., 1993; O’Farrell et al.. 1991). Men and women who have unprotected sexual contact, particularly with persons known to be at risk for HIV infection, are at increased risk of becoming infected themselves. Analyses of expected trends in HIV infection suggest that, by 1995, infection rates among non-IDU heterosexual men and women may result in a doubling of AIDS cases attributable to heterosexual transmission (Brookmeyer, 1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sexually Transmitted Diseases Because a history of sexually transmitted diseases has been documented as being a risk factor for HTV infection (CDC, 1992; Fennema et al., 1993; O'Farrell et al., 1991), it is important to have basic knowledge about these diseases-both in terms of etiology and incidence trends-so that the mechanisms by which STDs increase risk of HIV infection are more clearly understood. Bacterial STDs Along with the HTV epidemic, the United States is also facing syphilis, gonorrhea, and chancroid epidemics (Aral & Holmes, 1991). Since 1985, the number of primary and secondary syphilis cases reported in the US has been increasing. In 1990 alone, 50,223 cases of syphilis were reported; this was 9% greater than the number of cases reported in 1989. The incidence rate of 20 cases per 10,000 persons is 75% higher than that of 1985, and represents the highest incidence since 1949 (CDCd, 1991). The distribution of syphilis in the population is reflected in data from blood samples collected between 1976 and 1980 as part of the National Health and Nutrition Examination Survey (NHANES II). These data show that youth, single marital status, low education, low socioeconomic status, and urban residence were all associated with a positive syphilis test. However, the greatest difference in incidence was seen between ethnic groups. The NHANES II data show that the rate of a positive syphilis test was five times higher among African-Americans than among whites, even after controlling for all other factors (Aral & Holmes, 1991). Although African-Americans represent Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. only 12% of the total U.S. population, CDC data from 1990 examining incidence rates across the United States show that the percentage of syphilis cases among African- Americans reported from public health care sources was higher than the percentage of cases among whites for all regions o f the nation (CDCe, 1991). These data also indicate that one of the most notable trends in syphilis incidence has been the substantial increase in cases involving African-American heterosexual men and women (CDCd, 1991). Like syphilis, the reported incidence for gonorrhea in the U.S. has been higher among men than among women, and higher for African-Americans and Hispanics than for whites (Aral & Holmes, 1991). In 1988, the incidence rate for gonorrhea among non-Hispanic whites was 54 per 100,00 persons, while the rate for African-Americans was 34 times higher (1,801 per 100,00 persons). The rate for Hispanics (201 per 100,000 persons) was lower than the rate for African-Americans, but was still four times higher than the rate for whites (Moran et al., 1989). Although public health efforts (started in the mid-seventies) aimed at curbing the spread of gonorrhea have been successful in achieving an overall decrease in incidence, the racial differences in incidence rates have widened since 1984, when incidence among African-Americans increased dramatically (Aral & Holmes, 1991). Of particular concern is the fact that certain strains o f gonorrhea have developed resistance to traditional penicillin treatment, and thus must be treated with other antimicrobial drugs (Aral & Holmes, 1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chancroid incidence has also risen since 1984, especially in the inner-city and migrant-worker populations of some of the larger U.S. cities such as Los Angeles, Dallas, and New York City (Aral & Holmes, 1991). Much of the concern over the increase in chancroid cases (from 665 in 1984 to 4,714 in 1989) stems from the fact that chancroid infection is the most common cause of genital ulcerations which, if exposed to HIV, may facilitate transmission of the virus (Aral & Holmes, 1991). Indeed, studies of prostitutes and male STD clinic patients in Nairobi have shown that genital ulcers were associated with a greatly increased risk of sexually acquired HTV infection (Aral & Holmes, 1991). Like gonorrhea, the bacterium which causes chancroid has developed resistance to many antimicrobial drugs. In HIV-exposed persons, chancroid treatments (which have previously been found to be highly effective) often fail; the continuing infection allows the bacterium, as well as HIV, to spread (Aral & Holmes, 1991). Another bacterial STD, chlamydia, has become more common than syphilis, gonorrhea, or chancroid. Because chlamydia often has very mild symptoms that can go undetected until severe damage is done to the female organs, it has been recognized as an important cause of reproductive health problems in women. Prevalence of cervical chlamydial infection has been reported at levels of 4% to 8% in hospital outpatient departments (Phillips et al., 1989; Schachter et al., 1975), 6% to 23% in family planning clinics (CDC, 1985), 27% in incarcerated women (Holmes et al., 1993), and as high as 30% in female STD clinic patients (CDC, 1985). Unlike gonorrhea, chlamydial infection can be foimd in persons of all racial, ethnic, and socioeconomic groups. 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Because chlamydial infection has only recently become a federally mandated reportable disease, incidence rates may actually be conservative (H.R.C. Mason, personal communication. May 1993). Also, asymptomatic chlamydial infection is far more common than asymptomatic gonorrhea. Indeed, it is common to find cervical chlamydial infection in up to 5% of middle-class pregnant women and female college students, whereas gonococcal infection is typically found in less than 1% of the same population (Aral & Holmes, 1991). Like chancroid, cervical chlamydial infection can also cause ulcerations (a result of the inflammation of the infected area), which may increase the risk o f sexually transmitted HIV infection. Viral STDs Other STDs, such as genital herpes, genital warts, and hepatitis B, are caused by viruses and, thus are not curable with standard antimicrobial treatments. From the mid- 1960's to the onset o f the AIDS epidemic, genital herpes and genital warts seemed to be the most rapidly spreading STDs. This presumption was supported by results from random blood serum samples, collected in the U.S. in the late 1970s; these samples were tested for presence of the herpes simplex virus type 2 (HSV-2) antibody. The results suggested that approximately 25 million people had been infected with HSV-2 by 1978; this estimated prevalence was about 10 times higher than what might have been projected from the reported number of physician consultations for treatment of genital herpes. Moreover, the data showed that HSV-2 infection was three times more prevalent in African-Americans than whites and that, among African-Americans, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. women were more likely than men to have HSV-2 antibodies. Presence of antibodies was also found to be more common among inner-city residents and those of lower socioeconomic status (Aral & Holmes, 1991). However, one of the problematic aspects (with regard to disease recognition and treatment) of genital herpes is that few HSV-2 infections ever seem to cause recognizable symptoms. One study (ref. in Aral & Holmes, 1991) found that only one quarter to one third of those with HSV-2 antibodies were aware of past symptomatology consistent with genital herpes. Indeed, new cases of HSV infection that arise are usually found in persons who are exposed to the virus by sexual partners with asymptomatic infection. During the 1970s, an association between genital herpes and cervical cancer was observed (Aral & Holmes. 1991). However, it was argued that HSV-2 infection may only be a marker for high risk sexual behaviors that could also transmit other STDs which may, in fact, be the true cause of cervical cancer. Thus, more attention was given to genital warts, which are caused by strains of the Human Papilloma Virus (HPV). Due to advances in DNA technology, more than 60 HPV strains have been identified, several of which cause genital infection. Like HSV-2 infection, only a few HPV infections result in obvious genital warts; those which do are only rarely found in cervical cancer tissue. Conversely, those types which do not cause visible warts— especially types 16,18, and 31— are strongly associated with precancerous lesions or invasive cancers of the cervix, vagina, vulva, penis, and anus. Physician consultations for genital warts increased from 169,000 in 1966 to more than 2 million in 1988; the 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. incidence of subciinical HPV infections, including those associated with genital cancers, also seem to be rising at a similar rate (Aral & Holmes, 1991). Currently, genital and anal HPV infections seem to be the most prevalent STD in the United States. For example, a recent study reported that, among samples of women recruited from Seattle-area universities and STD clinics, HPV type 16 was found in the Pap test specimens of 22% of the college students and 44% of the STD clinic patients (ref. in Aral & Holmes, 1991). Findings such as these seem to support the hypothesis that a large proportion of sexually active adults seem to be infected. Like herpes, HPV can be treated, but not cured. HPV lesions can be eliminated or suppressed by toxic agents, laser therapy, cryotherapy, or electrocautery, but the virus is still present in the patient. Hepatitis B is also an important viral sexually transmissible disease. It is estimated that there are approximately 200 million carriers of the virus worldwide (Aral & Holmes, 1991). In the United States, new infections rose from 200,000 in 1978 to 300,000 in 1987; approximately half of the infected persons suffer from symptomatic acute hepatitis (Cates & Toomey, 1990). Despite the availability of an effective vaccine and the steep decline in hepatitis B infections attributable to homosexual contacts (from 21% to 9% in the period from 1982-1987), the proportions attributable (in that same period) to heterosexual contact rose from 15 to 22 percent; similarly, the proportions attributable to intravenous drug 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. use rose from 15 to 28 percent. Heterosexualiy-spread hepatitis B is most common among persons of lower socioeconomic status (Aral & Holmes, 1991). The STD-HTV Connection In light of the high prevalence rates for the above-mentioned STDs, levels of HIV seroprevalence in those presenting for treatment of sexually transmitted diseases is of great importance. Studies examining HTV seroprevalence in STD clinic populations have found seroprevalence rates ranging from 4% to 4.9% (Cannon et al., 1989; Erickson et al., 1990; Quinn et al., 1990). To date, several studies have shown a high correlation between HTV and syphilis seropositivity (Cannon et al., 1989; DeBuono et al., 1988; Hull et al., 1988; Quirm et al., 1988; Rabkin et al., 1987). O f the total of 11,528 persons in the five above-referenced studies who were screened for syphilis, 724 were found to be seropositive for syphilis; of these, 159 (or 22'".) were also found to be seropositive for HTV. Thus, the relative risk for HIV seropositivity was 6.8 for persons seropositive for syphilis (Moran et al., 1989). Another study of STD clinic patients found that the HIV seroprevalence rate for men was higher than that for women (5.6% vs 3.6%). Of those patients who tested seropositive for syphilis, 24.3% were also HTV infected, whereas 3.5% of patients with a nonreactive syphilis test were foimd to be HTV infected (Quinn et al., 1990). Schoenbaum et al. (1990) also found a strong positive association between positive serologic tests for syphilis and HIV infection, but found that the odds ratio for women with reactive syphilis serologies (OR 45.5, 95% Cl 5.3, 387.6) was much higher than the odds ratio for syphilis-seropositive men (OR 2.6, 95% 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CI 1.2, 5.8). This higher rate for women may be due to the fact that maie-to-female transmission of an STD (including HTV) is much easier than female-to-male transmission. In addition to the biological bases for the connection between STDs and increased risk of HTV infection, a history of STD infection serves as a marker for high- risk behaviors, such as inconsistent or lack of condom use, which also facilitate exposure to HIV. Studies of the sexual behaviors of STD clinic attenders, as well as other at-risk populations, may shed some light on the degree to which sexually active persons are changing (or not changing) their sexual behaviors to minimize their risk of HIV infection. One study with a sample of 601 homosexual men attending an STD clinic found that over 20% of the subjects had engaged in more than 23 episodes of unprotected anal sex during the previous 4-month period, and that higher frequency of anal sex was associated with lower condom use rates (Doll et al., 1991). Another study which compared women attending a family-planning clinic to women attending an STD clinic found that the STD clinic women were more likely than the family-planning clinic women to have had more than five sexual partners in the past five years (22% vs 13%). In addition, the women attending the STD clinic more commonly reported sexual partners who were either bisexual or IV drug users; almost 45% of these women, compared to 30% of the family-planning clinic women, reported participation in at least one high-risk activity (e.g., FV drug use, sexual intercourse with persons from high risk groups or persons known to be HIV infected) (Wasser et al., 1989). 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Studies of behavioral risk factors in other samples also demonstrate behavioral patterns similar to those found in STD clinic samples. Cochran and her colleagues (1990) examined sexual risk reduction behaviors in two matched samples of unmarried young adult university students. The first cohort was sampled in early 1986 and the second cohort was sampled in the fall of 1987; both cohorts were sampled from the same southern California university. They found that, although the second sample reported greater levels of concerns about contracting STDs, they were no more likely than the first sample to have changed their behaviors to reduce their risk of HTV infection; indeed, 52% of the first sample and 44% of the second sample reported no changes toward lower-risk behaviors (Cochran, Keidan, & Kalechstein. 1990). Catania et al. (1992) sampled heterosexual men and women from various parts of the United States— including cities that are considered to be high risk due to larger numbers o f AIDS cases and larger Hispanic and African-American populations— in order to obtain a prevalence estimate of AIDS-related risk factors and prevention activities. In addition to asking about risk factors (e.g., sexual intercourse with more than one partner, sex with a risky partner, blood transfusion, injecting drug use) in the past 5 years, they obtained information on condom use for those who were sexually active. They found that approximately 15% to 31% of the general national sample and 20% to 41% of the high-risk city sample were at some risk for HIV infection in the past 1-5 years. Only 17% of those reporting multiple sex partners and 12.6% of those reporting a risky sex partner (e.g., a partner who was nonmonogamous, an injecting 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. drug user, a hemophiliac, or who had tested HIV+) indicated consistent condom use. In general, condom use among the at-risk heterosexuals was low, with the less educated and women who had a risky sex partner being even less likely to use condoms. A second assessment on this same national cohort approximately one year later (Catania et al., 1995) showed rather interesting results in regard to STD/HIV risk factors, preventive behavior, and changes in behavior over time. Catania and his colleagues found no significant decreases in HIV risk behavior, either in the national sample or the high-risk city sample. They also found a 4% (McNemar X^23.88, p=0.001) increase in the number of heterosexuals who had multiple sex partners over the 1 year time span, with men (OR=3.88, 95% CI=2.3, 6.5) and "single" persons (OR=2.05, 95% Cl=1.07, 3.92) reporting multiple sex parmers at both assessments (Catania et al., 1995). Interestingly (and, perhaps, fortunately), the researchers also saw a slight increase in consistent condom use among those respondents with secondary (or casual) partners at both assessments (McNemar X ^8.20, /7=0.004), although inconsistent use with primary partners still existed. Younger (18-29 years of age) women and "singles" were more likely than women in their 30s (OR=4.2,95% Cl=1.94, 9.31) and women in their 40s (OR=7.2, 95% Cl=3.32, 15.76) to have reported changes in their risk status in that, at Time 1, they reported no risk factors and, at Time 2, they reported having multiple sex partners (Catania et al., 1995). Perhaps one of the most interesting studies o f behavior change in at-risk groups examined the self-reported sexual risk behaviors of STD clinic patients before and after 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Earvin "Magic" Johnson's announcement on November 7, 1991 o f his HTV infection (CDC, 1993). Study participants were patients at a suburban Maryland public STD clinic who were recruited as part o f a 29-week cross-sectional study; Magic Johnson's announcement was given during the 14th week of the study and, for these analyses, the subjects were divided into a "pre-armouncement" (n=186) and a "post-announcement" (n=97) group. The subjects were predominantly African-American and male; thirty-one percent had had one or more STDs and 78% had had 2 or more sexual partners in the previous year. There were no significant differences between the two groups in terms of demographic characteristics, self-reported STD history, or sexual behaviors. A comparison of the two groups found no significant differences in condom use during vaginal sex with either steady or non-steady partners in the previous 3 months. However, significantly fewer subjects in the "post-announcement" group reported "one night stands" or 3 or more sexual partners of the opposite sex during that same period of time. Thus, the greater impact of Magic Johnson's announcement seems to have been that o f decreasing number and type of sexual contacts among the heterosexual subjects rather than increasing condom use (CDC, 1993). The fact that a great number of at-risk persons are not protecting themselves from the risk of HIV, as well as the projected increases in incidence of HTV infection in non-IDU heterosexual men and women (CDC, 1992), leads to the inevitable question; what is it that keeps people from changing their behavior? 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. in . THEORETICAL MODEL TO BE EXAMINED In the attempt to answer this question, we propose a model which examines the direct effects of behavioral factors (previous STD history and pattern of risky sexual behavior), motivational factors (denial), and cognitive factors (appraisal of STD seriousness, appraisal of HIV seriousness and STD/HTV knowledge) on perceptions of personal vulnerability to both future STD infection and perceived vulnerability to HIV infection, and intentions to adopt risk-reducing behavior among STD clinic attenders, a population at increased risk of HIV infection. We will also examine the degree to which perceived vulnerability mediates the relationship between the antecedent variables and intentions for behavioral risk-reduction. The majority of prior research i n perceived vulnerability has focused primarily on the examination of main effects mocels of behavior, that is, the influence of individual variables on perceptions of risk. However, it is unlikely that complex cognitive processes can be explained as easily as an "a causes b" type of model. Rather, such processes are more likely to be influenced by a variety of factors acting in concert. Thus, we will attempt to fill this gap in the literature by examining the degree to which the aforementioned cognitive and behavioral variables interact with motivational factors to influence perceived vulnerability. The following discussion will review the literature which has examined each of the theoretical components of the proposed model. Since quite a few of these studies have examined these components in more generalized settings (other than STD clinics), 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. we will review the findings, discuss the results with respect to an STD/HIV context, and then discuss the manner in which we would expect the theoretical components to relate to perceived vulnerability in such a setting. Perceived Vulnerability and Unrealistic Optimism One reason why people do not change risky behaviors may lie in the area of perceived vulnerability, i.e., the individual's perception o f his or her own risk of acquiring a disease or illness. For example, if people perceive themselves to be vulnerable to, or "at-risk" of becoming infected with HTV, they may change their own behavior to reduce their risk of infection. Conversely, if they do not think they are vulnerable to the disease, they may see no reason to alter their behavioral patterns. This variability in perceptions of vulnerability leads to the next step in the theoretical line of inquiry: what causes individuals to feel vulnerable to an illness? Although perceptions of vulnerability are considered to be an integral predictor variable in explanatory behavioral models (Becker, 1974; Rogers, 1983), not as much is known about the origins of personal vulnerability. The research which has been done on this issue suggests that several factors may contribute to perceptions of vulnerability. Behavioral Factors Previous experience with disease One of the factors which may affect one's feelings o f vulnerability to illness is having had prior experience with that same illness. This relationship between these two variables is intuitively logical: if an individual has had a particular illness before, he or 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. she may feel prone, or vulnerable, to having the disease again. Similarly, individuals who have had no prior experience with a disease may indeed feel that the disease will never affect them and, thus, will feel little or no vulnerability. The research examining this matter shows that the previous experience- perceived vulnerability relationship does indeed operate in an intuitive fashion. In his examination of factors contributing to optimistic biases of personal probability of negative (illness and non-illness related) life events, Weinstein (1980) found that previous experience was positively correlated (r=0.42) with lower degrees of optimistic bias toward negative events (measured as the mean rating across the negative events). A subsequent study of optimistic bias toward health-related problems (Weinstein, 1982) replicated these findings, although the correlation between previous experience and optimistic bias was slightly lower (r=0.38). Since lower degrees of optimistic bias translate into higher perceived vulnerability, more personal experience with an event is positively associated with more accurate perceptions of vulnerability. Although the above findings are consistent with an intuitive hypothesis, one cannot assume that people will view their past experience as objectively as they should. A particularly interesting review of the effect of personal experience on the adoption of self-protective behavior states that the assessment of one's past experience can be extremely variable (Weinstein, 1989). For example, one may tend to downplay or even forget a particular experience with the passage of time, thus any immediate feelings of vulnerability resulting from the event are also likely to dissipate. Similarly, an 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. individual who believes that risk is cyclical may believe that the probability of the event re-occurring is lower (Slovic, Kunreuther, & White, 1974). Also, in the case that an individual does feel prone to re-experiencing the event, he or she may downplay any potentially negative consequences that may result from the event happening again. On the other hand, previous experience may positively influence one's future probability judgments, thus elevating one's perceptions of vulnerability to the recurrence of the same event or the new occurrence of similar events (Perloff, 1983). Based on the above discussion, it should not be assumed that people will always use their past as an objective marker o f their future risk for disease. Indeed, an estimate of vulnerability which is derived from one's previous experience with an illness or negative event may largely depend of the characteristics of that illness or event. For example, if an individual believes that the factors leading up to a certain negative event were extraordinary and. thus, are not likely to normally occur again, he or she may rate his vulnerability to such an event happening in the future as being relatively low. On the other hand, if the individual realizes that the factors leading up to the event are naturally occurring or highly likely to occur in his everyday life, he or she may believe that is it quite probable for the event to happen again in the future. Thus, with regard to a disease (e.g., STD) which is contracted through relatively ordinary behavior (e.g., sexual activity), one would expect a positive association between previous experience with that disease and perceived vulnerability to future incidence of that disease. 2 0 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Recent risky behavior Like previous experience, one's most recent behavior can also be a factor which influences one's perceptions of vulnerability. For example, if an individual takes into account their own practices of a healthy diet and regular exercise, he or she may estimate that his vulnerability to obesity-related problems is low. Similarly, someone who does not have a healthy diet and does not exercise at all may estimate his or her vulnerability to obesity-related disorders as being higher than average. Although these examples make intuitive sense, the underlying assumption here is that people view their own behavior impartially and then make accurate estimates of their own vulnerability based on these objective assessments. It is more likely, however, that people have a biased view of their current behavior and subsequently make biased predictions about their vulnerability. Studies which have examined people's recent behavior in relation to perceptions of risk seem to support this hypothesis. For example, James and her colleagues (1991) found that, while the majority of their sample (heterosexual and homosexual STD clinic attenders) reported having engaged in sexual behaviors which place them at risk o f HIV infection (e.g., frequent reporting of multiple parmers, high risk sexual behavior, and over 50% of the sample with an STD), only 19% of the sample considered themselves to be "personally at risk" for the disease. Another study found that their sample o f gay men tended to underestimate the riskiness of their sexual practices; men classified by their behavior (engaging in practices which allow for the efiScient exchange of blood 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and/or semen) as being in the high-risk group rated the riskiness of their behavior as only slightly higher than the low-risk group (Bauman & Siegel, 1987). The above results seem to support the hypothesis that people do not objectively assess their own behavior, and thus underestimate their vulnerability to disease. Indeed, this type of underestimation is likely to occur when people are trying to appraise their vulnerability to a disease they have never had before. For example, in both of the above studies, the subjects were asked to rate their vulnerability to HIV infection, a disease with which they are not likely to have had experience. If the studies had examined subject's perceived vulnerability to a disease with which they were familiar (e.g., in James et al.'s study, another STD), vulnerability ratings based on appraisals of recent behavior might have been more accurate due to subjects' realization that behavioral patterns which have already put them at risk once before may indeed do so again. Therefore, one would expect that individuals who continue patterns of high-risk behavior in the past will rate themselves as having greater vulnerability. Nonetheless, it should not be forgotten that, although objective appraisals of risk can result in more objective appraisals of vulnerability, the acknowledgment of one's own risk may result in feelings o f fear or anxiety. Individuals who have such feelings may then seek ways in which to assuage those fears, and the way in which they choose to do so may then reflect on subsequent appraisals of risk behavior and vulnerability. Such a process implies that other factors may be involved in the process of looking at one's own behavior and creating estimates of risk. 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Motivational Factors Denial Denial may be a factor which could help to explain the puzzling findings seen for previous behavior and estimates o f vulnerability. The term "denial,” which is defined as "the negation of something in word or act" (Lazarus, 1983), is familiar to most people as a Freudian expression denoting an individual's altered perception of his or her reality, usually one in which the negation involves an impulse, feeling or thought, or an external demand or situation. Although the word "denial" can be used to describe a wide array of psychotic, neurotic, or normal psychological processes, the common feature of denial is "the tendency to minimize or fail to appraise correctly an undesirable personal characteristic or event" (Janis, 1983). With respect to risk assessments from previous behavior, denial may be the manner in which people alleviate the uncomfortable feelings of fear or anxiety that may come from acknowledging their realistic risk of disease. Thus, risk assessments which are influenced by denial may result in biased estimates of vulnerability. Several studies have found more biased risk estimates in situations of greater threat (Lehman & Taylor, 1987; Perloff, 1983), thus supporting a motivational explanation to lowered perceptions of vulnerability through the mechanism of defensive denial. This mechanism may be better understood as an "emotion-focused" coping strategy in which emotional distress produced by threatening events is assuaged or regulated by the denial process (Lazarus, 1983). In a such a context, denial can be adaptive because it helps to re-establish 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. emotional equilibrium by the minimization of the threat (Janis, 1983). However, the minimization of the threat may have serious repercussions for one's physical well-being, especially if the denial of risk is accompanied by continued high-risk behavior. Gladis et al. (1992) examined the relationship between denial (measured by the Repression-Sensitization scale) and perceived vulnerability to AIDS in a sample o f high school students. They found a main effect for denial in that repressors (based on their R-S score) perceived less AIDS risk than sensitizers. Further, they found that denial moderated the relationship between past sexual behavior and perceived risk: past risky behavior and perceived risk were positively correlated for sensitizers and neutrals, but negatively correlated for repressors. That is, repressors did not seem to take their own high-risk behaviors into account when estimating their own vulnerability (Gladis et al., 1992). Since several previous studies have validated the R-S scale as a marker for motivated denial (Andrew, 1970; Cook, 1985; Olson & Zanna, 1979; Shipley, Butt, Horwitz, and Farbry, 1978), Gladis and her colleagues concluded that motivated denial may very well have been responsible for the misperceptions of vulnerability shown by the repressors in their sample. The motivational explanation is further supported by another study of perceived risk in which a relationship was found between psychological distress and perceived riskiness of sexual behavior. In their sample of homosexual men, Bauman and Siegel (1987) examined the degree to which anxiety (measured by the Anxiety Symptom dimension of the Brief Symptom Inventory [BSI]) regarding perceived risk for AIDS 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. may be managed through defensive denial. They found that, as anxiety decreased, the tendency for the men to underestimate their risk increased, thus suggesting that the underestimation o f risk (denial) may have been used to manage anxiety. In this study, however, emotional distress was not measured in terms of defensive denial, so it cannot be assumed that the anxiety measure used was a proxy measurement for denial. Nonetheless, it can be argued that anxiety and denial are both potential responses to threat, and therefore Bauman's and Siegel's results do support the motivational hypothesis for denial. Thus, based on this support, one would expect a main effect for denial in that greater degrees of denial are associated with lower perceived vulnerability to disease. Similarly, following from Gladis's work, one would also expect denial to moderate the relationship between past behavior/experiences and perceived vulnerability. Cognitive Factors Appraisal of disease seriousness Another potential factor which may influence an individual's perceptions of vulnerability is the person's appraisal of the seriousness o f a disease, or the degree to which the individual believes the disease to have serious repercussions in terms of his/her own health. These perceptions are, in part, based on an individual's pattern of illness behavior, which is defined as "the ways in which symptoms may be differentially perceived, evaluated, and acted (or not acted) upon by different kinds of persons" (Ross, 1987). Illness behavior can vary across a wide spectrum, ranging from hypervigilance 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. or hypochondriasis over disease to the other extreme of underreaction or denial that the illness even exists. In some cases, illness behavior may be less a matter of individual (personality- based) coping strategy and more a matter of the nature of the disease in question. Appraisals of a disease's seriousness may be dependent on the type of disease. For example, patients with STDs may not acknowledge that venereal diseases are indeed illnesses; rather, STDs may be seen as a nuisance or a minor casualty of a particular lifestyle (Ross, 1987). Similarly, characteristics of the disease— such as its duration and treatment— may influence appraisals of seriousness, as well as perceptions of vulnerability to future disease. If a specific disease is seen as not being serious, an individual's feelings of vulnerability can be affected in two ways. For example, if individuals perceive a disease as being curable or at least treatable, they may feel that the disease is relatively unimportant; therefore, because they feel no real threat from the disease, they will not feel vulnerable to negative health consequences from the disease. Further, they may not even feel vulnerable to disease onset. Alternatively, individuals may feel somewhat vulnerable to a specific disease based on realistic factors, such as hereditary predisposition or environmental exposure but, because the disease may be treatable, they may not feel that the disease is serious enough to warrant any preventive action. Conversely, if the disease in question is seen as being very intimidating because it is incurable, debilitating, or fatal, the probability that the individual will perceive the disease as being serious is much greater than if the disease were perceived as benign. If 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the individual perceives the disease as being more serious, he or she may experience greater feelings of vulnerability to serious health consequences and future onset. Although appraisals of seriousness have not been extensively examined in the literature, a study by Van der Velde and his colleagues (1994) examined perceived severity in relation to unrealistic optimism and perceived AIDS-related risk in four samples of varying actual risk. They found that, while perceived severity was not related to perceptions of risk in the low-risk (persons with multiple private partners from non-AIDS risk groups) and high-risk (persons with multiple partners, including prostitution contacts) heterosexual samples, there was a negative correlation with comparative risk in the high-risk homosexual male sample. That is, higher levels of perceived severity of .AIDS were related to lower levels of optimistic bias; in terms of vulnerability, lower optimistic bias would indicate more realistic perception of vulnerability to AIDS. Although the high-risk homosexual sample in Van der Velde et al.'s study had more realistic perceptions of their own vulnerability based on appraisals of the severity of AIDS, the fact that the correlation between these two variables was not found for the high-risk heterosexuals is rather curious. It doesn't seem likely that these two groups would rate the severity of AIDS differently; indeed, most people would seem to agree that the disease is quite severe. Rather, it seems likely that, because the homosexual community was one of the first groups to be stricken by the AIDS epidemic and because many in the gay community have witnessed the loss o f many friends and loved ones, 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. most of the people in that community are quite aware that AIDS can and does affect them. On the other hand, because many heterosexuals have not been affected by the AIDS epidemic, they still do not think of themselves as being vulnerable to the disease. Thus, with respect to the homosexual sample in the above mentioned study, it may be the case that perceived seriousness is highly influenced by past experience (not necessarily one's own) with an illness. That is, past experience has proven to these subjects that AIDS is a serious disease to which they cannot deny their vulnerability. However, with respect to both the low- and high-risk heterosexual samples, the probable lack of past experience with AIDS may facilitate their ability to deny that AIDS is a disease to which they are vulnerable; hence, the lack of association between perceived seriousness and ratings of risk. Thus, denial may help to explain why the perceived severity— perceived vulnerability relationship was different for these two groups. Following from Van der Velde et al.'s findings, one would expect a main effect for appraisals of disease seriousness, where greater ratings of a disease's seriousness are positively associated with greater ratings of perceived vulnerability to that disease. Additionally, one would also expect denial to moderate the relationship in that the correlation between appraisal of seriousness and perceived vulnerability is lower for those higher on the denial dimension. Knowledge about Disease Informational factors may also play a role in influencing perceptions of vulnerability. Accurate knowledge about a disease— such as how the disease develops or 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. is contracted, or the prognosis for affected individuals--may help people make more objective estimates of their own vulnerability to the illness, and may also prompt the adoption of preventive behaviors. For example, an individual who is aware of the behavioral factors that are linked to cancer incidence may realize that her normal health conscious behaviors will probably reduce her likelihood of developing cancer; having acknowledged this, she may feel less vulnerable to ever having cancer. But what happens if an individual's knowledge about a particular disease is inaccurate? Erroneous knowledge about a disease-such as incorrect modes of transmission— may also contribute to estimates of vulnerability. However, such estimates are likely to be as incorrect as the knowledge on which they were based. For example, a sexually-active person who believes HIV/AJDS to be a disease affecting only gay men is not likely to feel vulnerable to HIV infection if he or she does not identify with that particular group. Similarly, the same person is not likely to feel vulnerable if he or she does not know that HIV can be spread through non-gay sexual activity. Thus, the misinformation leads to an underestimation of risk, as well as a low probability of adopting risk-reducing behaviors. Similarly, knowledge about the incidence of the disease in society may also contribute to an individual's estimates of vulnerability to that illness. If, for example, a woman knows that there is an influenza epidemic and that several co-workers have been stricken by the flu virus, she may estimate a moderate to high likelihood of becoming ill with the flu. Thus, the woman's estimate about her vulnerability is based on relatively 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. accurate knowledge of the disease's incidence. However, lack of knowledge about disease incidence can bias estimates of vulnerability. For example, if a man does not know that syphilis infections have reached epidemic proportions in his city, he may think syphilis to be a somewhat rare disease; because the man believes syphilis to be an uncommon disease, he may not think that he is likely to be exposed and, thus, will not feel vulnerable to it. The issue of whether or not knowledge about a disease affects people's assessment of their vulnerability and/or risk-reducing behavior is an especially important one where STDs and AIDS are concerned because these two diseases remain almost entirely preventable. Thus, it would be valuable to know whether or not accurate knowledge about STD/HIV etiology and transmission affect individuals' perceptions and actions. Intuitively, one would expect that people who have a high (or accurate) level of knowledge about a disease would be less likely to engage in behaviors which put them at risk for that disease, and that those with less or inaccurate knowledge would continue to practice high-risk behaviors. Surprisingly, studies which have examined the effects of STD and AIDS-related knowledge have found that knowledge does not strongly (if at all) influence either perceptions of vulnerability or actual preventive behavior. For example, Johnson et al. (1992) found the AlDS-related knowledge scores to be quite high in their sample of young collegiate African-American men and women. However, despite the generally high level of knowledge, over two-thirds of the sample continued 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to engage in behaviors which put them at risk for HTV infection. Although this finding is not consistent with the way in which one would expect knowledge to influence behavior, Johnson and his colleagues had another finding which does fit the intuitive hypothesis: women and men who had multiple sex parmers (a risk factor for HTV infection) knew less about AIDS etiology. For example, these individuals were more likely to believe that AIDS is caused by a bacteria or the same virus that causes venereal disease. Thus, Johnson and his colleagues posit that young adults with multiple sex partners may not consider themselves to be vulnerable to AIDS because they believe that the disease is caused by conditions that are treatable. Another study by Geringer et al. (1993) examined the relationship between knowledge about condoms/STDs and actual condom use. They found that despite high knowledge scores, none of the variables reflecting knowledge about condoms and STDs was significantly related to the rate of actual condom use. They also found that, for women, three of the knowledge items (two of which were related to STD knowledge and one which was related to condom use in the prevention of pregnancy) were only significantly related to condom use with a new partner. This finding is consistent with those of other studies (e.g.. Linden et al., 1990) which have found that general knowledge was not related to condom use except in a few areas (Geringer et al., 1993). Based on these findings, one would expect knowledge to have little effect on perceived vulnerability. However, the reasoning for this is unclear. In the instance of the Johnson et al. (1992) study, why would educated young men and women with high 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. levels of AIDS-related knowledge continue to practice risky behaviors? Again, the answer to this inconsistency may lie in the area of denial in that denial may be a psychological barrier which diminishes the effect of knowledge on perceptions of risk and subsequent behavior. For example, the correlation between knowledge and perceived vulnerability is likely to be lower for persons who are high in denial. Conversely, the correlation between the two variables is likely to be higher for persons who score low in the denial dimension. Thus, denial functions as a moderator of the knowledge-vulnerability relationship. Perceived Vulnerability and Behavior Since perceived vulnerability is a factor in noted theoretical models of health behavior (e.g., the Health Belief Model, Protection Motivation Theory), the relationship between perceptions of vulnerability and actual (or intentions toward) preventive behavior has been addressed in the literature. Two reviews of the Health Belief Model have given perceived vulnerability a considerable degree of support as a predictor of behavior. The earlier of the two reviews (Becker, 1987) found vulnerability to be a significant predictor in 10 of the 11 studies examined, while the latter review (Janz & Becker, 1984) found vulnerability to be a significant predictor in 20 of the 26 studies examined. Similarly, studies (of various non-clinical samples) which have examined perceived vulnerability as a variable in the Protection Motivation Theory model have mostly found favorable support for its predictive abilities (Maddux & Rogers, 1983; Mulilis & Lippa, 1990; Wurtele & Maddux, 1987). 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Although there have been studies which have examined perceived vulnerability in relation to STD/HIV, only a few have done so using at-risk samples such as homosexual men or STD clinic patients. Three of these studies found that perceived vulnerability, either as part of the Health Belief Model (Montgomery et al., 1989) or alone (Lindan et al., 1990; Weinstock et al., 1993), were not associated with risk reduction behavior, such as condom use. Partial support for perceived vulnerability as a predictor was found by Van der Velde & Van der Pligt (1991), who found that vulnerability (examined as part of PMT) was significantly related to intentions for condom use among gay subjects, but not heterosexual subjects. Similarly, Geringer et al. (1993) found that perceived risk of getting an STD related to condom use for women, but not for men. Only one study (James, Gillies, & Bignell, 1991) found that perceived risk of HIV infection was associated with changes in sexual behavior, such as using condoms more often. Although there have been studies which have examined the relationship between perceived vulnerability to STD/HTV and behavior in at-risk samples, few studies have examined the antecedents of perceived vulnerability in such samples. This study, therefore, seeks to examine the antecedents of perceived vulnerability in a sample of STD clinic attenders, a group which is at elevated risk for HTV infection. T V . THE STUDY This study seeks to understand the origins of perceived vulnerability by exploring the degree to which cognitive (appraisal of disease seriousness, STD/HIV 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. knowledge), motivational (denial), and behavioral factors (previous STD history, pattern o f risky sexual behavior) influence such perceptions among STD clinic attenders. We propose to examine the degree to which cognitive, motivational, and behavioral variables directly influence perceptions of vulnerability, as well as the degree to which denial moderates the relationship between cognitive/behavioral variables and perceived vulnerability. For a visual representation of the proposed main effects model, see Figure 1 below. Prevfoui STD HUtpry P itte rn of risky sexual beltsvior Perceived Vuinerabiiily to fu tu re STD infection Perceived Vulnerability to negative consequences from future infection Intentions for Risk Reduction Appraisal of STD Seriousness Perceived Vulnerability to HIV infection Theoretical Model Figure 1 Hypotheses Regarding Main Effects Model Hypothesis 1: Previous STD history and perceived vulnerability We hypothesize that persons who have had STDs in the past will see themselves as being likely to get another STD in the future, but are not likely to feel vulnerable to having serious health consequences resulting from a STD infection. Thus, we expect a 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. positive correlation between previous STD history and vulnerability to future STD infection, and a weaker correlation between STD history and perceived negative consequences from a future STD infection. We hypothesize that the correlation between previous STD history and perceived vulnerability to HIV infection will be weaker than that between STD history and STD infection, mainly because it is suspected that people see STDs and HTV as being two very different types of diseases and may not acknowledge the risk relationship between the two illnesses. Hypothesis 2: Pattern of risky sexual behavior and perceived vulnerability A positive relationship is expected between a pattern of risky sexual behavior and perceived vulnerability to future STD infection because people who have patterns of behavior that have presently predisposed them to STD infection are likely to acknowledge that such patterns of behavior may also do so in the future. Similarly, we expect a positive, but weaker correlation between pattern of risky sexual behavior and perceived vulnerability to negative consequences from future STD infection; although people may acknowledge that their behavior puts them at risk, they are less likely to think that they will suffer grave consequences from that risk. With respect to HTV, we predict that the pattern of risky sexual behavior will be positively, but weakly correlated with perceived vulnerability to HIV infection because, although individuals may see their behavior as predisposing them to STDs, they may 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. believe that HTV infection is not in the same disease classification as an STD. Thus, they may not feel that their behavior puts them at risk for HIV. Hypothesis 3: Denial and perceived vulnerability We hypothesize that denial will be negatively correlated to perceived vulnerability to future STD infection, perceived negative consequences from future STD infection, and perceived vulnerability to HTV infection. That is, we expect that individuals who have higher levels of denial will see themselves as having low levels of vulnerability to both diseases. Hypotheses 4 and 5: Appraisal of STD/HIV seriousness and perceived vulnerability With respect to appraisal of STD seriousness, it is hypothesized that individuals who do not perceive STDs as being illnesses that should be taken seriously will rate themselves as being less vulnerable to future STD infection. Thus, we expect a positive correlation between appraisal of STD seriousness and perceived vulnerability to future STDs, and a stronger (also positive) correlation between seriousness and perceived vulnerability to negative consequences from a future STD infection. A similar positive relationship is hypothesized for perceived seriousness of HIV disease and perceived vulnerability to HIV infection. No predictions will be made for the relationship between either (1) appraisal of STD seriousness and perceived vulnerability to HIV or (2) appraisal of HIV seriousness and perceived vulnerability to future STD infection/negative consequences from future 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. STD infection. Because the average person may not mentally categorize HIV/AIDS with other, more common sexually transmitted diseases, he or she may not acknowledge the risk relationship between the two illnesses. Not enough is known about the way in which people view STDs and HIV/AIDS, thus predictions regarding the association between these two diseases cannot be made. Hypothesis 6: STD/HIV knowledge and perceived vulnerability Although the literature indicates that the relationship between knowledge and perceived vulnerability is not strong enough to significantly predict either negative or positive directionality, we will examine whether or not knowledge about STD and HIV is correlated with perceived vulnerability in this sample. If knowledge has an important influence on perceived vulnerability, we would expect to find an inverse relationship between the two variables. For example, higher levels of knowledge may indicate that an individual feels she knows how to protect herself, thus resulting in a lower perception of vulnerability. Thus, the two variables will be negatively correlated. Alternatively, if knowledge is not important in relation to vulnerability, no correlation would be found. Hypothesis 7: Perceived vulnerability to STD/HTV and intentions for risk reduction We hypothesize that perceived vulnerability to a future STD infection is positively related to intentions to adopt risk-reducing behaviors; that is, individuals who perceive themselves to be vulnerable to another STD will indicate greater intentions to 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. reduce their risk of such infection by means of condom use and/or the reduction in the number of sexual partners. Similar positive relationships are hypothesized for I) perceived vulnerability to negative consequences from future STD infection and behavioral intentions, and 2) perceived vulnerability to HTV infection and behavioral intentions. Hypotheses Regarding Moderated Effects Model One of the shortcomings of most of the previous research pertaining to perceived vulnerability is that only single influences are tested as explanations of vulnerability. However, the development of such perceptions are likely to be much more complex than can be explained in a single-influence type of model. Indeed, it is more likely that perceptions of risk are influenced by a variety of factors acting together, rather than separately. Thus, we seek to go beyond the findings of other studies by examining the extent to which motivational factors interact with cognitive and behavioral factors to influence perceived vulnerability to future STD infection, negative consequences from future STD infection, and HTV infection. For a visual representation of the moderating model, see Figure 2 below. 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. P revious STD History DenUI ^ Perceived V ulnerebility ^ to fu tu re STDs Ocnisl Perceived Vulnersbllity to HIV infection Appreisal of HIV Serioueness Denial Knowledge Denial M oderated E ffects Model Figure 2 Hypothesis IM: Denial as a moderator of risky sexual behavior and vulnerability We wish to conceptually replicate the findings of Gladis et al. (1992) by examining the extent to which denial moderates the relationship between patterns of risky sexual behavior (past sexual behavior) and perceived vulnerability to future STD infection and HIV infection. We predict that compared to sensitizers, repressors (on the R-S scale) will deny the fact that their previous risky behavior puts them at risk, and will thus rate their vulnerability as being lower. Thus, for deniers, the correlation between patterns of risky sexual behavior and perceived vulnerability to future STD infection will be weaker than for those who score low on the denial dimension. A similar pattern is expected for sexual behavior and perceived vulnerability to HIV infection. 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Hypothesis 2M: Denial as a moderator of previous STD history and vulnerability We predict that denial will be a significant moderator of the relationship between previous STD history and perceived vulnerability to future STD infection. That is, we expect that the association between previous STD history and vulnerability will be weaker for those scoring high on the denial dimension (repressors) and stronger for those who score low on the denial dimension (sensitizers). We predict the same pattern of results for denial's moderation of the relationship between previous STD history and perceived vulnerability to HTV infection. Hypotheses 3M and 4M: Denial as a moderator of appraisal of STD/HTV seriousness and vulnerability With respect to appraisal of STD seriousness, we predict that the association between the STD seriousness and perceived vulnerability will be weaker for those who score high on the denial dimension (repressors) than for those who score low on the denial dimension (sensitizers). Similarly, we predict that the correlation between appraisal of HIV seriousness and perceived vulnerability to HIV infection will be weaker for repressors than for sensitizers. Hypothesis SM: Denial as a moderator of knowledge and vulnerability Finally, we predict that denial will be a significant moderator of the relationship between STD/HTV knowledge and perceived vulnerability to future STD infection in that the relationship between the two variables will be weaker for repressors than for 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sensitizers. A similar pattern is expected for STD/HTV knowledge and perceived vulnerability to HTV infection. V. METHODS Data Collection Site Data were collected at the Ruth Temple Health Center in Los Angeles, California. This clinic predominantly serves individuals from lower SES groups, providing free medical care to its clients. Because of its location in the metropolitan Los Angeles area, the clinic serves individuals who are predominantly from the African- American and Latino communities. The STD clinic is in operation from 8 a.m. to 4 p.m., Monday through Friday. Permission for entry into the clinic was received from the clinic's supervisor, Portia Choi, MD, MPH (who was also the District Health Officer [DHO] for that area) and from Gary Richwald, MD, MPH, the head of Los Angeles County's Sexually Transmitted Diseases Program. Approval of the protocol and informed consent was received from the Institutional Review Board of the University of Southern California. Data Collection Personnel Intake data were collected in the Ruth Temple STD clinic by a bilingual male research assistant (RA), who was trained in data collection procedures and hired to serve every day during clinic hours. Follow-up data were collected via telephone interview by two bilingual RAs, one male and one female, who contacted the participants at their homes (or at the telephone number which they had specified at the 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. time of their recruitment into the study) and administered the follow-up survey. For the sake of avoiding potentially uncomfortable situations for study participants, men were contacted by a male RA and women were contacted by a female RA. Subject Recruitment and Questionnaire Administration Intake Assessment Individuals eligible for study participation included men and women of any ethnic background presenting for preliminary treatment (i.e., visiting the clinic for an initial examination, not for a follow-up exam) of a sexually transmitted disease; persons attending the clinic only for HTV testing were not recruited. Eligible participants had to be English or Spanish-speaking, over 18 years of age, and able to provide informed consent. Based on power and sample size analyses, it was estimated that a subject population of approximately 340 individuals were needed in order to have adequate power to obtain correlations among study variables ranging from 0.20 to 0.40 for both cross-sectional and longitudinal analyses. In an effort to obtain a random sample of people from the clinic population, the data collector used a floor plan of the clinic waiting area in which each seat in the waiting area was numbered. Then, using a random number chart, he chose a seat from the floor plan and approached the individual sitting in that seat for study participation. However, because the individuals in the clinic are often required to see a number of providers prior to seeing the physician and, therefore, moved around to a variety of seating locations during the time spent in the waiting area, the floor plan method of 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. random selection was abandoned in favor of a systematic sampling method, where every third person who entered the clinic waiting area was approached for study participation. Individuals who chose to participate were administered informed consent, in which they consented to one confidential questionnaire and a telephone follow-up. All subjects were given a copy of their informed consent and were encouraged to contact the P.I. at any time (the P.I.'s office phone number was provided on the consent form) if they had any questions. After completing the informed consent, the RA recorded the participant's name, address, phone number, date of birth, gender, and ethnicity in the master list of study participants. All participants were informed that the personal data in the master list was entirely confidential, was not used for any purpose other than record keeping, and that it would be kept under lock and key in the P.I.'s office. For the sake of maintaining as much confidentiality as possible, study participants were asked to generate a five-digit number (each number to be chosen at random) which, when combined with the date of their survey, was used as his or her unique subject ID code. This ID code was also recorded in the master list. Following enrollment into the master list, the RA briefly reviewed the survey with the individual, answered any questions that were asked, and left the person alone to self-administer the questionnaire. Spanish speakers had the option of choosing either an English-language or a Spanish-language questionnaire, the choice being dependent on the language with which they felt most comfortable. Although the RA was trained to administer the questionnaire to any individuals who were unable to read or whose 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vision problems prevented them from correctly reading the questionnaire, none of the individuals needed assistance in completing the questionnaire. Individuals declining participation were asked to give a very limited amount of demographic information (age, gender, ethnicity, education) for the purposes of comparing participants to those not wishing to participate. Analyses comparing decliners to study participants indicated that decliners were not significantly different from participants with regard to age, gender, ethnicity, and educational level. Follow-up Assessment For the follow-up survey, study participants were contacted by an RA of their same gender at the telephone number they had specified at the time of their study enrollment. For each individual, a total of six re-contact attempts over the course of 1 week were made. If the subject was successfully contacted and his or her identity was confirmed (using the subject's name, self-generated ID number, and birth date), the RA proceeded with the telephone follow-up interview. Upon completion of the phone interview, the RA debriefed the participant, answered any questions the individual may have had, and thanked him or her for participating in the study. If the subject could not be located or if the person's identity was not confirmed (either by their ID code or birth date and personal information from the master list), the individual was coded as lost to follow-up. 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. VI. MEASURES In this section, we will describe in detail the measures used in the study for both the intake and follow-up assessments. Additionally, we will discuss the results pertaining to the construction and validation of study variables prior to their inclusion into univariate analyses and tests of the theoretical model. When applicable, the results obtained here will be compared to those reported by others in prior studies. The means (of the proportion scores), standard deviations, and skewness of each of the measures are presented in Table 1. Table 1. Means, Standard Deviations. Ranges, and Skewness for Study Variables._________________ Mean (SD) Range Skewness STD history 1.57(0.49) 1.00 -0.28 Sexual risk behavior (intake) .74 (.55) 2.50 .57 Denial .51 (.11) 0.76 -.39 STD knowledge' .90 (.14) 0.86 -1.71 AIDS knowledge'* .88 (.12) 0.80 -1.82 STD seriousness (concern/threat) 4.06 (.83) 3.33 -.61 STD seriousness (treatability/curability) 2.46 (.85) 4.00 .14 AIDS seriousness (concern/threat) 4.38 (.83) 4.00 -1.22 AIDS seriousness (treatability/curability) 4.41 (.94) 4.00 -1.93 Perceived vulnerability, STD infection 2.66(1.74) 6.00 1.05 Perceived vulnerability. 2.65(1.71) 6.00 .85 negative STD consequences (Table I continued on next page) 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 1, continued) Mean (SD) Range Skewness Perceived vulnerability, HIV infection 2.42(1.65) 6.00 1.27 Intentions (condom use) 3.82(1.65) 4.00 -.94 Intentions (reduction in no. partners) 3.33 (1.36) 4.00 -.33 Sexual risk behavior (follow-up)_________.80 (.38)_____________ 1.88_________________ 1.67 Note: Means for denial, STD,itnowledge, and HIV knowledge are based on the proportion scores, which range from 0 to 1 . The means and standard deviations for the first 13 variables are based on the intake analytic sample, N=306. The mean and standard deviation for sexual risk at follow-up are based on the analytic sample at Time 2, N=126. * Participants who answered all 7 STD knowledge items correctly: 55.2% (n=169). Participants who answered 6 of 7 STD knowledge items correctly: 28.4% (n=87). " Participants who answered all 1 5 HIV/AIDS knowledge items correctly: 19.9% (n=61). Participants who answered 14 of 15 HIV/AIDS knowledge items correctly: 36.0% (n=l 12). Survey Instruments At the intake assessment, all study participants were asked to complete a 20- page questionnaire asking them to provide demographic information about themselves (e.g.. age, ethnicity, job status, education level, etc.) as well as information pertaining to their past and most recent sexual behavior and health beliefs. Separate questionnaires for males and females were used in data collection in order to include response options for the specific sexual behavior items which accommodated persons with same-gender as well as opposite gender partners. For example, the option "not applicable; my partner is male" was available on male questionnaires for the items pertaining to vaginal 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sex, and the option "my partner is female; we had sex using dildos or toys" was available on female questionnaires for women who had female partners. In all other respects and for all other items, the male and female versions of the survey were identical. A sample of the male intake questionnaire is presented in Appendix I. For the follow-up assessment, subjects were asked about their sexual behavior in the six weeks since leaving the STD clinic. The questions asked were identical to the specific sexual behavior questions asked at the Time 1 assessment, except the frame of reference was "in the last six weeks since leaving the clinic" rather than "in the last 30 days." As in the initial assessment, subjects were asked to provide details about their sexual encounters with both steady and casual partners. The follow-up questionnaire is presented in Appendix 2. Intake Measures Previous STD History In order to assess subjects' past use of the Ruth Temple STD clinic, all subjects were asked: "Is this your first time in this STD clinic?" (l=no, 2=yes). Additionally, subjects were asked whether or not they had ever thought they had an STD prior to the current visit. Those who responded "yes" were subsequently asked whether or not they went to a doctor or to a clinic to have the condition examined, and whether or not the condition was diagnosed as a sexually transmitted disease. Dichotomous response formats (l=no, 2=yes) were used for these questions. Due to missing data for the latter two questions pertaining to visiting a physician for STD treatment, only the question 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. pertaining to prior STD infection ("Before now, have you ever thought you had a sexually transmitted disease [STD]?") was used in subsequent analyses. Prior Sexual Risk Behavior In order to assess prior STD/HIV risk behaviors, we explored several dimensions of sexual behavior' which could conceivably contribute to an individual’s risk for STD/HTV infection. More specifically, we examined individual's sexually active status and the number of partners they had had in the month prior to the STD clinic visit where they completed the intake assessment. We also asked study participants to give detailed information about sexual partners they currently have, including the relationship status of those partners (steady versus casual), vaginal and anal activity with those specific partners, and the frequency of condom use during vaginal and anal sex with those partners. In the assessment of sexual risk behavior, participants were asked whether or not they had a steady parmer (e.g., a boyfriend, girlfriend, or spouse), whether or not they had a casual or non-steady partner (someone who is not identified as a girlfriend or ' For the purposes of this study, assessment of sexual behavior was limited to vaginal or anal sex, as these behaviors have a higher risk (compared to oral sex, etc.) for disease transmission. Questionnaires were assigned according to the participant's gender, and behavioral assessment items were written to include all possible combinations of sexual partners and sexual activities. For example, female questionnaires included response options which accommodated lesbian/bisexual women with female partners. Similarly, male questionnaires included a "does not apply" response option for gay/bisexual men who were responding to items on vaginal sex. Therefore, the assessments of risk behavior could be made for each participant, regardless of gender orientation and the gender of the partner. 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. boyfriend, but with whom the participant engages in sexual activity), the gender of those partners, and details about their most recent sexual encounter with that person. From this information, we were able to derive the following information about each participant: • Sexual activity status (sexually active versus not sexually active) • Number of sexual partners in the month prior to the clinic visit • Occurrence of vaginal sex with a steady partner in the last 30 days (l=no, 2=yes) • Occurrence of anal sex with a steady partner in the last 30 days ( I =no, 2=yes) • Occurrence of vaginal sex with a casual partner in the last 30 days (l=no, 2=yes) • Occurrence of anal sex with a casual partner in the last 30 days ( l=no, 2=yes) • Frequency of condom use when having vaginal sex with a steady partner (I=always/100% of the time to 5=never/0% of the time). • Frequency of condom use when having anal sex with a steady partner (I=always/100% of the time to 5=never/0% of the time). • Frequency of condom use when having vaginal sex with a casual partner (I=always/100% of the time to 5=never/0% of the time). 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. • Frequency of condom use when having anal sex with a casual partner (I=always/100% of the time to 5=never/0% of the time). The ten items presented above were used to construct a general measure which would assess participants' sexual risk behavior. These items and their respective response formats, as they appear in the questionnaire, are presented in Appendix 3. The assessment of sexually active status was based on the corroboration of number of sexual parmers in the last month with reported sexual activity with steady and/or casual parmers. This method was used because of observed inconsistencies between self-reported number of sexual parmers and self-reported actual behaviors. Thus, for individuals to be categorized as not sexually active, they had to meet three requirements: 1) no self-reported sex parmers in the past month, 2) no self-reported steady parmer or no sexual activity with steady partners, and 3) no self-reported casual parmer or no sexual activity with casual partners. All other persons were categorized according to self-reported behaviors. Those who were not sexually active were given a score of 0, while sexually active persons were given a score of 1 . The assessment of number of sexual parmers was based on self-reported number of sexual partners in the last month corroborated by the indication of sexual partners (steady and casual) with whom actual sexual activity occurred. Participants were then categorized into four groups (no parmers, I parmer, 2 partners, 3 or more parmers) based on the number of partners they reported. 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The third through sixth variables assessed whether or not an actual occurrence of vaginal or anal sex with a steady and casual partner took place in the last month. These variables used the items in the assessment of specific sexual behaviors which read: "Have you had (vaginal/anal) sex with this person in the last 30 days?" Persons for whom the activity occurred were given a score of "1" (per activity that occurred); if no activity occurred, or if the individual did not have a partner with whom to engage in that activity, the individual was given a score of "0." The seventh through tenth variables which contributed to the general risk index assessed the average frequency of condom use with steady and casual partners. More specifically, these variables utilized the items which asked (for both steady and casual partners): "When you have (vaginal/anal) sex with this person, how often is a condom used?" Response options for these items ranged (in 25% increments) from "100% of the time" to "none (0%) of the time." The additional response option of "not applicable" was available for those persons who did not practice the behavior in question. Each of the items had a scoring range from 0 to 4, with the lowest risk score (0) being given to those with no partner, those who had a partner but who did not engage in the behavior with that partner, and those who reported always using condoms with their partner for the behavior in question. The assignment of risk scores proceeded along the 25% increments of condom use, with the highest score (4) being given to those reporting that they never use condoms when engaging in the behavior with their partner. 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sexual behavior risk indices were made by standardizing and averaging each of the ten sexual risk variables, and creating risk scores from the means of the items. Exploratory correlation analyses indicated no significant differences between persons with steady parmers and those with casual partners in terms of associations with other study variables. Therefore, steady and casual partner groups were pooled in order to maintain optimal sample size and power for the creation of sexual risk scores and subsequent analyses using those scores. An overall risk score was created by taking the mean of all ten of the aforementioned items. An alternate overall score, which was based on sexual behaviors and condom use with both steady and casual partners, was created by taking the mean of items 3-10 and excluding the measure of sexual activity status and number of parmers. Several subscores were also created from different combinations of the ten items. The first subscore, which focused on risk attributable to multiple sex parmers, utilized the item pertaining to the number of sex partners the participants had in the last month. The second subscore pertained to use of condoms for vaginal sex. and utilized the two items which assessed frequency o f condom use during vaginal sex with both steady and casual partners. The third subscore was similar in that it focused on condom use during anal sex, and used the two frequency items for anal sex with steady and casuzil parmers. The fourth subscore focused on the occurrence of vaginal sex in general, and utilized the two items assessing vaginal sex with steady and/or casual parmers in the past 30 days. The fifth subscore was similar to the fourth, but focused on 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the occurrence of anal sex and used the two items pertaining to anal sex with steady and/or casual partners in the past 30 days. The variable assessing sexually active status was not used as a subscale because being sexually active may not be an accurate indicator of risk. That is, because sexual activity status can change over time, both in activity and in personal definition, it is a more subjective (and, therefore, less reliable) assessment of risk. Further, because the definition of "being sexually active" is changeable over time, the assessment of an individual's sexual activity status at the time of survey completion yields no information about the time during which he or she became infected with the STD. Thus, only the more objective indicators of sexual risk- such as actual behavior and quantifiable number of partners— were used in the creation of the sexual risk subscores. As previously stated, each score and subscore used variables which had been standardized to mean of 0 and a standard deviation of 1 ; the scores themselves were made by taking the means of the relevant variables. The two overall risk scores and the five subscores were each tested separately in regression analyses as the measure of sexual risk behavior and compared with each other in regard to their performance in the regression models. Denial To measure denial, a shortened version of the Repression-Sensitization scale (Epstein & Fenz, 1967) was used. This scale measures a trait-based style of coping with threatening information; the shortened version, which is a 29-item true/false inventory, 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. was developed to reduce the high correlation found between Byrne's (1961) MMPI- derived scale and measures of anxiety (Gladis et al., 1992). The items that comprise the scale are presented in Appendix 4. Each true-false item is scored 0 for a sensitizer response and 1 for a repressor response, with possible scores ranging from 0 to 1 and the mean of the summed items constituting an individual's score on the R-S scale. In keeping with Gladis et al.'s (1992) procedure, subjects at the top quartile of the scale were labeled "repressors,” those at the lowest quartile were labeled "sensitizers,” and those in the middle were labeled "neutrals"; the classification of subjects into the aforementioned three categories approximates the manner o f subject classification in the original MMPI-derived scale (Byrne, 1961). Both the categorical and continuous measures were used in data analyses. Frequency analyses were used to examine patterns o f missing data responses for each of the R-S items, and based on the observation that the missing data points were randomly distributed across all 29 items it was determined that there was no systematic bias with regards to missing data. With regard to the categorical measure of denial, individuals' scores on the scale were computed by calculating the mean score of the 29 items on the scale and then, following Gladis et al.'s (1992) methodology, dividing the mean score distribution into quartiles. The lowest quartile represented the sensitizers (22% of the sample, n=83), the middle two quartiles represented the neutrals (47.5% of the sample, n=178), and the topmost quartile represented the repressors (30.4% of the sample, n=l 14). A plot of the 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. R-S mean scores indicated that the responses were quite evenly distributed with a only very slight skew (please refer to Table 1), thus the categorization o f the mean scores into quartiles was based on a relatively normal distribution. Using the continuous measure o f denial, we examined the internal consistency of the R-S scale using both Cronbach’ s alpha (which was used by Gladis et al., 1992) and the Spearman correlation coefficient. The Spearman coefficient was originally used by Byrne (1961) to test the reliability of the R-S scale, and uses the method of split-half reliability, in which the mean scores of the odd and even R-S items are computed and the test of reliability is based on the correlation of the two mean scores. However, the use of the Spearman coefficient is not entirely appropriate as it is a rank-ordering procedure used for the analysis of ordinal data, whereas the R-S scale (as well as the knowledge scales, which are discussed in a subsequent section) requires a procedure for interval level data. Thus, we tested the reliability of the R-S scale using both Cronbach's alpha and Pearson correlations based on the split-half procedure. Our analyses show that the R-S scale has a reliability of 0.684 when using Cronbach's alpha; this reliability result was higher than that found by Gladis et al. (a=0.65), who tested the scale with a sample of adolescents (Gladis et al., 1992). Using the split-half procedure, the means of the odd and even variables had a Pearson correlation of 0.521 (p>0.0001). On the surface, these correlations seem low in comparison to the Brown-Spearman consistency coefficient (using split-half procedures) of 0.88 obtained by Byrne (1961), who tested the scale with university 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. undergraduates. The differences in the obtained reliability are most likely not due to skewness since, as discussed earlier, the distribution of the variable was relatively normal. Rather, the differences may be due to the types o f tests which were performed. Because the Pearson, Cronbach, and Brown-Spearman tests are derived from different formulas and, therefore, are bound to yield varying correlation coefficients depending on the other variables in the formulas, results obtained from these calculations cannot be compared. Thus, we cannot assume that an observed low correlation coefficient is equivalent to low reliability. Differences between reliability results obser\'ed here and those observed by other researchers may also be due. in part, to the types of samples which were used to test the scale. For example, samples of adolescents, university undergraduates, and inner-city STD clinic attenders are extremely different, and one would not expect such samples to interpret and respond to questions in a similar fashion. Appraisal of Seriousness of STDs and HTV Six items were used to assess individuals' perceptions o f the seriousness of sexually transmitted diseases. The items assessed individuals' beliefs about the seriousness of the particular STD, concern about the STD's effect on their overall health, the degree to which they feel threatened by the STD, the degree to which they think the STD is treatable, and the degree to which they think the STD is curable. Examples of these items are presented in Appendix 5. 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Because all sexually transmitted diseases are not alike, subjects were asked similarly phrased questions regarding the appraisal of seriousness of three different STDs: syphilis, gonorrhea, and genital warts/herpes. All response scales were in a 5- point Likert format, with responses ranging from "1 = not serious at all" to "5 = extremely serious." This response format is similar to the one used by Weinstein (1982) is his examination of perceived seriousness. Although H TV infection is a "sexually transmitted disease," the nature of H TV infection as an incurable and eventually fatal disease qualitatively separates it from other STDs. Indeed, it may be the case that individuals do not mentally classify HTV infection with other, more typical STDs. Therefore, appraisals of seriousness of STDs and HIV/AIDS may be different and, therefore, must be measured separately. Questions similar to those previously described for STDs were used to measure appraisals of seriousness of HIV/AIDS. Again, the response scales for these items were in 5-point Likert formats, with responses ranging from "1 = not serious at all" to "5 = extremely serious." For each set of seriousness variables (four sets comprised of six items each), Pearson correlation analyses were performed to determine the interrelationship of the items. For each of the four correlation analyses, a similar pattern was observed in which certain items, by virtue of their significant interrelatedness, seemed to form distinct groups within the set of six items. For example, two items were highly and significantly 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. associated with each other but had little or no correlation with the other four items in the six-item set. In order to further investigate this pattern, factor analyses (using a varimax rotated factor pattern) were performed on each of the four sets of items. As suspected, for syphilis, gonorrhea, and genital warts/herpes, two distinct factors were identified. The first factor, which is best described as pertaining to perceptions of concern or threat, was comprised of three items, while the second factor, which pertained more to perceptions about the treatability or curability of the disease in question, was comprised of two items. The one remaining item in each of these three disease categories, which asked individuals to state their level o : agreement or disagreement to the phrase "Getting (disease) is no big deal,” did lot fit into either of the two identified factors and, thus, was dropped from all subsequent analyses. The factor patterns (using the standard regression coefficients) for syphilis, gonorrhea, and genital warts/herpes are shown in Table 2 below. With respect to perceived seriousness of HIV/AIDS, factor analyses were unsuccessful due to the restricted variance of the items, thus the convergence criteria were not able to be satisfied (it did not converge). However, the correlation analyses indicate the same relationship patterns for this set of items, therefore the factor patterns for these variables were assumed to be similar to those found for each of the other three sets of variables. 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2. Varimax rotated factor patterns for perceived seriousness variables. Syphilis In your opinion, how curable is syphilis? (SYPH4): In your opinion, how treatable is syphilis? (SYPH3); How concerned are you that syphilis can affect your overall health? (SYPH5): In your opinion, how serious is a STD like syphilis? (SYPHl): How personally threatened are"/ou by a STD like syphilis? (SYPH6): Agree/disagree; Getting syphilis is "no big deal"? (SYPHl); Gonorrhea In your opinion, how curable is gonorrhea? (G0N4): In your opinion, how treatable is gonorrhea? (G0N3): How concerned are you that gonorrhea can affect your overall health? (G0N5): Treatment/cure Concern/threat .78820 .77478 -.05805 .03116 .03543 .01844 .07433 -.05281 .73666 .54537 .52124 .30048 Concern/threat Treatment/cure In your opinion, how serious is a STD like gonorrhea? (GON1 ): How personally threatened are you by a STD like gonorrhea? (G0N6): Agree/disagree: Getting gonorrhea is "no big deal"? (G0N2): Genital warts/heroes In your opinion, how curable are warts/herpes? (WARTS4): In your opinion, how treatable are warts/herpes? (WARTS3): How concerned are you that warts/herpes can affect your overall health? (WARTS5): In your opinion, how serious are STDs like warts/herpes? (WARTSI): How personally threatened are you by STDs like warts/herpes? (WARTS6): Agree/disagree: Getting warts/herpes is "no big deal"? (WARTS2): .05306 -.06251 .76807 .59049 .60781 .27807 Treatment/cure .90412 .81491 -.02098 .04467 .03244 .24669 .83746 .81791 -.00626 -.01468 .01370 -.00272 Concern/threat .00057 .01098 .86031 .54361 .67707 .21709 Note: Factor patterns are shown using the standard regression coefficients as factor loadings. 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In light of the results of the factor analyses, two seriousness scales were formed for each STD and the reliability of each o f these new scales was tested using Cronbach's alpha. Although the distributions were somewhat skewed, the overall reliability coefficients were, for the most part, quite strong. The raw alpha reliabilities for the seriousness indices pertaining to perceptions of concern/ threat ranged from 0.558 (for perceived threat of HTV/AIDS) to 0.715 (perceived threat of genital warts/herpes). From these three smaller STD indices, a general index pertaining to perceptions of concern/threat was created by combining the three indices for syphilis, gonorrhea, and genital warts/herpes and using the mean score of the three indices. Correlation analyses and tests of reliability (using Cronbach's alpha) were performed for this index, and the reliability was found to be strong (a=0.893). Pearson correlations and Cronbach's alphas for the disease-specific and the concern/threat indices are shown in Table 3. Table 3. Pearson correlations and Cronbach's alphas for seriousness indices pertaining to STD and AIDS concern/threat. Survey items in this factor: # 1 . In your opinion, how serious are sexually transmitted diseases (STDs) like____________? #5. How concerned are you that STDs like____________can seriously affect your overall health? #6. How personally threatened do you feel by STDs like___________ ? Syphilis SYPHl SYPH5 SYPH6 SYPHI 1.0000 SYPH5 .4182 1.0000 (Table 3 continued on next page) 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 3, continued) SYPHl SYPH5 SYPH6 SYPH6 .2608 .4114 1.0000 Cronbach's alpha for raw variables: 0.6024; for standardized variables: 0.6315 Gonorrhea GONl G0N5 GON6 GONl 1.0000 G0N5 .4692 1.0000 G0N6 .3333 .4886 1.0000 Cronbach's alpha for raw variables: 0.6723; for standardized variables: 0.6939 Genital warts/herpes WARTS 1 WARTS5 WARTS6 WARTS 1 1.0000 WARTS5 .4182 1.0000 WARTS6 .2608 .4114 1.0000 Cronbach's alpha for raw variables: 0.6024; for standardized variables: 0.6315 For general C all of the abovel STDs SYPHILIS GONORRHEA WARTS SYPHILIS 1.0000 GONORRHEA .7693 1.0000 WARTS .6651 .7709 1.0000 Cronbach's alpha for raw variables: 0.8930; for standardized variables: 0.8927 HlV/AlDS HIVl H1V 5 HIV6 HIV I 1.0000 HIV5 .3728 1.0000 HIV6 .1460 .4869 1.0000 Cronbach's alpha for raw variables: 0.5584; for standardized variables: 0.6021 6 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. For the seriousness indices pertaining to disease treatability or curability, the alphas obtained were somewhat stronger than those found for the perceived threat indices, and ranged from 0.706 (for perceived treatability/curability of HIV/AIDS) to 0.851 (for perceived treatability/curability of genital warts/herpes). As was done for the concern/threat indices, a general treatability/curability index was formed by using the mean score of the three individual STD treatment/curability indices. Tests of reliability (using Cronbach's alpha) for this index obtained an alpha reliability of 0.556, a much weaker value than that obtained for the concern/threat index. This low alpha may be due to individuals' distinguishing between potential treatability and curability of bacterial STDs (syphilis and gonorrhea) versus that of viral STDs (genital warts/herpes). Pearson correlations and Cronbach's alphas for the disease-specific and the combined treatability/curability indices are shown in Table 4 below. The two general STD indices— concern/threat and treatability/curability— and the two HIV/AIDS indices were used in the data analyses. For each of the scales, higher scores indicate higher perceived senousness. Table 4. Pearson correlations and Cronbach's alphas for seriousness indices pertaining to STD and AIDS treatability and curability. Survey items in this factor #3. In your opinion, how treatable are (STDs) like #4. In your opinion., how curable are STDs like (Table 4 continued on next page) 6 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 4, continued) Syphilis SYPH3 SYPH4 SYPH3 1.0000 SYPH4 .6051 1.0000 Cronbach's alpha for raw variables: 0.7544; for standardized variables: 0.7540 Gonorrhea GON3 G0N4 G0N3 1.0000 G0N4 .6862 1.0000 Cronbach's alpha for raw variables: 0.8118: for standardized variables: 0.8139 Genital warts/heroes WARTS3 WARTS4 WARTS3 1.0000 WARTS4 .7419 1.0000 Cronbach's alpha for raw variables: 0.8513; for standardized variables: 0.8518 For general fall of the above) STDs SYPHILIS GONORRHEA WARTS SYPHILIS 1.0000 GONORRHEA .4348 1.0000 WARTS .2919 .2019 1.0000 Cronbach's alpha for raw variables: 0.5572; for standardized variables: 0.5735 HIV/AIDS HIV3 HIV4 HIV3 1.0000 HIV4 .5673 1.0000 _______ Cronbach's alpha for raw variables: 0.7058; for standardized variables: 0.7239 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Knowledge of STDs/HTV Knowledge about STDs was assessed using seven items which asked about STD transmission, symptoms, parmer notification, time when treatment should be sought, and treatment adherence (Geringer et al., 1993). Knowledge about HIV/AIDS was assessed with a 15-item scale which asked about methods of HIV transmission, methods to prevent transmission, and AIDS symptoms (Nyamathi et al., 1993). Both scales had a True-False response format. Overall knowledge levels (for each scale) were examined by creating a score of the proportion of correct responses out of the total number of items answered. The proportion score was used in order to accommodate instances o f missing data on some o f the items on the knowledge scales. That is, using a summary score might underestimate individuals' levels of knowledge because items where there was missing data would still be included in the denominator of the score whereas a proportion score bases the denominator only on those items for which there is complete data. Frequency analyses were used to examine potential systematic bias for missing responses, and observations of the frequencies showed no systematic pattern of missing data. Further, correlation analyses showed that intervariable relationships when using summary scores were not substantially different from the inter-variable relationships when the proportion scores were used. Therefore, the use of the proportion scores was thought to be justified and these scores were used in subsequent analyses. The items for the STD 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. knowledge scale are presented in Appendix 6 and the items for the HIV knowledge scale are presented in Appendix 7. Tests of reliability for these two scales were performed using both Cronbach’ s alpha and the Spearman's split-half reliability procedure. The 7-item STD knowledge scale (Geringer et al., 1993) was not originally designed to form a construct and, thus, was not tested for reliability by the original investigators. However, tests of reliability were done here. Using Cronbach's alpha to examine the reliability o f this scale, we obtained a strong (raw) reliability coefficient of 0.8174. We obtained a similarly strong Spearman correlation of 0.6977 when using the split-half procedure. Tests of reliability were also performed for the 15-item HTV/AIDS knowledge scale previously used by Nyamathi et al. ( 1993). For this scale, we obtained a (raw) Cronbach's alpha of 0.9089 and a Pearson correlation of 0.8349. These values are higher than the Cronbach's alpha obtained by Nyamathi (a=0.77) in their study population of African-American and Latina women. These high coefficients are apparent despite the fact that the distributions of the knowledge variables are skewed, as study participants tended to have very high levels of both STD knowledge and AIDS knowledge (please refer to Table 1). Also, the stronger coefficients may be due in part to the greater heterogeneity of this sample in that the present sample has both men and women, and greater variance with regard to education and age. 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Perceived vulnerability to future STD infection and HIV infection Perceived vulnerability to these illnesses was assessed using items modeled after those used by Weinstein and others to measure perceived vulnerability, unique invulnerability, and unrealistic optimism (Burger & Bums, 1988; Perloff & Fetter, 1986; Weinstein, 1980, 1983, 1984). Subjects were asked to make three ratings of their own vulnerability. The first item asked the subject for a direct, non-comparative rating of his/her own risk (e.g., "The chance of me getting an STD in the future is..."). The response scale for the non-comparative rating was in a 5-point Likert format, with responses ranging from "1 = not likely at all" to "5 = extremely likely." The other two items asked the subject to make comparative assessments of his/her risk versus that of (I) other people in that clinic and (2) other people of the subject's own age and gender. Response scales for the comparative items were also in a 5-point Likert format, with responses ranging from " 1 = My risk is much less than others" to "5 = My risk is much greater than others." Thus, a total of nine items (three subscales consisting of three items per scale) were used to assess perceived vulnerability to future STD infection, perceived vulnerability to negative consequences from future STD infection, and perceived vulnerability to HIV infection. These items are presented in Appendix 8. Analyses using Cronbach's alpha were performed to determine the degree to which the three variables within each of the vulnerability dimensions were related, and Cronbach's alpha was used to test the reliability of each subscale. In general, each of 6 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the three vulnerability dimensions had very high internal reliability, with alphas ranging from 0.908 to 0.956. Intentions for risk reduction Subjects were asked to rate their likelihood of adopting STD and HIV risk- reducing behaviors in the future, such as using condoms with all sexual partners and reducing the number of sex partners. With regard to condom use, subjects were asked about likelihood within the next month to purchase condoms or get them from the clinic, to initiate condom use with steady/casual sex partners, and to respond positively to condom use initiated by steady/casual partners. Response formats for these items were 5-point Likert scales, with responses ranging from "1 = not likely at all" to "5 = extremely likely." A fourth item, also regarding condom use, asked subjects to estimate the percent of the time in the next month that they would use condoms when they had sex. The response scale for this item was a 5-point Likert-type scale ranging from "100%" to "0%." These items are presented in Appendix 9. Subjects with more than one partner were also asked to respond to items regarding (1) intentions to reduce their overall number of sexual partners, (2) intentions to limit their sexual relations to main/steady partners only, and 3) intentions to not have sex with anyone at all. Response scales for these three items were in a 5-point Likert format, with responses ranging from "1 = not likely at all" to "5 = extremely likely." All nine, when put together to form a larger construct, had a (raw) alpha reliability of 0.8359. However, because six of the items pertained specifically to 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. condom use and three of the items pertained specifically to sexual parmers, the total set of items were submitted to a factor analysis (varimax rotation), which revealed that there were indeed two factors present. The first factor, which consisted of the six items pertaining to intentions for future condom use, had an alpha (raw) reliability of 0.8714. The second factor, which consisted of the three items pertaining to subjects' intentions to reduce their number of sex parmers, had a raw alpha reliability of 0.7007; with the deletion of the item pertaining to celibacy— which had the lowest factor loading of the three items— the alpha increased to 0.7815, therefore the second factor was maintained with only the two items which contributed most to its strong alpha reliability. Based on the results of the factor analyses, two overall indices pertaining to intentions for risk reduction behavior were created and used in subsequent analyses. For the longitudinal analyses of sexual behavior at Time 2, each intentions factor was "matched" with the appropriate behavior(s) in the regression analyses. That is, the intentions for condom use factor was used in the regression analyses pertaining to condom use during vaginal and/or anal sex with both steady and new partners in the six weeks since the participants' clinic visit (and Time I assessment). Similarly, the intentions for reduction in the number of sexual parmers was used in the analyses pertaining to the number of sexual parmers reported by the participants at the Time 2 assessment. The factor patterns (using the standard regression coefficients), correlations, and Cronbach's alphas for these two indices are shown in Table 5. 6 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5. Varimax rotated factor patterns, correlation matrices, and Cronbach's alphas for behavioral intentions variables. Factor patterns fusing standardized regression coefficients) [n the next month, how likely are you to... * Buy condoms/get them from clinic? (INTI) Ask a new partner to use condoms? (INT2) Asking your steady partner to use condoms? (INT3) Use condoms if you new parmer asks you to? (INT4) Use condoms if your steady partner asks you to? (INT5) Use condoms when having anal/vaginal sex? [% of time] (INT6) Cut down on the number of people you have sex with? (INT7) Choose to have sex with only one partner? (INT8) Choose to not have sex with anyone? (FNT9) Condom use No. of partners Correlation matrix. Factor 1 INTI INT2 INTI 1.0000 INT2 .5027 1.0000 INT3 .6001 .5328 INT4 .4691 .6985 INT5 .5136 .5853 INT6 .4746 .5303 INT3 1.0000 .3985 .5765 .5084 .71305 .79652 .73611 .59819 .70247 .67102 -.06972 .07922 .03699 INT4 1.0000 .6470 .5246 .00234 -.04920 -.03790 .19943 .06530 -.01333 .88950 .72689 .36594 INT5 INT6 1.0000 .5303 1.0000 Cronbach's alpha for raw variables: 0.8714; for standardized variables: 0.8754 (Table 5 continued on next page) 69 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 5, continued) Correlation matrix. Factor 2 [NT7 INT8 INT9 1NT7 1.0000 [NTS .6493 1.0000 1N T9 .3605 .3146 1.0000 Cronbach's alpha for raw variables: 0.7007; for standardized variables: 0.7034 _______Cronbach's raw alpha with INT9 deleted: 0.7815; for standardized variables: 0.7874________ Follow-up Measures Actual sexual behavior Sexual behavior at foliow-up was assessed through telephone interviews with study participants who were able to be contacted for follow-up. Individuals were asked about their sexual behavior with both steady and casual partners since leaving the clinic 6 weeks before. The questions asked at the follow-up assessment were identical to those asked in the intake questionnaire (described previously) in that they pertained to protected and unprotected vaginal and anal intercourse with both types of sexual partners. Similarly, the procedure for coding and categorizing individuals according to level of risk behavior was identical to the procedure used in the creation of the Time 1 sexual behavior risk index and, like the Time 1 index, a mean risk score was created by standardizing and averaging the scores from each of the individual ten risk variables. This overall mean was used in the longitudinal analyses of the theoretical models. 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. STD diagnosis During the telephone follow-up assessment, study participants were also asked whether or not they had been diagnosed with a sexually transmitted infection during their clinic visit six weeks prior (when they completed the intake questionnaire). The response format for this variable was dichotomous; participants who had not been diagnosed with an STD received a score of 1, while those who had been diagnosed with an STD received a score of 2. This variable was used in the longitudinal analyses in the examination o f sexual behavior reported at Time 2. v n . RESm.TS. PART I: COMPARISON OF SAMPLE CHARACTERISTICS AND MAJOR STUDY VARIABLES In this section, we will present results from analyses pertaining to the examination o f the subject population, univariate relationships, and more complex relationships among study variables. Further, we will provide a detailed description of the approaches taken in the testing of the theoretical models, as well as the results of those analyses. Sample Characteristics A total of 707 individuals presenting for treatment at the Ruth Temple Health Center STD Clinic were approached for inclusion in the study. Of those approached, 313 (44.3%) persons declined study participation. Those who declined were asked to give brief demographic information (age, ethnicity, gender, education) in order to compare study participants with those declining participation. Analyses of these data 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. showed no significant differences between participants and decliners on any of the four demographic variables. After 10 weeks of full-time data collection, 394 individuals from the clinic were recruited into the study. Demographic information pertaining to the sample is shown in Table 6. The total sample was comprised o f 196 women and 198 men, the majority of whom were English speakers or who chose to complete the questionnaire in English (77.2%) as opposed to Spanish (22.8%). The majority of study participants identified themselves as African-American (65.7%), with a sizeable minority (25.4%) self- identifying as Latino/a. Seven individuals self-identified as being of mixed ethnicity (e.g., Latino/Caucasian mix); for the sake of facilitating analyses, we cross-referenced their self-reported ethnicity with the ethnicity written on the master list by the data collector, and these individuals were re-categorized according to their ethnic group as noted on the master list. There were six individuals for whom information of ethnicity was missing. The age distribution of the sample ranged from 18 to 66 years of age, and the mean age for the entire sample was approximately 32 years of age. Table 6. Demographic characteristics of the study sample ÇN=394).____________________________ Percentage Gender Women 49.7 Men 50.3 (Table 6 continued on next page) 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 6, continued) Language of questionnaire English Spanish Age 24 years of age or under 25-30 years of age 31-40 years of age 41 years of age and over Ethnicity (after re-categorization) African-American Latino Non-Latino White Other Marital Status Single/never married Single and living with partner Married and living with parmer Separated or divorced from parmer Widowed/other Highest grade completed Jr. high school or less High school diploma High school and some college A.A. degree or higher Percentage 77.2 22.8 29.7 21.8 33.0 15.5 67.2 25.7 3.0 4.1 56.6 9.5 11.1 20.8 2.0 19.5 38.8 27.2 13.2 (Table 6 continued on next page) 73 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 6, continued) Percentage Employment status Employed full or part-time 43.6 Unemployed 32.1 Receiving public assistance 1 7.7 Student 5.9 Other _ 0.8 Note: Due to missing data, percentages may not total to 100%.______________________________ Slightly more than half (56.6%) of the individuals in the study indicated that they were single or had never been married; 9.5% of the sample indicated that they were single and currently living with a parmer. Approximately 11% (43 of 394) of the sample were married and currently living with their parmers, and 20.5% were divorced or separated from their parmers. The remaining percentage of persons indicated that they were widowed (1.0%), had missing data for this item (1.3%), or indicated "Other" as their chosen response (1.0%). With regard to sexual orientation, the majority of the sample self-identified as heterosexual, with only 6.8% of the men identifying as gay or bisexual and 7.9% of the women identifying as lesbian or bisexual. With regard to level of education, study participants were asked to indicate the highest grade of school they had completed. Of the 389 (of the total 394) who responded, 19.5% indicated having completed junior high school or fewer years of schooling, 38.8% indicated having had completed high school, and 27.2% indicated 74 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. having had completed some college courses in addition to receiving their high school diploma. Slightly over 13% of the sample indicated having received an Associate of Arts (A.A.) degree or higher. With regard to employment status, 43.6% of the sample were employed full or part time; this is contrasted to a comparable-and slightly higher- percentage of the sample who indicated that they were unemployed or receiving Social Security, disability payments, or other forms of state and federal assistance. Approximately 6% of the sample indicated that they were full or part time students who were not working (as opposed to the students who were working full or part time in addition to attending classes, and who were categorized as being employed). There were no significant gender differences in language preference (English vs Spanish), age, self-identified ethnicity, marital status, or education level. However, for employment status, analyses indicated that the women of the sample were more likely to be unemployed and/or receiving public assistance (Xr=9.699, p=0.0A6). Analyses investigating the association of ethnicity with sociodemographic variables showed no effect for age, but did show significant differences for other variables. For example, there were significant (but not surprising) effects for language preference (more Latinos chose to complete a Spanish-language questionnaire; X^280.280./?< 0.0001), marital status (with more Latinos living with a partner or spouse; X*=74.165,/K 0.0001), level of education (with African-Americans having higher levels of education than Latinos; X ^578.622,p< 0.0001), and employment (with Latinos being more likely to have full or part-time employment compared to African-Americans; X ^ l 1.228, /?=0.024). 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. With regard to the longitudinal cohort, only 32% of the sample (N=126) were successfully recontacted for the follow-up assessment despite repeated attempts to locate study participants. Univariate analyses (Mantel-Haenszel Chi-Square analyses) were used to compare those successfully recontacted with those lost to follow-up. These analyses indicated no differences between the two groups with regard to age, education, and gender. However, African-Americans were significantly more likely to be lost to follow-up than Latinos and non-Latino Whites (X^5.752,p=0.016). Differences between the longitudinal cohort and those lost to follow-up with regard to the theoretical variables were examined using a T-test comparison of group means. This examination revealed two particularly interesting findings. Firstly, the participants who were lost to follow-up had significantly higher mean levels of vulnerability to HIV infection (M=2.651, SD= 1.842) than those in the longitudinal cohort (M=2.051, SD=1.228; t= -3.495, /?=0.0005). Secondly, those lost to follow-up had a lower mean level of perceived seriousness with regard to AIDS treatability and curability (M=4.292, SD= 1.104) compared to the longitudinal cohort (M=4.478. SD=0.819); this finding was marginally significant (t=1.722,/?=0.086). There were no other significant group differences found for any of the other theoretical variables. Sexual Behavior at Intake and Follow-up As a whole, the sample reported a mean of 7.3 sexual parmers (ranging from 0 to 400 parmers) in the last year and 2.6 parmers (ranging from 0 to 100) in the last month. In terms of current sexual parmers, 49 individuals (13.17%) indicated that they had 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. neither a steady nor a casual partner, 172 individuals (46.24%) reported having a steady partner only, 41(11.02%) persons reported having casual partners only, and 110 (29.57%) persons reported having both steady and casual partners. Slightly over half (57%) of the intake sample indicated that they had been treated in the past for sexually transmitted disease infection and 38.1% of those who were able to be recontacted reported having been diagnosed with a sexually transmitted disease during their clinic visit. Another 19% of the follow-up sample reported having been diagnosed with an "other" condition, which could have been an STD not mentioned in the follow-up interview or another condition (such as a yeast infection or an irritation) which is not necessarily considered an STD. Approximately 77% of the sample reported having been sexually active in the six weeks following their clinic visit. More detailed information pertaining to sexual behavior is presented in Table 7. Chi-Square analyses (not presented in the table, but discussed below) were used to examine associations between sexual behaviors and gender and ethnicity. Only those participants who had complete data on all pertinent variables were included in these and all subsequent analyses. Non-Latino Whites were excluded from the analytic sample, as there were too few individuals in this ethnic category to meaningfully compare with the larger samples of Latinos and African-Americans. 77 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 7. Sexual behavior of sample at intake (N=306) and follow-up (N=126) assessments. Intake Follow-up (%) (%) Sexually active at time of assessment No 11.8 22.6 Yes 88.2 77.4 Number of sexual partners in the prior month (or since leaving clinic, for follow-up) None — 14.1 0.9 One 56.0 58.5 Two 18.5 9.4 Three or more 11.4 31.1 Use of condoms for vaginal sex with steady parmer* 100% of the time 30.8 27.0 75% of the time 11.5 12.2 50% of the time 12.2 13.5 25% of the time 15.1 9.5 0% of the time 30.5 37.8 Use of condoms for vaginal sex with casual parmer* 100% of the time 74.0 30.0 75% of the time 6.1 0.0 50% of the time 6.4 20.0 25% of the time 3.4 0.0 0% of the time lO.l 50.0 (Table 7 continued on next page) 78 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Intake Follow-up (%) (Table 7, continued) Use of condoms for anal sex with steady partner* 100% of the time 68.3 89.3 75% of the time 5.5 0.0 50% of the time 4.1 0.0 25% of the time 3.7 0.0 0% of the time 18.5 10.7 Use of condoms for anal sex with casual parmer* 100% of the time 90.8 100.0 75% of the time 1.4 0.0 50% of the time 2.5 0.0 25% of the time 1 .1 0.0 0% of the time 4.2 0.0 Vaginal sex with steady parmer in the last month (or since clinic visit, for follow-up)'’ No 20.7 8.2 Yes 79.3 91.8 Anal sex with steady parmer in the last month (or since clinic visit, for follow-up)'’ No 85.3 90.4 Yes 14.7 9.6 Vaginal sex with casual parmer in the last month (or since clinic visit, for follow-up)'’ No 14.2 0.0 Yes 85.8 100.0 (Table 7 continued on next page) 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Intake Follow-up (%) (Table 7, continued) Anal sex with casual partner in the last month (or since clinic visit, for follow-up)'’ No 86.2 100.0 _______Yes_______________________________ IT8_________________ 00________________ Note: * For items pertaining to condom use, those for whom the behavior was not applicable (e.g., they had no casual parmer, they do not practice anal sex, etc.) were categorized with those who responded "no" to the item. Percentages for the occurrence of vaginal and anal sex are given for those with data on those items; those with missing data or who did not have a specific type of partner were not included in the percentage. Sexual behavior at intake At the initial assessment, the great majority of the sample reported being sexually active. Slightly over half o f the sample reported only one sexual parmer in the previous month, and males were more likely than females to have a higher number of parmers (Mantel-Haenszel 16.749, p<Q.OOO\). Condom use during vaginal sex with steady parmers tended to vary, with 30% of the sample indicating that they always used condoms and another 30% indicating that they never used condoms. Latinos were more likely than African-Americans to never use condoms for vaginal sex with steady partners (Mantel-Haenszel X ^ 6 .154, p=0.013). There was quite a bit of consistent condom use reported for vaginal sex with casual partners, and females were more likely than males to report using condoms 100% of the time (Mantel-Haenszel X^IO.437, /7=0.001). Use of condoms during anal sex with steady parmers was also quite high, 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. although there was a significant minority (18.5%) who indicated that they never use condoms during anal sex. There was a significant ethnicity effect for this variable, as Latinos were more likely than African-Americans to indicate never using condoms for this behavior (Mantel-Haenszel X-=12.887, p<0.0001). The pattern of results for anal sex with casual partners was similar to that for steady partners, and a significant gender effect was observed here in that men were more likely than women to report condom use during this behavior with casual partners (Mantel-Haenszel X^3.865,p=0.049). In general, vaginal sex with both steady and casual partners was more prevalent than anal sex with those partners. With respect to anal sex, African-Americans were less likely to report having anal sex with either steady (X^6.805, p=0.009) or casual (X'=4.710, p=0.030) partners. There were no significant gender effects for these variables. Sexual behavior at follow-up The patterns for sexual behavior at follow-up were, for the most part, similar to the patterns observed at intake. Approximately three-quarters of the sample were sexually active after their clinic visit, and Latinos were more likely than African- Americans to report sexual activity (X^5.525, p=0.019). Approximately 58% of the sample reported one sexual partner since their clinic visit, with Latinos reporting higher numbers of partners (Mantel-Haenszel 14.990,/r=0.0001). Condom use for vaginal sex with steady partners decreased (30.8% at intake compared to 27.0% at follow-up), although African-Americans were more likely than Latinos to report condom use 100% 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of the time (Mantel-Haenszel X^8.544, /;=0.0031). This finding may be partially explained by the difference in number of partners reported at Time 2 by African- Americans as compared to Latinos in that initiating condom use may be easier with steady partners than with new (or numerous) partners. A similar pattern was seen for vaginal sex with casual partners, although there were no significant effects for gender or ethnicity. Again, occurrence o f vaginal sex during the follow-up time period was more prevalent than anal sex, and sexual activity was more likely to occur with steady partners; Latinos were more likely than African-Americans to have reported vaginal sex with a steady partner (X-=14.361, /7=0.0001). Although most o f the sexual activity occurred with steady partners, men were more likely than women to have reported vaginal sex with a casual partner in the time since leaving the clinic (X ^.3 3 9 , p=0.0375). Univariate Analyses of Time 1 Study Variables Correlation analyses were performed in order to determine the interrelationships among theoretical model variables. The results of these analyses are presented in Table 8 (correlations of demographic and theoretical model variables) and Table 9 (correlations among theoretical variables only). The continuous measures for the theoretical variables (except for STD history, which is dichotomous), age, and education were used in all subsequent analyses. 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 8. Correlations between demographic variables and major study variables. Ethnicity Gender Age Education STD history -.03 .09 .11 -.02 Sexual risk (intake) -.20'" .13' -.13' -.19'" Denial -.05 .09 .08 -.02 STD knowledge .13" -.01 -.03 .23'" AIDS knowledge .09 -.05 .02 .24'" STD threat .43"' -.02 -.03 .22"' STD treatability .00 -.12' -.14' -.04 AIDS threat .43"' .04 .02 .24'" AIDS treatability -.00 .03 -.13' .16'" PV , future STD -.02 .03 .04 -.09 PV , negative consequences -.15" .01 .00 -.16" PV , HIV .02 .00 .12" -.04 Intentions for condom use .11 -.06 -.12' .10 Intentions for reduction in number of sexual parmers -.04 -.09 -.05 .00 Sexual risk (follow-up) -.19"' .15 .06 -.11 Note: The correlations among the first 1 3 variables are based on the intake analytic sample, N=306. The correlations for sexual risk at follow-up are based on the analytic sample at Time 2, N=126. Ethnicity was coded as follows: l=Latino, 2-African-American; gender was coded as follows: l=women, 2=men. " p< .05, " p< .01, •” p<.005 Relationships between demographic factors and theoretical variables With regard to associations between demographic factors and theoretical variables, there were several noteworthy findings (see Table 8). Ethnicity was 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. negatively associated with sexual risk behavior, both at intake (r=-0.20, p<.005) and at follow-up (r=-0.19,/K.005), indicating that Latinos tended to have higher sexual risk scores than did African-Americans. Ethnicity was also positively associated with STD concern/threat (r=0.43,p<.005) and AIDS concern/threat (r=0.43,/;<.005), with African-Americans indicating greater perceived threat for these two diseases. Interestingly, African-Americans also indicated less perceived vulnerability to negative consequences from future STD infection (r=-0.l5,p<.005). Age was significantly and negatively correlated with sexual risk behavior at intake (r= - 0.13,/7<.05), indicating that younger persons had higher levels of risk behavior. However, age was also inversely associated with intentions for condom use and intentions for reduction in number of partners in that younger persons had higher scores on the intentions variables. Education was also significantly and inversely correlated with sexual risk behavior at intake (r=-0.19,/7<.005) and perceived vulnerability to negative consequences from future STD infection (r=-0.16,p<.01), indicating that lower education levels are associated with higher risk behavior and greater perceived vulnerability. Not surprisingly, education was positively and significantly associated with STD knowledge (r=0.23,/7<.005) and AIDS knowledge (r=0.24, p<.005) in that higher levels of education are associated with higher knowledge scores for these two diseases. Education was also positively associated with three of the four seriousness 84 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 73 CD ■ D O Q . C g Q . ■ D CD o ' 3 O 3 CD 8 ë~ 3 " Ï 3 CD 3. 3 " CD CD ■ D O Q . C a o 3 ■ D O CD Q . ■ D CD C/) T able 9. Pearson correlations among major study variables. 1 2 3 4 5 6 7 8 9 10 II 12 13 14 15 1, STD history 1.00 2. Sexual risk (intake) 13" 1.00 3, Denial -.04 .01 1.00 4. STD knowledge -.01 -.08 .14'" 1.00 1 5. AIDS knowledge .01 -.08 I I ' .4 6 '" 1.00 6, STD threat .03 -.08 -.06 .1 8 "' .1 7 '" 1.00 7. STD treatability -.07 -.02 -.08 -.01 -.04 -0 9 1.00 8. AIDS threat .02 -.09 -.09 .3 1 "' .2 7 "' .6 7 '" -.01 1.00 9. AIDS treatability -.03 -.05 .04 .3 2 '" 29". .08 .15" .15" 1.00 10. PV, future STD .06 .04 -.04 -.03 -.07 .01 -.08 .07 -.17"' 1.00 II. PV, neg. conseq. .10 .04 -.05 -.08 -1 5 " -.03 -.06 -.02 -.16"' .79"' 1.00 12. pv.mv .17" .08 -.10 -.03 -.05 .04 -.10 .04 -.19"' .6 6 "' .6 5 '" 1.00 13. Intentions (condoms) .09 -.22"' -.06 .09 .13' .20'" .04 .13' .04 .05 .09 .04 1.00 14. Intentions (partners) .21 .05 .13 .20 .26 .03 -.01 -.02 .17 .13 -.02 .16 .25 1.00 15. Sexual risk (follow-up) .20' .21' .13 .07 .06 -.03 .03 -.08 .03 .00 .05 .12 -.16 .03 1.00 Note; The corrélations among variables 1-13 are based on Ibe intake analytic sample, N^306. The correlations Jbr intentions for reduction in numbers o f sexual partners [noted above as Intentions (partners)) are based on the subset o f the sample who indicated more than one partner and who responded to those particular intentions items (N =5I). The correlations for sexual risk at follow-up are based on the analytic sample at Time 2 (N=126). ’ p< .05. * * p< .01, * ’* p< .005 00 L n variables: STD concern/threat (r=0.22,/?<.005), AIDS concern/threat (r=0.24,/K.005), and seriousness pertaining to AIDS treatability and curability (r=0.16,/7<.005). Associations among theoretical model variables Correlation analyses were performed in order to examine inter-relationships among the major theoretical variables. Although these results are presented in table form (please refer to Table 9), several significant findings are discussed here. Please note that correlations presented for sexual risk behavior at follow-up are based on the longitudinal sample (N=126) rather than the cross-sectional sample (N=306). Of the two theoretical variables assessing participants' past behavioral risk for STD/HIV infection, prior STD history was positively associated with sexual risk behavior assessed at intake (r=0.13. p<.05), and both of these variables were associated with sexual risk behavior at follow-up (prior STD history: r=0.20, p<.05; sexual risk behavior at intake: r=0.21, p<.05). Prior STD history was also associated with higher perceptions of vulnerability to HIV infection (r=0.17,p<.01), whereas sexual risk behavior was associated with fewer intentions for risk reduction behavior as it pertained to condom use (r=-0.22, p<.005). STD knowledge and AIDS knowledge were significantly and positively associated with each other (r=0.46,/7<.005). Additionally, STD knowledge was associated with three of the four seriousness variables: STD concern/threat (r=0.18, p<.005), AIDS concern/threat (r=0.3 l,p<.005), and AIDS treatability/curability (r=0.32, p<.005); the same pattern of results was found with regard to the association 86 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. between AIDS knowledge and these three seriousness variables. These associations suggest that higher levels of knowledge are related to more realistic assessments o f diseases seriousness, both in terms of the perceived level of threat the disease may pose to individuals and the degree to which the disease is treatable or curable. Interestingly, STD knowledge and AIDS knowledge were also significantly associated with denial (STD knowledge: r=0.14,/7<.005; AIDS knowledge: r=0.1 l,p<.05), indicating that those with higher scores on the denial scale are also likely to have higher scores on the knowledge scales. Although the association between these variables and denial cannot speak to causation, it does suggest that high levels of accurate information about a disease may provoke feelings of threat strong enough to trigger defensive denial. STD concern/threat was significantly and positively associated with both AIDS concern/threat (r=0.67, /j<.005) and intentions for risk reduction behavior (r=0.20, p<.005). AIDS concern/threat and AIDS treatability/curability were positively and significantly correlated with each other (r=0.l5,p<.01) but, interestingly, STD concern/threat and STD treatability had a negative intercorrelation which was not significant. The directionality of the association between the two STD seriousness variables suggests that, although STDs may be perceived as being potentially threatening in terms of their effects on health, they are not perceived to be serious diseases with regard to their treatability and curability. 87 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. All three perceived vulnerability variables were strongly and positively correlated, with the highest correlation being between perceived vulnerability to future STD infection and perceived vulnerability to negative consequences from future STD infection (r=0.79, /?<.005). With regard to the associations between the intake variables and Time 2 sexual risk behavior, a separate correlation analysis was done using the analytic follow-up sample (N=126). The results o f this analysis indicated that only two variables— prior STD history and sexual risk behavior at intake— were significantly and positively associated with risk behavior at follow-up. The negative relationship between intentions for condom use and Time 2 risk behavior was only marginally significant (p=.07) but interesting nonetheless, as fewer intentions for using condoms (assessed at intake) were associated with a higher risk score at follow-up. VIII. RESULTS. PART II: SUBSTANTIVE ANALYSES Data Analytic Approaches to Testing Theoretical Models Test of the main effects hypotheses Four regression analyses were used to test the main effect hypotheses (see Table below). In the first regression model, intentions for risk reduction will be regressed on the behavioral (previous STD history, patterns of risky sexual behavior), motivational (denial), cognitive (appraisal o f STD/HIV seriousness, STD/HTV knowledge), and the perceived vulnerability variables. In order to protect against collinearity effects, the two perceived vulnerability variables pertaining to STD infection were combined (the mean of the two combined scores was used) into one general "STD-related perceived 88 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vulnerability" variable. Demographic variables were also entered into the model as independent predictors. This analysis serves as the test for Main Effects Hypothesis 6: perceived vulnerability as a predictor of behavioral intentions. Additionally, this analysis provided the direct residual paths for the other predictor variables to behavioral intentions. Regression #1A Regression #1B Regression #2 Regression #3 Regression #4 Dependent Variable Independent Variables Intentions for risk reduction (STD-specific model) • Previous STD history • Sexual risk behavior • Denial • STD seriousness (concern/threat) • STD seriousness (treatability/curability) • STD knowledge • Perceived vulnerability to STD infection/negative consequences Intentions for risk reduction (HIV-specific model) • Previous STD history • Sexual risk behavior • Denial • AIDS seriousness (concern/threat) • AIDS seriousness (treatability/curability) • HIV/AIDS knowledge • Perceived vulnerability to HIV infection Perceived vulnerability to future STD infection • Previous STD history • Sexual risk behavior • Denial • STD seriousness (concem/threat) • STD seriousness (treatability/curability) • STD knowledge Perceived vulnerability to negative consequences from future STD infection • Previous STD history • Sexual risk behavior • Denial • STD seriousness (concem/threat) • STD seriousness (treatability/curability) • STD knowledge Perceived vulnerability to HIV infection • Previous STD history • Sexual risk behavior • Denial • AIDS seriousness (concem/threat) • AIDS seriousness (treatability/curability) • HIV/AIDS knowledge 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Regression equations 2 and 3 tested Main Effects Hypotheses 1-5 which are specific to STD infection. In Equation 2, perceived vulnerability to future STD infection was regressed on the behavioral, motivational, and cognitive variables. Similarly, Equation 3 regressed perceived vulnerability to negative consequences from future STD infection on the antecedent and demographic variables. These analyses served to determine the degree to which each of the antecedent variables predicts perceptions of vulnerability. Regression equation 4 tested Main Effects Hypotheses 1-5 which are specific to HIV infection. This model was similar to Regressions 2 and 3 in that perceived vulnerability was regressed on the antecedent and demographic variables. For each of the above regression analyses, any variables whose regression coefficients were found to be significant were said to be a unique predictor of the dependent measure. Please refer back to Figure 1 for a representation of the model. Analysis of the path model In order to test the hypotheses regarding the effect of perceived vulnerability as a mediating variable, we need the path coefficients for Paths A and B (see the example in Figure 3), which are the paths used in the formula which tests Denial (IV) InunU ona for rlak réduction (DV) P erceiv ed v u inerabillljr to fu tu re STD infecU on (WED) M ediation Model Figure 3 the significance of the mediator. If the mediating model is significant, we should find 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. significant coefficients for Paths A and B. Also, we should find that the strength of the direct path C (i.e., the strength of the path coefficient) from the IV to the DV is diminished by the presence of the mediating variable in the model (Baron & Kenny, 1986). Each of the necessary path coefficients could be obtained from the four regression equations outlined above. For example. Equation 1, which regresses intentions for risk reducing behavior on the cognitive, motivational, behavioral, and vulnerability variables, provided the coefficients for Path B. which is the path from the mediators (perceived vulnerability) to the dependent variable. Equations 2-4, in which each o f the perceived vulnerability variables are the dependent measures, provided the coefficients for the path from the independent variables to the mediators (Path A). Three separate tests of mediation were done for each of the three vulnerability variables. The significance of the mediator will be examined with the following formula (Sobel, 1982). In this formula, the numerator is the product of the unstandardized regression coefficients from Path A (the path from the IV to the mediator, noted as a) and Path B (the path from the mediator to the DV, noted as b). The denominator also uses the unstandardized regression coefficients from Paths A and B, as well as their standard errors (the standard error from Path A is noted as s,; the standard error from Path B is noted as jy). This formula tests the indirect effect of the independent variable on the dependent variable via the mediator. 91 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. / = - a *b s u f I * Figure 4 Test of the moderating models Two regression analyses were used to test each of the moderating hypotheses (see example in Figure 4). In the first regression (Step 1 in the example), perceived vulnerability (DV) was regressed on the independent variables in order to obtain the path coefficient for the main effects. The second regression model (Step 2) had the independent variables (IVS) which were present in the first model and the interaction term entered sequentially, so that the effect of the interaction term could be obtained after controlling for the main effects. Separate analyses were done for each of the moderation hypotheses pertaining to perceived vulnerability to 1 ) future STD infection, 2) negative consequences of future STD infection, and 3) HIV infection. If the moderating effect is significant, we should expect to find that the strength of the relationship between the independent variable and the dependent variable to be stronger or weaker depending on the condition of the moderator. In other words, we would expect to find a significant interaction effect between the independent variable 92 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and the moderator. Using the example in Figure 4 (above), one would expect a significant main effect between risky sexual behavior and perceived vulnerability. That is, one would expect that an individual with a pattern of risky behavior would perceive himself to be more vulnerable to future STD infection. However, if denial is a significant moderator, the relationship between risky behavior (IV) and perceived vulnerability (DV) would_be weaker under conditions o f greater denial than under conditions of lower denial. In other words, if the individual is a denier (scores higher on the denial dimension), his denial is likely to reduce any effect that an appraisal of his own risky behavior would have on the ratings of his vulnerability. Cross-Sectional Analyses; Tests of the Path Model As described in the previously presented table (p. 89), four multiple regression analyses were needed to test the relevant path models. The results fi-om the main analyses (Equations 1-4) are shown beginning in Table 10. The results from the analyses in which the sexual risk subscales were substituted in the model in place of the overall sexual risk variable are briefly discussed in the text following the discussion of the main analyses; the unstandardized regression coefficients (and standard errors) for the various sexual risk subscores are presented (for comparative purposes) in tables immediately following the table presentation of the main analysis. For example, the results of the main analysis for the first regression model. Equation 1 A, are presented in Table 10. The results of the re-analyses of Equation 1 A— that is, the regression coefficients for each of the different sexual risk subscores— are presented in Table 11. 93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Regression analyses were performed separately for the different partner subgroups (i.e., steady only, casual only, steady and casual); however, since the results of these analyses did not substantially differ from one another, all partner groups were combined for the major analyses. Thus, the data presented here are based on the results of the analyses using the pooled sample. Table 10. Results from multiple regression analyses of Equation I A: Intentions for future condom use and reduction in number of paitners regressed on theoretical variables, STD-specific models (cross- sectional analytic sample; N=306). Regression model b SE P Equation 1 A(a): Predictors of intentions for future condom use (/?■ = 0.1464) Ethnicity -.037 .173 .830 Gender .002 .124 .989 Age -.017 .006 .009 Education .042 .075 .579 Prior STD history .310 .124 .013 Sexual risk behavior -.639 .136 .000 Denial -.413 .586 .481 STD knowledge .449 .447 .316 STD seriousness (concem/threat) .234 .082 .005 STD seriousness (treatability/curability) .062 .073 .394 Perceived vulnerability to future STD infection/negative consequences .052 .037 .168 (Table 10 continued on next page) 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 10, continued) Regression model b SE P Equation 1 A(b): Predictors of intentions for future reduction in number of sex partners = 0.0485) Ethnicity -.097 .238 .684 Gender -.089 .266 .738 Age -.008 .011 .450 Education -.060 .145 .678 Prior STD history -.174 .246 .479 Sexual risk behavior -.172 .223 .441 Denial .212 1.098 .847 STD knowledge 1.427 .770 .066 STD seriousness (concem/threat) -.114 .163 .487 STD seriousness (treatability/curability) -.143 .154 .354 Perceived vulnerability to fiimre STD infection/negative consequences -.040 .079 .611 Table 11. Results from re-analyses of regression model #IA, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample, N=306). Regression model b(S£) P R: F P Equation lA(a): STD-specific model regressing intentions for future condom use on theoretical variables, where the sexual risk score is comprised of... Original mean score with all 1 0 sexual risk variables -.639(.136) .0001 .146 4.585 .0001 1 . Mean score excluding sexually active status -.518(.127) .0001 .131 4.037 .0001 and no. partners in past month variables (Table 11 continued on next page) 95 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 11, continued) Regression model b (SE) p R * 2. Number of sex partners in the past month -.037(.080) .6484 .089 2.561 .0042 3. Mean score pertaining to use of condoms -.377(.083) .0001 .143 4.427 .0001 for vaginal sex with steady/casual parmers 4. Mean score pertaining to us£ of condoms -.260(.087) .0032 .119 3.458 .0002 for anal sex with steady/casual partners 5. Mean score pertaining to occurrence of .082(.071) .2511 .091 2.375 .0082 vaginal sex with steady/casual partners 6. Mean score pertaining to occurrence of -.144(.077) .0614 .103 2.762 .0021 anal sex with steady/casual partners _____ ____________________________ _____ The first set of analyses (regression models IA and IB) regressed intentions for risk reduction behavior on the behavioral (prior STD history, sexual risk behavior [overall 10-item mean score]), cognitive (STD and HTV knowledge, perceptions of STD and HIV seriousness), and motivation-based (denial) independent measures. Separate models were tested for the two different intentions factors (intentions regarding condom use and intentions regarding reduction in the number of sex partners) as the dependent variable; separate models were also tested using the STD-related and HIV-related independent variables so that each model was specific to STDs or HIV. The second and third sets of analyses (regression models 2 and 3) were specific to STD-related perceived vulnerability, in that perceived vulnerability to future STD infection and perceived vulnerability to negative consequences from STD infection were the respective dependent measures. The last set of analyses (regression model 4) regressed 96 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. perceived vulnerability to HIV infection on the aforementioned independent variables. All of the analyses controlled for age, education, ethnicity, and gender by including them in the model as predictor variables. Equation lA The first main set of analyses consisted of the models which examined intentions for condom use as the dependent measure. These data are presented in Table 10. In the first analysis (see Equation I A, Table 10) which was performed, intentions for future condom use was regressed on the model variables which pertained to STDs: that is, STD knowledge, STD seriousness pertaining to concem/threat. STD seriousness pertaining to treatability/curability, perceived vulnerability to future STD infection, and perceived vulnerability to negative health consequences arising from STD infection were included in the model with the behavioral and motivation variables. The complete model was statistically significant (F=4.585;/?=0.000l) and accoimted for approximately 15% of the variance in intentions for future condom use. Of the variables which comprised the model, the strongest predictor was sexual risk behavior (b=-0.6395, SE=0.1361,/7=0.0001). STD seriousness (concem/threat), prior STD history, and age were also significant predictors of intentions. The same model was then retested with intentions for reduction in the number of sex partners as the dependent variable. This model was not statistically significant and accounted for almost 5% of the variance in intentions for reduction in the number of partners. O f the variables in the model, only STD knowledge emerged as a marginally 97 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. significant predictor of intentions to reduce one's number of sexual partners (b=l .4272, SE=0.7700, /7=0.0662). Because sexual risk behavior was not a significant predictor of intentions, this model was not included in the following re-analyses of Equation lA using the sexual risk subscales. Re-analyses of equation lA All six of the re-aiialyzed models— that is, those which substituted one o f the risk subscales in place of the composite sexual risk variable— were significant and, for the most part, the variables which were significant in the primary model (which used the overall, 10-item mean risk score) were also significant in the re-analyzed models. The regression weights for the various sexual risk subscores obtained from these re-analyses are presented, for comparative purposes, in Table 11. The first of the six re-analyzed models, in which a mean score comprised of the seven variables dealing specifically with sexual behaviors with steady and casual partners was substituted for the lO-item composite score, was significant (F=4.037,p=0.0001) and accounted for 13% o f the variance in intentions for condom use. In this model, the measure of sexual risk behavior (b=-0.5189, SE=0.1273, p=0.0001 ) was a significant predictor of intentions for condom use. The second model, which used number of sexual partners in the prior month as the measure of sexual risk, was also significant (F=2.561, p=0.0042) and accounted for almost 9% of the variance. However, of all the variables included in the model, only seriousness pertaining to STD concem/threat and age were significant predictors. The 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. third model, in which the variables pertaining to condom use during vaginal sex were used to create a mean risk score, was highly significant (F=4.427, /;=0.0001) and accounted for 14% of the variance. In this model, sexual risk behavior (b—0.3766, SE=0.0830,/7=0.000I) was a significant predictor. Other significant variables were age, prior STD history, and STD seriousness (concem/threat). The fourth model,_which included a mean risk score based on variables pertaining to condom use during anal sex, was also highly significant (F=3.458, /7=0.0002), and accounted for 12% of the variance in the dependent measure. In this model, the same variables as in the third model were significant predictors of intentions for condom use. The fifth model, which used a mean score of variables measuring the occurrence of vaginal sex with steady/casual partners, was significant (F=2.375, p=0.0082) and accounted for approximately 9% of the variance. In this model, the sexual risk measure did not approach significance. The sixth model, which substituted the mean score of the measures for occurrence of anal sex with steady/casual partners in place of the composite risk score, was significant and accounted for 10% of the variance (R-=0.1025, F=2.762,p=0.G02I). In this model, sexual risk behavior (b= -0.1441, SE=0.0767, p=0.0614) was a marginally significant predictor of intentions. As seen in these re-analyses, the measure of sexual risk behavior was not always a significant predictor of intentions for future condom use. However, the comparison of the strength of the regression coefficients for the subscores indicate that the sexual risk measures used in the third and fourth revised models— condom use for vaginal sex with 99 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. steady/casual partners [mean score] and condom use for anal sex with steady/casual partners [mean scorej-carry most of the predictive weight in the composite 10-item measure of sexual risk. Equation IB The second intentions model (see Equation IB, Table 10) was specific to HTV infection and included the HIV-related knowledge, seriousness, and perceived vulnerability variables. Like Equation lA, the model was tested twice: first, with intentions for future condom use as the independent variable, and the second time with intentions for reduction in the number of sexual parmers as the independent measure. These results are presented in Table 12. The first test o f the model, with intentions for future condom use as the outcome measure, was statistically significant (F=4.114, p=0.0001), and accounted for approximately 13% of the variance. For this model, sexual risk behavior (b=-0.6599, SE=0.1376,/7=0.0001) emerged as a significant predictor of intentions. Table 12. Results from multiple regression analyses of Equation IB: Intentions for future condom use and reduction in number of parmers regressed on theoretical variables, HIV-specific models (cross- sectional analytic sample; N=306). Regression model b SE p Equation I B(a): Predictors of intentions for future condom use (^* = 0.1334) Ethnicity .072 .175 .679 Gender -.004 .125 .974 (Table 12 continued on next page) 100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 12, continued) Regression model b SE P Age -.020 .007 .002 Education .017 .077 .826 Prior STD history .312 .126 .014 Sexual risk behavior -.659 .138 .000 Denial -.466 .591 .431 AIDS knowledge 1.106 .561 .049 AIDS seriousness (concem/threat) .104 .085 .221 AIDS seriousness (treatability/curability) -.036 .069 .603 Perceived, vulnerability to HIV infection .034 .038 .380 Equation 1 B (b): Predictors of intentions for future reduction in number of sexual partners {R- = 0.0765) Ethnicity .124 .237 .604 Gender .021 .250 .933 Age -.002 .011 .858 Education -.062 .147 .674 Prior STD history -.165 .236 .486 Sexual risk behavior -.152 .226 .502 Denial .245 .008 .806 AIDS knowledge .542 .013 .594 AIDS seriousness (concem/threat) -.349 .150 .021 AIDS seriousness (treatability/curability) .189 .127 .139 Perceived vulnerability to HIV infection -.008 .075 .913 101 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The second test of the HIV-specific model regressed intentions for reduction in the number of sexual partners on the theoretical and vulnerability variables. This model was not statistically significant. In this model, sexual risk behavior was not a significant predictor of intentions and, therefore, this model was not included in the re- analyses of Equation IB using the sexual risk subscales. Re-analyses of equation JB Each of the six re-analyzed models containing the sexual risk subscales were statistically significant (see Table 13) and, as in the re-analyses for equation lA, the pattern of variables which were significant within each model differed. For the model using the mean score of the items specific to sexual behavior with steady/casual parmers, sexual risk behavior (b=-0.5313, SE=0.1289,/?=0.0001) was found to be a significant predictor variable. In the second model, which used number of sex parmers in the prior month as the measure of sexual risk, sexual risk behavior was not a significant predictor of intentions for future condom use. Table 13. Results from re-analyses of regression model #1B, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample, N=306). Regression models b (SE) p F p Equation IB: HIV-specific model regressing intentions for future condom use on theoretical variables, where the sexual risk score is comprised of... Original mean score with all 10 sexual risk variables -.659(. 138) .0001 .133 4.114 .0001 (Table 1 3 continued on next page) 102 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 13, continued) Regression model b(S£) 1 . Mean score excluding sexually active status -.531(. 129) and no. partners in past month variables 2. Number of sex partners in the past month -.016(.080) 3. Mean score pertaining to use of condoms -.399(.085) for vaginal sex with steady/casual partners 4. Mean score pertaining to use of condoms -.282(.088) for anal sex with steady/casual parmers 5. Mean score pertaining to occurrence of .072(.073) vaginal sex with steady/casual partners 6. Mean score pertaining to occurrence of -.144(.079) anal sex with steady/casual parmers____________________ .0001 .117 3.529 .0001 .8422 .071 1.995 .0288 .0001 .134 4.093 .0001 .0015 .108 3.124 .0005 .3239 .078 1.994 .0293 .0699 .086 2.273 .0115 The third model, in which the mean score of the variables pertaining to use of condoms for vaginal sex were used as the measure of risk, was significant and accounted for approximately 13% of the variance; in this model, sexual risk behavior (b=-0.399I, SE=0.0845.p=0.0001) was a statistically significant predictor of intentions for condom use. Similarly, sexual risk behavior was a significant predictor variable in the fourth model, which used the mean scored pertaining to use of condoms for anal sex as the measure of risk. The fifth model, in which occurrence of vaginal sex was used as the sexual risk behavior item, was significant, but had a lower goodness of fit that either of the first two models. In this model, sexual risk behavior was not significant. However, sexual risk 103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. behavior was marginally significant (b=-0.1442, SE=0.0793,p=0.0699) in the sixth model, which utilized occurrence of anal sex as the sexual risk measure. Equation 2 The second main regression model to be tested regressed perceived vulnerability to future STD infection on the behavioral, cognitive, and motivational variables; these results are presented in Tÿ)le 14. This model was not statistically significant. Further, none o f the six re-analyzed models were statistically significant and, in five of the six models, none of the other variables within the models were significant. The only significant finding occurred in the re-analysis of the sixth model, which used occurrence of anal sex with steady/casual partners as the measure of risk (see Table 15). In this model, sexual risk behavior (the risk subscore) emerged as a significant predictor of perceived vulnerability to future STD infection (b=0.3388, SE=0.1203, p=0.0052). This finding seems to indicate that the behavior of anal sex has a contributing role in the development of perceptions of vulnerability to STD infection, although its predictive strength was not strong enough to make the overall model statistically significant. Table 14. Results from multiple regression analyses of Equation 2: Perceived vulnerability to future STD infection regressed on theoretical variables, STD-specific model (cross-sectional analytic sample; N=306). Regression model b SE p Equation 2: Predictors of perceived vulnerability to future STD infection = 0.0229) Ethnicity .070 .290 .809 (Table 14 continued on next page) 104 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 14, continued) Regression model b SE P Gender .060 .208 .773 Age .005 .011 .676 Education -.214 .125 .088 Prior STD history .160 .208 .441 Sexual risk behavior -.089 .228 .694 Denial -.819 .980 .404 STD knowledge .011 .748 .988 STD seriousness (concern, threat) .024 .137 .864 STD seriousness (treatability/curability) -.166 .122 .175 Table 15. Results from re-analyses of regression model #2, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample, N=306). Regression models b(5£) P R- F P Equation 2: Model regressing perceived vulnerability to future STD infection on theoretical variables, where the sexual risk score is comprised of... Original mean score with all 1 0 sexual risk variables -.089(.228) .6942 .023 .692 .7321 I. Mean score excluding sexually active status and no. partners in past month variables .0830211) .6947 .023 .692 .7322 2. Number of sex parmers in the past month -.0250129) .8482 .021 .601 .8131 3. Mean score pertaining to use of condoms .0660138) .6324 .022 .653 .7676 for vaginal sex with steady/casual partners (Table 15 continued on next page) 105 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 15, continued) Regression model biSE) P R: F P 4. Mean score pertaining to use of condoms for anal sex with steady/casual partners -.065(.I43) .6497 .015 .421 .9359 5. Mean score pertaining occurrence of vaginal sex with steady/casual partners -.058(.ll6) .6190 .022 .573 .8351 6. Mean score pertaining occurrence of anal sex with steady/casual partners .339(.l20) .0052 .046 1.283 .2401 Equation 3 The third main model, which regressed perceived vulnerability to negative health problems arising from STD infection, was marginally significant (F=1.683, p=0.0839), although none of the variables within the model reached significance as predictors. These results are shown in Table 16. With regard to the subscales, the same pattern as that shown with the re-analyses for regression model 2 was observed here. Of the six re-analyses (see Table 17), only the model in which occurrence of anal sex with steady/casual partners was used as the sexual risk predictor measure approached significance (F=1.854,p=0.0518). Within that model, the sexual risk variable was the only significant predictor of perceived vulnerability to negative health problems arising from STD infection (b=0.3110, SE=0.1170,p=0.0083). As in Equation 2, the behavior of anal sex seems to have influence in the development of perception of vulnerability and was most likely the variable within the composite risk score which allowed the main model to reach even marginal significance. 106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 16. Results from multiple regression analyses of Equation 3; Perceived vulnerability to negative consequences from future STD infection regressed on theoretical variables, STD-specific model (cross- sectional analytic sample; N=306). Regression models b SE P Equation 3: Predictors of perceived vulnerability to negative consequences from STD infection = 0.0540) Ethnicity — -.483 .279 .085 Gender .052 .201 .795 Age -.000 .010 .980 Education -.202 .121 .096 Prior STD history .326 .200 .105 Sexual risk behavior -.215 .219 .328 Denial -.903 .944 .339 STD knowledge -.484 .721 .502 STD seriousness (concem/threat) .087 .132 .509 STD seriousness (treatability/curability) -.112 .117 .339 Table 17. Results from re-analyses of regression model #3, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample, N=306). Regression models b {SE) p R- F P Equation 3: Model regressing perceived vulnerability to negative consequences from STD infection on theoretical variables, where the sexual risk score is comprised of... (Table 17 continued on next page) 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 17, continued) Regression model b SE P Original mean score with all ten sexual risk variables -.215(.219) .3285 .054 1.683 .0839 1 . Mean score excluding sexually active status and no. partners in past month variables -,076(.204) .7113 .051 1.597 .1067 2. Number of sex partners in the past month ■023(.125) .8537 .046 1.368 .1944 3. Mean score pertaining to use of condoms for vaginal sex with steady/casual parmers -.037(.133) .7830 .048 1.487 .1435 4. Mean score pertaining to use of condoms for anal sex with steady/casual partners -.182(.138) .1876 .039 1.158 .3194 5. Mean score pertaining occurrence of vaginal sex with steady/casual parmers -.068(.113) .5495 .046 1.258 .2549 6. Mean score pertaining occurrence of anal sex with steady/casual parmers .311(.117) .0083 .065 1.854 .0518 Equation 4 The fourth main regression model— in which perceived vulnerability to HIV infection was the independent measure— did reach statistical significance (F=2.760, p=0.0029) and accounted for almost 9% of the variance. In this model (shown in Table 18), sexual risk behavior was not a significant predictor of perceptions of vulnerability to HIV infection. 108 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 18. Results from multiple regression analyses of Equation 4: perceived vulnerability to HIV infection regressed on theoretical variables, HIV-specific model (cross-sectional analytic sample; N=306). Regression models b SE P Equation 4: Predictors of perceived vulnerability to HIV infection {R~ Ethnicity -.009 = 0.0856) .265 .973 Gender — -.071 .190 .710 Age .017 .009 .083 Education -.039 .116 .737 Prior STD history .483 .189 .011 Sexual risk behavior .138 .209 .509 Denial -1.399 .894 .119 AIDS knowledge -.022 .852 .979 AIDS seriousness (concem/threat) .125 .128 .330 AIDS seriousness (treatability/curability) -.303 .105 .004 Re-analyses of Equation 4 Each of the six re-analyzed models also reached statistical significance (see Table 19). However, in the first five models, the sexual risk behavior subscores were not significantly associated with perceived vulnerability to HTV infection. It is only in the sixth model, which used the occurrence of anal sex as the measure of risk, that sexual risk behavior (b=0.3732, SE=0.1131, p=0.0011) emerged as a significant predictor of perceived vulnerability. The observed pattern of findings show that the sexual risk behavior measure is only particularly strong as a predictor when it contains 109 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the anal sex subscore, which was significant by itself as a subscore and which carries most of the predictive weight in the composite risk score. Table 19. Results from re-analyses of regression model #4, contrasting various risk behavior subscores as the measure of sexual risk behavior (cross-sectional analytic sample, N=306). Regression models b(SE) P R- F P Equation 4: Model regressing perceived vulnerability to HIV infection on theoretical variables, where the sexual risk score is comprised of... Original mean score with all 10 sexual risk variables .138(.209) .5095 .086 2.760 .0029 1 . Mean score excluding sexually active status and no. partners in past month variables .258(.193) .1813 .089 2.908 .0017 2. Number of sex parmers in the past month -.072(.119) .5444 .079 2.476 .0074 3. Mean score pertaining to use of condoms for vaginal sex with steady/casual partners ■ 149(.l28) .2425 .088 2.800 .0025 4. Mean score pertaining to use of condoms for anal sex with steady/casual partners -.028(.132) .8341 .083 2.559 .0056 5. Mean score pertaining occurrence of vaginal sex with steady/casual partners -029(.109) .7918 .080 2.271 .0146 6. Mean score pertaining occurrence of anal sex with steady/casual partners .373(.ll3) .0011 .112 3.359 .0004 Strangely, one of the other variables which was found to be a significant predictor variable in the analyses of Equation 4— seriousness pertaining to AIDS treatability/curability— was inversely associated with perceived vulnerability; that is, those who perceived HIV to be relatively treatable and curable (a low score on the seriousness measure) were more likely to have stronger perceptions of vulnerability to HIV infection in their future. This finding contradicts the hypothesized manner in 110 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. which these two variables were thought to be associated, and has no intuitively apparent explanation as of yet. Tests of the Mediating and Moderating Models Because none of the three perceived vulnerability variables were significant predictors of intentions for risk reduction in either of the two models tested, it was not likely that they could be significant mediating variables for the relationships between the other independent variables and the main dependent measure (intentions). Therefore, tests of mediation were not performed. With respect to moderating effects, interaction terms tested by making eight new variables, each of which was the cross-product of denial with one of the cognitive or behavioral variables. For example, the variable which examined the interaction between denial and patterns of risky sexual behavior was created using the cross-product of denial and the mean score of the sexual behavior risk index. Each interaction term was entered individually into four separate regression models which regressed (respectively) each of the three vulnerability variables (equations 2-4 in the regression table) and the intentions variables (equations lA and IB in the regression table) on the main independent measures. Thus, in essence, each of the interaction terms was tested five times. Of all the interaction terms tested, none emerged as significant predictors of the vulnerability variables or either of the two intentions variables. The only interaction to have even a minimal effect was the interaction of denial % prior STD history, which was 1 1 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. only marginally significant (b=1.917, SE=1.126,p=0.089) when included in the model regressing intentions for risk reduction on 1) prior STD history, 2) the mean score o f the sexual risk index, 3) denial, 4) STD knowledge, 5) seriousness pertaining to STD concern/threat, and 6) seriousness pertaining to STD treatability/curability. Longitudinal Analyses Two multiple regression analyses were used to test the longitudinal model which regressed Time 2 sexual behavior risk on the main study variables, the perceived vulnerability variables, and intentions for risk reduction behavior (condom use and reduction in number of sex partners). The analyses were conducted using the subset of study participants who were able to be contacted and included in the follow-up (N=126). Sexual risk behavior at Time 2 was operationalized using an overall mean risk score based on the participants' self-reported (at Time 2) sexual behaviors. Like the cross-sectional analyses of the theoretical model, each regression model was specific to either STDs or HIV/AlDS, and each model was specific to either of the two intentions variables (condom use or reduction in number of partners), thus creating a total of four separate models. Each of the four models was run twice, with the first run controlling for sexual behavior at the intake assessment and the second run not controlling for sexual behavior; because the results of these two runs were not appreciably different, the results from the regressions which controlled for sexual behavior at intake are presented. These analyses also controlled for demographic factors by including them in the model as predictor variables. 112 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. As shown in Table 20, the first model (Equation I Aa) was specific to STD infection and intentions for condom use; this model was found to be marginally significant (F=l .710, p=0.0809) and accounted for approximately 21% of the variance in sexual risk behavior at follow-up (R^O.2104). Of the variables included in the model, only intentions for future condom use was a significant predictor, with lower intentions predicting higher sexual risk scores at follow-up (b=-0.1763, SE=0.0757, p=0.0224). Table 20. Results from multiple regression analyses predicting sexual risk behavior at follow-up, STD- specific models (longitudinal sample; N=126). Regression models b SE P Equation tA(a): Predictors of sexual risk behavior at Time 2 assessment, model specific to STD-related variables and intentions for future condom use (R- = 0.2104) Ethnicity -.233 .234 .322 Gender .246 .167 .144 Age .000 .009 .999 Education .056 .102 .582 Prior STD history .264 .163 .110 Sexual risk behavior .122 .169 .470 Denial .917 .822 .267 STD knowledge .589 .557 .294 STD seriousness (concern/threat) .071 .121 .560 STD seriousness (treatability/curability) .099 .096 .304 (Table 20 continued on next page) 113 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 20, continued) Regression models b SE P Perceived vulnerability to future STD infection/negative consequences .037 .056 .503 Intentions for future condom use -.176 .076 .022 Equation I A(b): Predictors of sexual risk behavior at Time 2 assessment, model specific to STD-related variables and intentions for future reduction in number of sexual partners (R * = 0.4063) Ethnicity -.566 .508 .277 Gender .535 .417 .213 Age -.012 .019 .519 Education .125 .252 .625 Prior STD history -.543 .469 .259 Sexual risk behavior 1.041 .440 .027 Denial 2.701 1.926 .174 STD knowledge 3.040 1.423 .044 STD seriousness (concern/threat) -.279 .322 .394 STD seriousness (treatability/curability) .351 .213 .113 Perceived vulnerability to fumre STD infection/negative consequences .073 .134 .592 Intentions for reduction in number of sex parmers -.321 .162 .059 The second of the STD specific models (Equation 1 Ab, also shown in Table 20) included the variable pertaining to intentions for reduction in the number of sexual parmers. This model was not statistically significant (F=1.311,/7=0.2775). In this model, several of the independent variables included in the regression equation emerged 114 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. as significant, with the strongest predictor being prior sexual risk behavior (b= 1.0410, SE=0.440l,p= 0.0268). STD knowledge (b=3.0396, SE=1.4231,/ t =0.0436) was also statistically significant, and intentions for reduction in number of sex partners was marginally significant (b=-0.3214, SE=0.1619,p=0.0592). The third model (Equation IBa), which was specific to HIV/AIDS and included the variable pertaining to future condom use, was significant (F=l .946, p=0.04l4) and accounted for approximately 23% of the variance (R^O.2327). These data are presented in Table 21. As in the STD-specific model, intentions for future condom use emerged as the only significant predictor (b=-0.2029, SE=0.0753, p=0.0087) of risk behavior at follow-up. No other variables approached significance. Table 21. Results from multiple regression analyses predicting sexual risk behavior at follow-up, HIV- specific models (longitudinal sample; N=126 ) Regression models b SE P Equation I B(a): Predictors of sexual risk beiiavior at Time 2 assessment, model specific to HIV-related variables and intentions for future condom use ((?- = 0.2327) Ethnicity -.192 .241 .427 Gender .176 .167 .296 Age .001 .009 .897 Education -.000 .102 .997 Prior STD history .226 .166 .175 Sexual risk behavior .243 (Table 21 continued on next page) .172 .161 115 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 21, continued) Regression models b SE P Denial .469 .697 .503 AIDS knowledge .924 .943 .330 AIDS seriousness (concern/threat) .091 .125 .467 AIDS seriousness (treatability/curability) .149 .095 .122 Perceived vulnerability to HIV infection .054 .066 .417 Intentions for future condom use -.203 .075 .009 Equation lB(b): Predictors of sexual risk behavior at Time 2 assessment, model specific to HIV-related variables and intentions for future reduction in number of sex partners {R- = 0.3336) Ethnicity -.483 .537 .377 Gender .259 .403 .527 Age .009 .019 .666 Education -.146 .256 .573 Prior STD history -.294 .403 .472 Sexual risk behavior .927 .426 .039 Denial 1.521 1.574 .343 AIDS knowledge -.701 2.587 .788 AIDS seriousness (concern/threat) .115 .306 .711 AIDS seriousness (treatability/curability) .419 .239 .091 Perceived vulnerability to HIV infection .086 .158 .593 Intentions for reduction in numbers of sex partners -.247 .163 .142 116 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The fourth model (Equation IBb, Table 21) was specific to HIV and included the variable pertaining to intentions for reduction in number of sex partners. This model was not statistically significant. In this model, the only significant predictor variable was risk behavior at Time 1 (b=0.9267, SE=0.4256,p=0.0395). Additional Analyses of Longitudinal Data Matching intentions variables with specific Time 2 behaviors In order to more closely examine the relationships between the two intentions variables and Time 2 sexual risk behavior, we performed more extensive analyses on the longitudinal data. However, instead of using an overall Time 2 sexual risk score as the independent variable, sexual behavior subscores were used as the outcome measures of interest. These subscores were: (1) self-identifying as sexually active at follow-up, (2) number of sexual partners in the prior month (the time between intake and follow-up assessments), and (3) frequency of condom use for vaginal sex for steady and casual partners. Subscores pertaining to anal sex were not analyzed due to the extremely low frequency of anal sex in the follow-up cohort; also, subscores were created only for those on whom there were sufficient data pertaining to Time 2 sexual behavior. Further, intentions variables were "matched" to the specific sexual behaviors to which they most intuitively corresponded. For example, intentions for reduction in number of sexual partners was "matched" in a regression model with the sexual risk subscore pertaining to number of partners in the prior month; similarly, intentions for future condom use was "matched" in a regression model with the sexual risk subscore pertaining to frequency 117 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of condom use with a steady partner. As in the prior regression analyses, each of the models was nm twice, once in which the model was specific to STD infection and once in which the model was specific to HIV infection. Additionally, all other demographic and theoretical variables’ were included in the models as predictors. The results of these analyses are shown in Table 22. The first regression model matched intentions for reduction in the number of sex partners with the sexual behavior subscore pertaining to individuals' reporting being sexually active at the Time 2 assessment; Model I a was specific to STD infection and Model lb was specific to HIV infection. In both models, intentions for reduction in the number of sex partners were not found to be a significant predictor of self-reported sexually active status at Time 2. Of the other demographic and theoretical variables included in the regression models, the only significant predictors of Time 2 sexual activity were ethnicity and STD seriousness, which were found to be significant in the STD-specific model only. - In order to reduce collinearity, the two STD-specific perceived vulnerability variables (to future infection and to negative consequences from future STD infection) were combined to form one perceived vulnerability variable specific to STD infection. The results of regression analyses using this variable were not appreciably different from analyses which utilized the two distinct perceived vulnerability variables. Thus, for the sake of simplicity, the combination variable was included with the other theoretical model variables in the regression analyses which follow. 118 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 22. Regression analyses of longitudinal data in sexual risk behaviors at Time 2 are corresponding intentions variables. matched with Regression models b SE P Model la: Predictors of reporting sexually active status at Time 2 assessment, model specific to STD-related variables and matched with intentions for reduction in numbers of sex partners = 0.4104) Ethnicity -1.272 .586 .041 Gender ~ .108 .482 .825 Age .019 .021 .296 Education .316 .291 .289 Prior STD history .565 .543 .309 Sexual risk behavior .262 .508 .611 Denial 1.662 2.224 .462 STD knowledge 1.328 1.643 .427 STD seriousness (concern/threat) .021 .372 .955 STD seriousness (treatability/curability) .558 .246 .033 Perceived vulnerability to future STD infection/negative consequences .207 .155 .196 Intentions for reduction in number of sex parmers -.267 .187 .166 Model lb: Predictors of reporting sexually active status at Time 2 assessment, model specific to HIV-related variables and matched with intentions for reduction in numbers of sex parmers {R} = 0.1729) Ethnicity -.856 .685 .223 Gender .071 .514 .892 Age .019 .025 .442 Education .116 .326 .725 (Table 22 continued on next page) 119 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 22, continued) Regression models b SE P Prior STD history .411 .514 .432 Sexual risk behavior .277 .543 .615 Denial .032 2.007 .988 HIV knowledge 1.206 3.299 .718 HIV seriousness (concern/threat) -.010 .390 .979 HIV seriousness (treatability/curability) -.007 .304 .979 Perceived vulnerability to HIV infection .069 .202 .734 Intentions for reduction in number of sex partners -.187 .208 .375 Model 2a; Predictors of number of sex partners reported at Time 2 assessment, model specific to STD-related variables and matched with intentions for reduction in numbers of sex partners (R^ = 0.6933) Ethnicity 2.187 .423 .000 Gender .352 .347 .321 Age -.018 .015 .243 Education -.199 .210 .350 Prior STD history -1.001 .391 .017 Sexual risk behavior .231 .366 .535 Denial -.672 1.603 .679 STD knowledge -.532 1.184 .657 STD seriousness (concenVthreat) -.776 .268 .008 STD seriousness (treatability/curability) -.348 .177 .062 (Table 22 continued on next page) 120 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 22, continued) Regression models b SE P Perceived vulnerability to future STD infection/negative consequences -.307 .112 .012 Intentions for reduction in numbers of sex partners -.027 .135 .843 Model 2b: Predictors of number of sex partners reported at Time 2 assessment, model specific to HIV-relatetMariables and matched with intentions for reduction in number of sex partners (/?’ = 0.3872) Ethnicity 1.819 .611 .007 Gender .132 .459 .776 Age -.025 .022 .277 Education .072 .291 .805 Prior STD history -.324 .459 .487 Sexual risk behavior -.539 .484 .277 Denial 1.726 1.791 .345 AIDS knowledge 2.222 2.945 .458 AIDS seriousness (concern/threat) -.684 .348 .061 AIDS seriousness (treatability/curability) -.328 .272 .239 Perceived vulnerability to HIV infection -.019 .179 .915 Intentions for reduction in numbers of sex partners -.009 .185 .960 Model 3 a: Predictors of frequency of condom use during vaginal sex with steady partner, model specific to STD-related variables and matched with intentions for future condom use = 0.2711) Ethnicity -1.053 .402 .Oil Gender .276 .287 .339 (Table 22 continued on next page) 121 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 22, continued) Regression models b SE P Age .010 .016 .545 Education .333 .176 .062 Prior STD history .457 .282 .108 Sexual risk behavior .306 .290 .296 Denial -1.355 1.415 .341 STD knowledge .919 .959 .341 STD seriousness (concern/threat) .096 .209 .647 STD seriousness (treatability/curability) .175 .166 .295 Perceived vulnerability to future STD infection/negative consequences .092 .096 .339 Intentions for future condom use -.371 .130 .006 Model 3b: Predictors of frequency of condom use during vaginal sex with steady partner, model specific to HIV-related variables and matched with intentions for future condom use (R ^ = 0.3245) Ethnicity -1.146 .405 .006 Gender .106 .282 .707 Age .014 .016 .364 Education .274 .172 .115 Prior STD history .263 .278 .348 Sexual risk behavior .559 .289 .056 Denial -2.261 1.171 .057 HIV knowledge .635 1.587 .690 HIV seriousness (concern/threat) .188 .211 .375 (Table 22 continued on next page) 122 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 22, continued) Regression models b SE p HIV seriousness (treatability/curability) .148 .161 .362 Perceived vulnerability to M T V infection .157 .112 .164 Intentions for future condom use -.409 .127 .002 Model 4a: Predictors of frequency of condom use during vaginal sex with new parmers, model specific to STD-related variables and matched with intentions for future condom use (R!‘ = 0.1196) Ethnicity .322 .602 .595 Gender .167 .429 .281 Age -.001 .024 .955 Education .215 .263 .416 Prior STD history .437 .422 .303 Sexual risk behavior .282 .435 .519 Denial 2.041 2.118 .338 STD knowledge 1.269 1.436 .379 STD seriousness (concern/threat) -.219 .312 .486 STD seriousness (treatability/curability) .222 .248 .374 Perceived vulnerability to future STD .064 .143 .655 infection/negative consequences Intentions for future condom use -.263 .195 .182 (Table 22 continued on next page) 123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 22, continued) Regression models b SE P Model 4b: Predictors of frequency of condom use during vaginal sex with new partners, model specific to HIV-related variables and matched with intentions for future condom use (/?’ = 0.1408) Ethnicity .525 .621 .401 Gender .428 .433 .326 Age .002 .024 .948 Education .110 .264 .678 Prior STD history .491 .428 .255 Sexual risk behavior .444 .444 .320 Denial 1.468 1.799 .417 AIDS knowledge 1.471 2.437 .548 AIDS seriousness (concern/threat) -.232 .324 .475 AIDS seriousness (treatability/curability) .428 .247 .087 Perceived vulnerability to HIV infection .057 .172 .739 Intentions for future condom use -.301 .194 .126 The second regression model to be examined matched the same dependent variable (intentions for reduction in number of sex partners) with the number of sexual partners reported at the Time 2 assessment. As in the first set of regressions, intentions for reduction in number of sexual partners was not a significant predictor of number of sexual partners self-reported at Time 2 in either regression model. However, other theoretical variables were found to be significant. For example, in the STD-specific 124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. model (Model 2a), ethnicity was positively associated with number of sex partners reported at Time 2, whereas prior STD history, both seriousness variables (concern/threat and treatability/curability, respectively), and the perceived vulnerability variable were negatively associated with self-reported number of sex partners. In the HIV-specific model (Model 2b), ethnicity was also found to be positively associated with number of sexual partners, whereas only one other variable— AIDS seriousness (concern/threat)— was found to be negatively (and marginally) associated with number of parmers. The third regression model matched frequency of condom use during vaginal sex with a steady partner (as the dependent variable) with intentions for future condom use. In both the STD-specific model (Model 3a) and the HIV-specific model (Model 3b), intentions for future condom use was a highly significant predictor of frequency of condom use with steady partners. Interestingly, this relationship was negative, meaning that greater intentions to use condoms were associated with less actual (self-reported) condom use with steady partners at Time 2. This finding may reflect the fact that a substantial percentage of the follow-up sample reported being sexually active at Time 2 with only one parmer; given the nature of the individuals' relationship with their steady partner (e.g., spouse, intimate lover), the use of condoms may not have been believed to be necessary or desirable, despite the participants' stated intentions at Time 1. Other theoretical variables were also found to be significant predictors of condom use. In the STD-specific model (Model 3a), for example, ethnicity was negatively associated with 125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. condom use, while educational level was positively associated. In the HIV-specific model (Model 3b), both ethnicity and denial were negatively associated with condom use, while prior sexual risk behavior (that is, the sexual risk index at Time 1) was positively associated with condom use. The final regression model tested matched a similar dependent variable, firequency of condom use during vaginal sex with a new (or casual) partner with intentions for future condom use. In both the STD-specific model (Model 4a) and the HIV-specific model (Model 4b), intentions for fiiture condom use did not predict condom use with new sexual partners. Further, none of the other theoretical or demographic variables included in either model were significant. The influence of STD diagnosis on Time 2 sexual behavior Further analyses of the longitudinal data were conducted to examine the influence of STD diagnosis— that is, whether or not individuals were actually diagnosed with a sexually transmitted infection during their clinic visits— on subsequent sexual risk behavior, assessed at Time 2. Multiple regression models were used to examine both main effects of STD diagnosis on Time 2 sexual risk, as well as potential interaction effects of STD diagnosis with other theoretical model variables (e.g., STD diagnosis x gender, STD diagnosis x denial, etc.). The outcome variable in these regression models- -sexual risk at follow-up— was operationalized using an overall mean risk score, as well as subscores pertaining to (1) the occurrence of sexual activity in the prior month (i.e., the time between intake and follow-up assessments), (2) the fi-equency of condom use 126 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. with steady and casual partners, and (3) the occurrence of vaginal sex with steady and casual partners. Subscores pertaining to anal sex were not included due to the extremely low frequency of anal sex in the follow-up cohort. Regression analyses indicated no significant main effects for STD diagnosis on Time 2 sexual risk behavior; analyses utilizing the Time 2 sexual risk subscores yielded the same result. Tests of the STD diagnosis x demographic variable interactions on Time 2 sexual risk (overall score and risk subscores) also showed no significant effects, as did the tests of the STD diagnosis x theoretical model variable interactions. IX. RESULTS. PART ni: ANCILLARY ANALYSES OF THE STUDY VARIABLES In an attempt to further explore the data, additional analyses were performed. More specifically, these analyses serve to further elucidate and expand upon some of the results obtained through the multiple regression analyses. Further Exploration of Variable Interactions As presented earlier, analyses were done to test the interaction of denial and the theoretical model variables on the perceived vulnerability variables. However, we have not yet examined the degree to which denial may interact with the perceived vulnerability variables to influence other outcomes of interest, namely intentions for risk reduction and subsequent sexual risk behavior at Time 2. Similarly, we have not yet examined the degree to which demographic variables interact with theoretical model variables to influence perceived vulnerability, intentions for risk reduction, and 127 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. subsequent risk behavior. Therefore, using the same procedure as before with regard to the creation o f interaction terms and the modeling of the terms in the regression equations, we examined these potential interaction relationships using multiple regression analyses. Interaction of denial and perceived vulnerability For these analyses, three interaction terms were created; (1) denial x perceived vulnerability to future STD infection, (2) denial x perceived vulnerability to negative consequences from future STD infection, and (3) denial x perceived vulnerability to HIV infection.^ Each of these three interactions were included in regression models which (respectively) examined the predictors of three different dependent measures: (1) intentions for future condom use, (2) intentions for reduction in number o f sex partners, and (3) sexual risk behavior assessed at Time 2. Thus, there were nine different regression models which were ultimately tested. Each of the models included the demographic and other theoretical variables in the regression equation, and each analysis controlled for main effects. In the first series of analyses, intentions for future condom use was regressed upon denial x P V sro, denial x PVpRos, denial x PVpgv. In each of these analyses, the interactions of denial x the perceived vulnerability variables were not significant. These non-significant results were also seen in the second series of analyses, in which ^ For the sake of simplicity in reporting (and ease in reading) the results of these analyses, the perceived vulnerability variables will be abbreviated as follows: perceived vulnerability to future STD infection = PVym , perceived vulnerability to negative consequences from STD infection = P V pR og , perceived vulnerability to HIV infection = PV„,v • 128 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. intentions for reduction in number of sex partners was the dependent measure. Therefore, based on these findings, it can be concluded that there are no significant interaction effects for denial and any of the three perceived vulnerability variables on intentions for risk reduction. The third series of analyses examined the interaction effects of denial x the perceived vulnerability variables on Time 2 sexual risk behavior. For these analyses, sexual risk at Time 2 was examined in two ways. First, the models were run with an overall mean sexual risk score (the same as was used in the longitudinal analyses) as the dependent measure. Then, the models were resubmitted using individual sexual behavior subscores— such as frequency of condom use during vaginal sex with a steady partner— as the different respective dependent measures. Thus, Time 2 sexual behavior was analyzed both generally (using the overall score) and specifically (using the individual items). In the analyses using the overall Time 2 sexual risk score, none of the denial x perceived vulnerability interactions were significant. Further, none of the interactions were significant in the regression models utilizing the individual sexual risk behaviors as the dependent measures. Interactions of demographic and theoretical variables Additional analyses were performed in order to test the degree to which demographic and theoretical model variables interact to influence the perceived vulnerability variables, intentions for risk reduction behavior, and actual sexual risk 129 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. behavior at Time 2. In each o f the following analyses, interaction terms were modeled in the same manner as was described above. Each of the cognitive (STD concern/threat, STD treatability/curability, HTV concern/threat, HIV treatability/curability, STD knowledge, HIV knowledge), motivational (denial), and behavioral variables (previous STD history, pattern of risky sexual behavior) were crossed with each of the four main demographic variables (education, gender, age, and ethnicity), thus creating a total o f 36 interaction models to be tested against six dependent measures (the two intentions variables, the three perceived vulnerability variables, and Time 2 sexual risk behavior). As before, each interaction term was tested using multiple regression analyses, and the term was entered into its respective regression equation after controlling for main effects. Each model was specific to either STDs or HTV. Data presented below (and in the tables) are the unstandardized regression coefficients and standard errors. Regressions on Time 1 outcome variables Of the interaction models tested, only five reached significance. Two of these five significant models pertained to the interaction of gender x prior sexual risk behavior (measured by the standardized mean sexual risk score). In the STD-specific model in which intentions for condom use was the dependent variable, the gender x sexual risk behavior interaction was significant (b=-0.576, SE=0.257,p=0.026), suggesting that men who have lower mean levels of sexual risk behavior had significantly stronger intentions to use condoms in the future. Other variables which were significant in this model were age, prior STD history, and STD seriousness pertaining to concern/threat. 130 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Similarly, the gender x sexual risk behavior interaction was a significant predictor of intentions for future condom use in the HTV-specific model (b==-0.555, SE=0.259, p=0.033)\ other significant variables in this model were age, prior STD history, and AIDS knowledge. These results are shown in Table 23. Table 23. Regression models in which interactions pertaining to gender x were significant (N=306). prior sexual risk behavior Regression models b SE P Model la: Predictors of intentions for future condom use. model specific to STD-related variables = 0.1334,F=3.200,p=0.0001) Ethnicity .057 .178 .746 Gender -.023 .129 .858 Age -.019 .007 .005 Education .015 .077 .846 Prior STD history .323 .127 .011 Sexual risk behavior .712 .694 .306 Denial -.499 .599 .405 STD knowledge .404 .456 .377 STD seriousness (concern/threat) .245 .084 .004 STD seriousness (treatability/curability) .067 .074 .369 Gender x sexual risk behavior -.576 .257 .026 Model 1 b: Predictors of intentions for future condom use. model specific to HIV-related variables (R^ = 0.1192, F=2.814, p=0.0006) Ethnicity .193 .179 .284 Gender -.039 .131 .765 (Table 23 continued on next page) 131 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 23, continued) Regression models b SE P Age -.022 .007 .002 Education .003 .078 .964 Prior STD history .330 .128 .010 Sexual risk behavior — .517 .695 .457 Denial -.548 .601 .362 AIDS knowledge 1.018 .569 .075 AIDS seriousness (concern/threat) .095 .087 .273 AIDS seriousness (treatability/curability) -.042 .070 .550 Gender x sexual risk behavior -.555 .259 .033 Two of the three remaining significant interactions pertained to STD-specific regression models and interaction terms which included the perceived seriousness variables. These data are shown in Table 24. The first of these models regressed perceived vulnerability to future STD infection on the theoretical model variables and the gender x STD seriousness (treatability/curability) interaction term. In this equation, gender x STD seriousness (treatability/curability) was positively and significantly associated with perceived vulnerability, which suggests that men who have stronger perceptions o f seriousness pertaining to the treatment of STD infections are more likely to have greater perceptions of vulnerability to future STD infections (b=0.513, SE=0.252, /7=0.043). In the second regression model, both the main effect of perceived 132 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. seriousness (concern/threat) and the interaction of ethnicity x STD seriousness (concem/threat) were significant predictors of perceived vulnerability to negative consequences from future STD infection. These findings suggest that, while stronger concerns about the seriousness of STD infection are positively associated with perceived vulnerability to future negative consequences (b=1.556, SE=0.727,p=0.033), these perceptions of STD seriousness and vulnerability to future negative consequences may be stronger among Latinos than African-Americans (b=-0.728, SE=0.361, /?=0.045). There were no other significant variables in either of these two models. Table 24. Regression models in which interactions pertaining to perceived seriousness variables were significant (N=306). Regression models b SE p Model la: Predictors of perceived vulnerability to future STD infection, model specific to STD-related variables (/?’ = 0.0378, F=0.816, p=0.6519) Ethnicity .116 .860 .893 Gender -1.227 .648 .059 Age .004 .034 .917 Education -.224 .366 .540 Prior STD history .155 .209 .458 Sexual risk behavior .075 .213 .727 Denial -.899 .985 .362 STD knowledge -.197 .757 .795 STD seriousness (concem/threat) .021 .137 .877 (Table 24 continued on next page) 133 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 24, continued) Regression models b SE P STD seriousness (treatability/curability) -.946 .694 .174 Gender x sexual risk behavior .514 252 .043 Model lb: Predictors of perceived vulnerability to negative consequences from future STI^infection, model specific to STD-related variables (/?’ = 0.0719, F=1.6I0,/7=0.0756) Ethnicity 2.252 1.436 .118 Gender .867 .988 .381 Age -.015 .047 .755 Education -.254 .619 .681 Prior STD history .293 .200 .144 Sexual risk behavior -.008 .206 .967 Denial -1.037 .955 .279 STD knowledge -.237 .732 .746 STD seriousness (concem/threat) 1.556 .728 .033 STD seriousness (treatability/curability) -.116 .118 .324 Ethnicity x STD seriousness (treatability/curability) -.728 .361 .045 The final interaction model regressed intentions for condom use on the theoretical variables and the gender x denial interaction term (STD-specific model) These data are shown in Table 25. In this model, the interaction of gender x denial was significantly and positively associated with intentions for condom use (b=2.614, SE=1.256, /7=0.038). Although this finding suggests that men with higher levels of 134 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. denial are more likely to have stronger intentions to use condoms in the future, the result is most likely spurious due to the great deal of error in the measure of denial, as seen by examining the standard error of the main effect for denial (b—1.344, SE=2.938, /7=0.648). It is most likely that the significant main effect for gender is the prominent contributing factor explaining why the gender x denial interaction is significant. Additional significant main effects were found for prior STD history, prior sexual risk behavior, and STD seriousness (concern/threat). Table 25. Regression model in which the interaction pertaining to gender x denial was significant. (N=306). Regression models b SE P Predictors of intentions for condom use, model specific to STD-related variables (R- = 0.1311,F=3.136,p=0.0001) Ethnicity -.106 .908 .907 Gender -1.345 .667 .045 Age .013 .032 .662 Education .384 .412 .352 Prior STD history .286 .127 .025 Sexual risk behavior -.459 .129 .001 Denial -1.344 2.938 .648 STD knowledge .535 .469 .256 STD seriousness (concern/threat) .239 .083 .005 STD seriousness (treatability/curability) .053 .074 .475 Gender x denial 2.614 1.256 .038 135 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In addition to these significant findings, there were three regression equations in which the interaction terms were marginally significant. In the first equation, the interaction of ethnicity x AIDS seriousness (concern/threat) was found to be marginally and inversely associated with perceived vulnerability to HTV infection, suggesting that Latinos who experience greater perceptions of concern or threat regarding AIDS may be more likely than their African-American counterparts to feel vulnerable to future HIV infection (b=-0.572, SE=0.309, p=0.065). In the second regression model, in which perceived vulnerability to HIV infection was the outcome of interest, the interaction of education x prior sexual risk behavior was marginally but positively associated with the dependent measure (b=0.445, SE=0.235,/ t =0.059), suggesting that individuals with higher levels of both education and sexual risk behavior may be more likely to feel vulnerable to HIV infection in the future. The final interaction, denial x gender, was marginally associated with intentions for future condom use (b=2.407, SE=1.257, p=0.056) in the HIV-specific regression equation. Like the significant interaction between denial x gender described earlier, this finding is most likely spurious due to the high amount of error in the measure of denial (in this model: b=-l .655, SE=2.948, p=0.575). Regressions on Time 2 sexual behaviors Regression analyses were performed in order to examine the effects of variable interactions on Time 2 sexual risk behavior. In these analyses, both the overall mean sexual risk behavior score and sexual risk subscores (pertaining to specific sexual 136 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. behaviors) were utilized as outcome measures. As with the previous regression analyses pertaining to Time 2 behavior, subscores pertaining to anal sex were not analyzed due to the extremely low frequency of anal sex in the follow-up cohort. Additionally, the three significant interactions pertaining to denial are not reported because, due to the extreme amount o f error in the denial measure, the findings are likely to be spurious and their interpretations meaningless. Therefore, with the exception of these three denial-related interactions, all other significant findings are reported below. In total, there were seven significant interactions. The first two interactions which were found to be significant pertained to the interaction of age x STD history in the regression models in which the occurrence of vaginal sex at Time 2 was the outcome of interest. This interaction was significant in both the STD-specific model (b=-0.059, SE=0.297, p=0.050) and the AIDS-specific model (b=-0.062, SE=0.029,/7=0.039), suggesting that younger individuals who have had prior STDs are more likely to report having had vaginal sex in the time following their clinic visit. These data are shown in Table 26. Table 26. Regression models pertaining to Time 2 behavior, in which the age x STD history interactions were significant (N=126). Regression models b SE p Model la: Predictors of the occurrence of vaginal sex reported at Time 2, model specific to STD-related variables = 0.2002, F= 1.376, p=0.1854) Ethnicity -1.131 1.087 .302 (Table 26 continued on next page) 137 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 26, continued) Regression models b SE P Gender .071 .855 .934 Age .102 .051 .047 Education .645 .598 .283 Prior STD history 1.761 1.425 .220 Sexual risk behavior .213 .252 .401 Denial 1.102 .128 .393 STD knowledge .805 .863 .354 STD seriousness (concern/threat) .025 .187 .894 STD seriousness (treatability/curability) .182 .147 .221 Age X prior STD history -.059 .029 .050 Model lb; Predictors of the occurrence of vaginal sex reported at Time 2, model specific to AIDS-related variables (/?’ = 0.2147, F= 1.503, p=0.1300) Ethnicity -.965 1.130 .396 Gender -.007 .849 .994 Age .112 .050 .028 Education .594 .593 .320 Prior STD history 1.994 1.422 .165 Sexual risk behavior .399 .255 .122 Denial .589 1.100 .594 AIDS knowledge .828 1.444 .568 AIDS seriousness (concern/threat) .061 .201 .760 AIDS seriousness (treatability/curability) .272 .146 .065 Age X prior STD history -.062 .029 .039 138 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There were three significant interactions in regression equations in which the outcome measure was frequency o f condom use during vaginal sex with a steady partner. These data are shown in Table 27. The first of these significant interaction terms (in an STD-specific model) was the interaction of education x STD seriousness (concern/threat), which was negatively associated with frequency of condom use (b=- 0.470, SE=0.206, /?=O.O20, suggesting that individuals who had higher educational levels but lower levels of concern over the seriousness of STDs reported higher frequencies of condom use during vaginal sex with their steady partners. Similarly, in an AIDS-specific model, the interactions of AIDS seriousness (treatability/curability) x age (b=-0.038, SE=0.018,/?=0.036) and AIDS seriousness (treatability/curability) x ethnicity (b=0.968, SE=0.421,p=0.024) were significant predictors of condom use. Table 27. Regression models in which interactions pertaining to perceived seriousness variables were significant predictors of frequency of condom use during vaginal sex with steady partners (N=126). Regression models b SE p Model la: Predictors of frequency of condom use during vaginal sex, model specific to STD-related variables (/?’ = 0.2827. F=2.168, /7=0.0166) Ethnicity -.950 1.961 .629 Gender .501 1.557 .749 Age -.026 .077 .732 Education 2.115 .835 .013 Prior STD history .492 .287 .091 Sexual risk behavior .718 (Table 27 continued on next page) .285 .014 139 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 27, continued) Regression models b SE P Denial -1.126 1.488 .452 STD knowledge .902 1.013 .376 STD seriousness (concern/threat) 1.010 1.031 .331 STD seriousness (treatability/curability) .177 .165 .288 Education x STD seriousness (concern/threat) -.470 .206 .025 Model lb: EVedictors of frequency of condom use during vaginal sex, model specific to AIDS-related variables {R- = 0.2839, F=2.I80, p=0.0159) Ethnicity -5.479 1.834 .004 Gender 2.470 1.777 .169 Age .199 .079 .015 Education 1.202 1.089 .273 Prior STD history .311 .291 .289 Sexual risk behavior .761 .289 .011 Denial -2.223 1.249 .079 AIDS knowledge -1.511 1.671 .369 AIDS seriousness (concern/threat) .115 .227 .614 AIDS seriousness (treatability/curability) .930 1.002 .356 Age X prior STD history -.038 .018 .036 Ethnicity x prior STD history .968 .421 .024 The two remaining significant findings pertained to the interaction of age x AIDS knowledge. In the first regression equation, age x AIDS knowledge was a 140 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. significant predictor of frequency of condom use during vaginal sex with a steady partner b=-0.038, SE=0.018,p=0.036); this regression equation is shown in Table 28. This interaction term was also found to be a significant predictor of a combined frequency of condom use variable, which combined frequency of use during vaginal sex with steady partners and frequency of use with new partners (b=-0.338, SE=0.169, p=0.049), suggesting that^ounger individuals who have a higher level o f AIDS knowledge are more likely to report using condoms more often when having vaginal sex with both steady and new partners. Table 28. Regression model in which the age x AIDS knowledge interaction was a significant predictor of frequency of condom use during vaginal sex with steady partners (N=I26). Regression models b SE P Predictors of frequency of condom use during model specific to AIDS-related variables (R- = vaginal sex, : 0.2967. F=2.320,p=0.0l01) Ethnicity -9.443 4.647 .046 Gender 2.017 3.129 .521 Age .462 .159 .005 Education 3.321 2.368 .165 Prior STD history .395 .290 .177 Sexual risk behavior .905 .292 .003 Denial -2.046 1.236 .102 AIDS knowledge 9.436 8.022 .243 AIDS seriousness (concern/threat) .070 .228 .759 AIDS seriousness (treatability/curability) .050 .189 .792 Age X prior STD history -.473 .175 .008 141 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Examining Curvilinear Distributions of Theoretical Model Variables In the attempt to explain why the perceived vulnerability variables were not significant predictors of intentions for risk reduction or Time 2 sexual behavior, regression analyses examining potential curvilinear effects were performed. In essence, it was hypothesized that if a given variable had a nonlinear distribution, a statistical analysis of the data— such j s multiple regression— that is reliant on linear distributions in order to determine variable relationships would not be able to glean any significant association between nonlinearly and linearly distributed variables. Therefore, in order to examine a nonlinear or curvilinear variable using regression analyses, it is necessary to transform it into a quadratic variable by creating a term which is the square of the nonlinear variable in question. For example, if a perceived vulnerability variable is nonlinear and we wish to transform it, the variable must be multiplied against itself (PV * PV) in order for it to become a quadratic term. The new quadratic term, which has a linear distribution, can then be entered into a regression model and tested with multiple regression analyses. This procedure was employed to test the curvilinear effects of the three perceived vulnerability variables. The new quadratic variables which were created from each of the vulnerability variables were entered into separate multiple regression models as predictors of intentions for condom use, intentions for reduction in number of partners, and Time 2 sexual risk behavior, thus creating a total of nine regression models. Regression equations pertaining to the STD-related perceived vulnerability 142 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. variables contained the other theoretical model variables which were specific to STDs, whereas the equation pertaining to the HIV-specific vulnerability variable contained the other theoretical model variables which were specific to HIV/AIDS. The same procedure was used to create a quadratic term from the continuous denial variable; this new quadratic term was tested in six separate regression analyses as a predictor of 1) perceived vulnerability to_STD infection, 2) perceived vulnerability to negative consequences from future STD infection, 3) perceived vulnerability to HTV infection, 4) intentions for condom use, 5) intentions for reduction in number of sex partners, and 6) Time 2 sexual risk behavior. Each model controlled for demographic variables by entering them into the regression equation as predictors. Similarly, the original (prior to quadratic transformation) vulnerability and denial variables were included in the regression models. Of the nine vulnerability-related regression models tested, none of the quadratic vulnerability variables were found to be significant predictors of intentions for risk reduction or Time 2 sexual risk behavior. Likewise, none of the six denial-related regressions showed the new quadratic term for denial as a significant predictor of perceived vulnerability, intentions for risk reduction, or Time 2 risk behavior. Re analysis using a quadratic term created from the categorical measure of denial yielded similar non-significant results. 143 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. X. DISCUSSION. PART I: SUMMARY OF FINDINGS This study of sexually transmitted disease (STD) clinic attenders in Los Angeles, California, examined the degree to which behavioral (prior STD infection, sexual risk behavior), cognitive (STD and HIV knowledge, perceptions of STD and HIV seriousness), and motivational (denial) variables influence perceptions of vulnerability to future STD infection, :o negative consequences resulting from STD infection, and to HIV infection. Further, we examined the degree to which behavioral, cognitive, and motivational factors and perceptions of vulnerability influenced intentions for risk reduction behavior and actual sexual behavior assessed six weeks following their intake assessment at the STD clinic. We also collected detailed information pertaining to recent sexuai practices with steady partners and casual (non steady) partners, including occurrence )f vaginal and anal sex with these partners and the frequency with which condoms are used for these activities with these specific partners. In general, the majority of the imdy participants were, by their behavioral self- report, able to be defined as sexually active. At intake, slightly over half of the sample reported that they had had only one sex partner in the month prior to their intake assessment and 57% reported having had prior STD infection. The vast majority of the sample (approximately 79%) reported having had vaginal sex with steady partners in the month prior to assessment; of those who practiced this activity, only 30.8% reported consistent condom use, and another 30.5% of respondents indicated that they never used 144 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. condoms for this activity. Anal sex was not as prevalent in this sample, with less than 15% of respondents indicating that they had practiced anal sex with a steady and/or a casual partner. Nonetheless, the frequency of condom use for this activity was higher, with 74% reporting consistent use with steady partners and almost 91% reporting consistent use with casual partners. At follow-up, 38% of the sample reported having been diagnosed with an STD during their clinic visit, and 77% indicated that they had been sexually active in the time since leaving the clinic. More than half of the sample (58.5%) reported having had sex with only one parmer in the time since they had visited the clinic. However, a significant percentage of the follow-up sample (41%) indicated having had three or more sex partners in that same time frame. The majority of the sexual activity which was reported at follow-up was vaginal sex with steady partners. Indeed, no one reported having had anal sex with a casual parmer at the follow-up assessment. Rates of condom use for both vaginal and anal sex with both steady and casual parmers were slightly lower than was observed at intake. With regard to the other theoretical variables, the study participants had relatively high scores on both the STD and the AIDS knowledge scales, and were quite evenly distributed across the repression-sensitization scale (the measure of denial). The sample had somewhat high mean scores for three of the four seriousness variables (STD concern/threat, AIDS concern/threat, and AIDS treatability/curability), and mid-range 145 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. mean scores for the three perceived vulnerability items and for the measure of intentions for condom use. With respect to the analyses of the theoretical models, there were several interesting findings. Of the behavioral, motivational, and cognitive variables hypothesized to influence intentions for condom use, prior STD history, sexual risk behavior, and age were consistent predictors in both of the models with intentions for condom use as the dependent measure. STD concern/threat emerged as a significant predictor in the first model (which was specific for STD infection) and AIDS knowledge was marginally significant in the second model (which was specific to HIV/AIDS). Of the sexual risk subscores tested, the subscores pertaining to condom use for vaginal sex with steady/casual partners and anal sex with steady/casual partners were highly significant, and carried most of the predictive weight in the composite sexual risk mean score. Of the models which tested each of the three perceived vulnerability variables as dependent measures, the pattern of results was somewhat different. None of the study variables emerged as significant predictors of perceived vulnerability to future STD infection or perceived vulnerability to negative consequences arising fi’ om STD infection. When perceived vulnerability to HTV infection was the dependent measure, however, prior STD history and AIDS treatability/curability emerged as significant predictors of vulnerability; age was a marginally significant predictor. Similar patterns of results were observed for the re-analyses of the models using the sexual risk 146 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. subscores, and the comparison of the regression coefficients of the various risk subscores indicated that the variables measuring the occurrence of anal sex with steady/casual parmers carried most of the predictive weight of the composite sexual risk score. Longitudinal analyses with a subset of the sample who were able to be contacted and included in the longitudinal cohort indicated that intentions for condom use was a consistent predictor (in both the STD-specific and HTV/AIDS-specific models) of sexual risk behavior at follow-up. Similarly, in the analyses utilizing sexual risk behavior subscores for Time 2 behavior, intentions for condom use was significantly associated with frequency of condom use during vaginal sex with a steady partner. Further, sexual risk at Time 1 was a significant predictor of risk behavior at follow-up (in an AIDS- specific model which included the intentions for reduction in number of partners variable). There was no observed effect of STD diagnosis on Time 2 behavior. Of the additional interaction terms that were tested, eight were found to be significant predictors of perceived vulnerability, intentions for risk reduction, and Time 2 sexual risk behavior. More specifically, there were five significant findings involving perceived seriousness variables (both concern/threat and treatability/curability) interacting with age, gender, ethnicity, and level of education. Sexual risk behavior at Time 1, prior STD history, and AIDS knowledge also interacted with demographic variables to influence intentions for condom use and risk behavior at follow-up. 147 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Although the results of this study do not confirm the hypotheses pertaining to perceived vulnerability as a predictor of behavioral intentions and subsequent risk behavior, they do lend support to several other theoretical hypotheses, such as the role of past behavior and perceptions of seriousness in the assessment of personal risk for disease and intentions for subsequent behavior. These findings also lead to further questioning about the role of perceived vulnerability as a predictive variable for behavioral intentions and actual behavior. The findings seen here can be examined from one or more perspectives with regard to the conditions which may have contributed to significant or non-significant results, or reasons why variables hypothesized to be significant were not significant (and vice versa). In the following sections, we will discuss the findings from two of these perspectives: methodological aspects which may explain some of the results observed and theoretical interpretations which may help to understand or re-examine the results. XI. DISCUSSION. PART II; METHODOLOGICAL INTERPRETATION OF FINDINGS Several methodological factors may have contributed to the surprising non significant results obtained in this study. Perhaps one of the major factors which may have led to null results is the issue of scale reliability. Although we were able to obtain quite high reliability for several of the theoretical variables (such as the perceived vulnerability variables and intentions for risk reduction), some of the other scales had 148 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. reliability coefficients that were somewhat low. For example, the raw Cronbach's alpha for STD treatability/curability was 0.5572, a low coefficient compared to the alpha obtained for AIDS treatability/curability (a=0.7058). The low reliabilities are likely to have inhibited the predictive power of some of the variables which were anticipated as being significant predictors of both perceived vulnerability, risk reduction intentions, and sexual risk behavior at follow-up. What is still somewhat perplexing is the fact that the alpha we obtained for the measure of denial (a=0.68) was stronger than the reliability coefficient (a=0.65) obtained by Gladis et al. (1992), who was able to obtain significant main effects and moderating effects for denial using the exact same measure. Further work with this scale in other samples may help to determine if the differences in the success of the scale are due to other methodological issues or other factors which are more theoretically based. Another factor which may have contributed to the null findings is the range restriction and skewness observed in some of the theoretical variables. For example, the measures for the STD and HIV/AIDS knowledge items had distributions that were quite skewed because, in general, the sample had a relatively high degree of knowledge in each of these subjects and were able to answer most (if not all) of the items correctly. While it was certainly gratifying to see that the sample was knowledgeable about STDs and HlV/AlDS, the fact that the distributions for these variables were so skewed does not lend these variables as much predictive power as they may have had given a more normal distribution and more variance. 149 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. However, it may not be the case that skewness is entirely to blame for some of the null results obtained here. The AIDS treatability/curability measure had a mean of 4.41 (from a possible range of I to 5) and a skewness of -1.93. Indeed, this skewness is higher than that observed for any other variable. However, despite the skewness, AIDS treatability/curability did emerge as a significant predictor of perceived vulnerability to HIV infection. Therefore, the skewed distribution for this variable does not seem to have been an impediment to its predictive ability in the regression model. The factor which was probably the most detrimental to the predictive ability of several of the theoretical variables was the fact that some of the measures had very high standard errors and, because they were not accurate measures, they may have functioned poorly as predictors in a regression model. An example of this can be seen most clearly in the examination of the denial variable in the four regression equations which were used to test the theoretical models predicting risk reduction intentions and perceived vulnerability (please refer to Table 10). In each of these models, the standard error for denial is over 0.500 and, in the case o f the model predicting perceived vulnerability to future STD infection, the SE is as high as 0.980. Although Gladis and her colleagues (1992) do report regression coefficients for their regression analyses, they do not report the standard errors of those coefficients. Thus, a comparison o f the measure between their study and this one could not be done. Nonetheless, the consistently high standard errors obtained in this study indicate that the measure of denial is fiaught with measurement error and is, therefore, weak; given this weakness, it is not surprising that 150 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. we did not obtain a main effect for denial in any of the regression models. Further, the error in this measure puts into question the reliability of significant findings, such as interactions, which included the denial measure. Future testing and methodological research needs to be done with these instruments in order to determine if they are the most appropriate for the measurement of their respective theoretical constructs. XII. DISCUSSION. PART III: THEORETICAL INTERPRETATfON OF FINDINGS The results obtained in this study are quite interesting in that variables which were thought to be highly influential in terms of their predictive ability were, in actuality, non-significant in multiple regression models. Indeed, it was highly surprising that none of the three perceived vulnerability variables produced a significant effect on either intentions for risk reduction behavior or Time 2 sexual risk behavior. One would have expected the effect to be present, particularly in light of the fact that vulnerability to STD infections may have been more salient in participants' minds due to their presence for treatment at an STD clinic; their desire to avoid such clinic visits in the future would have influenced intentions for risk reduction and subsequent sexual risk behavior. This desire to avoid future infection may have indeed been what was occurring on a cognitive level, as evidenced by the significant association of STD seriousness with risk reduction intentions and Time 2 behavior. What is interesting to note in this case, though, is that the effect of these variables is direct rather than mediated through another cognitive process, namely, the 151 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. assessment of personal risk factors which contributes to the development of perceptions of vulnerability. This finding both supports and contradicts previous research on perceived vulnerability. For example, Weinstein (1980; 1982; 1987) posits that cognitive variables— such as assessments of prior behavior/experience and assessments of perceived seriousness of certain illnesses or diseases— are influential in individuals' assessments of personal risk, i.e., perceived vulnerability to disease. This hypothesis suggests that the assessment of personal vulnerability is necessary for the development of intentions for behavior change and must occur prior to the assessment of what to do about the risk (e.g., intentions to change risk behavior) and prior to any actual behavior following the assessment of risk. In other words, a mediating model is suggested and, although the results from this study support the importance of behavioral and cognitive variables which can contribute to one's risk assessments (e.g., AIDS treatability/curability emerging as a significant predictor of perceived vulnerability to HIV infection), we do not find strong support here for the mediating hypothesis, which suggests that an objective, thorough assessment of vulnerability may not be absolutely necessary for the development of risk reduction intentions and subsequent changes in behavior. Another surprising finding was the lack of effect seen for denial. The effects of denial on perceptions of vulnerability was quite strongly supported by Gladis et al. (1992), and their findings suggested a moderating model hypothesis which denial moderates the effects of other variables which are thought to strongly influence 152 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. perceived vulnerability to H TV infection. In order for this moderation effect to occur, however, a main effect for denial on the dependent measure is desirable; Gladis and her colleagues were able to find a significant main effect for denial in their sample of adolescents. In contrast, our study did not find a significant main effect for denial, despite the fact that the actual measure had adequate reliability and a relatively normal distribution. However, methodological aspects (such as high measurement error) could very well have affected this variable's predictive ability in a regression model. Because we found no main effect, we were not able to find any significant results in the tests of the moderating models and, therefore, could not lend support to the moderating hypotheses put forth by Gladis and her colleagues (1992). One potential explanation of this lack of effect may pertain to the nature of the samples in both Gladis's study (high school adolescents) and this study (STD clinic attenders). These two samples are extremely different on a variety of levels, and these differences may have contributed to the degree to which denial was operating in the assessment of vulnerability to disease. Although our sample of STD clinic attenders contained individuals who were young adults, the sample as a whole was quite a bit older (mean age=32 years old) than Gladis's sample of high school students (mean age=15 years old). Similarly, the social context in which data from the two samples was collected was extremely different: a high school campus as compared to an inner- city STD clinic. The developmental and contextual differences of the two groups of individuals is quite likely to influence the degree to which denial is used as a cognitive 153 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. defense mechanism in the assessment of personal risk. For example, a 15 year old adolescent who is being asked about his vulnerability to HIV infection may have a variety of reasons why he does not feel vulnerable, even if he is sexually active. He, like other adolescents, may have the feelings of indestructibility that are common to that developmental stage of life. Further, the context of this adolescent-that of a high school campus-is one which may not prompt salient thoughts about disease risk, particularly not risk for HIV infection. Even if this adolescent is at risk, his developmental and contextual frame of reference is more likely to prompt a denial reaction for his actual risk, which would then lead to unrealistically lowered assessments of vulnerability. In contrast, a 32-year old individual who is presenting in an inner-city STD clinic is more likely to have the developmental capacity to know that negative things can happen to him; after all, he is in an STD clinic awaiting treatment for a suspected venereal disease. Further, the medically-focused atmosphere of the clinic itself is likely to prompt thoughts about disease and illness, particularly if educational materials (such as pamphlets, posters, and videos) present in the clinic environment are issuing warnings about disease risk to those in the waiting area. It seems less likely that denial could be employed in a setting such as a clinic, in which the concepts of disease and illness are so salient. Therefore, given these circumstances, one would not expect a main effect for denial. 154 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. If this study was implemented in a more context-neutral setting, such as a community center or another social gathering place, we might have foimd a main effect for denial in predicting perceived vulnerability to disease. However, at this time, our findings cannot lend support to the hypotheses posed by Gladis's and her colleagues (1992). Future research in more context-neutral settings (such as community centers or other social gathering places) may be able to determine the degree to which denial functions as a defense mechanism. xni. DISCUSSION. PART IV: DIRECTIONS FOR FUTURE RESEARCH Although the findings from this study did not strongly support the relationship between perceived vulnerability to disease and behavioral intentions/actual behavior, the whole area of perceptions o f vulnerability and the assessment of personal risk needs continued examination, particularly in real-world environments where individuals are dealing with assessments of their health and their personal risk for a variety of events and stressors (which may or may not be disease-related) in the course of their daily life. Our findings pertaining to the relationship between AIDS treatability/curability and perceived vulnerability to HTV infection is of particular relevance for future research, as it suggests that appraisals of seriousness of a disease— and, therefore, appraisals of one's personal risk to that disease-may be dependent on factors pertaining to the nature of the illness itself. Because HIV infection is incurable and eventually fatal, one would think that infection with this virus would be taken more seriously as a health threat when compared to a "traditional" STD infections which, because they are 155 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. usually successfully treated and cured, may not influence perceptions of vulnerability. Although traditional sexually transmitted diseases are indeed treatable and (in most cases) curable, the morbidity associated with STD infection and the increased risk for HIV infection is indeed something to be taken seriously, particularly because it has such a high prevalence in and such detrimental effects on the overall health of minority communities. Additionally, as more successful treatments (such as protease inhibitors) become available for HIV infection, the degree to which this disease is seen as life- threatening may decrease and, therefore, an HIV infection may not be viewed with the same degree of seriousness as it was in the early and mid-80s, when treatments for the disease were scarce. Future research should examine the perceived seriousness of disease as something which has the potential to have a greater impact on clinic attenders' perceptions of personal risk and subsequent sexual behavior, and should examine in greater depth other reasons why disease and illness may have more serious or less serious cognitive associations for individuals. Like all research investigations, this study had several limitations which must be addressed. Although we were able to accrue a sample which was large enough to provide adequate power for most of the analyses performed, there was a substantial percentage o f the individuals approached who declined study participation. Early on in the study, it was determined that the high décliner rates were due to the fact that individuals did not want to permit access to their medical files. This was understandable, given the mistrust that many individuals— particularly communities of 156 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. color-have of health professionals in general and medical researchers in particular, and the section of the protocol which called for abstraction of data from medical charts was dropped. Although the omission of this part o f the study substantially improved (doubled) the previous participation rate and the total sample at the end of ten weeks' data collection was larger than the estimated sample, there were still many who refused participation. Another limitation of the study pertained to the high attrition rate obtained in the attempt to contact study participants for follow-up assessment. Although it is acknowledged that STD clinic samples can, for various reasons, be a challenging population with which to work, it was hoped that the explanation (during subject recruitment) of the project as a longitudinal study— and the willingness of subjects to consent to being re-contacted for a second assessment— would yield a larger longitudinal sample. Anecdotal information shared by researchers who have worked with STD clinic-based samples suggests that the 32% follow-up rate obtained in this study was indeed comparable to rates obtained by other studies (personal communication, Susan Buchbinder, M.D., September 1995). However, more research needs to be done to determine how such samples— which are truly at increased risk of STD re-infection and HIV infection— can be studied longitudinally so that sexual behavior patterns (and the changes in those patterns) over time can be examined more closely in relation to risk behavior. Further, such research needs to include extensive collaboration with STD clinic patients themselves, as they are the best resources for providing information about 157 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ways in which clinic attenders can be motivated to participate and remain in research studies. They can also provide valuable information about the types of services other clinic attenders like themselves may find worthwhile, and suggest ways in which researcher-subject relationships can be more mutually beneficial and equitable. Limitations aside, this study does offer several other relevant contributions to research. Because this study included men and women from a multi-ethnic urban community, the findings from this study are likely to be generalizable to other urban minority samples, particularly those who are seeking medical care in public health clinics. Also, because the sample population was predominantly heterosexual, the results from this study may be particularly generalizable to the study of HTV risk behaviors in heterosexual higher risk samples. This latter population is particularly important, as the data pertaining to HIV incidence continue to indicate a growing trend in the percentage of new HIV infections which are attributable to heterosexual contact. Further, because "traditional" STDs and HIV can both be transmitted through unprotected sexual contact— and because syphilis, gonorrhea, and AIDS (combined) account for two-thirds of all reportable infectious diseases in the U.S. (Yankauer, 1994)- -the. cognitive relationship between STDs and HIV needs further examination in terms of individuals' perceptions of their own sexual behavior and the potential risks attached to those behaviors. As previously mentioned, there are likely to be a whole host of other psychological and sociological factors which influence STD infection, perceptions of 158 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vulnerability to STDs and HIV, intentions for risk reduction behavior, and sexual practices. One of these factors may be found in the examination o f the social context of minority communities. For example, STD infection may be related to contexts determined by socioeconomic status. A recent study of San Francisco African- American and white adolescents found that adolescents residing in impoverished, working-class neighborhoods were more likely to be diagnosed with gonorrhea and chlamydia that those in non-working class, higher SES neighborhoods (Ellen et al., 1995). Even after controlling for socioeconomic factors, African-American adolescents still had higher rates of gonorrhea and chlamydia than did white adolescents. Based on these findings, the researchers hypothesized that other, non-behavioral factors such as geographic segregation may maintain higher levels of infection in the immediate community firom which minority adolescents are establishing their sexual networks, thus accounting for the higher STD rates (Ellen et al., 1995). This study by Ellen et al. (1995) shows that, even after controlling for the traditional socioeconomic factors which can account for increased risk of disease in minority communities, the risk for persons of color remains higher than that for non- Hispanic whites, and that the increased risk may be due to environmental factors which increase the overall disease risk for the community as a whole. Given that many urban minorities do not have the resources or the power to significantly change their environment in ways that will lower the disease risk, it may be unrealistic to expect assessments o f vulnerability that are objective and separate from an environmental 159 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. context. Indeed, it may be the case that behavioral research is placing too little emphasis on the impact of socioenviroomental factors— such as endemic disease prevalence rates and social and cultural norms— on sexual health. Future research should focus more attention on the degree to which contextual and environmental factors influence sexual behavior choices and perceptions about sexual behavior and sexual health. Along these lines, another factor which may be particularly relevant for understanding perceived vulnerability and risk behavior in underserved and indigent urban minority communities is the impact of environmental events and situations on psychological well-being, and how those psychological, intrapersonal states influence sexual health and sexual decision-making. More specifically, the sense of hopelessness or futility which comes from individuals' assessment of their own environment and their role in that environment may have an important impact on individuals' sexual behavior, their perceptions about their own behavior, and their perceptions about their vulnerability to disease. In his book. In the shadow o f the epidemic: Being HIV-negative in the a^e o f AIDS, psychotherapist Walt Odets describes this sense of hopelessness among HIV- negative gay men in San Francisco who are coming to terms with the impact of the AIDS epidemic not only on their community, but on their own identity. Odets points out that, because these men are often so overwhelmed by the fact that cohort after cohort of their friends and peers have died, they harbor strong feelings of "survival 160 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. guilt" because they have remained HIV-negative. Further, because these men are aware that such a great percentage of their community is HIV infected (approximately 50% in San Francisco's gay community, as estimated by the San Francisco Department of Health), they may feel that their own seroconversion is inevitable and "only a matter o f time." The feelings of guilt and depression these men have from trying to cope with the AIDS epidemic may be exacerbated by their own existing psychological issues, such as an internalized sense of homophobia and guilt about their sexual identity or "lifestyle" (Odets, 1995). Given all of these factors, gay men who are trying to make sense out of the horrid toll the epidemic has taken and continues to take on their community may question the importance of staying healthy and HIV-negative or may develop a sense of apathy about safer sexual practices. Odets illustrates this point with the example of an HIV-negative man who is told by his physician that his cholesterol levels are too high: ...And she said, 'Well, these are among the highest levels I've seen in my entire medical practice,' and she was asking about my diet, and so on, and finally I said, 'You know, I don't care what my cholesterol is, and if I think about having a heart attack in 10 or 20 years, it just doesn't mean anything in the context o f HIV.' To me this was like living in a war, and she's saying to me, 'You know you should really think about getting a nose job.' (p. 116) The concept of hopelessness can certainly apply to those living in underserved communities. Urban minority communities have "epidemics" other than AIDS— such as racism, violence, drug and alcohol abuse, poverty, malnutrition, and unemployment- with which they must cope and in which they must try to co-exist. Such epidemics 161 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. undoubtedly have a greater influence on the daily lives of many minority men and women, so much so that is it not surprising that health, particularly sexual health, would take a back seat to the importance of merely surviving in the midst of these social problems. Indeed, when one considers that, for many young African-American and Latino men, simply reaching adulthood is something which cannot be taken for granted (even without HIV), how likely is it that using condoms when having sex would be seen as an important issue? Social contexts in minority commimities has been examined by Kalichman and his colleagues (1995). In their study of inner-city STD clinic attenders, they asked their study participants to rate (in order o f importance) a variety of social problems, including AIDS. They found that, while AIDS was perceived as being a more serious problem than alcoholism, housing, and child care, it was a less serious problem than unemployment, discrimination, drug abuse, crime, and teen pregnancy. These results support the idea that a larger, macro-level contextual basis is needed for examinations of disease and illness, particularly in the cases of urban and indigent populations where the pervasiveness of other societal level problems arc salient and more important than adopting preventive behaviors for health maintenance and promotion. In communities where options for better lives and sources of pleasure and comfort may be few, sexual behavior may be one of the ways of reaffirming a life force that, in the outside world, seems nonexistent. Like Odets's samples of gay men, sex may be a way to "self-medicate" the negativity of everyday life. Risk taking behavior, 162 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sexual or otherwise, may be a way by which individuals in such underserved communities "stay in the present" rather than look ahead to a future which may seem bleak. While "staying in the present" may ordinarily provide a psychological stronghold for remaining grounded in reality, the self-focused nature of this coping strategy may prove to be very unfortunate when combined with risk behaviors-such as unsafe sex and drug abuse— which can have long-term negative consequences for the individuals practicing the behaviors and others around them. Although it is unrealistic and downright impossible for health interventions to solve broader societal problems, it is absolutely crucial for researchers to understand the communities they serve, as well as the impact of the larger community environment on the individuals living there. In essence, it is important for future research to understand and work witfiin the psychosocial paradigms that already exist in communities rather than imposing theoretical paradigms that may be inappropriate for the communities in question. In his book, Odets (1995) also raises the point that sexual behavior is much more than simply a recreational activity or, for heterosexuals, a method of procreation. In some cases, sexual activity has the power to confirm and reaffirm an identity, as in the case of gay men and lesbians. In other cases, sex is a way of bonding with another individual, using the physical intimacy to foster a psychological and spiritual connection. In all cases, however, sexuality is a component of human nature. The reprimands placed upon all forms of sexuality (save heterosexual marriages and celibacy) by a predominantly conservative society, as well as the implicit message o f 163 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. many prevention campaigns that safer sex is purely a function o f rational risk assessment and personal assertion, do not take into account the more interpersonal and intrapersonal meanings of sexuality and sexual expression (Odets, 1995). Now that it is well known that the simple dissemination of knowledge is not the answer to the AIDS epidemic, we, as prevention researchers must, by necessity, look deeper into the many complex issues that are associated with and influence sexual behavior. It is the present challenge of STD and HIV prevention research to try to address these issues from a more human perspective and in a manner which is culturally sensitive in our continuing efforts to curb the AIDS epidemic. 164 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. REFERENCES Andrew, J.M. (1970). Recovery from surgery, with and without preparatory instruction, for three coping styles. Journal of Personality and Social Psychology, l i , 223-226. Aral, S.O. and Holmes, K.K. (1991). Sexually transmitted diseases in the AIDS era. Scientific American. 264. 62-69. Baron, R.M. and Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psvchoiogv. 51. 1173-1182. Bauman, L.J. and Siegel. K. (1987). Misperceptions among gay men of their risk for AIDS associated with their sexual behavior. Journal of Applied Social Psychology. 17.329-350. Becker, M.H. (1974). The health belief model and personal health behavior. Health Education Monographs. 2 (whole No. 4). Brookmeyer, R. (1991). Reconstruction and future trends of the AIDS epidemic in the United States. Science. 253. 37-42. Burger, J.M. and Bums. L.B. (1988). The illusion of unique invulnerability' and the use o f effective contraception. Personality and Social Psychology Bulletin. J4, 264- 270. Byrne, D. (1961). The Repression-Sensitization scale: rationale, reliability, and validity. Journal of Personality. 29. 334-349. Cannon, R. et al. (1989). Syphilis is strongly associated with HIV infection in Baltimore STD clinic patients independent of risk group. Presented at 5th International Conference on AIDS, Montreal, Canada, June 4-9 (abstract Th.A.0.18). Catania, J.A., Coates, T.J., Greenblatt, R.M., Dolcini, M.M., Kegeles, S.M., Puckett, S., Corman, M., and Miller, J. (1989). Predictors of condom use and multiple partnered sex among sexually active adolescent women: implications for AIDS- related health interventions. Journal of Sex Research. 26. 514-524. 165 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Catania, J.A., Coates, T.J., Stall, R., Turner, H., Peterson, J., Hearst, N., Dolcini, M.M., Hudes, E., Gagnon, J., Wiley, J., and Groves, R. (1992). Prevalence o f AIDS- related risk factors and condom use in the United States. Science. 258. 1101- 1106. Catania, J.A., Binson, D., Dolcini, M.M., Stall, R., Choi, K.H., Pollack, L.M., Hudes, E., Canchola, J., Phillips, K., Moskowitz, J.T., and Coates, T.J. (1995). Risk factors for HIV and other sexually transmitted diseases and prevention practices among US heterosexual adults: changes from 1990 to 1992. American Journal of Public Health. 85(111 1492-1499. Cates, W. and Toomey, K.E. (1990). Sexually transmitted diseases: an overview of the situation. Primarv Care. 17111 1-27. Centers for Disease Control. (1985). Chlamydia trachomatis infection: policy guidelines for prevention and control. Morbiditv and Mortality Weekly Report. M, 53S-74S. Centers for Disease Control. (1991a). Morbidity attributable to HTV infection/AIDS. Morbiditv and Mortality Weeklv Report. 40. 41-44. Centers for Disease Control. (1991b). The HIV/AIDS epidemic: the first 10 years. Morbiditv and Mortality Weeklv Report. 40. 357. Centers for Disease Control. (1991c). Update: Acquired Immunodeficiency Syndrome— United States, 1981-1990. Morbiditv and Mortality Weeklv Report. 40, 358-369. Centers for Disease Control. (199Id). Primary and secondary syphilis— United States, 1981-1990. Morbiditv and Mortality Weeklv Report. 40(19’ ). 314-323. Centers for Disease Control. (199 le). Regional and temporal trends in the surveillance of syphilis. United States, 1986-1990. Morbiditv and Mortality Weekly Report. 4fi(SS-3), 29-33. Centers for Disease Control. (1992a). The second 100,000 cases of Acquired Immunodeficiency Syndrome— United States, June 1981-December 1991. Morbiditv and Mortality Weekly Report. 41(2). 28-29. 166 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Centers for Disease Control. (1992b). 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. Morbiditv and Mortalitv Weeklv Report. 4i(no. RR-I7). Centers for Disease Control. (1993). Sexual risk behaviors of STD clinic patients before and after Earvin "Magic" Johnson's HIV-infection annoimceraent— Maryland, 1991-1992. Morbiditv and Mortalitv Weeklv Report. 42(31 45-48. Centers for Disease Control. (1994). AIDS among racial/ethnic minorités— United States, 1994. Morbiditv and Mortalitv Weeklv Report. 43(35), 644-647,653- 655. Centers for Disease Control. (1995a). Update: Acquired Immunodeficiency Syndrome-United States, 1994. Morbiditv and Mortalitv Weeklv Report. 44(41 64-67. Centers for Disease Control. (1995b). Update: .AIDS among women— United States, 1994. Morbiditv and Mortality Weeklv Report. 44(51 81-84. Chu, S.Y., Buehler, J.W.. and Berkelman, R.L. (1990). Impact of the human immunodeficiency virus epidemic on mortality in women of reproductive age, United States. Journal of the American Medical Association. 264.225-229. Cochran, S.D., Keidan, J., and BCalechstein, A. (1990). Sexually transmitted diseases and Acquired Immimodeficiency Syndrome (AIDS): changes in risk reduction behaviors among young adults. Sexuallv Transmitted Diseases. 12(2), 80-86. Cohen, J. (1977). Statistical power analysis for the behavioral sciences. New York: Academic Press, Inc. Cook, J.R. (1985). Repression-sensitization and approach-avoidance as predictors of response to a laboratory stressor. Journal of Personalitv and Social Psvchoiogv. 49, 759-773. DeBuono, B.A. et al. (1989). Counseling and antibody testing to prevent HIV infection. JAMA. 261. 2502. Doll, L.S., Byers, R.H., Bolan, G., Douglas, J.M., Moss, P.M., Weller, P.D., Joy, D., Bartholow, B.N., and Harrison, J.S. (1991). Homosexual men who engage in high-risk sexual behavior: a multicenter comparison. Sexuallv Transmitted Diseases. 18(3). 170-175. 167 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ellen, J.M., Kohn, R.P., Bolan, G., Shiboski, S., and Kxieger, N. (1995). Socioeconomic differences in sexually transmitted disease rates among Black and White adolescents, San Francisco, 1990-1992. American Journal of Public Health. 85(1 n . 1546-1548. Epstein, S. and Fenz, W.D. (1967). The detection of areas of emotional stress through variations in perceptual threshold and physiological arousal. Journal of Experimental Research in Personalitv. 2. 191-199. Erickson, B., Wasserheit, J.N., Rompalo, A.M., Brathwaite, W., Classer, D., and Hook, E.W. III. (1990). -Routine voluntary HIV screening in STD clinic clients: characterization of infected clients. Sexuallv Transmitted Diseases. 17(4). 194- 199. Fennema, J.S.A., van Ameijden. E.C.J., Coutinho, R.A., van Doomum, G.J.J., Henquet, C.J.M., and van den Hoek, J.A.R. (1993). HTV prevalence among clients attending a sexually transmitted diseases clinic in Amsterdam: the potential risk for heterosexual transmission. Genitourinarv Medicine. 69. 23-28. Geringer, W.M., Marks, S.. Allen. W.J., and Armstrong, K.A. (1993). Knowledge, attitudes, and behavior related to condom use and STDs in a high risk population. Journal of Sex Research. 30(11. 75-83. Holmes, M.D., Salyer, S.M., Bickell, N.A., Vermund, S.H., Hanff, P.A.. and Phillips, R.S. (1993). Chlamydial cervical infection in jailed women. American Journal of Public Health. 83(41. 551-555. Hull. H.F., Bettinger, C.J., Gallaher, M.M., Keller, N.M., Wilson, J., and Mertz, G.J. (1988). Comparison o f HIV-antibody prevalence in patients consenting to and declining HIV-antibody testing in an STD clinic. JAMA. 260. 935-938. James, N.J., Gillies, P.A., and Bignell, C.J. (1991). AIDS-related risk perception and sexual behavior among sexually transmitted disease clinic attenders. International Journal o f STD and AIDS. 2, 264-271. Janis, l.L. (1983). Stress inoculation. In S. Breznitz (Ed.), 77ze dem a/o/stresj (pp. 35- 76). New York: International Universities Press, Inc. Janz, N.K. and Becker, M.H. (1984). The Health Belief Model: A decade later. Health Education Ouarterlv. 11. 1-47. 168 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Johnson, E.H., Gant, L., Hinkle, Y.A., Gilbert, D., Willis, C., Hoopwood, T. (1992). Do African-American men and women differ in their knowledge about AIDS, attitudes about condoms, and sexual behaviors? Journal of the National Medical Association. 84.49-64. Kalichman, S.C., Adair, V., Somlai, A.M., and Weir, S.S. (1995). The perceived social context of AIDS: study of irmer-city sexually transmitted disease clinic patients. AIDS Education and Prevention. 7(4), 298-307. Lazarus, R.S. (1983). Costs and benefits of denial. In S. Breznitz (Ed.), The denial of stress (pp. 1-30). -New York: International Universities Press, Inc. Lehman, D.R. and Taylor, S.E. (1987). Date with an earthquake: coping with probable, unpredictable disaster. Personalitv and Social Psvchoiogv Bulletin. 11, 546-555. Lindan, C., Kegeles, S., Hearst, N., Grant, P.. and Johnson, D. (1990). Heterosexual behaviors and factors that influence condom use among patients attending a sexually transmitted disease clinic— San Francisco. Morbiditv and Mortality Weeklv Report. 19, 685-689. Maddux, J.E. and Rogers, R.W. (1983). Protection motivation and self-efficacy: a revised theory of fear appeals and attitude change. Journal of Experimental Social Psvchoiogv. 19, 469-479. Montgomery, S.B., Joseph, J.G., Becker, M.H., Ostrow, D.G., Kessler, R.C., and Kirscht, J.P. (1989). The Health Belief Model in understanding compliance with preventive recommendations for AIDS: how useful? AIDS Education and Prevention. 1(4), 303-323. Moran, J.S., Aral, S.O., Jenkins, W.C., Peterman, T.A., and Alexander, E.R. (1989). The impact of sexually transmitted diseases on minority populations. Public Health Reports. 104(6), 560-565. Mulilis, J.P. and Lippa, R. (1990). Behavioral change in earthquake preparedness due to negative threat appeals: a test of protection motivation theory. Journal of Applied Social Psvchoiogv. 20. 619-638. Odets, W. (1995). In the shadow o f the epidemic: Being HTV negative in the age o f AIDS. Durham, North Carolina: Duke University Press. 169 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. O'Farrell, N., Windsor, I., and Becker, P. (1991) HIV-1 infection among heterosexual attenders at a sexually transmitted diseases clinic in Durban. South African Medical Journal. 80. 17-20. Olson, J.M. and Zanna, M.P. (1979). A new look at selective exposure. Journal of Experimental Social Psvchoiogv. 15. 1-15. Perloff, L.S. (1983). College students perceptions of vulnerability to divorce. Unpublished manuscript, Univ. of Illinois, Chicago. Perloff, L.S. and Fetzer. BrK. (1986). Self-other judgements and perceived vulnerability to victimization. Journal of Personality and Social Psvchoiogv. 3, 502-510. Phillips, R.S., Hanff, P.A., Holmes, M.D., Werthheimer, A., and Aronson, M.D. (1989). Chlamydia trachomatis cervical infection in women seeking routine gynecological care: criteria for selective testing. American Journal of Medicine. 515-520. Quinn, T.C. et al. (1988). Human immunodeficiency virus infection among patients attending clinics for sexually transmitted diseases. New England Journal of Medicine. 318. 197-203. Quinn, T.C., Cannon, R.O., Classer, D., Groseclose, S.L., Brathwaite, W.S.. Fauci, A.S., and Hook, E.W. III. (1990). The association of syphilis with risk of Human Immunodeficiency Virus infection in patients attending sexually transmitted disease clinics. Archives of Internal Medicine. 150. 1297-1302. Rabkin, C.S., Thomas, P.A., Jaffe, H.W., and Schultz, S. (1987). Prevalence of antibody to HTLV-III/LAV in a population attending a sexually transmitted diseases clinic. Sexuallv Transmitted Diseases. 14. 48-51. Rogers, R.W. (1983). Cognitive and psychological processes in fear appeals and attitude change: a revised theory of protection motivation. In J. T. Cacioppo & R. E. Petty (Eds.), Social psychophvsiology (Vol. 2, pp. 153-176). New York: Guilford. Ross, M.W. (1987). Illness behavior among patients attending a sexually transmitted disease clinic. Sexuallv Transmitted Diseases. 14(31 174-179. 170 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Schachter, J., Hill, E.C., King, E.B., Coleman, V.R., Jones, P., and Meyer, K.F. (1975). Chamydial infection in women with cervical dysplasia. American Journal of Obstetrics and Gvnecologv. 123. 753-757. Schoenbaum, E.E., Webber, M.P., Vermund, S., and Gayle, H. (1990). HTV antibody in persons screened for syphilis: prevalence in a New York emergency room and primary care clinics. Sexuallv Transmitted Diseases. 17^41 190-193. Shipley, R.H., Butt, J.H., Horwitz, B., and Farbry, J.E. (1978). Preparation for stressful medical procedure: effect of amount of stimulus preexposure and coping style. Journal of Consulting and Clinical Psvchoiogv. 46,449-507. Slovic, P., Kunreuther, H., and White, G.F. (1974). Decision processes, rationality, and adjustment to natural hazards. In G.F. White (Ed.), Natural hazards: Local, national, g l o b a l New York: Oxford. Sobel, M.E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhart (Ed.), Sociological methodology 1982 (pp. 290- 312). San Francisco: Jossey-Bass. Solomon, M.Z. and DeJong, W. (1989). Preventing AIDS and other STDs through condom promotion: a patient education intervention. American Journal of Public Health. 79, 453-458. Stein. M.D., Leibman, B., Wachtel, T.J., Carpenter, C.C.J., Fisher, A., Durand, L., O'Sullivan. P.S., and Mayer, K.H. (1991). HIV-positive women: reasons they are tested for HIV and their clinical characteristics on entry into the health care system. Journal of General Internal Medicine. 6, 286-289. Taylor, S.E. (1983). Adjustment to threatening events: a theory of cognitive adaptation. American Psvchologist. 18, 1161-1173. Taylor, S.E., Lichtman, R.R., and Wood, J.V. (1984). Attributions, beliefs about control, and adjustment to breast cancer. Journal of Personalitv and Social Psvchoiogv. 46,489-502. Taylor, S.E. and Brown, J.D. (1988). Illness and well-being: a social psychological perspective on mental health. Psvchological Bulletin. 103(2). 193-210. 171 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Taylor, S.E., Kemeny, M.E., Aspinwall, L.G., Schneider, S.G., Rodriguez, R„ and Herbert, M. (1992). Optimism, coping, psychological distress, and high-risk sexual behavior among men at risk for Acquired Immunodeficiency Syndrome (AIDS). Journal of Personalitv and Social Psvchoiogv. 63(31. 460-473. Thomas, S.B., Gilliam, A.G., Iwrey, C.G. (1989). Knowledge about AIDS and reported risk behaviors among black college students. American Journal of College Health. 38. 8-13. Van der Velde, F.W. and Van der Pligt. J. (1991). AIDS-related health behavior: coping, protection-motivation, and previous behavior. Journal of Behavioral Medicine. J 4 ,429-451. Van der Velde, F.W., Van der Pligt, J.. and Hooykaas, C. (1994). Perceived AIDS- related risk: accuracy as a fiuiction of differences in actual risk. Health Psvchoiogv. 13. 25-33. Wasser, S.C., Aral, S.O., Reed. D.S., and Bowen. G.S. (1989). Assessing behavioral risk for HIV infection in famih -planning and STD clinics: similarities and differences. Sexuallv Transmitted Diseases. 16(41. 178-183. Weeks, M.R., Schensul. J.J., Williams. S.S., Singer, M., and Grier, M. (1995). AIDS prevention for .African-American and Latina women: building culturally and gender-appropriate interventions. AIDS Education and Prevention. 7(3), 251- 263. Weinstein, N.D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psvchoiogv. 19(5), 806-820. Weinstein, N.D. (1982). Unrealistic optimism about susceptibility to health problems. Journal of Behavioral Medicine. 1(4), 441-460. Weinstein, N.D. (1983). Reducing unrealistic optimism about illness susceptibility. Health Psychology. 2, 11-20. Weinstein, N.D. (1984). Why it won't happen to me: perceptions of risk factors and susceptibility. Health Psychology. 3(51.431-457. Weinstein, N.D. (1987). Unrealistic optimism about susceptibility to health problems: conclusions from a community-wide sample. Journal of Behavioral Medicine. lû(5), 481-500. 172 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Weinstein, N.D. (1989). Effects of personal experience on self-protective behavior. Psychological Bulletin. 105. 31-50. Weinstein, N.D. and Nicolich, M. (1993). Correct and incorrect interpretations of correlations between risk perceptions and risk behaviors. Health Psvchoiogv. 12,235-245. Weinstock, H.S., Lindan, C., Bolan, G., Kegeles, S.M., and Hearst, N. (1993). Factors associated with condom use in a high-risk heterosexual population. Sexuallv Transmitted Diseases. 20. 14-20. Wurtele, S.K. and Maddux, J.E. (1987). Relative contributions of protection motivation theory components in predicting exercise intentions and behavior. Health Psychology. 6,453-466. Yankauer, A. (1994). Sexually transmitted diseases: a neglected public health priority. American Journal of Public Health. 84 ri2~): 1894-1897. 173 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix I. Intake questionnaire, male version 174 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. M-Code: M-Date: UNIVERSITY OF SOUTHERN CALIFORNIA SCHOOL OF MEDICINE HEALTH ATTITUDES SURVEY 175 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Age: years What ethnic group do you most identify with? (please check one) _______ Non-Hispanic White _______ Hispanic / Latino Afncan-American / Black _______ Asian _______ Native American Other: 3. Marital status: _______ Single / never married _______ Never married and living with parmer/lover _______ Married and living with spouse _______ Married, but separated from spouse _______ Divorced _______ Widowed Other: 4. How much schooling have you completed? (please check one) _______ Elementary school only (Grades 1-6) _______ Finished junior high school (Grades 7-9) _______ High school (Grades 10-12) or G.E.D. _______ High school and some college courses _______ A.A. degree or technical school n-aining _______ B.A. or B.S. degree _______ M.A., M.S.. Ph.D., or other graduate degree 5. Which of the choices below best describes your current work situation? (please check one) _______ Working full or part time _______ Unemployed _______ Retired _______ On public assistance or Social Security _______ Out of work because of illness or disability _______ Full or part time student Other: 6. Is this your first time in this sexually transmitted diseases (STD) clinic? No Yes 176 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7. Before now, have you ever thought you had a sexually transmitted disease (STD)? No (go to question 8) _______ Y es — > Did you go to the doctor or to a clinic and have it checked out? No Yes -> Was it diagnosed as a sexually transmitted disease (STD)? No Yes 8. When you have sex, who do you usually have sex with? Men only Mostly men Men & women Mostly women Women equally only 9. In an average month, how often do you usually have sex? Every day ________ About 2-3 times a week ________ About once a week Every few weeks About once a month or less 10. How many different people did you have anal or vaginal sex with in the last year? Write number here:________ 11. How many different people did you have anal or vaginal sex with in the last month? Write number here:________ 12. When you have anal or vaginal sex with a steady partner, such as a girlfriend or boyfriend, how often do you use a condom? _______ Always (100%) _______ Most of the time (75%) _______ About half the time (50%) _______ Sometimes (25%) _______ Never (0%) 177 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13. When you have anal or vaginal sex with a new partner or someone other than a girlfriend or boyfriend, how often do you use a condom? _______ Always ( 100%) _______ Most of the time (75%) ________About half the time (50%) _______ Sometimes (25%) _______ Never (0%) 14. In the next month, how likely are you to either buy condoms or get them from the clinic? (Circle one number) 1 - 2 3 4 5 not at all likely slightly likely moderately very likely extremely likely likely 15. In the next month, how likely is it that you will ask your steady partnerfs) to use a condom when you have sex? (Circle one number) 1 2 3 4 5 not at all likely slightly likely moderately very likely extremely likely likely 16. In the next month, how likely is it that you will ask a new partner(s) to use a condom when you have sex? (Circle one number) 1 2 3 4 5 not at all likely slightly likely moderately very likely extremely likely likely 17. In the next month, how likely is it that you will use a condom if your steadv partner asks you to? (Circle one number on the rating scale below) 1 2 3 4 5 not at all likely slightly likely moderately very likely extremely likely likely I S. In the next month, how likely is it that you will use a condom if a new partner asks you to? (Circle one number on the rating scale below) 1 2 3 4 5 not at all likely slightly likely moderately very likely extremely likely likely 178 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19. In the next month, after completing your treatment, what percent of the time do you think you will use condoms when you have anal or vaginal sex? (Check one) All of Most of Half of Some of None of the time the time the time the time the time (100%) (75%) (50%) (25%) (0%) If you currently have more than one sex partner, answer the next three questions and then go on to the next page. If you currently have only one sex partner, go to the next page now. 20. In the next month, how likely is it that you will cut down on the number of people you have sex with? (Circle onemumber on the rating scale below) 1 2 3 4 5 not at all likely slightly likely moderately very likely extremely likely likely 21. In the next month, how likely is it that you will choose to have sex with only one person? 1 2 3 4 5 not at all likely slightly likely moderately very likely extremely likely likely 22. In the next month, how likely is it that you will choose to not have sex with anyone? 1 2 3 4 5 not at all likely slightly likely moderately very likely extremely likely likely GO TO THE NEXT PAGE 179 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23. If you and someone who had a sexually transmitted disease (such as gonorrhea or chlamydia) had sex without using a condom, do you think you would get a sexually transmitted disease? No Yes 24. If you are taking medicine for a sexually transmitted disease (such as gonorrhea or chlamydia) and the symptoms go away, are you completely cured of the sexually transmitted disease? No Yes 25. Do you think that you would usually be able to tell if someone a sex partner was infected with a sexually transmitted disease (STD)? No Yes 26. Do you think that people who have had sexually transmitted diseases (STDs) in the past have a greater chance of getting a sexually transmitted diseases (STDs) in the future? No Yes 27. Do you think that people who have had sexually transmitted diseases (STDs) in the past have a greater chance of being exposed to HIV, the AIDS virus? No Yes 28. Do you think that you would usually be able to tell if a sex partner has HIV? No Yes GO TO THE NEXT PAGE 180 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. These next questions will ask you to think about your partner(s) and the types of things you do with them. 29. Do you have a steady partner, like a girlfriend or boyfriend? No (GO TO QUESTION 40)________Yes (CONTINUE WITH THE QUESTIONS ON THIS P.\GE) 30. Is this person: male female 31. Have you had anal or vaginal sex with this person? No Yes 32. About how long have you been having sex (anal or vaginal) with this person? ______ One week or less ______ More than a week but less than a month ______ More than a month but less than 6 months Six months or more 33. The last time you had anal or vaginal sex with this person, were you high or buzzed on alcohol or drugs before you had sex? No Yes 34. Have you had VAGINAL sex with this person in the last 30 days? _______No Yes Not applicable: my partner is a male 181 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35. The last time you had VAGINAL sex with this person, did you use a condom? _______No Yes Not applicable: my ______ I’ ve never had partner is a male vaginal sex with this person 36. When you have VAGINAL sex with this person, how often is a condom used? Would you say? ______ Every time (100%) ______ Almost every time (75%) ______ Sometimes (50%) ______ Almost never (25%) ______ Never (0%) ______ Not applicable: my partner is a male ______ I've never had vaginal sex -vith this person 37. Have you had ANAL sex with this person in the last 30 days? _______No \ es 38. The last time you had ANAL sex wit: this person, did you use a condom? _______No \ .-s I've never had anal sex with this person 39. When you have ANAL sex with this person, how often is a condom used? Would you say? ______ Every time (100%) ______ Almost every time (75%) ______ Sometimes (50%) ______ Almost never (25%) ______ Never (0%) ______ I've never had anal sex with this person. GO TO THE NEXT PAGE 182 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40. Have you had anal or vaginal sex with anyone who is ngta steady partner in the last 30 days? ______ No (GO TO QUESTION 52) ______ Yes (CONTINUE WITH THE QUESTIONS ON THIS PAGE) 41. In the last 30 days, how many partners did you have anal or vaginal sex with? Write number here:____________ 42. How many of these partners were men? Write number here: __________ 43. How many of these partners were women? Write number here:____________ If you have had sex with more than one partner in the last 30 days, think of the person you had sex with most recentiv who is NOT a steadv partner when you answer the following questions. 44. Is this person male or a female? male female 45. The last time you had sex with this person, were you high or buzzed on alcohol or drugs before you had sex? No Yes 46. Have you had VAGINAL sex with this person in the last 30 days? _______ No Yes Not applicable: my partner is a male 47. The last time you had VAGINAL sex with this person, did you use a condom? _______No Yes Not applicable: my ______ I've never had partner is a male vaginal sex with this person 183 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 48. When you have VAGINAL sex with this person, how often is a condom used? Would you say? _______Every time (100%) _______Almost every time (75%) _______Sometimes (50%) _______ Almost never (25%) _______Never (0%) _______Not applicable: my parmer was male _______I've never had vaginal sex with this person 49. Have you had ANAL-sex with this person in the last 30 days? No Yes 50. The last time you had ANAL sex with this person, did you use a condom? _______ No Yes I've never had anal sex with this person 51. When you have ANAL sex with this person, how often is a condom used? Would you say? _______Every time (100%) _______Almost every time (75%) _______Sometimes (50%) _______Almost never (25%) _______Never (0%) _______I've never had anal sex with this person GO TO THE NEXT PAGE 184 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. These questions ask how you see your own risk of getting a sexually transmitted disease (STD), such as syphilis, gonorrhea, herpes, etc. If you already know you have an STD, answer the questions in terms of getting an STD in the future. Please respond by circling qqs number on the scale that best describes how you feel. Please answer as honestly as possible. 52. How would you rate your chances of getting a sexually transmitted disease (STD) in the future? 1 2 3 4 5 6 7 much below below slightly average slightly above much above average average below above average average average average 53. Compared to other people in this clinic, how would you rate your chances of getting a sexually transmitted disease (STD) in the future? 1 2 - 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as others in above average average average this clinic average 54. Compared to other men of your same age and race/ethnic group, how would you rate your chances of getting a sexually transmitted disease (STD) in the future? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as other men above average average average of my race/age average 55. How likely do you think you are to experience serious health problems resulting from a sexually transmitted disease (STD) in the future? 1 2 3 4 5 6 7 much below below slightly average slightly above much above average average below above average average average average 56. Compared to other people in this clinic, how would you rate your chances of experiencing serious health problems resulting from a sexually transmitted disease (STD) in the future? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as others in above average average average this clinic average 57. Compared to other men of your same age and race/ethnic group, how would you rate your chances of experiencing serious health problems resulting from a sexually transmitted disease (STD) in the future? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as other men above average average average of my race/age average 185 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58. If you continued doing the same things you usually do when you have sex, what do you think are your chances of getting a sexually transmitted disease (STD) in the future? 1 2 3 4 5 6 7 much below below slightly average slightly above much above average average below above average average average average These questions will ask how you see your own risk of getting infected with HTV, the AIDS virus. Answer the questions by circling gne number on the response scale that best describes how you feeI.Again, please answer as honestly as possible. If you know you are HIV-positive, skip these questions and go to the next page now. 59. How would you rate your chances of getting infected with the AIDS virus in the future? 1 2 3 4 5 6 7 much below below slightly average slightly above much above average average below above average average average average 60. Compared to other people in this clinic, how would you rate your chances of getting infected with the AIDS virus in the future? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as others in above average average average this clinic average 61 . Compared to other men of your same age and race/ethnic group, how would you rate your chances of getting infected with the AIDS virus in the future? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as other men above average average average of my race/age average 62. If you continued doing the same things you usually do when you have sex, how would you rate your chances of getting infected with the AIDS virus in the future? 1 2 3 4 5 6 7 much below below slightly average slightly above much above average average below above average average average average GO TO THE NEXT PAGE 186 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. These questions will ask how you think about sexually transmitted diseases (STDs) such as syphilis. Answer the questions by circling gne number on the response scale that best describes how you feel. 63. In your opinion, how serious is a sexually transmitted disease (STD) like syphilis? 1 2 3 4 5 not at all slightly moderately very extremely serious serious serious serious serious 64. To what extent do you agree or disagree with this statement: "Getting syphilis is no big deal." 1 2 3 4 5 I totally disagree I slightly disagree I neither agree I slightly agree 1 totally agree nor disagree 65. In your opinion, how treatable is syphilis? In other words, can medicines get rid of symptoms of sexually transmitted diseases (STDs) like syphilis? I 2 3 4 5 not at all a little bit moderately pretty very treatable treatable treatable treatable treatable 66. In your opinion, how curable is syphilis? In other words, can medicines completely cure someone of a sexually transmitted disease (STD) like syphilis? 1 2 3 4 5 not at all a little bit moderately pretty very curable curable curable curable curable 67. How concerned are you that a sexually transmitted disease (STD) like syphilis can seriously affect your overall health? I 2 3 4 5 not at all slightly moderately very extremely concerned concerned concerned concerned concerned 68. How personally threatened do you feel by a sexually transmitted disease (STD) like syphilis? I 2 3 4 5 not at all slightly moderately very extremely threatened threatened threatened threatened threatened GO TO THE NEXT PAGE 187 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. These questions will ask how you think about sexually transmitted diseases (STDs) such as gonorrhea. Answer the questions by circling gne number on the response scale that best describes how you feel. 69. In your opinion, how serious is a sexually transmitted disease (STD) like gonorrhea? 1 2 3 4 5 not at all slightly moderately very extremely serious serious serious serious serious 70. To what extent do you agree or disagree with this statement: "Getting gonorrhea is no big deal." I 2 3 4 5 I totally disagree I slightly disagree 1 neither agree I slightly agree I totally agree _ nor disagree 71. In your opinion, how treatable is gonorrhea? In other words, can medicines get rid of symptoms of sexually transmitted diseases (STDs) like gonorrhea? 1 2 3 4 5 not at all a little bit moderately pretty very treatable treatable treatable treatable treatable 72. In your opinion, how curable is gonorrhea? In other words, can medicines completely cure someone of a sexually transmitted disease (STD) like gonorrhea? 1 2 3 4 5 not at all a little bit moderately pretty very curable curable curable curable curable 73. How concerned are you that a sexually transmitted disease (STD) like gonorrhea can seriously affect your overall health? 1 2 3 4 5 not at all slightly moderately very extremely concerned concerned concerned concerned concerned 74. How personally threatened do you feel by a sexually transmitted disease (STD) like gonorrhea? 1 2 ' 3 4 5 not at all slightly moderately very extremely threatened threatened threatened threatened threatened GO TO THE NEXT PAGE 188 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. These questions will ask how you think about sexually transmitted diseases (STDs) such as genital warts and herpes. Answer the questions by circling one number on the response scale that best describes how you feel. 75. In your opinion, how serious are sexually transmitted diseases (STDs) like genital warts or herpes? 1 2 3 4 5 not at all slightly moderately very extremely serious serious serious serious serious 76. To what extent do you agree or disagree with this statement: "Getting genital warts or herpes is no big deal." 1 - 2 3 4 5 I totally disagree 1 slightly disagree I neither agree I slightly agree 1 totally agree nor disagree 77. In your opinion, how treatable are sexually transmitted diseases (STDs) like genital warts or herpes? In other words, can medicines get rid of symptoms of warts or herpes? 1 2 3 4 5 not at all a little bit moderately pretty very treatable treatable treatable treatable treatable 78. In your opinion, how curable are sexually transmitted diseases (STDs) like genital warts or herpes? In other words, can medicines completely cure someone of warts or herpes? 1 2 3 4 5 not at all a little bit moderately pretty very curable curable curable curable curable 79. How concerned are you that sexually transmitted diseases (STDs) like genital warts or herpes can seriously affect your overall health? 1 2 ' 3 4 5 not at all slightly moderately very extremely concerned concerned concerned concerned concerned 80. How personally threatened do you feel by sexually transmitted diseases (STDs) like genital warts or herpes? 1 2 3 4 5 not at all slightly moderately very extremely threatened threatened threatened threatened threatened GO TO THE NEXT PAGE 189 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. These questions will ask how you think about HIV, the AIDS virus. Answer the questions by circling one number on the response scale that best describes how you feel. 81. In your opinion, how serious is HIV? I 2 3 4 5 not at all slightly moderately very extremely serious serious serious serious serious 82. To what extent do you agree or disagree with this statement: "Getting HIV is no big deal." I 2 3 4 5 I totally disagree I slightly disagree I neither agree I slightly agree 1 totally agree nor disagree 83. In your opinion, how treatable is HIV? In other words, can medicines get rid of symptoms of HIV? I 2 3 4 5 not at all a little bit moderately pretty very treatable treatable treatable treatable treatable 84. In your opinion, how curable is HIV? In other words, can medicines completely cure someone of HIV? 1 2 3 4 5 not at all a little bit moderately pretty very curable curable curable curable curable 85. How concerned are you that HIV can seriously affect your overall health? 1 2 ' 3 4 5 not at all slightly moderately very extremely concerned concerned concerned concerned concerned 86. How personally threatened do you feel by HIV? 1 2 3 4 5 not at all slightly moderately very extremely threatened threatened threatened threatened threatened 87. What do you think are the chances that vou have been exposed to HIV through having sex? 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 190 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 88. What do you think are the chances that anv of vour sex partners in the last year were infected with HIV? 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 89. The following statements tell how people may feel or act in their daily lives. If you think a statement describes how you usually feel or act, circle the T for "True". If you think a statement does flfit describe how you usually feel or act, circle the F for "False". Please answer as truthfully as possible. T F I get angry sometimes. T F I am more of a happy-go-lucky person than a deep thinker. T F I have very few quarrels with members of my family. T F I find discussions about sex slightly annoying. T F Bad words often come into my mind and I cannot get rid of them. T F Sometimes when I am not feeling well I am angry. T F I have a habit of counting things that are not important, such as bulbs on electric signs. T F I try to plan in advance what to do if threatening situations arise. T F Sex education should not be a part of the high school curriculum. T F I get so mad that I feel like beating or smashing things. T F I am often troubled with disturbing thoughts. T F People have too much sex on their minds. T F Most of the people who know me would say I am a cheerful person. T F I almost never think of things too bad to talk about. T F I sometimes tease animals. T F When I leave home 1 tend to worry about such things as whether I left the stove or iron on. T F I like to let people know where I stand on things. T F I sweat easily even on cool days. T F I rarely wonder what hidden reason a person may have for doing something nice for me. T F I work under a great deal of tension. T F I am easily awakened by noise. T F I tend to get along well with people and am liked by almost everybody. T F I have daydreams that I make a fool of someone who knows more than 1 do. T F I think of ways to get even with certain people. T F When things go wrong, I cannot rest until I've corrected the situation. T F I usually have to stop and think before I act even in unimportant matters. 191 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. T F I frequently find myself worrying about something. T F Most nights 1 go to sleep without thoughts or ideas bothering me. T F I tend to keep on with things until others lose their patience with me. 90. Please read the following statements. If your think the statement is true, circle the T for true. If you think the statement is not true, circle the F for false. T F You can get sexually transmitted diseases from having sex. T F If you are taking medicine for a sexually transmitted disease (STD) and the symptoms go away, stop taking the medicine. T F If you do not see any signs of an infection, you do not have a sexually transmitted disease (STD). T F The more sex parmers you have, the greater the chances of getting a sexually transmitted disease (STD). T F Most health clinics treat sexually transmitted diseases (STDs). T F If you think you have a sexually transmitted disease (STD) but the symptoms go away, don't go to see a doctor. T F You should go to see a doctor if vour panner has a sexuallv transmitted disease (STD). GO TO THE NEXT PAGE 192 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91. Please read the following statements. If your think the statement is true, circle the T for true. If you think the statement is not true, circle the F for false. T F AIDS is caused by a virus. T F People with AIDS usually die. T F People can get AIDS from sharing needles. T F AIDS is passed from mother to baby. T F Persons with AIDS lose weight. T F I know places where I can get tested. T F People can get AIDS from unprotected sex (sex without a condom). T F People with AIDS have fevers and infections. T F People can get AIDS from donating blood. T F People can protect themselves from AIDS by using condoms. T F aI c Is^ 5^^ AIDS from eating food prepared by a person with T F People can get AIDS when sneezed on by persons with AIDS. T F People can get AIDS from a toilet used by a person with AIDS. T F Cleaning needles (works) with water is enough to kill the AIDS virus. T F People can get AIDS from sharing dirty needles (works) with someone with AIDS. GO TO THE NEXT PAGE 193 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IF YOU HAVE ANY COMMENTS OR SUGGESTIONS ABOUT ANY ASPECT OF THIS STUDY, PLEASE FEEL FREE TO WRITE THEM IN THE SPACE BELOW. THANK YOU FOR YOUR PARTICIPATION IN THIS STUDY! 194 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 2. Follow-up questionnaire 195 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TELEPHONE FOLLOW-UP SURVEY Contact attempt # 1 : date: Successful? Y N Contact attempt #2: date: Successful? Y N Contact attempt #3 : date: Successful? Y N 1. Lost to follow-up? Y N 2. Date o f contact: / / 3. M-ID code: 4. Date o f Birth: / / 5. Age: vears 6. Gender: M F 7. What ethnic group do you most identify with? (CHECK ONE) Non-Hispanic White Asian-American Hispanic / Latino Native American Afhcan-American / Black Other: 8. Identity of participant confirmed? N o (END SURVEY H ER E) Yes (CONTINUE W ITH SURVEY) Now, Pm going to ask you to think back to when you were at Ruth Temple’ s STD clinic on (insert clinic visit date here). Please answer "yes” or "no" to each of the following questions. 9. When vou were at the clinic a few weeks ago, did the doctor tell vou that you had an'STD? No (GO T O Q1 2) Yes I'm going to read a list of some common STD infections. As I read through the list, please answer "yes" or no" if the doctor at the clinic diagnosed you with that infection when you were at the clinic that day. 10. Did the doctor tell you that you had: syphilis? gonorrhea? chlamydia? chancroid? herpes? gemtal warts? non-go nococcal urethritis, or NGU? another STD that I didn't mention? N Y N/A (no STD N Y N/A (no STD N Y N/A (no STD N Y N/A (no STD N Y N/A (no STD' N Y N/A (no STD' N Y N/A (no STD N Y N/A (no STD' 196 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11. Did the doctor give you a shot, some pills, or something else to treat the infection? No Yes Now, I'm going to ask you about the sexual contacts you may have had since you were last at the clinic on (insert clinic visit date here). 12. How many different people did you have anal or vaginal sex with in the time since you were at the clinic? Write number: (if NO SEX with anyone, go to end; otherw ise, CONTINUE with SURVEY) 13. How many of these people were men? Write number here: 14. How many of these people were women? Write number here: 15. Of all these partners, how many did you have VAGINAL sex with? Write number here:________ 16. Of all the persons you had VAGINAL sex with in the past 6 weeks since vou were at the clinic, with how many of those persons did you use a condom? Would you say (READ RESPONSE OPTIONS TO SUBJECT)... All of the partners ( 100%) None of the partners (0%) Most of the partners (75%] N/A: did not have vaginal sex About half the partners (50%) Some of the partners (25%) 17. Of all these partners, how many did you have ANAL sex with? Write number here:_________ 18. Of all the persons you had ANAL sex with in the past 6 weeks since you were at the clinic, with how many of those persons did you use a condom? Would vou say (READ RESPONSE OPTIONS TO SUBJECT)... All of the partners ( 100%) None of the partners (0%) Most of the partners (75%) N/A: did not have anal sex About half the parmers (50%) Some of the parmers (25%) SEX WITH STEADY PARTNER 19. Do you have a steady partner, like a girlfriend or boyfriend? Yes No (GO TO Q26) 20. Is this person: male female 197 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21. Have you had VAGINAL sex \\ith this person in the time since you were at the clinic? No Yes N/A: partner is male 22. The last time you had VAGINAL sex with this person, did you use a condom? No Yes N/A: ____ never had vaginal partner is male sex with this person 23. Of all the times you had VAGINAL sex with this person in the past 6 weeks since you were at the clinic, how often did you use a condom? Would you say (READ RESPONSE O PTIO NS TO SUBJECT)... Always ( 100%) Never (0%) Most of thejime (75%) N/A: partner is a male About half the time (50%) Never nad vaginal sex w/this person Sometimes (25%) 24. Have you had ANAL sex with this person in the time since you were at the clinic? No Yes 25. The last time you had ANAL sex with this person, did you use a condom? No Yes ____ never had anal sex with this person 26. Of all the times you had ANAL sex with this person in the past 6 weeks since you were at the clinic, how often did you use a condom? Would vou sav (read RESPONSE O PTIO N S TO SUBJECT)... Always ( 100%) _____ Sometimes (25%) Most of the time (75%J _____ Never (0%) About half the time (50%) _____ Never had anal sex with this person SEX W ITH NON-STEADY PARTNER 27. Have you had anal or vaginal sex with arwone who is not a steady partner in the past 6 weeks since you were at the clinic? No (GO TO END) _______ Yes If you have had sex with more than one partner in the time since you were at the clinic, think of the person you had sex with most recently who is NOT a steady partner when you answer the following questions. 28. Is this person male or a female? _______ male _______female 29. Have you had VAGINAL sex with this person in the time since you were at the clinicY _______ No Yes N/A: partner is a male 198 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30. The last time you had VAGINAL sex with this person, did you use a condom? No Yes N/A: I’ ve never had vaginal partner is a male sex with this person 31. Of all the times you had VAGINAL sex with this person in the past 6 weeks since you were at the clinic, how often did you use a condom? Would you say.. Every time ( 100%) ____ Never (0%) Almost eveiT time (75%) ____ N/A: partner was a male Sometimes (50%) ____ Never had vag. sex with this person Almost never (25%) 32. Have you had ANAL sex with this person in the time since you were at the clinic? _______ No _ _Yes 33. The last time you had ANAL sex with this person, did you use a condom? _______ No _______ Yes Never had anal sex with this person 34. Of all the times you had ANAL sex with this person in the past 6 weeks since you were at the clinic, how often did you use a condom? Would you say... Every time ( 100%) ____ Almost never (25%) Almost everv time (75%) ____ Never (0%) Sometimes (50%) Never had anal sex with this person END Thank you for your participation in this study! We are very grateful for the time you've shared with us. Do you have any questions you'd like to ask? (DEBRIEF individual, answer any QUESTIONS THE SUBJECT MAY HAVE. IF NO QUESTIONS, THANK THEM AGAIN .A N D END INTERVIEW). 199 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 3. Items used in the measurement of sexual risk behavior. 1. Is the study participant sexually active? 0 = No 1= Yes 2. Number of sexual partners in the past month 0 = no partners 1= I partner 2 = 2 partners 3 = 3 or more partners 3. When you have vaginal sex with this (steady) partner, how often do you use a condom? 0 = 100% of the time (or participant has no partner for this activity) 1 = 75% of the time 2 = 50% of the time 3 = 25% of the time 4 = 0% of the time 4. When you have vaginal sex with this (new/casual) partner, how often do you use a condom? 0 = 100% of the time (or participant has no partner for this activity) 1 = 75% of the time 2 = 50% of the time 3 = 25% of the time 4 = 0% of the time 5. When you have anal sex with this (steady) partner, how often do you use a condom? 0 = 100% of the time (or participant has no partner for this activity) 1 = 75% of the time 2 = 50% of the time 3 = 25% of the time 4 = 0% of the time (appendix 3 continues) 200 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 3 (cont'd) 6. When you have anal sex with this (new/casual) partner, how often do you use a condom? 0 = 100% of the time (or participant has no partner for this activity) 1 = 75% of the time 2 = 50% of the time 3 = 25% of the time 4 = 0% of the time 7. Have you had vaginal sex with this (steady) partner in the past 30 days? 0 = No 1 = Yes 8. Have you had anal sex with this (steady) partner in the past 30 days? 0 = No 1 = Yes 9. Have you had vaginal sex with this (new/casual) partner in the past 30 days? 0 = No 1 = Yes 10. Have you had anal sex with this (new/casual) partner in the past 30 days? 0 = No 1 = Yes All items are standardized and averaged prior to construction of scales. Risk index #1 = mean of Items 1-10 Risk index #2 = mean of Items 3-10 Risk subindex A = Item 1 Risk subindex B = Item 2 Risk subindex C = mean of Items 3 and 4 (frequency of condom use during vaginal sex with steady and new partner) Risk subindex D = mean of Items 5 and 6 (frequency of condom use during anal sex with steady and new partner) Risk subindex E = mean of Items 7 and 9 (occurrence of vaginal sex with steady and new partner) Risk subindex F = mean of Items 8 and 10 (occurrence of anal sex with steady and new partner) 201 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 4. The Repression-Sensitization scale (Epstein & Fenz, 1967). Instructions: The following statements tell how people may feel or act in their daily lives. If you think a statement describes how you usually feel or act, circle the T for "True". If you think a statement does ngt describe how you usually feel or act. circle the F for "False". Please answer as truthfully as possible. T F 1 get angry sometimes. T F 1 am more of a happy-go-lucky person than a deep thinker. T F 1 hâve very few quarrels with members of my family. T F 1 find discussions about sex slightly annoying. T F Bad words often come into my mind and I cannot get rid of them. T F Sometimes when I am not feeling well I am angry. T F I have a habit of counting things that are not important, such as bulbs on electric signs. T F I try to plan in advance what to do if threatening situations arise. T F Sex education si ould not be a part of the high school curriculum. T F 1 get so mad tha: I feel like beating or smashing things. T F I am often troubled with disturbing thoughts. T F People have too much sex on their minds. T F Most of the people who know me would say I am a cheerful person. T F 1 almost never think of things too bad to talk about. T F 1 sometimes tease animals. T F When I leave home I tend to worry about such things as whether I left the stove or iron on. T F I like to let people know where I stand on things. T F I sweat easily even on cool days. T F I rarely wonder what hidden reason another person may have for doing something nice for me. T F I work under a great deal of tension. (appendix continues) 202 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 4 (cont'd). T F I am easily awakened by noise. T F I tend to get along well with people and am liked by almost everybody. T F I have daydreams that I make a fool of someone who knows more than I do. T F I think of ways to get even with certain people. T F When things go wrong, I cannot rest until I've corrected the situation. T F I usually have to stop and think before I act even in unimportant matters. T F I frequently find myself worrying about something. T F Most nights I go to sleep without thoughts or ideas bothering me. T F I tend to keep on with things until others lose their patience with me. 203 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 5. Items used to measure perceived seriousness of disease. This scale was used to measure perceived seriousness to 4 sexually transmitted diseases: syphilis, gonorrhea, genital warts/herpes, and HIV. Although the following example is specific to HTV, the items used for the measurement of the other three STDs differed only in the specific STD which was printed in bold type. Instructions: These questions will ask how you think about HTV, the AIDS virus. Answer the questions by circling ong number on the response scale that best describes how you feel. 1. In your opinion, how serious is HTV? 1 2 3 4 5 not at all slightly _ moderately very extremely serious senous serious serious serious 2. To what extent do you agree or disagree with this statement: "Getting HTV is no big deal." 1 2 3 4 5 I totally disagree I slightly disagree I neither agree I slightly agree I totally agree nor disagree 3. In your opinion, how treatable is HIV? In other words, can medicines get rid of symptoms of HIV? 1 2 3 4 5 not at all a little bit moderately pretw very treatable treatable treatable treatable treatable 4. In your opinion, how curable is HIV? In other words, can medicines completely cure someone of HIV? 1 2 3 4 5 not at all a little bit moderately pretty very curable curable curable curable curable 5. How concerned are you that HIV can seriously affect your overall health? 1 2 3 4 5 not at all slightly moderately very extremely concerned concerned concerned concerned concerned 6. How personally threatened do you feel by HIV? 1 2 3 4 5 not at all slightly moderately very extremely threatened threatened threatened threatened threatened 204 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 6. STD knowledge items (Geringer et al., 1993) Instructions: Please read the following statements. If your think the statement is true, circle the T for true. If you think the statement is not true, circle the F for false. T F You can get sexually transmitted diseases from having sex. T F If you are taking medicine for an STD and the symptoms go away, stop taking the medicine. T F ^ f you do not see any signs of an infection, you do not have an T F The more sex parmers you have, the greater the chances of getting an STD. Most health clinics treat STDs. T F If you think you have an STD but the symptoms go away, don't go to see a doctor. T F You should go to see a doctor if your partner has an STD. 205 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 7. HIV/AIDS knowledge items (Nyamathi et al., 1993). Instructions: Please read the following statements. If your think the statement is true, circle the T for true. If you think the statement is not true, circle the F for false. T F AIDS is caused by a virus. T F People with AIDS usually die. T F People can get AIDS from sharing needles. T F ~ AIDS is passed from mother to baby. T F Persons with AIDS lose weight. T F I know places where you can get tested. T F People can get AIDS from unprotected sex (sex without a condom). T F People with AIDS have fevers and infections. T F People can get AIDS from donating blood. T F People can protect themselves from AIDS by using condoms. T F People can get AIDS from eating food prepared by a person with AIDS. T F ^ g ^ e can get AIDS when sneezed on by persons with T F P e ^ e can get AIDS from a toilet used by a person with T F Cleaning needles (works) with water is enough to kill the AIDS virus. People can get AIDS from sharing dirty needles (works) with someone with AIDS. 206 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appendix 8. Items used to measure perceived vulnerability. Perceived vulnerability to future STD infection 1 . How would you rate your chances of getting an STD in the future? 1 2 3 4 5 6 7 much below below sli^tly average slightly above much above average average below above average average average average 2. Compared to other people in this clinic, how would you rate vour chances of getting an STD in the future? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as others in above average average average this clinic average 3. Compared to other men of your same age and race/ethnic group, how would you rate your chances of getting an STD in the future? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as other men above average average average of my race/age average Perceived vulnerability to negative health problems resulting from STD infection 1 . How likely do you think you are to experience serious health problems resulting from an STD in the future? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as others in above average average average this clinic average 3. Compared to other men of your same age and race/ethnic group, how would you rate your chances of getting infected with the AIDS virus in the friture? 1 2 3 4 5 6 7 much below below slightly about the same slightly above much above average average below as other men above average average average of my race/age average 207 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Perceived vulnerability to HIV infection in persons at risk for the disease: An examination of STD clinic patients
PDF
The influence of future orientation on the relationship between condom use and violence among homeless/runaway youth: Implications for the prevention of HIV and violence
PDF
Predictors of adolescent depressive symptomatology
PDF
Receipt of and perceived difficulty in accessing needed services among HIV-seropositive men and women
PDF
Stressful life events, social support, and smoking behaviors in United States And Chinese adolescents: A cross-cultural comparison
PDF
Adolescent stress and styles of coping: Predictors and moderators of psychological distress
PDF
The relationship between high school anti-smoking policies and adolescent cigarette use
PDF
An investigation of caregiver sense of control and coping strategies in moderating the effects of caregiving demands
PDF
Patterns and risk factors for adolescent smoking progression
PDF
Self-help and problem-proneness as predictors of weapon carrying among adolescents
PDF
Antioxidants and risk of myocardial infarction and cancer in a cohort of middle-aged finnish men: the kuopio ischemic heart disease risk factor study
PDF
Diet, body size and menarche in a multiethnic cohort
PDF
What's age got to do with it? Gay identity-related factors as mediators of the relationship between age and HIV risk in gay and bisexual men
PDF
Artery wall injury and LDL-cholesterol in early atherosclerosis: The Los Angeles atherosclerosis study
PDF
Prevention of adolescent substance use: An investigation of cross -level interaction effects of self, peer, family and school level risk and protective factors using multilevel modeling.
PDF
Emotional intelligence and smoking risk factors in early adolescents
PDF
A randomized controlled trial of provider independent interventions to increase mammography screening among HMO members who are overdue for a mammogram
PDF
Implicit cognition and dissociative experiences as predictors of adolescent substance use
PDF
Aggression and victimization as predictors of drug use in early adolescence
PDF
Correlates of smoking behaviour among Egyptian adolescents
Asset Metadata
Creator
Ruiz, Monica Susanne
(author)
Core Title
Perceived vulnerability to HIV infection in persons at risk for the disease: An examination of STD clinic patients
Degree
Doctor of Philosophy
Degree Program
Preventive Medicine
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
health sciences, public health,OAI-PMH Harvest,psychology, behavioral
Language
English
Contributor
Digitized by ProQuest
(provenance)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c17-278847
Unique identifier
UC11352851
Identifier
9733129.pdf (filename),usctheses-c17-278847 (legacy record id)
Legacy Identifier
9733129.pdf
Dmrecord
278847
Document Type
Dissertation
Rights
Ruiz, Monica Susanne
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
health sciences, public health
psychology, behavioral