Close
About
FAQ
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
/
Depression severity, self-care behaviors, and self-reported diabetes symptoms and daily functioning among low-income patients receiving depression care
(USC Thesis Other)
Depression severity, self-care behaviors, and self-reported diabetes symptoms and daily functioning among low-income patients receiving depression care
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Depression Severity, Self-Care Behaviors, and Self-Reported Diabetes Symptoms and Daily
Functioning among Low-income Patients Receiving Depression Care
By
Hyunsung Oh
Dissertation Submitted in Partial Fulfillment
of the Requirements for the Degree of
Doctor of Philosophy
School of Social Work
University of Southern California
August 2014
1
Abstract
This dissertation examined three research questions focused on relationships between depression
and self-care behaviors, self-reported diabetes symptoms, and daily functioning among diabetes
patients receiving depression care. Depression among diabetes patients has been believed to
influence disease management. Many researchers have developed and tested depression care
models for diabetes patients that have different formats and contents. However, many of the
previous clinical trials found that patients assigned to innovative depression care groups did not
report significantly increased self-care behaviors. Yet, this group had significantly improved
depressive symptoms, self-reported diabetes symptoms, and daily functioning. Despite the
contradicting evidence, few studies have examined potential underlying mechanisms and
relationships across these variables.
To bridge this literature gap, three individual studies were conducted to better understand
equivocal findings in previous clinical trials. These studies involved secondary data analyses of
data (N = 387) collected from the Multifaceted Diabetes and Depression Program (MDDP)
randomized clinical trial (RCT) that tested the effectiveness of socio-culturally adapted
collaborative depression care for low-income, predominantly Hispanic diabetes patients in
safety-net clinics. In this trial, data on the frequency of healthy diet, exercise, self-blood glucose
monitoring (SBGM) and foot care, and body mass index (BMI) were obtained. The value of
depression care for diabetes patients was assessed. Depression care for diabetes patients is often
believed to improve self-care behaviors and diabetes symptoms by reducing apathy, a common
cardinal symptom. However, belief is rarely based on empirical evidence, and findings from
recent studies provide contradicting implications. The dissertation is aimed to provide the
groundwork for future debates and rigorous empirical tests.
2
Study 1 assessed whether problem-solving therapy (PST) receipt, depressive symptoms change,
and the interaction of these two independent variables predict patient diabetes self-care
improvement. Depression is a prevalent comorbidity associated with lower adherence to
recommended self-care behaviors and suboptimal diabetes outcomes. This association is
attributable to feelings of hopelessness, lack of interest in daily jobs, and less optimism observed
among depressed patients. Unfortunately, few RCTs demonstrated significantly increased self-
care behaviors in an intervention group that showed significant decline in depressive symptoms.
However, this lack of evidence for this association lends to further questioning about whether
reduced depressive symptoms observed during depression care are associated with self-care
behaviors. To examine this question, regression-based analyses of data in a RCT are necessary.
Also, PST for diabetes patients was aimed at identifying emotional distress associated with
diabetes self-care management behaviors in an effort to reduce emotional distress and improve
self-care management. We also examined the effect of PST on self-care behaviors. The last
question asked whether PST amplified the effect of declined depression on self-care behaviors.
For this question, the moderating effect of PST was examined. Secondary data analysis was
conducted with data (N = 387) collected from a RCT that tested collaborative depression care for
diabetes patients in safety-net clinics. Because both the intervention group and enhanced usual
care group showed a notable decrease in depressive symptoms, we analyzed these groups
together with statistical control of group membership. PST receipt, depression symptoms change
during 12 months after baseline, and the interaction of these two variables were regressed on
each self-care behavior change at 12 (N = 281), 18 (N = 249), and 24 (N = 235) months. A
bilingual social worker provided PST sessions scheduled between 8 and 12 times. Depression
was measured with the Patient Health Questionnaire-9 (PHQ-9). For the self-care behavior
3
frequency, the Summary of Diabetes Self-Care Activities (SDSCA) was used to measure weekly
frequency of the following behaviors: healthy diet, exercise, SBGM, and foot care. In addition,
BMI was included as a self-care behavior. Because each self-care behavior had lower inter-
correlation, each behavior was regressed by depressive symptom change and demographic and
clinical confounders. For analysis, multivariate regression analysis was conducted with SPSS
21.0. Three notable findings were found. First, PST receipt was not associated with concurrent
and prospective increased self-care behaviors. Second, decreased depression was associated with
more frequent healthy diet at the 12- (p < .01), 18- (p < .05), and 24-month follow-up (p < .05)
and increased foot care at the 12- (p < .05) and 24-month follow-up (p < .01). Finally, the
interaction between PST receipt and depressive symptoms change was significantly associated
with decreased foot care at the 12-month follow-up (p < .05) and decreased SBGM at the 18-
month follow-up (p < .05), suggesting that patients receiving PST had decreased depressive
symptoms, lower frequency of foot care, and SBGM. Future depression care should incorporate
standard self-care management education programs that have demonstrated positive impacts on
health behaviors, health status, and healthcare use.
Study 2 aimed to investigate whether improved self-care behaviors, which are often observed
among patients in depression care, predict better or worse depressive symptoms concurrently and
prospectively. In several clinical trials testing the effectiveness of depression care, patients
increased self-care behavior frequency; however, the extent of the increases were inconsistent
across studies. The increase in self-care behavior seems attributable to patient activities initiated
during psychotherapy where emotional distress from diabetes management was addressed.
However, it is not clear whether the increased self-care behavior has any effect on prospective
depression risk. It is likely that more frequent self-care behavior might increase depression risk
4
because patients may feel more distress from the additional self-care responsibilities, which may
require cognitive self-regulatory resources or giving up pleasurable activities (e.g., avoiding
meeting friends or abiding by an exercise plan). Furthermore, previous studies have bolstered the
significant effect of emotional distress on depression risk. However, better self-care behavior
could also reduce the risk of depression because increased self-care behavior may reduce
diabetes symptoms and the risk of complications. Also, some self-care behavior, such as
exercise, was found to reduce depressive symptoms and has been implemented in depression
care. To investigate which scenario was more likely, this study investigated whether more
frequent self-care behavior achieved during depression care was associated with depression risk,
concurrently and prospectively. Secondary analyses were conducted with MDDP. The extent of
self-care behavior adherence was measured weekly for factors including healthy diet, regular
exercise, SBGM, foot care, and BMI. Changes in self-care diet, exercise, SBGM, foot care, and
BMI during 12 months since baseline were regressed on either clinical depression status,
determined by the PHQ-9, or depression severity, which was measured at the 12- (N = 281), 18-
(N = 249), and 24-month follow-up (N = 235). By adding the baseline values of depression into
the statistical models, we focused on the variations in depressive symptom changes since trial
enrollment. A sensitivity analysis of continuous and categorical self-care behavior variables for
three groups, including the lower, middle, and upper 33%, were examined. Continuous predictor
results found that one unit change in diet during the previous 12 months was associated with a
19% reduced risk of clinical depression at the 12-month follow-up (p < .05). A similar degree of
reduced risk was found for the variable of increased exercise, measured at the 18- (p < .05) and
24-month follow-up (p < .05). When categorical self-care behaviors were regressed on
depression outcomes, the explained variances of models improved by 2% to 5%. More frequent
5
exercise consistently predicted decreased depression, prospectively; only a concurrent effect of
increased healthy diet on depression was found. Finally, significant associations between foot
care and SBGM and depression outcomes were found inconsistently. Increased frequency of
exercise predicted a significant reduction in both the odds of clinical depression and depressive
symptoms, prospectively. Results suggest dual benefits of exercise for both diabetes outcomes
and depression. According to the findings, more frequent self-care behavior does not appear to
be a risk factor for clinical depression risk or depression severity.
Study 3 assessed the mediational role of patient self-care behavior associated with depression,
self-reported diabetes symptoms, and daily functioning. A theoretical model proposed that self-
care behaviors partially explained the significant effect of depression on diabetes symptoms. A
few observational studies have demonstrated empirical findings supporting this theory; however,
the same conceptual model has not been assessed in clinical trials. This study conducted
secondary analyses of the MDDP sample to test the mediational role of self-care behaviors,
including the frequency of healthy diet, exercise, SBGM, foot care, and BMI. For this study,
change in depression symptoms at 6 months post-baseline, change in self-care behavior at 12
months post-baseline, and change of self-reported diabetes symptoms and daily functioning at 18
months post-baseline were examined. Path analysis in the structural equation modeling scheme
and regression-based mediation analysis with the bootstrapping method were conducted to
examine unadjusted and adjusted mediation tests for each self-care behavior. In the tests for
unadjusted mediation effects in path analysis models of significant effects of change in diet on
diabetes symptoms (p < .01) and daily functioning (p < .05), no significant relationship was
found. In models controlled for potential confounders, decreased depression severity was
associated with increased exercise (p < .01), decreased self-reported diabetes symptoms (p <
6
.01), and increased daily functioning (p < .01). Exercise was associated with improved daily
functioning (p < .05), and SBGM was associated with reduced self-reported diabetes symptoms
(p < .05). As a result, only exercise significantly mediated the association between depression
and daily functioning (p < .05). Future clinical trials of the widely accepted conceptual model are
needed that use diverse measures of self-care behaviors, differentiating them by objective and
subjective constructs, and that examine biomarkers related to self-care behaviors, such as A1C.
This dissertation attempted to elucidate the value of depression care for diabetes patients.
Depression care for diabetes patients is often believed to improve self-care behaviors and
diabetes symptoms by reducing one of the cardinal symptoms: apathy. However, the belief rarely
relies on empirical evidence, and findings from recent studies provided contradicting
implications. I hope the findings in this dissertation provide the groundwork for more rigorous
empirical research.
7
Dedication
Social work students are always taught the importance of life changes and developmental
tasks. Many clients seeking help often experience recent major life events. Single moms would
not have sought help without having had a new baby. Government policies are constructed to
assist those people who just had life changes or major events. For instance, Medicare gained its
legitimacy due to older adults having more physical illnesses and less income. Also, the Program
for Women, Infants, and Children provides food to recipients to assist in healthy early life
developments of unborn and newborn babies. Thus, understanding the nature of life changes by
major developmental tasks is very important for social work students and social work researchers
to understand.
I am fortunate to have been given precious opportunities to struggle with some
developmental tasks: pregnancy, birth of baby, and raising kids during this PhD program. Unique
contexts came into play because I was an immigrant who had lower English proficiency, less
income as a PhD student, and increased stress coming from meeting requirements in an intensive
PhD program. As Dr. Leonard Pearlin described in his papers, the effect of objective stressors
was amplified by low resources related to my backgrounds and life domains, resulting in serious
consequences in mental and physical health. For instance, I had temporary blindness during the
second semester while I was overwhelmed by school courses and dealing with bureaucratic
systems involving health insurance for our baby’s delivery. I was frustrated by the fact that we
live in a world where even a very informed social work researcher fails to adeptly address crises
often followed by major developmental changes. I could feel our clients’ distress and frustration s
when they attempt to seek help from social workers.
8
I am glad to have gone through those crises during the past four years and meet all the
requirements in this program. These successes are attributable to people who have supported me
in many ways. In school, Ms. Malinda Sampson and Dr. Michalle Mor-Barak were very
instrumental when I needed to deal with personal and academic issues. Without their help, I
could not have successfully solved these issues. Also, my cohort, Anthony, James, Hsun-Ta, Liz,
and Amy, were very important friends who listened to me, relieving my emotional distress
related to school issues. Finally, Pey, Charli, Diana, and Drs. Soydan and Kaplan gave me very
critical assistance I needed for professional growth to be an independent scholar. With their help,
I could prevent making mistakes in data management, writing, and career planning. In addition,
many professors were helpful while I was developing research ideas and implementing research
plans. Particularly, Dr. Lawrence Palinkas was an important mentor on campus when Dr.
Kathleen Ell stayed home. I attribute my achievements in school to the assistance of these
people.
Also, many in my personal life were very instrumental in my completing the PhD
training. My friends stayed with me through all the good and bad news. They sympathized with
my predicaments and achievements. Also, my peers in church helped me go through the family
issues and my developmental tasks coming from the birth of my two children. Finally, I am very
appreciative to my family. My parents and parents-in-law should be acknowledged for this
achievement. They prayed for me so much, and their prayers were fundamental to my
achievements, as well as important as other types of support. Also, my brother-in-law and his
wife, Mr. and Mrs. Lee, and my cousin, Tae-Sung Oh, in Los Angeles helped my family and I
settle into this new country. Also, my sister, So-Jung Oh, has been a very good friend through
9
this doctoral program, as well as through my life. Without these people, I could not have finished
this program.
I would like to acknowledge some people who have been so important to me during the
past four years and probably in my whole life. First, Dr. Kathleen Ell will remain my hero for her
influence in my finding research topics and in modeling to be a researcher. Her life is a precursor
that triangulated her social work research and her personal life to improve human rights and the
well-being of underserved people. In addition, her life-threatening condition and age never
stopped her from seeking an innovative research agenda, which was driven by her steady passion
for helping needy families. I feel a great responsibility to spread her contributions to other
researchers and practitioners to facilitate better practice for low-income patients with chronic
illnesses and comorbid depression. However, the most important people are my wife, Jinhee, and
my kids, Celine and Clara. Without their supports and existence, I wouldn’t be studying the
topics I am working on. They gave me opportunities to understand the seriousness of life
stressors and the importance of social contexts for our clients. In addition, they loved me to the
extent that I had never experienced before, and this love was the best resource for me to be able
to focus on research.
10
Table of Contents
Abstract ................................................................................................................................1
Dedication ............................................................................................................................7
List of Tables .....................................................................................................................12
List of Figures ....................................................................................................................14
Chapter 1: Introduction ......................................................................................................15
Chapter 2: Depression Symptoms Change, Receipt of Problem-Solving Therapy, and
Self-Care Behaviors among Low-Income, Predominantly Hispanic Diabetes
Patients ...................................................................................................................23
INTRODUCTION .......................................................................................................23
METHODS ..................................................................................................................27
Sample and Procedure....................................................................................................... 27
Measures ........................................................................................................................... 33
Analyses ............................................................................................................................ 34
RESULTS ....................................................................................................................35
DISCUSSION ..............................................................................................................42
Chapter 3: Are Changes in Self-Care Behavior Frequency During Depression Care
Associated with the Risk of Depression?...............................................................48
INTRODUCTION .......................................................................................................48
METHODS ..................................................................................................................53
Sample and Procedure....................................................................................................... 53
Measures ........................................................................................................................... 56
11
Analysis............................................................................................................................. 58
RESULTS ....................................................................................................................59
DISCUSSION ..............................................................................................................67
Chapter 4. Mediation of Self-Care Behaviors in the Relationship between Depression,
Diabetes Symptoms, and Daily Functioning .........................................................73
METHODS ..................................................................................................................79
Sample and Procedure....................................................................................................... 79
Measures ........................................................................................................................... 82
Analysis............................................................................................................................. 84
RESULTS ....................................................................................................................86
DISCUSSION ..............................................................................................................96
Chapter 5. CONCLUSION ..............................................................................................101
REFERENCES ................................................................................................................111
12
List of Tables
Table Page
1. Demographic and clinical characteristics for analyzed samples and drop-
outs……………………………………………………………………………
29
2. Comparisons of demographic and clinical characteristics depending on PST
receipt status (N = 281)……………………………………………………………
31
3. Relationship among depression, PST receipt, and self-care behaviors at 12-
months follow-up……………………………………………….............................
37
4. Relationship among depression measured at 12-months follow-up, PST receipt,
and self-care behaviors at 18-months follow- up………..........................................
39
5. Relationship among depression measured at 12-months follow-up, PST receipt,
and self-care behaviors at 24-months follow- up………..........................................
41
6. Demographic and clinical characteristics depending on the extent of self-care
behavior changes………………………………………………………………......
55
7. Relationship between the extent of change in each self-care behavior, clinical
depression diagnosis, and the severity of depression……………………………...
61
8. Relationship between change of self-care behaviors during 12 months post-
baseline and depression outcomes at 24-month follow-up………………………
64
9. Relationship between self-care behaviors at 12 months and depression outcomes
at 24-month follow-up, controlled for baseline values…………………………....
66
10. Description for demographic and clinical variables for all patients, analyzed
patients, and excluded patients…………………………………………………….
81
13
11. Mean comparisons of self-care behaviors and self-reported diabetes symptoms
and daily functioning depending on the extent of change in depressive symptoms
during the first 6 months since enrollment in the MDDP (lower 50% vs. upper
50%) (N = 226)……………………………………………………………………
88
12. Evaluating the mediation of diet change (baseline - 12 months) in the association
between change of depression (baseline - 6 months) and self-reported diabetes
symptoms and functional impairment (baseline - 18 months)………………….....
91
13. Evaluating the mediation of exercise change (baseline - 12 months) in the
association between change of depression (baseline - 6 months) and self-reported
diabetes symptoms and functional impairment (baseline - 18 months)…………...
92
14. Evaluating the mediation of SBGM change (baseline - 12 months) in the
association between change of depression (baseline - 6 months) and self-reported
diabetes symptoms and functional impairment (baseline - 18 months)…………...
93
15. Evaluating the mediation of foot care change (baseline - 12 months) in the
association between change of depression (baseline - 6 months) and self-reported
diabetes symptoms and functional impairment (baseline - 18 months)…………...
94
16. Evaluating the mediation of BMI change (baseline - 12 months) in the
association between change of depression (baseline - 6 months) and self-reported
diabetes symptoms and functional impairment (baseline - 18 months)…………...
95
14
List of Figures
Figures Page
1. Titles and research models for individual studies.……………………………… 17
2. Study 1’s analytic models ……………………………….……………………… 25
3. Study 2’s analytic models ……...……………………………............................. 51
4. Path model of mediation analysis …………..……….......................................... 78
5. Unadjusted results from path analyses with each self-care behavior as a
mediator …………………………………….………..........................................
89
15
Chapter 1: Introduction
Hispanics are unequally affected by diabetes mellitus and its strongest predictor: obesity
(Braveman, Cubbin, Egerter, Williams, & Pamuk, 2010; Kirk et al., 2008; Narayan, Boyle,
Thompson, Sorensen, & Williamson, 2003). According to a meta-analysis, Hispanics were
estimated to have 0.46% higher A1C than other racial/ethnic groups, and this difference was
consistent in studies within different settings and diverse analysis methods (Kirk et al., 2008).
The unequal distribution is exacerbated by low socioeconomic status (SES). Low SES often
results in limited health care access and failure to engage in health-promoting behaviors
(Bowser, Utz, Glick, & Harmon, 2010; Cusi & Ocampo, 2011; Narayan et al., 2003; U.S. Census
Bureau, 2011; Umpierrez, Gonzalez, Umpierrez, & Pimentel, 2007). Hispanic patients with
diabetes are less likely to visit physicians and maintain self-care management behaviors, such as
diet, exercise, and self-blood glucose monitoring (SBGM) (Ali, Stone, Peters, Davies, & Khunti,
2006; Cusi & Ocampo, 2011; Skaff, Mullan, Fisher, & Chesla, 2003). In a population-based
study, Mexican American’s risk of retinopathy was twice that of non-Hispanic Whites (Harris,
Klein, Cowie, Rowland, & Byrd-Holt, 1998). Poor diabetes management appears to increase
financial burden for Hispanics, whose average income is lower than other racial/ethnic groups
(U.S. Census Bureau, 2011). In 2007, the direct cost of diabetes was estimated $116 billion, and
its indirect cost, including loss of productivity, functional impairment, and lower extremity
amputation, was estimated at $58 billion a year (Center for Disease Control and Prevention,
2012). These data demonstrate a strong need to enhance patient self-care management.
Comorbid depression is prevalent among diabetes patients, and its rate is 1.77 times
higher than that among healthy controls (Ali et al., 2006; Pan et al., 2010; Stuart & Baune,
2012). Depression is often recurrent due to social and emotional stress in managing diabetes
16
(Fisher et al., 2008). Depression is associated with adverse health outcomes, including diabetes
complications, disability, and all-cause and coronary heart disease mortality (Black, Markides, &
Ray, 2003; Egede, Nietert, & Zheng, 2005; Egede & Hernández-Tejada, 2013; Huang et al.,
2012; Von Korff et al., 2005a). Black et al. (2003) demonstrated significantly increased hazard
ratios for macro- and microvascular complications, disability, and mortality among groups with
diabetes and either minimal or minor depression.
Depression is believed to contribute to undermined self-care behaviors and to be
responsible for adverse health outcomes. Katon (2003) proposed that self-care behaviors may
serve mediating roles in the relationship between depressive and diabetes symptoms and daily
functioning (p. 218). Decreased motivation to follow prescribed self-care may increase
depression symptoms. Depressed patients often lack satisfying work and leisure activities, are
less optimistic, and feel helpless (Detweiler-Bedell, Friedman, Leventhal, Miller, & Leventhal,
2008; Gonzalez et al., 2008a; Katon, 2003). Moreover, depressed diabetes patients are less likely
to follow overall self-care behaviors, appointment keeping, diet, medication, exercise, and
SBGM (Gonzalez et al., 2008a). Depression also undermines patient motivation to engage in
self-care behaviors prescribed by physicians and dieticians (Katon et al., 2009b; Katon, 2008).
For instance, the significant relationship between depression and reduced medication adherence
measured after 3 months was significantly explained by the level of self-efficacy (Schoenthaler,
Ogedegbe, & Allegrante, 2009). Many observational studies suggest a significant association
between depressive symptoms and recommended diabetes management behaviors
(Ciechanowski, Katon, & Russo, 2000; Ciechanowski, Katon, Russo, & Hirsch, 2003; Egede,
Ellis, & Grubaugh, 2009; Gonzalez et al., 2007; Gonzalez et al., 2008a; Gonzalez et al., 2008b;
Lin et al., 2004; Wagner, Tennen, & Osborn, 2010). Also, the severity of depression and the
17
extent of adherence to self-care behaviors had a dose-response relationship across exercise,
healthy diet amount, and diet choices (Ciechanowski et al., 2003). Reducing depression
symptoms may improve self-care behaviors.
However, there is little evidence that depression care can effectively facilitate self-care
behaviors (Heckbert et al., 2010; Lin et al., 2006; Markowitz, Gonzalez, Wilkinson, & Safren,
2011). Findings in observational studies suggest that improved self-care behaviors and quality of
life indicators should be simultaneously observed in groups with lower depressive symptoms if
Katon’s (2003) assumption is correct. For example, the adherence to exercise and diet were not
different between intervention and enhanced usual care (EUC) groups, whereas intervention
group patients experienced more than a 50% decrease in depression severity and higher mean
scores in self-rated quality of life and global functioning (Katon et al., 2010). Another study also
showed no improvement of healthy diet and recommended exercise frequency in an intervention
group that showed significantly improved depression (Lin et al., 2006). Several explanations for
inconsistent findings have been reported (Detweiler-Bedell et al., 2008; Lustman, Griffith,
Freedland, Kissel, & Clouse, 1998; Markowitz et al., 2011).
Efforts to understand the complex relationships among depression, self-care behaviors,
and quality-of-life indicators are necessary to develop better depression care. For this goal, there
are three research questions this doctoral dissertation wanted to examine. To empirically
examine these questions, secondary analyses were conducted with data in the Multifaceted
Diabetes and Depression Program (MDDP), a randomized clinical trial to test the effectiveness
of a socio-culturally adapted collaborative depression model for low-income Hispanic patients
with diabetes in public safety-net clinics (Ell et al., 2009). Collaborative depression care was
originally developed to increase detection of depression among primary care patients and to
18
increase patient’s adherence to depression care by offering both antidepressant medication (AM)
and evidence-based problem-solving therapy (PST) in many collaborative depression care trials
(Hegel et al., 2002; Katon, Unützer, Wells, & Jones, 2010; Unützer et al., 2002). Collaborative
depression team care was found to be effective in reducing depressive symptoms and increasing
daily functioning among racial/ethnic minorities (Bao et al., 2011; Miranda et al., 2003) and
diabetes and cancer patients in safety-net clinics (Ell et al., 2009; Ell et al., 2010; Ell et al., 2011;
Ell et al., 2008). In addition to the efficacy on clinical outcomes, two cost-benefit analyses
demonstrated a good level of cost-benefit ratio of this depression care (Hay, Katon, Ell, Lee, &
Guterman, 2012; Simon et al., 2007).
Figure 1. Titles and research models for individual studies.
19
Figure 1 demonstrates a title and a scheme for a research model in these three individual
studies. The first MDDP study aimed to examine whether depression remission among patients
who received PST was associated with increased self-care behavior frequency. In an earlier
observational study, higher depressive symptoms significantly predicted less frequent self-care
behaviors, prospectively (Gonzalez et al., 2008b). However, this finding did not tell us whether
decreased depressive symptoms, resulting from depression care is associated with change in self-
care behavior frequency. Also, it is intriguing to examine the effect of PST on self-care
frequency.
During PST sessions in a collaborative depression care model, patients learn how to
clearly understand problems in the social and interpersonal domains, generate better alternatives,
optimally assess possible consequences, and carry out a solution effectively and systematically
(Areán, Hegel, & Reynolds III, 2001; Nezu, Nezu, Felgoise, McClure, & Houts, 2003; Nezu,
Nezu, & D'Zurilla, 2013). PST trains patients to have rational, positive, and constructive set or
cognitive appraisals to problems that routinely arise (Nezu et al., 2003; Nezu et al., 2013).
Although there are substantial differences across models, trainings targeting problem-solving
skills have been used often to address self-care behaviors (Areán et al., 2001; Hill-Briggs, 2003;
Hill-Briggs & Gemmell, 2007). This secondary analysis focused on actual receipt of PST, rather
than artificial assignment to the intervention study arm that offered PST. Clinical studies often
report outcomes by groups that were assigned randomly at enrollment. Although patients
assigned to an intervention group are offered depression treatment, not all patients actually
receive the treatment (Ell et al., 2010; Katon et al., 2004a; Katon et al., 2010). Thus, participating
in PST may increase the self-care behavior frequency.
20
Finally, it is not clear whether decreased depressive symptoms after receiving PST has a
different effect on self-care behavior frequency compared with other patients. We hypothesized
that depressive symptoms decline occurring in a group receiving PST would demonstrate more
frequent self-care behaviors because of PST’s focus on the facilitation of successful diabetes
management. To empirically test this hypothesis, the interaction effect between depression
change and receipt of PST was examined in a regression model, controlling for demographic and
clinical confounders.
As the second study, it is important to know whether more frequent self-care behavior,
often triggered by participating in a diabetes management training, is associated with future
depressive symptoms. Depression care clinical trials often demonstrated improved self-care
behaviors in an intervention group as compared with baseline values, although the improvement
did not reach statistical significance compared with a control group (Katon et al., 2010; Lin et al.,
2006; Williams et al., 2004). Therapeutic components designed to activate self-care behaviors
are often taught while depressive symptoms are treated (Baumeister, Hutter, & Bengel, 2012;
Katon et al., 2010; Lamers, Jonkers, Bosma, Knottnerus, & van Eijk, 2011). Since emotional
distress related to diabetes management influences depressive symptoms (Gonzalez, Fisher, &
Polonsky, 2011), therapists often help patients solve the distress through developing action plans
for adequate diabetes management (Baumeister et al., 2012; Ell et al., 2009; Katon et al., 2010).
As a result, patients often demonstrate increased frequency of self-care behaviors (Katon et al.,
2010; Lin et al., 2006; Williams et al., 2004). However, it is plausible that a more frequent
schedule for self-care behaviors may burden patients because they are now being asked to
commit more time and energy. Keeping up with intensive self-care behaviors appears to require
self-regulation, which has limited capacity (Muraven & Baumeister, 2000). Particularly, when
21
environmental support has not changed, patients with increased self-care behavior frequency
might face situations where they use more self-regulation resources to keep up with changes in
their self-care behaviors. According to a survey among certified diabetes educators, diabetes
patients may need up to 2 hours a day to follow recommended self-care behaviors by the
American Diabetes Association (Russell, Suh, & Safford, 2005). The changed expectation on
self-care behavior frequency might cost emotional well-being by depleting limited self-control
resources (Muraven & Baumeister, 2000). Thus, it is plausible that a change in self-care behavior
frequency during depression care may negatively affect depression outcomes. However, it is also
possible that increased self-care behaviors might predict a reduction in depressive symptoms. For
instance, exercise is a significant therapy for healthy depression patients, and studies
demonstrated significantly improved diabetes symptoms and depression among diabetes patients
(Antero Kesaniemi et al., 2001; Detweiler-Bedell et al., 2008). Also, improved self-care
behaviors may lower the risk of diabetes complications and disability, and this effect may also
decrease diabetes-specific emotional distress, which is related to depressive symptoms (Renn,
Feliciano, & Segal, 2011). In other words, better diabetes outcomes resulting from better self-
care behaviors may reduce concerns with diabetes management. However, empirical assessment
is needed for the possible effect of more frequent self-care behavior on the risk of depression,
concurrently and prospectively. The second study examined this research question through
secondary analysis with data in the MDDP.
Finally, the third study aimed to test a possible mediating role of self-care behaviors in
the association between depression and health outcomes. Many depression care trials assumed
the conceptual linear path model in Katon (2003), which proposed that depression severity
decline may contribute to increased self-care behaviors and also results in better diabetes and
22
functioning outcomes. In other words, more frequent self-care behavior would link the improved
depressive symptom triggered by depression care with the observed decreased diabetes
symptoms and better physical functioning in clinical trials. This assumed mediation path appears
persuasive when findings in observational studies were examined (e.g., American Diabetes
Association, 2013; Gonzalez et al., 2008a). However, when clinical trial’s data were examined,
empirical evidence was too immature to validate this belief. Intervention groups in clinical trials
consistently showed notable improvements in severity and remission rate of depression, as well
as quality of life indicators and self-reported diabetes symptoms compared with control groups,
whereas significantly more frequent self-care behaviors were inconsistently observed in the same
trials (Ell et al., 2010; Katon et al., 2004a; Katon et al., 2010; Lin et al, 2006; Lustman et al.,
1998; Morgan et al., 2013; Williams et al., 2004). To validate Katon’s (2003) path model,
regression analysis based on methodological methods in Baron and Kenny (1986) becomes
necessary. Thus, Study 3 examined the mediating roles of self-care behaviors by using mediation
analysis with the bootstrapping method.
23
Chapter 2: Depression Symptoms Change, Receipt of Problem-Solving Therapy, and Self-Care
Behaviors among Low-Income, Predominantly Hispanic Diabetes Patients
INTRODUCTION
Self-care behaviors are highly recommended to patients with diabetes patients (American
Diabetes Association, 2013). Self-care behaviors include self-managing efforts not only for
keeping one’s blood glucose level in the recommended range (e.g., regular exercise, SBGM, and
adhering to medication), but also for preventing possible complications (e.g., foot care)
(American Diabetes Association, 2013; Castaneda et al., 2002; Polonsky et al., 2011). Good self-
care behaviors are associated with better diabetes outcomes (Goldhaber-Fiebert, Goldhaber-
Fiebert, Tristan, & Nathan, 2003). For instance, a group that learned structured SBGM
demonstrated 1.2% lower A1C levels when compared to its baseline A1C level among non-
insulin treated patients (Polonsky et al., 2011). Six-month resistance training for strengthening
muscles also showed a 1.2% decline in A1C levels at the 6-month assessment (Castaneda et al.,
2002). A prospective study that followed diabetes patients up to 10 years demonstrated that an
intensive self-care behaviors intervention had significant effects on weight, waist circumference,
physical fitness, and A1C level (Look AHEAD Research Group, 2013).
Depression has a significant association with overall self-care behaviors among diabetes
patients (Ciechanowski et al., 2003; Heckbert et al., 2010; Ludman et al., 2004; Schoenthaler et
al., 2009). A meta-analysis demonstrated that depressed diabetes patients were less likely to
follow overall self-care behaviors (Gonzalez et al., 2008a). Similar findings were found in
studies of the general population; for example, a 15% higher prevalence of sedentary lifestyle
was found when depression was diagnosed (Egede et al., 2005). Katon (2003) proposed a
conceptual path that self-care behaviors, including exercise, diet, and “monitoring symptoms or
24
signs of exacerbation,” are affected by depressive symptoms, resulting in undermined diabetes
(p. 218). Decreased motivation to follow prescribed self-care behaviors relates to the role of
certain cardinal symptoms in depression, including lacking interest in usual jobs, decreased
pleasure, and helplessness (Gonzalez et al., 2008a; Katon, 2003). Depression may undermine
patients’ motivation or self -efficacy to engage in self-care behaviors that are usually prescribed
by primary care physicians and dieticians (Katon et al., 2009b; Katon, 2008). Therefore, many
believe that decreased depressive symptoms are necessary for better self-care behaviors and
measured the extent of self-care behaviors as one of secondary outcomes in trials investigating
evidence-based depression care (Baumeister et al., 2012; Ell et al., 2010; Katon et al., 2004a;
Katon et al., 2010). In other words, it has been hypothesized that patients may demonstrate
increased self-care behaviors frequency once they have reduced depressive symptoms resulting
from depression care.
There are two mechanisms by which self-care behaviors are influenced by depression
care. First, as introduced above, reduced depressive symptoms may co-occur with increased
motivation to execute daily self-care behaviors. Following recommended self-care behaviors was
estimated to require more than two hours per day (Russell et al., 2005). To regularly meet this
burdensome daily task, depression care should reduce apathy associated with depression.
Another mechanism is that many psychotherapy treatments for depression teach patients to
overcome barriers for successful diabetes management because stress from meeting
recommended self-care behaviors is a significant factor contributing to depression (Gonzalez et
al., 2011). Thus, it is important to help patients identify social stressors and develop action plans
to remove diabetes-related stressors contributing to existing depressive symptoms. PST is a good
example of an activity that assists patients in understanding their problems in the social and
25
interpersonal domains, generating better alternatives, assessing possible positive and negative
consequences, and carrying out a solution effectively (Nezu et al., 2003; Nezu et al., 2013). PST
trains patients to have rational, positive, and constructive set or cognitive appraisals to problems
that routinely arise (Nezu et al., 2003; Nezu et al., 2013). As a result, patients can develop more
feasible action plans to successfully manage self-care behaviors and functional impairment
related to diabetes. Improved self-care behaviors may derive from the systematic ways to address
stressful barriers related to diabetes management (Hill-Briggs & Gemmell, 2007).
Nevertheless, it is not clear whether reductions in depression severity resulting from
psychotherapy for depression influenced future self-care behavior outcomes. Contrary to the
ample evidence in observational studies (Ciechanowski et al., 2003; Egede et al., 2009; Gonzalez
et al., 2007; Gonzalez et al., 2008a; Gonzalez et al., 2008b; Lin et al., 2004; Wagner et al., 2010),
limited evidence is available to determine whether decreased depression resulting from exposure
to PST lead to improved self-care behaviors. Most evidence has come from intervention studies
that reported mean differences of outcomes between intervention and comparison groups (Ell et
al., 2010; Katon et al., 2004a; Katon et al., 2010). Although significantly superior depression
outcomes were observed in an intervention group, no significant differences in adherence to
recommended diet and exercise plan were found (Lin et al., 2006). However, the null results are
not necessary to answer the question whether alleviated depression severity and having PST
increase self-care behavior frequency. First, null results in independent two-group mean
comparisons do not give data to examine whether decreased depression and use of PST relate to
increased self-care behaviors. To test the significance of correlations, regression analysis is a
more suitable analysis method. In addition, randomized clinical trials may not take into account
the extent of decreased depression and actual receipt of PST because intent-to-treat (ITT)
26
analysis includes all participants enrolled at baseline assessment, regardless of intervention
receipt and responding to following assessments.
Therefore, by using regression analysis, we investigated whether a change of depression
during post-baseline and actual receipt of PST are associated with the frequencies of self-care
behaviors, including diet, exercise, SBGM, foot care, and body mass index (BMI; Toobert &
Glasgow, 1994; Toobert, Hampson, & Glasgow, 2000). Figure 2 demonstrates the proposed
analytic models.
Figure 2. Study 1’s analytic models .
To measure the frequencies of the four self-care behaviors, the Summary of Diabetes
Self-Care Activities (SDSCA) was used. This measure has been used in depression intervention
studies that demonstrated no statistically significant differences in self-care behaviors between
the intervention and control groups (Ell et al., 2010; Lin et al., 2006; Williams et al., 2004). BMI
often reflects the quality of self-care behaviors and relates to diabetes outcomes (Chiu, Wray,
Beverly, & Dominic, 2010; Lin et al., 2006). Also, BMI measures the size of food portions,
which cannot be measured with the healthy diet subscale in the SDSCA. For this particular
study, a secondary analysis with a sample of patients participating in a study by Ell et al. (2010;
27
2011) was conducted. In the previous study, at 12-month assessment, 40.3% and 35.0% of
intervention and EUC groups were found to have depression remission, measured with the
Hopkins Symptom Checklist (Derogatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974; Ell et al.,
2010). Our study examined whether changes in depression during12 months since baseline and a
history of PST receipt after enrollment predicted increases in self-care behaviors and changes in
BMI, concurrently and prospectively. In addition, this study examined whether decreased
depressive symptoms among patients having PST have multiplicative effects on the frequencies
of self-care behaviors and BMI change. For this objective, interaction terms between these two
predictors were examined.
METHODS
Sample and Procedure
This study conducted a secondary analyses with data in the MDDP, which tested the
effectiveness of socio-culturally adapted depression care for low-income patients using safety-
net clinics located in a metropolitan area. Screenings for clinical depression was done for 1,803
eligible patients with the Patient Health Questionnaire-9 (PHQ-9); 27.3% (N = 492) of the total
eligible patients were asked to participate in the MDDP, and 387 patients provided informed
consent (Ell et al., 2011). Patients were randomly assigned into either EUC or collaborative
depression care (the intervention [INT] group) groups. The INT group received either
antidepressant medications or PST. Ell et al. (2010) demonstrated up to 58.5% and 79.3% of
patients in the INT received antidepressants and PST or other counseling services during the 18-
month follow-up. However, because of a possible spill-over effect, up to 32.5% of the EUC
reported they had received some depression treatment during the first 12 months of the study
period (Ell et al., 2010). Patients assigned to the EUC group were offered psycho-educational
28
materials for depression and diabetes management, and the results from depression screening
were sent to their primary care physician in the safety-net clinic. Because primary care
physicians had patients in these two groups, some form of depression care in the EUC was
possible. Due to the spill-over effect, significant patients in the EUC also demonstrated
depression remission at follow-ups (Ell et al., 2011). In addition, intervention studies tend to
have smaller sample sizes due to higher marginal costs for additional recruitment. Therefore, this
particular study analyzed data from the two study groups.
The MDDP collected data at baseline and at 6, 12, 18, and 24 months after the baseline
assessment. This secondary analysis analyzed patients who participated in 12-, 18-, and 24-
month follow-ups. Specifically, this study examined whether depression remission status predicts
the extent of self-care behaviors measured at 12-, 18-, and 24- month follow-up assessments.
Table 1 showed that 281, 249, and 235 patients were analyzed for each analysis and
demonstrates demographic and clinical characteristics of patients who were analyzed or dropped
out. The attrition rates of these follow-up assessments are 27.4%, 35.9%, and 39.3%,
respectively. To find possible factors associated with attrition, chi-square tests were conducted.
Patients with shorter years of education were more likely to participate in the 18-month
assessment (p < .05). Also, patients speaking only Spanish were also more inclined to participate
in the 12- and 18-month assessments (p < .05). Finally, patients who responded to the 24-month
follow-up demonstrated significantly higher scores in SBGM at baseline (p < .05). Other than
these characteristics, no statistically significant differences were found.
29
Table 1
Demographic and clinical characteristics for analyzed samples and drop-outs
12 months 18 months 24 months
Analyzed
(N = 281)
Drop-outs
(N = 106)
p Analyzed
(N = 249)
Dropouts
(N = 139)
p Analyzed
(N = 235)
Dropout
(N = 152)
p
N (%)
Mean (SD)
N (%)
Mean (SD)
N (%)
Mean (SD)
Age ≥50 206
(73.3%)
73 (68.9%) .39 183 (73.5%) 96 (69.6%) .41 170 (72.3%) 109
(71.7%)
.89
Intervention
group
142
(50.5%)
51 (48.1%) .67 128 (51.4%) 65 (47.1%) .42 124 (52.8%) 69 (45.4%) .16
Lower than
high school
a
236
(84.0%)
81 (76.4%) .08 211 (84.7%) 106 (76.8%) .05 194 (82.6%) 123
(80.9%)
.68
Some college
or graduates
a
13 (4.6%) 9 (8.5%) .14 11 (4.4%) 11 (8.0%) .15 12 (5.1%) 10 (6.6%) .54
Female
234
(83.3%)
84 (79.2%) .36 206 (82.7%) 112 (81.2%) .70 198 (84.3%) 120
(78.9%)
.18
Separated
b
104
(37.0%)
45 (42.5%) .33 91 (36.5%) 58 (42.0%) .29 85 (36.2%) 64 (42.1%) .24
Never married
b
35 (12.5%) 12 (11.3%) .76 30 (12.0%) 17 (12.3%) .94 29 (12.3%) 18 (11.8%) .88
Dysthymia
158
(56.2%)
56 (52.8%) .55 141 (56.6%) 73 (52.9%) .48 132 (56.2%) 82 (53.9%) .67
Only Spanish
c
244
(86.8%)
83 (78.3%) .04 217 (87.1%) 110 (79.7%) .05 204 (86.8%) 123
(80.9%)
.12
On insulin
treatment at
baseline
78 (27.8%) 29 (27.4%) .94 71 (28.5%) 36 (26.1%) .61 65 (27.7%) 42 (27.6%) 1.00
Baseline
PHQ-9
14.59 (2.88) 15.05 (3.13) .18 14.62 (2.87) 14.89 (3.10) .39 14.57 (2.85) 14.94
(3.10)
.23
Diet at
baseline
4.25 (1.74) 4.34 (1.90) .64 4.22 (1.73) 4.37 (1.89) .43 4.25 (1.71) 4.30 (1.91) .80
Exercise at
baseline
2.03 (2.37) 1.87 (2.34) .55 2.03 (2.35) 1.90 (2.38) .58 2.04 (2.33) 1.90 (2.41) .57
SBGM at
baseline
2.77 (2.82) 2.35 (2.75) .19 2.75 (2.79) 2.48 (2.82) .36 2.88 (2.86) 2.30 (2.69) .05
Foot care at
baseline
4.99 (2.79) 4.81 (2.86) .57 5.05 (2.74) 4.74 (2.92) .30 4.96 (2.80) 4.92 (2.83) .87
BMI at
baseline
32.95 (7.44) 32.78 (7.88) .85 32.96 (7.34) 32.79 (7.96) .83 33.10 (7.76) 32.60
(7.23)
.53
30
Table 2 demonstrates differences in demographic and clinical characteristics at baseline
by actual receipt of PST; 45.9%, 46.6%, and 48.1% patients participated in 12-, 18-, or 24-month
follow-ups, respectively, and received PST during 12 months after the baseline assessment. With
respect to differences in demographic and clinical variables, patients with a history of PST
receipt were likely to have at least a high school education (p < .05), show dysthymia at baseline
(p < .01), be an English speaker (p < .01), have increased depressive symptoms at baseline (p
< .05), and exhibit a smaller frequency of weekly exercise at baseline (p < .05). No other
significant difference in self-care behavior frequencies and BMI depending on the history of PST
were found.
31
Table 2
Comparisons of demographic and clinical characteristics depending on PST receipt status (N =
281)
PST receipt during 12 months post-baseline
Yes No p
N (%)
Mean (SD)
N (%)
Mean (SD)
Age ≥50
a
100 (77.5%) 106 (69.7%) .14
Lower than high school
99 (76.7% ) 137 (90.1% ) .002
Some college or graduates
11 (8.5% ) 2 (1.3% ) .004
Female
a
104 (80.6%) 130 (85.5%) .27
Separated
49 (38.0%) 55 (36.2%) .76
Never married
17 (13.2%) 18 (11.8%) .74
Dysthymia
a
82 (63.6% ) 76 (50.0% ) .02
Only Spanish
a
102 (79.1% ) 142 (93.4% ) .00
On insulin treatment at baseline
a
39 (30.2%) 39 (25.7%) .39
On insulin treatment at 12 months
a
46 (35.7%) 42 (27.6%) .15
On insulin treatment at 18 months
b
44 (37.9%) 42 (31.6%) .29
On insulin treatment at 24 months
c
45 (39.8%) 39 (32.0%) .21
Depression at baseline
a
15.02 ± 3.00 14.24 ± 2.74 .02
Depression at 12 months
a
7.07 ± 5.58 7.97 ± 5.47 .18
Diet at baseline
a
4.12 ± 1.73 4.36 ± 1.76 .26
Diet at 12 months
a
4.37 ± 1.59 4.34 ± 1.52 .41
Diet at 18 months
b
4.45 ± 1.52 4.35 ± 1.89 .59
Diet at 24 months
c
4.42 ± 1.73 4.37 ± 1.82 .83
Exercise at baseline
a
1.91 ± 2.33 2.13 ± 2.40 .44
Exercise at 12 months
a
2.02 ± 2.09 2.61 ± 2.36 .03
32
Exercise at 18 months
b
2.38 ± 2.44 2.97 ± 2.56 .07
Exercise at 24 months
c
2.66 ± 2.40 3.11 ± 2.37 .15
SBGM at baseline
a
2.79 ± 2.86 2.75 ± 2.79 .91
SBGM at 12 months
a
2.86 ± 2.53 2.68 ± 2.62 .56
SBGM at 18 months
b
3.37 ± 2.86 3.02 ± 2.76 .34
SBGM at 24 months
c
3.11 ± 2.79 2.94 ± 2.65 .64
Foot care at baseline
a
5.24 ± 2.60 4.78 ± 2.93 .17
Foot care at 12 months
a
4.85 ± 1.81 4.61 ± 2.17 .33
Foot care at 18 months
b
5.41 ± 1.94 5.00 ± 2.23 .12
Foot care at 24 months
c
5.39 ± 1.91 5.03 ± 2.19 .18
BMI at baseline
d
33.32 ± 7.56 32.64 ± 7.35 .45
BMI at 12 months
d
33.19 ± 7.44 32.62 ± 7.43 .53
BMI at 18 months
e
32.96 ± 7.44 32.34 ± 7.51 .52
BMI at 24 months
f
33.68 ± 7.77 32.03 ± 7.84 .11
Note:
a
N = 281: PST (Yes = 129, No = 152);
b
N = 249: PST (yes = 116, no = 133);
c
N = 235:
PST (yes = 113, no = 122);
d
N = 279: PST (Yes = 128, No = 151);
e
N = 247: PST (yes = 115,
no = 132);
f
N = 231: PST (yes = 110, no = 121)
33
Measures
1. Depression
Depression was measured by the PHQ-9 (Löwe, Kroenke, Herzog, & Gräfe, 2004)). The
PHQ-9 was originally developed to be used as a brief self-reported measure for patients using
primary care clinics (Kroenke, Spitzer, & Williams, 2001). It was administered to 6,000 patients
from 15 primary care clinics located nationally and was found to have adequate psychometric
properties (Kroenke et al., 2001). The PHQ-9 was also tested with various groups, including
older adults (Chen et al., 2010), the general population (Martin, Rief, Klaiberg, & Braehler,
2006), and university students (Adewuya, Ola, & Afolabi, 2006). The PHQ-9 has nine items
describing criteria for DSM-IV diagnosis. Respondents are asked to choose the frequency of
selected symptoms during the last 2 weeks, ranging from 0 (not at all) to 3 (nearly every day)
(Löwe et al., 2004). The severity of depression is found by adding up the responses to these nine
items, resulting in range from 0 (minimal depressive symptoms) to 27 (highest depressive
symptoms).
This measure has been used in randomized clinical trials (RCTs) that investigate the
effectiveness of collaborative depression care and other psychological and pharmacological
treatments addressing depression or general emotional distress (Katon et al., 2010; Katon et al.,
2010; Park, Katon, & Wolf, 2013). In these trials, the PHQ-9 has demonstrated significant
changes after the implementation of an intervention, suggesting good responsiveness to treatment.
2. Self-care behaviors
The self-care behaviors were measured with the SDSCA (Toobert et al., 2000). The
SDSCA asks the extent of daily adherence to specific self-care behaviors that are recommended
to diabetes patients. Patients provide information on these behaviors in the past week. This
34
measure has 11 items located in six types of self-care behaviors: a) general healthy eating (2
items), b) specific healthy eating (2 items), c) physical activities (2 items), d) regular SBGM (2
items), e) foot care (2 items), and f) smoking (1 item). Respondents are asked to rate each item
from 0 (never) to 7 (every day during the past week).
Aggregation of scores across subdomains are not recommended due to lower inter-item
correlation (Toobert et al., 2000). Another study also demonstrates apparently varied values of
inter-item correlation (Vincent, McEwen, & Pasvogel, 2008). Although the internal consistency
of the SDSCA was .68 for all items, item-to-total correlation ranged from -.22 (high-fat diet)
to .58 (SBGM). Of note, a negative correlation was found in an item asking about the
consumption of a high-fat diet, which seems consistent with the finding that more active diabetes
patients experience occasions when they are offered foods with high fatty contents and high in
carbohydrates (Early, Shultz, & Corbett, 2009). The SDSCA have been used in observation
studies (Ciechanowski et al., 2000; Ciechanowski et al., 2003; Von Korff et al., 2005b) and in
RCTs (Ell et al., 2010; Katon et al., 2010; Lin et al., 2006) for addressing depression in diabetes
patients.
Analyses
To examine the proposed association among change in self-care behavior frequency and
change of depressive symptoms during 12 months since baseline, the history of PST receipt, and
the multiplicative effect, a series of regression analyses were conducted. To keep temporal order
between the predictors and the outcomes, three sets of regression analyses were developed,
including a cross-sectional analysis and two prospective analyses. Specifically, self-care
behavior frequencies were measured at either 6 or 12 months after depression data were
regressed. Thus, self-care behavior frequency at 18- and 24-month follow-ups were used. With
35
these three analytic models, we assessed the length of lagged time between change in depression
severity and change in self-care behavior frequency. A significant association in the cross-
sectional analysis would suggest that changes in depression and self-care behavior frequency
may occur simultaneously. From the findings in the prospective analyses, we are able to know
whether any time lag between depressive symptoms change and activation of self-care behaviors
exist. For each set of regression analyses, five dependent variables were examined including
weekly frequency of healthy diet, exercise, SBGM, foot care, and BMI change since baseline
assessment. For testing the interaction term between change in depression and the history of PST
receipt, centering was conducted for the depression variable with the mean values at baseline and
12-month follow-up. We made the centering of the depression variable with the unadjusted mean
value for easy interpretation of parameters for the interaction term. Each model controlled for
possible confounders including age, study arm, education, gender, marital status, dysthymia,
baseline PHQ-9 score, language, and insulin treatment at baseline. These confounders were
selected because self-care behaviors appear to be affected by these variables. The alpha level .05
was used for statistical significance, and the alpha level .10 was also indicated in tables. For
these statistical analyses, SPSS 21.0 (Chicago, IL, USA) was used.
RESULTS
Table 3 demonstrates results from a cross-sectional study examining the proposed
relationship between change in depressive symptoms, PST receipt, interaction term between the
two predictors, and change in self-care behavior frequency and BMI change since baseline after
controlling for confounding variables. The proposed models explained 8% (Foot care) to 87%
(BMI) of variance for change occurred during 12 months after the baseline assessment. With
respect to key findings, the baseline value of each self-care behavior and BMI explained the
36
significant amount of variance at the 12-month assessment (p < .05), suggesting a strong
predictive power of previous levels of self-care behaviors. Notably, decreased depressive
symptoms were only associated with a more frequent consumption of a healthy diet (p < .01) and
recommended foot care (p < .05). In other words, patients with larger decreases in depression
reported significantly increased frequency in healthy diet and foot care during the same period.
However, contrary to prediction, PST receipt was not a significant predictor for changes in any
self-care behaviors and BMI during 12 months, possibly losing effect sizes by changes in
depressive symptoms controlled in the same model. Finally, a significant interaction term was
found in a model predicting the frequency of foot care (p < .05). In other words, patients, who
had received PST, engaged in recommended foot care 0.1 day per week more compared with
those who had not had the same psychotherapy when the extent of decline in depressive
symptoms is same.
37
Table 3
Relationship among depression, PST receipt, and self-care behaviors at 12-months follow-up
Unstandardized coefficient (S.E .)
Diet
(n = 281)
Exercise
d
(n = 281)
SBGM
d
(n = 281)
Foot care
d
(n = 281)
BMI
(n = 279)
Constant 3.44 (.44)*** .50 (.09)*** .31 (.09)** .61 (.07)*** 1.92 (1.14)
Age ≥50 .31 (.20) -.06 (.04) .04 (.04) .01 (.03) -.37 (.41)
Lower than high school
a
-.34 (.30)* -.02 (.06) .00 (.06) .01 (.05) 1.04 (.59)
Some college or graduates
a
-.60 (.49) -.02 (.10) .10 (.10) -.03 (.07) .96 (.95)
Female
-.02 (.24) -.08 (.05) -.04 (.05) -.04 (.04) -.43 (.48)
Separated
b
.02 (.19) .07 (.04) .01 (.04) .04 (.03) .26 (.38)
Never married
b
-.45 (.27) .01 (.06) .01 (.06) .02 (.04) .33 (.54)
Dysthymia
-.32 (.18) -.07 (.04) -.01 (.04) .01 (.03) .51 (.36)
Baseline PHQ-9 .01 (.03) -.00 (.01) -.00 (.01) -.00 (.01) -.13 (.06)*
Only Spanish
c
.50 (.30) .06 (.06) -.06 (.06) -.01 (.05) -.53 (.60)
On insulin treatment at 12
months
-.09 (.19) -.09 (.04)* .11 (.04)** .02 (.03) .00 (.37)
Baseline value for
dependent variable
.29 (.05)*** .17 (.05)** .37 (.05)*** .13 (.04)** .93 (.02)***
PHQ-9 at 12 month -.07 (.02)** -.01 (.01) -.00 (.01) -.01 (.00)* .01 (.04)
PST receipt -.08 (.18) -.05 (.04) .02 (.04) .03 (.03) -.11 (.36)
PHQ-9 at 12 month X PST
receipt
.03 (.03) .00 (.01) .01 (.01) .01 (.01)* -.06 (.06)
Explained variance R
2
= .23 R
2
= .12 R
2
= .24 R
2
= .08 R
2
= .87
R-square increase due to
interaction
△R
2
= .00 △R
2
= .00 △R
2
= .00 △R
2
= .01 △R
2
= .00
*p < .05; **p < .01; ***p < .001.
Note.
a
Reference group: high school diploma;
b
Reference group: currently married;
c
Reference group: English only
speaking or bi-lingual patients;
d
Variable was log transformed to satisfy the normality of residual.
38
Table 4 shows results from analyses where independent variables predicted self-care
behaviors measured 6 months afterward and at an18-month follow-up. Predictors explained 8%
(foot care) to 87% (BMI) of variances in change of self-care behavior frequency and BMI during
18 months. Baseline values for outcomes had robust influences on the frequency of these
behaviors at 18 months (p < .05). Other than BMI, diet and SBGM showed a larger predictive
power of baseline values compared with exercise and foot care. Notably, the frequency of
healthy diet showed significant associations with depressive symptoms change (p < .05),
suggesting increased frequency of healthy diet following decreased depressive symptoms. With
respect to PST receipt, no self-care behaviors and BMI were associated. Patients participated in
the PST might reduce the exercise frequency by .08 day. Finally, interactions between depression
and PST receipt was associated with increased SBGM, and this suggests significantly increased
SBGM frequency among patients who received the PST when they had the same level of
depression decline during 12 months (p < .05).
39
Table 4
Relationship among depression measured at 12-months follow-up, PST receipt, and self-care
behaviors at 18-months follow-up
Unstandardized coefficient (SE)
Diet
(n = 249)
Exercise
d
(n = 249)
SBGM
d
(n = 249)
Foot care
d
(n = 249)
BMI
(n = 247)
Constant 3.50 (.55)*** .45 (.11)*** .24 (.10)* .62 (.08)*** .45 (1.23)
Age ≥50 .06 (.25) -.03 (.01) -.09 (.05) .01 (.03) -.18 (.43)
Lower than high school
a
-.30 (.38) -.04 (.08) -.00 (.07) .08 (.05) .52 (.65)
Some college or graduates
a
-.61 (.62) .14 (.12) .13 (.12) .10 (.08) 1.43 (1.07)
Female
-.19 (.29) -.08 (.06) .03 (.06) .01 (.04) -.19 (.51)
Separated
b
-.05 (.24) .08 (.05) .01 (.05) .04 (.03) -.02 (.41)
Never married
b
-.40 (.34) -.04 (.07) .01 (.07) .02 (.05) -1.09 (.89)
Dysthymia
-.10 (.23) -.05 (.05) .05 (.04) .01 (.03) .14 (.39)
Baseline PHQ-9 .02 (.04) -.01 (.01) .01 (.01) .00 (.01) -.13 (.07)
Only Spanish
c
.06 (.38) .12 (.07) .06 (.07) -.08 (.05) .25 (.66)
On insulin treatment at 18
months
-.28 (.22) -.03 (.04) .11 (.05)* .05 (.03)+ .72 (.38)+
Baseline value for
dependent variable
.32 (.06)*** .13 (.06)* .37 (.06)*** .10 (.05)* .95 (.03)***
PHQ-9 at 12 month -.06 (.03)* -.01 (.01) -.01 (.01) -.00 (.00) .04 (.05)
PST receipt .21 (.23) -.08 (.04) .04 (.04) .02 (.03) .23 (.38)
PHQ-9 at 12 month X PST
receipt
.05 (.04) -.01 (.01) .02 (.01)* -.00 (.01) -.05 (.07)
Explained variance R
2
= .16 R
2
= .14 R
2
= .25 R
2
= .08 R
2
= .87
R-square increase due to
interaction
△R
2
= .01 △R
2
= .01 △R
2
= .02 △R
2
= .00 △R
2
= .00
*p < .05; **p < .01; ***p < .001.
Note.
a
Reference group: high school diploma;
b
Reference group: currently married;
c
Reference group: English only
speaking or bi-lingual patients;
d
Variable was log transformed to satisfy the normality of residual.
40
Finally, Table 5 demonstrates results from prospective analysis for predicting self-care
behaviors measured after 12 months. The explained variance for the changed foot care frequency
jumped to 12% compared with the previous models. Magnitudes for baseline value’s correlation
apparently increased in models regressed on diet, exercise, and foot care. In general, the
predictive power of previous behaviors decline as the time interval between two measurements
become longer. The change of depression was associated with an increased frequency of healthy
diet (p < .05) and foot care (p < .01), as was found in the cross-sectional analysis. PST receipt
was not found to have a significant association with any of the self-care behaviors and BMI. No
interaction effect was found either.
41
Table 5
Relationship among depression measured at 12-months follow-up, PST receipt, and self-care
behaviors at 24-months follow-up
Unstandardized coefficient (SE)
Diet
(n = 235)
Exercise
d
(n = 235)
SBGM
d
(n = 235)
Foot care
d
(n = 235)
BMI
(n = 231)
Constant 2.87 (.56)*** .42 (.10)*** .27 (.11)* .62 (.06)*** 2.44 (1.47)
Age ≥50 .13 (.25) -.08 (.05) .01 (.05) .04 (.03) .70 (.51)
Lower than high school
a
.11 (.36) .03 (.07) -.06 (.07) .11 (.04)** .15 (.75)
Some college or graduates
a
.44 (.58) .02 (.11) .19 (.12) .12 (.07) .76 (1.19)
Female
-.54 (.30) -.00 (.06) .04 (.06) -.07 (.04)* -.66 (.64)
Separated
b
-.02 (.24) .01 (.05) -.02 (.05) .01 (.03) .53 (.49)
Never married
b
-.15 (.34) -.06 (.07) .07 (.07) .10 (.04)* -.21 (.69)
Dysthymia
-.45 (.23)* -.03 (.04) -.02 (.05) -.03 (.03) -.40 (.46)
Baseline PHQ-9 .02 (.04) .00 (.01) .00 (.01) .00 (.01) -.04 (.08)
Only Spanish
c
.38 (.37) .09 (.07) .06 (.08) -.07 (.04) -.64 (.78)
On insulin treatment at 18
months
-.07 (.22) -.04 (.04) .11 (.05)* .02 (.03) .56 (.45)
Baseline value for
dependent variable
.39 (.06)*** .20 (.06)** .30 (.06)*** .17 (.04)*** .92 (.03)***
PHQ-9 at 12 month -.06 (.03)* -.01 (.01) .01 (.01) -.01 (.00)** -.03 (.06)
PST receipt .20 (.23) -.03 (.04) -.01 (.05) .01 (.03) .63 (.46)
PHQ-9 at 12 month X PST
receipt
-.00 (.04) -.01 (.01) -.01 (.01) .01 (.01) .06 (.08)
Explained variance R
2
= .24 R
2
= .15 R
2
= .18 R
2
= .20 R
2
= .84
R-square increase due to
interaction
△R
2
= .00 △R
2
= .00 △R
2
= .00 △R
2
= .01 △R
2
= .00
*p < .05. **p < .01. ***p < .001
Note.
a
Reference group: high school diploma;
b
Reference group: currently married;
c
Reference group: English only
speaking or bi-lingual patients;
d
Variable was log transformed to satisfy the normality of residual.
42
DISCUSSION
This study with secondary analyses attempted to expand our understanding of the effect
of decreased depressive symptoms on adherence to self-reported self-care behavior frequency
and BMI change among diabetes patients when these patients enrolled a depression care trial. In
addition, because PST sessions often address emotional distress from ill-managed diabetes
management, we tested whether PST had an independent impact on self-care behaviors when
controlled for the effect from decreased depressive symptoms. Finally, the multiplicative effect
of these two effects were also empirically tested. These questions were examined in three models
with cross-sectional and 6- and 12-month lagged prospective analysis.
Contrary to predictions, the empirical tests did not provide clear support for the
propositions inferred from theories and findings in the existing literature, requiring careful
interpretation. First, the change in depressive symptoms and baseline depressive symptoms may
have different relationship patterns, depending on the type of self-care behaviors. We found that
decreased depression only had significant association with healthy diet frequency change during
12, 18, and 24 months since the initial assessment and with recommended foot care frequency
change during 12 and 24 months. Previous studies consistently showed significant associations
between healthy diet frequency and depression level (Ciechanowski et al., 2003; Gonzalez et al.,
2007; Gonzalez et al., 2008a; Lin et al., 2004). This consistency from both observational and
clinical studies support the existence of a significant effect of reduced depression severity on
healthy diet frequency. However, a significant effect on foot care does not provide same level of
confidence because some observational studies demonstrated no relationship between depression
severity and foot care (Gonzalez et al., 2008a; Gonzalez et al., 2008b). Because SDCSA only
43
adopted a foot care subdomain in 2000, empirical evidence is still lacking to evaluate the
relationship.
However, a discrepancy in findings related to exercise and SBGM frequency appears to
exist between observational studies and clinical studies. A prospective study demonstrated
negative impacts with the frequency of exercise and SBGM with a reduction in depressive
symptoms, and the measurement of self-care behaviors were about 9 months apart (Gonzalez et
al., 2008b). A meta-analysis showed significant associations between these two behaviors and
the risk of depression (Gonzalez et al., 2008a). According to the findings, it appears that
increasing exercise and SBGM frequency may not be achieved merely by alleviating depressive
symptoms. Previous RCTs that investigated the efficacy of depression care often demonstrated
insignificant differences in frequencies of these two behaviors (Katon et al., 2010; Lin et al.,
2006; Williams et al., 2004). Our findings suggest that activating exercise and SBGM among
depressed patients is a complicated task, requiring deeper comprehension on factors and
mechanisms by which patients are motivated to implement recommended behaviors.
In this study, the frequency of a healthy diet had a robust association with depressive
symptoms change in three models with different time intervals between predictor and outcome.
There are two explanations for this finding. First, depressive symptoms may have had a direct
effect on the patient’s motivation to eat healthy foods. Depression’s cardinal symptom, apathy,
may have been reduced, causing patients to gain energy and motivation to cook healthier foods.
Also, alleviated symptoms may have helped patients seek a long-term objective, which is
successful for diabetes management, rather than meet an immediate desire, such as eating high-
fat foods (Appelhans et al., 2012; Cornil & Chandon, 2013). Another explanation is that
44
alleviated depression may have helped patients mobilize social support from family members,
which could have resulted in more frequent consumption of a healthy diet. Social support from
family members is associated with a better diet among diabetes patients (King et al., 2010).
Depressive symptoms often negatively influence chronic illness patients’ relationship s with
caregivers and other significant others, resulting in decreased social support (Bolger, Foster,
Vinokur, & Ng, 1996; Hammen, 2006; Oh, Ell, & Subica, 2014). In other words, the observed
increases in diet are indirectly developed through improved social support followed by better
relationships with caregivers and other significant others. Further investigations are suggested for
testing the validity of these two explanations.
Finally, PST does not seem sufficient to increase self-care behaviors, although it focuses
on addressing barriers for better self-care behaviors. The findings are not anomalous from
systematic reviews on the effects of models based on PST (Fitzpatrick, Schumann, & Hill-
Briggs, 2013; Hill-Briggs & Gemmell, 2007). As a post-hoc analysis, we also conducted another
model without the depression variable to examine whether changed depression would conceal
the PST’s correlati ons with outcomes. Insignificant associations between PST and self-care
behavior frequency were found (not shown). Also, only two significant interaction terms
between PST and depressive symptoms were found without any notable consistency. In other
words, PST may not affect the association between depression change during 12 months and the
frequency of self-care behaviors other than a couple of deviations. PST was initially developed
for healthy people with depression (Nezu et al., 2013) and was adapted for optimal use in a
primary care clinic (Hegel, Barrett, & Oxman, 2000). The MDDP held 8-12 PST sessions, with
possible booster sessions, up to one year after the patient’s fir st enrollment (Ell et al., 2010). Two
possible reasons seem to explain why PST was not associated with any self-care behavior. First,
45
the focus of PST on depression may subdue the possible benefit for self-care behaviors, although
PST helped patients clarify their self-care goals and activate self-care behaviors. For instance,
meeting friends is a good way to facilitate pleasurable activities addressing depressive symptoms
in PST. Unfortunately, many diabetes patients confront situations where they are asked to
consume unhealthy foods or sacrifice time for exercise and SBGM. Thus, addressing depression
and maladaptive self-care behaviors sometimes results in two conflicting activities that do not
align well with one another (Detweiler-Bedell et al., 2008). Second, PST was not developed to
increase self-care behavior frequency. Thus, strategies validated in rigorous research in previous
studies focusing on the effectiveness of self-management training should have supplemented
PST if the self-care behavior frequency was one of the objectives in collaborative depression
care. For decades, the literature documented evidence-based practices for each of the self-care
behaviors (American Diabetes Association, 2013; Polonsky et al., 2011; Toobert et al., 2000).
Because PST in the MDDP was developed for untreated depression among patients using
primary care clinics, expecting improvements in self-care behaviors may be too optimistic.
However, the chronic care model emphasizes the importance of psychosocial treatment for
diabetes management, and many wish that such psychosocial treatment can also help a patient’s
disease management (Wagner et al., 2001). Also, problem-solving skills were found to be
significant predictors for increased self-care behaviors (King et al., 2010). Therefore, further
research is suggested to innovate PST that can address both depression and low adherence to
self-care behaviors.
This study carries limitations, requiring readers to exercise caution with interpretation of
the findings. Because this is a prospective study, a significant number of patients dropped out
from the follow-up study. Examining any differences in demographic and focal variables at
46
baseline assessment, education, language, and SBGM at baseline were found to be significantly
different between the groups analyzed and those that dropped out. It is not unclear whether
different attrition rates by these three variables may introduce biases to results. Second, in theory,
self-reported frequency of specific self-care behaviors measured with the SDSCA might
introduce errors, such as recall bias, and increase type 2 errors to statistical inference (Gonzalez
et al., 2013). Direct observation and electronic monitoring through meters or sensor-attached
gadgets are available for reducing possible biases related to self-reported measures (Gonzalez et
al., 2013). However, few studies are available that answer the question, leaving mixed empirical
findings on the validity and sensitivity of self-reported frequency (Asimakopoulou & Hampson,
2005; Heisler, Smith, Hayward, Krein, & Kerr, 2003; Toobert et al., 2000). Further studies are
necessary to establish the validity of self-reported measures, particularly the SDSCA, by
comparing other methods, such as a blood glucose meter. In addition, the patient’s perception on
the competency with implementing self-care behaviors (Glasgow et al., 2005), self-efficacy, or
problem-solving skills for diabetes management (Hill-Briggs, 2003; Lorig, 1996) and the extent
of patient activation for chronic illness management (Hibbard, Greene, & Tusler, 2009) are
significant outcomes that might precede actual behavioral changes. It is possible that decreased
depressive symptoms are not a sufficient condition for increased self-care behaviors but are a
necessary condition. For example, social-ecologic supports for diabetes patients showed a
notable amount of variance in healthy eating and exercise (Glasgow, Strycker, Toobert, & Eakin,
2000; King et al., 2010). Therefore, it is important to understand the effect of depression on
multiple aspects of self-care behaviors, including behavioral, cognitive, and relational domains.
This study is one of few studies to extensively investigate the effect of depression
alleviation on frequency of each self-care behavior. The findings support the importance of the
47
chronic care model that emphasizes a whole-person focused, coordinated approach for chronic
illness patients. Although a significant variation was observed, addressing clinical depression is
important for maintaining a healthy diet because preparing and sharing healthy food within a
family and among friends often involve interpersonal interactions, which are largely influenced
by depression. Also, with intervention unequally focusing on clinical depression, it is hard to
expect increases in recommended self-care behaviors. Because the interaction between
depression and self-care behaviors are so intricate, further studies are necessary to answer more
sophisticated and informed research questions. As opposed to the significant clinical effects
found in RCTs that tested the effectiveness of self-management training for diabetes patients
(Minet, Møller, Vach, Wagner, & Henriksen, 2010), RCTs that examined the effectiveness of
depression care for depressed diabetes patients often demonstrate null findings with a few
anomalies (Bogner, Morales, de Vries, & Cappola, 2012; Harkness et al., 2010; Katon et al.,
2010; Markowitz et al., 2011). Also, further studies need to test the effect of PST if it adopts an
evidence-based approach for self-management for diabetes patients that demonstrates significant
improved self-care behaviors (Lorig, Ritter, Villa, & Piette, 2008; Lorig, Ritter, Villa, & Armas,
2009).
48
Chapter 3: Are Changes in Self-Care Behavior Frequency During Depression Care Associated
with the Risk of Depression?
INTRODUCTION
Comorbid depression is a debilitating condition among diabetes patients (Ali et al., 2006;
Pan et al., 2010; Stuart & Baune, 2012). The literature suggests a bidirectional relationship
between depression and diabetes (Renn et al., 2011). For diabetes patients, functional impairment
by diabetes-related complications and life modifications by recommended self-care behaviors
(e.g., monitoring carbohydrates consumption) often contribute to increased depressive symptoms
(Knol et al., 2007). Conversely, depression is found as an independent predictor for increased
micro- and macro-vascular complications, self-reported symptomatology, and increased A1C
levels (Black, 1999; Black et al., 2003; Coleman, Katon, Lin, & Von Korff, 2013; Egede et al.,
2009; Egede & Hernández-Tejada, 2013; Margaret et al., 2014). As a result, a 1.5-fold increase
of mortality among depressed diabetes patients was found in a recent meta-analysis (Park et al.,
2013).
One pathway by which depression and diabetes are associated is through self-care
behaviors for diabetes management. In a study in the Netherlands, diagnosis of type 2 diabetes
increased the risk of having clinical depression by 70%, whereas undiagnosed type 2 diabetes
patients did not demonstrated a similar increase, suggesting that the increased burden of
managing diabetes is a significant predictor for a higher prevalence of depression (Knol et al.,
2007). This finding was also shown in a U.S. study (Golden et al., 2008). The findings are
consistent with studies that found diabetes-related burden as a significant predictor for depressive
symptoms (Lloyd, Pambianco, & Orchard, 2010). The extent of emotional burdens specifically
related to diabetes management was the strongest predictor for A1C level change during an 18-
49
month follow-up, resulting in a null relationship of depressive symptoms and depression
diagnosis (Fisher et al., 2010). Also, depression works as a risk factor for diabetes management
because this condition often undermines the extent of self-care behaviors (Gonzalez et al., 2008a;
Gonzalez et al., 2008b). Apathy is a common cardinal symptom for depression diagnosis and
may explain the lower adherence to recommended self-care behaviors and prescribed medication
among chronic illness patients (DiMatteo, Lepper, & Croghan, 2000). Behavioral changes
related to depression is believed to be responsible for maladaptive self-care behaviors
(Ciechanowski et al., 2003; Egede et al., 2009; Gonzalez et al., 2007; Gonzalez et al., 2008a;
Gonzalez et al., 2008b; Lin et al., 2004; Wagner et al., 2010).
To address the negative impact of depression on diabetes management, many depression
treatments have been developed and tested in RCTs (Beatty & Lambert, 2013; Egede &
Hernández-Tejada, 2013; Glazier, Bajcar, Kennie, & Willson, 2006; Harkness et al., 2010; Hill-
Briggs & Gemmell, 2007; Markowitz et al., 2011; van der Feltz-Cornelis et al., 2010). Because
burdensome diabetes management is often cited as an important contributor for depression, many
of these interventions had activities assisting patients in achieving successful diabetes
management, including exercise (Piette et al., 2011), didactic education of diabetes management
(Huang et al., 2002), and PST (Ell et al., 2009; Katon et al., 2010; Lin et al., 2006; Williams et
al., 2004). For instance, the Team Care model for diabetes or heart disease patients taught
patients how to manage medication, persistent pain, blood pressure, blood glucose level, and
sleep, as well as treating depression (Katon et al., 2010). The MDDP, which recruited low-
income diabetes patients, also used PST to enhance a patient’s ability to manage his or her
diabetes symptoms and self-care behaviors for the disease (Ell et al., 2009; Ell et al., 2010).
50
This therapeutic approach seems reasonable because many diabetes patients report
emotional distress related to diabetes management, and this emotional burden has been found to
explain significant variation in depressive symptoms (Fisher et al., 2010). Emotional distress
related to diabetes management involves anxiety or helplessness coming from an inability to
manage diabetes symptoms and stress from overwhelming self-care behaviors (Delahanty et al.,
2007). In particular, many patients have demonstrated concerns with managing time for
prescribed self-care behaviors and often feel guilty when the planned management is not
implemented (Delahanty et al., 2007). As a result, it is important to help patients develop plans
that reduce emotional distress from functional loss and unfulfilled daily self-care behaviors. For
example, Lorig et al. (2008) found reduced health distress and changes in self-care behavior
frequency and diabetes outcomes among patients who were randomly assigned to either a 6-
week self-care management training, which was provided by a peer leader, or automated
telephone call reinforcing self-care behaviors (Lorig et al., 2008). Another intervention, which
targeted multiple domains in diabetes self-care behaviors, also showed significantly reduced
depression scores, although the extent of improvement is apparently smaller than that of studies
primarily focusing on clinical depression (Ell et al., 2010; Katon et al., 2004a; Katon et al., 2010;
Lorig et al., 2009). The chronic care model emphasizes the importance of a collaborative
approach for diabetes patients, who are often burdened with physical, psychological, and social
attributes (Wagner et al., 2001; Woltmann et al., 2012). Previous interventions have adhered to
the notion that diabetes patients with depression need integrated approaches that have dual
focuses on disease control and depression care. One observation study suggested a therapeutic
effect of adherence to diabetes-management on depression (Sacco et al., 2007). Therefore, it
appears that self-care behaviors and depression severity have a positive reciprocal upward spiral.
51
However, the literature in self-regulation theory suggests possible burdens of increased
self-care behaviors on patient’s emotion. In other words, increased self -care behaviors often
noticed in diabetes management programs or depression care interventions might increase the
risk of depression. According to self-regulation theory, cognitive and emotional resources for
self-regulation is limited (Baumeister, Bratslavsky, Muraven, & Tice, 1998; Muraven &
Baumeister, 2000; Vohs, Baumeister, & Ciarocco, 2005). If people are asked to engage in self-
regulation too much, their resources for self-regulation are often depleted, resulting in increased
impulsive and abusive responses to external stimulus (Bauer & Baumeister, 2011). In other
words, when study participants were asked to watch delicious food (e.g., cookies, cake),
cigarettes, and alcohol and allowed to consume these attractive objects, they consumed more
than they usually do without such manipulation (Bauer & Baumeister, 2011). The observations
seem to be relevant to diabetes patients who have to engage in self-regulation to follow
recommended self-care behaviors that are prescribed. Diabetes patients are often asked to adhere
to multiple self-behaviors, although the extent of types and intensity of behaviors vary by
diabetes progression and experience of previous complications (American Diabetes Association,
2013). Diabetes patients need to spend more than two hours for daily recommended self-care
behaviors (Russell et al., 2005). In addition, the loss of pleasurable activities due to more
investment on self-care behaviors also may increase depressive symptoms (Cabassa, Hansen,
Palinkas, & Ell, 2008; Gask, Macdonald, & Bower, 2011). To follow recommended self-care
behaviors, some patients avoid participating in dinners with family members because they are
concerned about not following their diet plan (Detweiler-Bedell et al., 2008; Early et al., 2009;
Gask et al., 2011). Therefore, it is plausible that increased frequency of self-care behaviors
gained in self-management training might function as a risk factor for depression.
52
However, the effect of increased self-care behaviors on depression risk has largely been
left unexamined. This study examined whether the change in self-care behavior frequency
predicts depression risk, concurrently and prospectively. Figure 3 shows analytic models that
were examined in Study 2. Specifically, this study conducted a secondary data analysis of data
obtained in a clinical trial that tested the effectiveness of collaborative depression care. There
were two study groups, including the INT group and EUC group. Whereas both groups were
exposed to depression care for 12 months, 6-month-interval follow-up assessments collected
self-reported self-care behavior frequency in diet, exercise, SBGM, and foot care. Change in
self-care behavior frequency was measured between baseline and 12-months follow-up when key
depression care was implemented. Thus, these data allow for the investigation of whether
increased self-care behaviors during depression care are associated with depression symptoms
measured cross-sectionally and at 6 and 12 months after treatment.
Figure 3. Study 2’s analytic models
53
METHODS
Sample and Procedure
This study conducted secondary analyses with data collected in the MDDP, which was
conducted in two safety-net primary clinics in Los Angeles, California. The MDDP aimed to test
the effectiveness of socio-culturally adapted version of collaborative depression care. Of the
patients approached for recruitment, 30.2% were found to have clinical depression, which was
determined by a self-reported depression measure, the PHQ-9. Furthermore, 78.7% of them (N =
387) participated in the study after providing written informed consent (Ell et al., 2010). Patients
with alcohol misuse assessed with a self-reported screener, other psychiatric illnesses, and acute
suicidal ideation were excluded despite meeting the major depressive disorder criteria (Ell et al.,
2010). Recruitment details are found in Ell et al. (2010).
To examine the predictive power of change in self-care behavior frequency while
depression care was implemented, three analytic models were developed (Figure 3). For these
models, change in frequency of healthy diet, recommended exercise, SBGM, foot care, and BMI
change were predictors for the risk of depression at 12-, 18-, and 24-month follow-ups. One
cross-sectional analysis and two prospective analyses were conducted that predicted depression
diagnosis 6 and 12 months after the time self-care behaviors were measured. Patients who
responded to 12-month follow-up and assessment in which depression was measured were
included in the analysis. A difference in attrition rate for each follow-up resulted in 281 (72.6%),
249 (64.3%), and 235 (65.7%) patients analyzed for cross-sectional analysis and 6 and 12
months prospective analyses. According to results from analyses comparing demographic and
clinical variables, there were significant differences in education, language, and baseline
frequency of SBGM (Oh & Ell, unpublished).
54
Table 6 describes the results of comparisons between two groups selected with clinical
depression, which was determined by cutoff score (≤10) in the PHQ-9. According to the cutoff,
33.5%, 38.6%, and 35.7% of study participants in the three analytic models were diagnosed with
probable clinical depression, and these levels showed significantly reduced overall depression
severity in the MDDP. Patients who had dysthymia at baseline were more likely to report clinical
depression at 12- (p < .05), 18- (p < .001), and 24-month follow-up (p < .001). Also, patients
assigned to the intervention arm were less likely to have clinical depression only at the 18-month
follow-up (p < .05) but not at the 12- and 24-month assessments. Finally, education, marriage
status, and language were associated with clinical depression risk at the 24-month follow-up (p <
.05).
Table 6 also demonstrates mean differences in depression, diet, exercise, SBGM, foot
care, and BMI at baseline among groups with or without clinical depression. Baseline depression
severity was associated with the likelihood of clinical depression (p < .05). However, no
significant differences in the frequency of each self-care behavior was found. At a glance, no
consistent pattern was noted other than exercise frequency, demonstrating that patients without
clinical depression during follow-ups had about 0.3 days more frequent weekly exercise at
baseline. In addition, predictors and change in frequency of each self-care behaviors were
examined by depression diagnosis status at the three follow-up periods. No significant
association was found. At a glance, although anomalies were found, patients without clinical
depression, in general, had larger gains in healthy diet and exercise. Interestingly, depressed
patients consistently had bigger gains in BMI change.
55
Table 6
Demographic and clinical characteristics depending on the extent of self-care behavior changes
Depression Status (PHQ-9 ≥ 1 0)
12-month follow-up 18-month follow-up 24-month follow-up
Yes
94 (33.5%)
No
187 (66.5%)
Yes
96 (38.6%)
No
153 (61.4%)
Yes
84 (35.7%)
No
151 (64.3%)
N (%) N (%) N (%) N (%) N (%) N (%)
Age ≥50 69 (73.4%) 137 (73.3%) 69 (71.9%) 114 (74.5%) 64 (76.2%) 106 (70.2%)
Intervention group
40 (42.6%) 102 (54.5%) 40 (41.7% )* 88 (57.5% )* 47 (56.0%) 77 (51.0%)
Lower than high
school
a
74 (78.7%) 162 (86.6%) 81 (84.4%) 130 (85.0%) 61
(72.6% )**
133
(88.1% )**
Some college or
graduates
a
6 (6.4%) 7 (3.7%) 5 (5.2%) 6 (3.9%) 9 (10.7%)* 3 (2.0%)*
Female
73 (77.7%) 161 (86.1%) 79 (82.3%) 127 (83.0%) 68 (81.0%) 130 (86.1%)
Separated
b
36 (38.3%) 68 (36.4%) 37 (38.5%) 54 (35.3%) 29 (34.5%) 56 (37.1%)
Never married
b
14 (14.9%) 21 (11.2%) 15 (15.6%) 15 (9.8%) 18
(21.4% )**
11 (7.3% )**
Dysthymia
62 (66.0% )* 96 (51.3% )* 72
(75.0% )***
69
(45.1% )***
62
(73.8% )***
70
(46.4% )***
Only Spanish
c
77 (81.9%) 167 (89.3%) 85 (88.5%) 132 (86.3%) 66
(78.6% )**
138
(91.4% )**
On insulin
treatment at
baseline
28 (29.8%) 50 (26.7%) 28 (29.2%) 43 (28.1%) 27 (32.1%) 38 (25.2%)
Baseline PHQ-9 15.46 ±
2.69***
14.16 ±
2.88***
15.07 ±
2.56*
14.34 ±
3.02*
15.19 ±
2.50**
14.23 ±
2.98**
Diet at baseline 4.03 ± 1.68 4.36 ± 1.77 4.25 ± 1.61 4.20 ± 1.80 4.02 ± 1.58 4.38 ± 1.77
Exercise at
baseline
1.76 ± 2.26 2.16 ± 2.41 1.88 ± 2.20 2.13 ± 2.44 1.82 ± 2.07 2.16 ± 2.46
SBGM at baseline 2.46 ± 2.76 2.92 ± 2.84 2.86 ± 2.82 2.68 ± 2.78 2.94 ± 2.79 2.84 ± 2.91
Foot care at
baseline
4.87 ± 2.80 5.05 ± 2.79 4.92 ± 2.62 5.14 ± 2.82 5.17 ± 2.60 4.84 ± 2.91
BMI at baseline 33.84 ± 6.49 32.50 ± 7.86 33.04 ± 6.60 32.92 ± 7.78 34.16 ± 8.02 32.51 ± 7.57
Change of diet for
12 months
0.02 ± 1.99 0.30 ± 1.80 .00 ± 1.86 .38 ± 1.84 .31 ± 1.93 .35 ± 1.79
Change of exercise
for 12 months
.20 ± 2.83 .36 ± 3.01 -.07 ± 2.87 .51 ± 2.96 -.07 ± 2.41 .54 ± 3.13
Change of blood
testing for 12
months
.27 ± 2.76 -.13 ± 3.04 -.32 ± 2.92 .09 ± 2.82 .26 ± 2.77 -.19 ± 3.15
Change of foot
care for 12 months
-.37 ± 3.11 -.23 ± 3.14 -.41 ± 3.24 -.33 ± 2.99 -.33 ± 2.95 -.10 ± 3.15
Change of BMI for
12 months
-.17 ± 2.48 -.04 ± 2.99 -.16 ± 2.23 -.15 ± 3.26 -.39 ± 2.10 -.05 ± 3.35
56
Measures
Self-care behaviors were defined as the frequency of recommended self-care behaviors
and BMI. BMI was calculated with self-reported weight and height. The frequency of
recommended self-care behaviors were measured with the SDSCA (Toobert & Glasgow, 1994;
Toobert et al., 2000). Patients were asked to recall how many times they implemented
recommended or discouraged self-care behaviors that are categorized into diet, exercise, SBGM,
foot care, and smoking during the previous seven days when patients were not sick. For this
study, smoking was not examined because too few patients were exposed to it. Although 10
items were used to measure the overall self-care behaviors, Toobert et al. (2000) suggested using
a value that summarizes each self-care behavior, since inter-correlations between subscales are
fairly low for a single construct. In the MDDP study, parameters from a bivariate correlation
analysis of the four subdomains ranged from .32 (SBGM – foot care) to .05 (exercise – SBGM).
An average score for each domain describes the number of days a patient implemented the
recommended self-care behaviors, ranging from 0 (lowest) to 7 (highest). Items asking about
unhealthy behaviors were reverse coded with a larger score, indicating more frequent adherence
to self-care behaviors. The SDSCA is viewed as one of the most reliable self-reported measures,
resulting in extensive usage in clinical trials validating interventions for diabetes patients (Ell et
al., 2010; Gregg, Callaghan, Hayes, & Glenn-Lawson, 2007; Katon et al., 2010; Lin et al., 2006;
Schillinger, Handley, Wang, & Hammer, 2009; Williams et al., 2004). The SDSCA
demonstrated good evidence for reliability, acceptable sensitivity to change of healthy diet and
recommended exercise, and concurrent validity when compared with other self-reported
measures for the same construct (Toobert et al., 2000; Weinger, Butler, Welch, & La Greca,
2005). In this study, a categorized ranked variable was developed and examined in analytic
57
models, as well as raw continuous scores from subdomains. Specifically, three categories, upper
33%, middle 33%, and lower 33%, were developed by ranking a score in each self-care behavior
and BMI. This manipulation may help measure effect sizes of substantial change in self-care
behaviors because the baseline category was controlled for in a statistical model.
To measure the severity of depression and probable clinical depression as a dependent
variable, the PHQ-9 was used (Löwe et al., 2004). The PHQ-9 asks nine items that are used for
diagnosis of major depressive disorder in the Diagnostic and Statistical Manual of Mental
Disorders, Fourth Edition (DSM-4) (Löwe et al., 2004). Respondents were asked to assess how
often they experienced emotional states described in each item. They were asked choose among
“not at all” (0), “ several days” (1), “ more than half the days” (2), or “ nearly every day” (3).
Answers to nine items were summed, ranging from 0 to 27. The PHQ-9 was developed to
increase primary care provider’s capacity to detect and monitor clinical depression (Kroenke &
Spitzer, 2002; Löwe et al., 2004; Spitzer, Kroenke, & Williams, 1999). Its psychometric
properties were extensively validated with patients recruited in primary care settings (Spitzer et
al., 1999). Also, several studies found its broader applicability to different groups by race, ethnic
group, age, and language (Huang, Chung, Kroenke, Delucchi, & Spitzer, 2006; Wittkampf,
Naeije, Schene, Huyser, & van Weert, 2007). As a diagnostic tool, a score of at least 10
demonstrates an optimal level of sensitivity and specificity when it is used in groups with a
higher prevalence of clinical depression, including low-income diabetes patients (Kroenke &
Spitzer, 2002; Wittkampf et al., 2007). In this particular study, two values from the PHQ-9 were
used. First, as a diagnostic measure, patients scoring at least 10 or higher were coded as having a
clinical depression, creating a dichotomous variable. Second, to measure the severity of
58
depressive symptoms, the raw score from the PHQ-9 was used in analytic models. In this study,
Cronbach’s α for 12 -, 18-, and 24-month assessments were .83, .83, and .84, respectively.
Analysis
To explore self-care behavior’s association with the likelihood of having clinical
depression and depressive symptoms, two analyses were conducted with predictors: categorized
and continuous self-care behavior variables. First, the chi-square test was conducted to examine
whether patients located in either upper, middle, or lower 33% had significantly different
prevalence rate of clinical depression at 12-, 18-, and 24-month follow-ups. Second, with the
severity of depressive symptoms as an outcome, ANOVA tests were conducted by the
categorized self-care behavior variables. ANOVA test provides an F-value for mean differences
of the three groups depicting the level of self-care behavior frequency and BMI. For all tests, a
significance level of α = 0.05 was used. Analyses used SPSS 21.0 (Chicago, IL, USA).
To test the concurrent and prospective association between self-care behaviors, likelihood
of having clinical depression, and the severity of depressive symptoms, two groups of analytic
models were developed and examined. The first group of analytic models used the raw scores
from subdomains in the SDSCA, showing the frequency of each self-care behavior weekly and
the BMI. Two statistical methods were used depending on the depression outcome. Logistic
regression was used to examine the association with clinical depression, determined by the cutoff
(≥10) in the PHQ-9. This analysis provided odds ratios (OR) for having clinical depression when
predictor values showed a one-unit change. OR and 95% confidence intervals (CI) were reported.
With the severity of depressive symptoms, multivariate regression was conducted. In this
analysis, the coefficient for a parameter, standard error (SE), and significance level for t-test
were reported in the results. The second group of analytic models had categorical predictors with
59
three domains: upper, middle, and lower 33%. This group demonstrated relative ORs and
coefficients for the middle or upper 33% with respect to self-care behavior frequency, when
compared to patients who ranked in the lower 33%. Like in the first group, logistic regression
and multivariate regression were conducted by the type of depression outcomes. All logistic and
multivariate regression analyses were controlled for demographic and clinical variables,
including age (<50 vs. 50+ years), study groups (INT vs. EUC), education (lower vs. high school
vs. above), gender, marital status (separated vs. marriage vs. never married), dysthymia at
baseline, the PHQ-9 score at baseline, primary language (Spanish vs. English only or bilingual),
and insulin treatment at baseline. Also, to examine the association between changes of both
predictors and outcomes after patient’s enrollment, baseline values for these variables were
controlled in the analytic models. This allowed for the examination of the association between
changes in self-care behaviors and in depressive symptoms since baseline assessment. For a
significance test, p < .05 was used. SPSS 21.0 (Chicago, IL, USA) was used for analyses.
RESULTS
Table 7 describes the likelihood of having clinical depression and the mean of depression
severity by three groups (upper, middle, and lower 33%) by ranking the change of frequency of
each self-care behavior and BMI during 12 months since baseline. The change was calculated by
subtracting the baseline value from a score at the 12-month assessment. The upper 33% group
had larger increases in the frequency of self-care behavior during 12 months compared with the
middle and lower 33%. No significant differences were found in either depression outcomes
among the three groups for different frequency of self-care behaviors. There was no dose-
response relationship between self-care behaviors and the risk of clinical depression or severity.
In the models predicting the risk of depression at 18-month follow-up, significantly lower
60
depressive symptom scores were observed in the upper and middle 33% compared with the
lower 33%, demonstrating a dose-response relationship between the two variables. Individuals
who exercised more frequently compared with the baseline survey had lower depressive
symptoms at 18-month follow-up (p < .05). Other than exercise, neither a significant difference
nor a dose-response relationship were observed (p > .05). Finally, when depression outcomes
measured at 24 months were examined, exercise was significantly associated with both the
likelihood of clinical depression (p < .05) and depression severity (p < .05). In depression
severity, the upper group demonstrated significant lower depressive symptoms compared with
the other two groups (p < .05). Other than exercise, no significant differences were found.
61
Table 7
Relationship between the extent of change in each self-care behavior, clinical depression
diagnosis, and the severity of depression
PHQ-9 ≥10 at 12 months PHQ-9 score at 12 months
Self-care behaviors change during 12
months
Yes
94 (33.5%)
N (%)
No
187 (66.5%)
N (%)
P Mean (SD) P
Diet Lower 33% 30 (34.5%) 57 (65.5%) .52 8.38 (5.48) .20
Middle 33% 32 (29.6%) 76 (70.4%) 6.95 (5.46)
Upper 33% 32 (37.2%) 54 (62.8%) 7.48 (5.62)
Exercise Lower 33% 32 (34.0%) 62 (66.0%) .67 7.50 (5.43) .17
Middle 33% 36 (36.0%) 64 (64.0%) 8.29 (5.59)
Upper 33% 26 (29.9%) 61 (70.1%) 6.77 (5.52)
SBGM Lower 33% 28 (28.6%) 70 (71.4%) .24 6.83 (5.03) .13
Middle 33% 36 (40.0%) 54 (60.0%) 8.46 (5.53)
Upper 33% 30 (32.3%) 63 (67.7%) 7.45 (5.96)
Foot care Lower 33% 31 (30.1%) 72 (69.9%) .56 7.08 (5.04) .54
Middle 33% 34 (37.4%) 57 (62.6%) 7.79 (5.99)
Upper 33% 94 (33.5%) 187 (66.7%) 7.87 (5.61)
BMI Lower 33% 30 (32.3%) 63 (67.7%) .77 7.60 (5.79) .39
Middle 33% 34 (36.6%) 59 (63.4%) 8.10 (5.52)
Upper 33% 30 (32.3%) 63 (67.7%) 6.98 (5.33)
PHQ-9 ≥10 at 18 months PHQ-9 score at 18 months
Self-care behaviors change during 12
months
Yes
96 (38.6%)
No
153 (61.4%)
P Mean (SD) P
Diet Lower 33% 33 (43.4%) 43 (56.6%) .26 8.46 (5.94) .42
Middle 33% 39 (40.6%) 57 (59.4%) 8.26 (5.95)
Upper 33% 24 (31.2%) 53 (68.8%) 7.32 (5.24)
Exercise Lower 33% 39 (47.6%) 43 (52.4%) .11 9.27 (5.95) .01
Middle 33% 32 (36.0%) 57 (64.0%) 8.18 (5.58)
62
Upper 33% 25 (32.1%) 53 (67.9%) 6.56 (5.41)
SBGM Lower 33% 36 (40.9%) 52 (59.1%) .84 8.51 (5.82) .34
Middle 33% 30 (38.0%) 49 (62.0%) 8.28 (5.44)
Upper 33% 30 (36.6%) 52 (63.4%) 7.28 (5.91)
Foot care Lower 33% 36 (38.3%) 58 (61.7%) .42 8.06 (5.59) .50
Middle 33% 27 (33.8%) 53 (66.3%) 7.49 (6.11)
Upper 33% 33 (44.0%) 42 (56.0%) 8.57 (5.52)
BMI Lower 33% 27 (32.5%) 56 (67.5%) .37 7.37 (5.42) .29
Middle 33% 37 (43.0%) 49 (57.0%) 8.74 (6.21)
Upper 33% 31 (39.2%) 48 (60.8%) 7.86 (5.47)
PHQ-9 ≥10 at 24 months PHQ-9 score at 24 months
Self-care behaviors change during 12
months
Yes
84 (35.7%)
No
151 (64.3%)
P Mean (SD) P
Diet Lower 33% 26 (36.9%) 41 (63.1%) .97 8.14 (5.59) .79
Middle 33% 34 (35.4%) 62 (64.6%) 7.60 (5.73)
Upper 33% 26 (35.1%) 48 (64.9%) 5.33 (7.54)
Exercise Lower 33% 34 (43.0%) 45 (57.0%) .03 8.78 (5.63) .01
Middle 33% 33 (39.3%) 51 (60.7%) 8.19 (5.19)
Upper 33% 17 (23.6%) 55 (76.4%) 6.04 (5.55)
SBGM Lower 33% 26 (31.3%) 57 (68.7%) .32 7.53 (5.56) .85
Middle 33% 25 (33.8%) 49 (66.2%) 7.66 (5.43)
Upper 33% 33 (42.3%) 45 (57.7%) 8.01 (5.71)
Foot care Lower 33% 28 (34.6%) 53 (65.4%) .78 7.94 (5.79) .76
Middle 33% 31 (38.8%) 49 (61.3%) 7.89 (5.69)
Upper 33% 25 (33.8%) 49 (66.2%) 7.34 (5.17)
BMI Lower 33% 32 (39.0%) 50 (61.0%) .33 7.93 (5.79) .85
Middle 33% 29 (38.7%) 46 (61.3%) 7.83 (5.90)
Upper 33% 22 (28.9%) 54 (71.1%) 7.43 (5.56)
63
Table 8 shows findings from logistic and multivariate regressions that examined
associations between the change of frequency of self-care behaviors during 12 months post-
baseline and the likelihood of clinical depression and depression severity at 12-, 18-, and 24-
month assessments, controlling for baseline depressive symptoms and other confounders.
Change in the frequency of a healthy diet was negatively associated with the likelihood of having
clinical depression (p < .05) and depressive symptoms (p < .001) at the 12-month follow- up.
The results suggest that patients eating a healthy diet more frequently during 12 months since
baseline were 19% less likely to have clinical depression, and one more day of healthy diet was
associated with a .81-unit decline of depressive symptoms. Other self-care behaviors were not
associated with depression risk. In analyses with depression risk measured at the 18-month
follow-up, change in frequency of recommended exercise during the 2 month post-baseline
demonstrated a significant association with the likelihood of having clinical depression (p < .05)
and depression severity (p < .01). Specifically, a one-day increase in the frequency of exercise in
a week predicted a 17% lower risk for clinical depression and a -.53–unit decline in depressive
symptoms. Other self-care behaviors did not demonstrate significant correlation with the risk of
depression 6 months afterward (p > .05). Finally, in the analyses predicting the risk of depression
at 24 months, an increase in exercise frequency was associated with the risk of having clinical
depression (p < .05) and with depressive symptoms (p < .01). Similar to the effect sizes
observed in the models with the risk of depression at the 18 month post-baseline period, one
more day of recommended exercise compared with baseline was associated with a 15% lower
risk of clinical depression and with a -.52–unit decline of depression severity at 24-month
follow-up.
64
Table 8
Relationship between change of self-care behaviors during 12 months post-baseline and
depression outcomes at 24-month follow-up
12 months 18 months 24 months
Change
from
baseline to
12 month
PHQ-9 ≥10
(n = 281)
PHQ-9
(n = 281)
PHQ-9 ≥10
(n = 249)
PHQ-9
(n = 249)
PHQ-9 ≥10
(n = 235)
PHQ-9
(n = 235)
OR (95% CI) B (SE) OR (95% CI) B (SE) OR (95% CI) B (SE)
Diet 0.81 (0.66-0.99)* -.81 (.23)*** 1.00 (0.80-1.25) -.14 (.27) 0.99 (0.77-1.26) -.03 (.26)
Exercise 0.92 (0.80-1.06) -.21 (.15) 0.83 (0.71-0.96)* -.53 (.17)** 0.85 (0.72-0.99)* -.52 (.16)**
SBGM 1.08 (0.94-1.23) .25 (.15 0.96 (0.83-1.11) -.22 (.17) 1.10 (0.95-1.28) .22 (.16)
Foot care 0.94 (0.80-1.10) -.05 (.18) 0.96 (0.81-1.14) .02 (.20) 1.00 (0.81-1.22) -.05 (.20)
BMI 0.99 (0.89-1.09) -.14 (.11) 0.99 (0.88-1.10) -.09 (.13) 0.97 (0.87-1.08) -.06 (.12)
Explained
variance
R
2
= .27 R
2
= .23 R
2
= .30
Note: Linear regression adjusted for age (<50 vs. 50+), study groups (INT vs. EUC), education
(lower vs. high school vs. above), gender, marital status (separated vs. marriage vs. never
married), dysthymia at baseline, PHQ-9 score at baseline, primary language (Spanish vs. English
only or bilingual), insulin treatment at baseline, and baseline values of predictors; OR, odds ratio;
SE, standard error; CI, confidence interval.
65
Table 9 shows the association between categorized self-care behavior frequency and the
risk of depression at 12-, 18-, and 24-month follow- ups. The reference group was the patients in
the lower 33%. The middle-frequency group with a healthy diet was associated with a .55 times
reduced risk of clinical depression (p < .05) and a 1.68-unit lower value for depressive
symptoms (p < .05) at the 12-month follow-up. The upper frequency group with a healthy diet
had a 2.04-unit lower value for depressive symptoms (p < .05) than the reference group at the
12-month follow-up. Of note, the middle frequency group in SBGM showed a .63 times lower
risk of clinical depression at the 12-month follow-up. In the 6-month prospective analysis, the
upper frequency group, in recommended exercise, had a .63 times lower risk of clinical
depression (p < .05) and a 3.41-unit lower value for depression severity (p < .01) than the
reference group. Also, the upper frequency group in SBGM had 2.18-unit lower value for
depressive symptoms than the lower-frequency group (p < .05). In the 12-month prospective
analysis, the upper frequency group had a .68 times lower risk of clinical depression (p < .05)
and a 3.48-unit lower value for depression severity (p < .001).
66
Table 9
Relationship between self-care behaviors at 12 months and depression outcomes at 24-month
follow-up, controlled for baseline values
12 months 18 months 24 months
Self-care
behaviors
at 12
month
PHQ-9 ≥10
(n = 281)
PHQ-9
(n = 281)
PHQ-9 ≥10
(n = 249)
PHQ-9
(n = 249)
PHQ-9 ≥10
(n = 235)
PHQ-9
(n = 235)
OR (95% CI) B (SE) OR (95% CI) B (SE) OR (95% CI) B (SE)
Middle
33% diet
0.45 (0.20-
0.98)*
-1.68 (.82)* 0.96 (0.43-2.15) .15 (.91) 1.05 (0.44-2.52) 1.13 (.91)
Upper
33% diet
0.58 (0.26-1.31) -2.04 (.88)* 1.70 (0.75-3.87) .87 (.97) 1.09 (0.43-2.78) .46 (.97)
Middle
33%
exercise
1.07 (0.50-2.29) -.19 (.81) 0.99 (0.48-2.05) -.44 (.90) 1.12 (0.49-2.54) -1.34 (.87)
Upper
33%
exercise
0.61 (0.27-1.35) -1.42 (.86) 0.37 (0.16-.86)* -3.41 (.98)** 0.32 (0.12-0.86)* -3.48 (.94)***
Middle
33%
SBGM
0.63 (0.27-1.46) -.55 (.86) 0.71 (0.31-1.64) -.79 (.97) 0.69 (0.27-1.81) -.71 (.95)
Upper
33%
SBGM
1.23 (0.52-2.90) 1.39 (.93) 0.45 (0.19-1.08) -2.18 (1.03)* 2.34 (0.82-6.68) 1.74 (1.01)
Middle
33% foot
care
0.37 (0.16-
0.84)*
-1.16 (.86) 0.58 (0.25-1.36) -.96 (.95) 1.06 (0.42-2.66) -.45 (.94)
Upper
33% foot
care
0.52 (0.24-1.16) -1.30 (.87) 0.75 (0.32-1.75) -.28 (.97) 0.73 (0.27-1.93) -.76 (.96)
Middle
33% BMI
2.22 (0.77-6.42) .33 (1.12) 2.30 (0.95-5.58) 1.08 (1.26) 0.48 (0.14-1.69) -.29 (1.20)
Upper
33% BMI
1.03 (0.26-4.10) -1.62 (1.43) 1.15 (0.39-3.39) 1.03 (1.57) 0.61 (0.13-2.88) -1.06 (1.55)
Explained
variance
R
2
= .29 R
2
= .29 R
2
= .35
Note: Reference group for each self-care behavior is a group in the middle (50%); When multiple cases
were tie, the mean of ranks for the ties was assigned; Linear regression adjusted for age (<50 vs. 50+),
study groups (INT vs. EUC), education (lower vs. high school vs. above), gender, marital status
(separated vs. marriage vs. never married), dysthymia at baseline, PHQ-9 score at baseline, primary
language (Spanish vs. English only or bilingual), insulin treatment at baseline, and baseline groups (upper,
middle, and lower 33%) in frequency of self-care behaviors; OR, odds ratio; SE, standard error; CI,
confidence interval.
67
DISCUSSION
Since many self-management or psychosocial interventions for diabetes patients aim to
facilitate more frequent healthy behaviors, understanding the effects of such interventions on
depression is important for the design of more effective interventions. Yet, the effects of changes
in self-care behavior frequency on depression were, until now, largely untested. In this study, the
findings from one cross-sectional and two prospective analyses demonstrate inconsistent results
regarding an association between change of self-care behaviors and the risk of depression,
depending on the type of behaviors. The only consistent finding is an association between
increased exercise and decreased depression risk. Also, a more frequent consumption of a
healthy diet only has a concurrent correlation with a lower risk of depression but not with
prospective depression risk. Changes in SBGM, foot care frequency, and BMI are not
significantly associated with depression risk, other than a couple of anomalies. For diet and
exercise, the hypothesized negative effect of limited strength for self-regulation is not supported
by empirical analyses. Rather, patients who improved adherence to recommended diet and
exercise plans demonstrate better concurrent or prospective depression outcomes.
The finding that exercise demonstrated robust therapeutic effects on the risk of clinical
depression and depressive symptoms is consistent with previous studies that recommended
exercise as a therapeutic activity for depression in diabetes patients and healthy controls
(American College of Sports Medicine, 2000; Dunn, Trivedi, & O'Neal, 2001; Hamer &
Stamatakis, 2014; Rieck, Jackson, Martin, Petrie, & Greenleaf, 201+3). Cognitive-behavioral
therapy often involves sessions where patients are asked to activate behavioral changes, such as
going outside for exposure to the sun or engaging in moderate physical activities, and these
behavioral changes were associated with lower depressive symptoms (Piette et al., 2011; Ryba,
68
Lejuez, & Hopko, 2014). In addition to corroborating previous findings of the benefits of
exercise for reducing depressive symptoms in the general population, our findings suggest that
exercise may be an effective and appropriate intervention for diabetes patients with clinical
depression. Sometimes, providers are concerned with the level of stress that diabetes patients
may feel when they are asked to change their lifestyle too much (Russell et al., 2005; Safford,
Russell, Suh, Roman, & Pogach, 2005). Changing exercise level is one of the important life
domains where patients are asked to change, and receiving a diagnosis of diabetes actually may
increase the extent of physical activity (Schneider et al., 2014). Thus, as found in Knol et al.
(2007), too much life change may cause emotional burden from diabetes management, resulting
in increased depressive symptoms. In contrast, our study showed stronger benefits from more
frequent exercise than emotional burdens from the lifestyle change. Therefore, we suggest that
providers encourage patients to increase the exercise frequency when they begin to exhibit
symptoms of clinical depression. Finally, increased exercise demonstrated a stronger effect on
subsequent depression risk until one year follow-up. The positive effect of increased frequency
of recommended exercise was more apparent in prospective studies, suggesting a causal
relationship. The existence of a rigorous dose-response relationship between the change in
exercise frequency and the risk of depression also supports the possible causal relationship
(Dunn et al., 2001).
Another interesting finding was a robust concurrent correlation between the change in
healthy diet frequency and the risk of depression without any significant results in prospective
analysis. In other words, patients having more frequent healthy diet at the 12-month follow-up
reported better depressive symptoms only at the same follow-up. This finding may suggest that
more frequent healthy diet may be an outcome of depression, rather than a cause for worsening
69
depressive symptom. The causal role of depression for poor diet is fairly well-established,
although several moderators affect the association (O’Conner & Conner, 2011). In other words,
more frequent healthy diet is not related to either higher or lower depression risk. Also, the cross-
sectional association may demonstrate the importance of social support in maintaining a healthy
diet. A healthy diet is particularly related to social support from family members and significant
others (Early et al., 2009; King et al., 2010; Pollard, Zachary, Wingert, Booker, & Surkan, 2014;
Wen, Shepherd, & Parchman, 2004). A qualitative study demonstrated processes by which
Latino diabetes patients were reinforced to change their dietary habits via interaction with family
members and community members who also have diabetes (Pollard et al., 2014). Also, diabetes
patients had to negotiate with family members to maintain a similar diet for diabetes
management (Early et al., 2009). However, patients with clinical depression often face
interpersonal conflicts because depressive symptoms undermine the ability for fruitful interaction
with others (Hammen, 2006). As a result, depression patients are more likely to experience
discordance with family members and chronic illness patients with comorbid depression, thus,
experiencing reduced social support (Hickey et al., 2005). In one study with diabetes patients,
depression was found to reduce social support (Sacco & Yanover, 2006). Therefore, it is possible
that an increased healthy diet frequency may be a result of reduced depressive symptoms, rather
than a predictor, as we assumed in this study. However, this speculation is left untested because
the MDDP did not collect the extent of social support change over follow-ups.
Finally, SBGM, foot care, and BMI change do not seem to have a significant effect on
concurrent and future depressive symptoms. In previous observational studies, these three
behaviors demonstrated weak or no significant associations with depressive symptoms when
compared with diet and exercise (Ciechanowski et al., 2000; Ciechanowski et al., 2003;
70
Gonzalez et al., 2007; Lin et al., 2004). Also, RCTs examining the effect of depression care
showed similar results: no group difference of change in frequencies of SBGM, foot care, and
BMI (Lin et al., 2006; Williams et al., 2004). Thus, these behaviors may not be associated with
depressive symptoms or diagnosis of depression and may be influenced by other factors, such as
diabetes symptoms and occurrence of complications.
This study has several limitations for conducting secondary analysis. First, the findings
have limited generalizability because this study was conducted in safety-net clinics where most
of the patients were low-income and Hispanic. Thus, the findings should be carefully interpreted
when applied to patients with private health insurance or higher socioeconomic statuses. Second,
we don’t have complete inf ormation about the reasons for changes in self-care behavior
frequencies. Not all patients were exposed to the same intervention. The MDDP initially
randomized enrolled patients to either INT or EUC. Because patients assigned to these two study
arms were treated in the same clinics, significant levels of spillover effects were noticed (Ell et
al., 2010). Also, because of the nature of RCT, among INT, only 79.3% received PST during 12
months, and 58.5% received antidepressant medications, resulting in many untreated patients (Ell
et al., 2010). Therefore, some observed changes in self-care behavior frequency might not be
solely triggered by the interventions. However, a significant number of patients typically drop
out of treatment, and this limitation to treatment exposure was reflected in our study. Also, the
significant spillover effect across the two study arms helped to increase analytic power by not
removing data from EUC. Second, because the MDDP did not focus on self-care behavior
improvement, the extent of change appears smaller than other diabetes-management targeted
clinical trials (e.g., Minet et al., 2010). It is possible that some activities attempting to reduce
depressive symptoms may negatively influence adherence to self-care behaviors (Detweiler-
71
Bedell et al., 2008). In this study, the frequency of foot care declined by the 6-month follow-up,
and the level was not recovered to the level at baseline assessment. Therefore, the reduced
frequency might be the reason for null findings on the magnitude of associations, concurrently
and prospectively. In other words, the findings in our study should be limited to depression care,
rather than self-management trainings that probably trigger larger effect sizes on self-care
behavior changes. It is plausible that the negative effect of too much self-care behavior change
might be observed if the extent of change exceeds a certain critical point. Therefore, we suggest
studies testing the effectiveness of self-care behavior training. Finally, multiple comparisons is a
significant limitation because α = 0.05 will be statistically significant on the basis of chance
alone.
In conclusion, more frequent exercise predicted a significantly lower risk of depression
for up to the 1-year follow-up. A healthy diet was also inversely associated with depression, but
this association was not significant when examined prospectively. There are numerous studies
that show that depression predicts diet pattern. However, there are also studies showing that an
unhealthy diet, such as high-fat diet, may reduce depressive symptoms (e.g., Palinkas, Wingard,
& Barrett-Connor, 1996). Therefore, further studies are recommended to determine whether the
less frequent unhealthy diet has any correlation with future depressive symptoms. However,
changes in SBGM and foot care frequency and BMI change do not appear to be significant
covariates for future depression. These findings can inform practice. First, the increased
frequency of healthy behavior does not appear to create emotional distress for diabetes patients.
The increase of exercise showed a strong anti-depressive effect in our study and other
observational studies. Thus, practitioners may need to be concerned not with the actual
frequency of self-care behaviors, but with patient’s perce ption of diabetes management. A couple
72
of good measures are available to measure the extent of perceived diabetes distress and data
collected that demonstrated a robust association with depression and diabetes outcomes (Aikens,
2012; Delahanty et al., 2007; Pouwer, 2009). In addition, exercise should be included in
depression care for diabetes patients, in addition to cognitive activities. Exercise has been
proposed as a key therapeutic behavior for chronic illness patients to improve both physical and
mental health outcomes (Detweiler-Bedell et al., 2008). However, activating exercise is also
affected by depressive symptoms. Therefore, we suggest further clinical trials to examine the
effect of hybrid models having evidence-based psychotherapy or pharmacotherapy and exercise
components.
73
Chapter 4. Mediation of Self-Care Behaviors in the Relationship between Depression, Diabetes
Symptoms, and Daily Functioning
Depression is a significant comorbidity contributing to more self-reported diabetes
symptoms (Ali et al., 2006; Ciechanowski et al., 2003; Goldney, Phillips, Fisher, & Wilson,
2004; Katon et al., 2009a; Ludman et al., 2004; Von Korff et al., 2005a; Von Korff et al.,
2005b). A cross-sectional study found that depression is associated with a higher risk of
reporting diabetes symptoms, ranging from 1.93 times (cold hands and feet) to 4.96 times
(daytime sleepiness) (Ludman et al., 2004). Also, self-reported diabetes symptoms have
exhibited a dose-response relationship with depression symptoms (Ciechanowski et al., 2003).
Similar findings have been observed when the relationship between depression and self-reported
daily functioning was investigated (Black et al., 2003; Ciechanowski et al., 2003; Egede,
Grubaugh, & Ellis, 2010; Goldney et al., 2004; McKellar, Humphreys, & Piette, 2004;
Paschalides et al., 2004). Patients with both diabetes and depression have reported significantly
higher functional impairment (Paschalides et al., 2004). According to a population-based study
that examined the subdomains of the physical composite scale (PCS) in the Short Form-36 health
survey (SF-36), depressed diabetes patients demonstrated a 20% lower score in all domains
within the PCS when compared with groups with either diabetes or depression only (Goldney et
al., 2004). Longitudinal studies have also supported the significant influence of depression on
future reports of diabetes symptoms and daily functioning (Black et al., 2003; Katon, Lin, &
Kroenke, 2007; McKellar et al., 2004).
Katon (2003) proposed a conceptual model that posited mediating roles of self-care
behaviors in the relationship between depressive symptoms and consequences of diabetes,
including diabetes symptoms and daily functioning (p. 218). Reduced motivation to adhere to
74
prescribed self-care behaviors is affected by some of the cardinal symptoms in depression,
including lacking interest in usual jobs, decreased pleasure, and helplessness (Egede, 2005;
Gonzalez et al., 2008a; Katon, 2003). A meta-analysis demonstrated that depressed patients with
diabetes were less likely to follow overall self-care behaviors, diet, exercise, SBGM, and
medication adherence (Gonzalez et al., 2008a). Depression may undermine patients’ self-
efficacy to engage in self-care behaviors that are usually prescribed by primary care physicians
and dieticians (Katon et al., 2009b; Katon, 2008). For instance, the significant relationship
between depression and medication adherence at 3 months follow-up was explained by the level
of self-efficacy (Schoenthaler et al., 2009). Also, the conceptual model is supported by findings
from observational studies that demonstrated significant associations between depressive
symptoms and some self-care behaviors (Ciechanowski et al., 2003; Egede et al., 2009; Gonzalez
et al., 2007; Gonzalez et al., 2008a; Gonzalez et al., 2008b; Wagner et al., 2010). For instance, a
population-based study demonstrated the lowest adherence to exercise in a group with major
depression and diabetes when compared with diabetes patients either with no or minor
depression (Egede et al., 2009). Also, the severity of depression and the extent of adherence to
self-care behaviors had a dose-response relationship across exercise, healthy diet amount, and
diet choices (Ciechanowski et al., 2003). However, less-consistent findings were found in SBGM
and foot care (Ciechanowski et al., 2003; Gonzalez et al., 2007; Gonzalez et al., 2008a; Lin et al.,
2004). Nonetheless, the impact of depressive symptoms on self-care behaviors is fairly
persuasive. Addressing depression seems paramount in helping patients follow self-care
behaviors that are routinely prescribed by health professionals (American Diabetes Association,
2013). Observational studies have provided direct evidence supporting the mediating roles of
self-care behaviors (Chiu et al, 2010; McGrady, Laffel, Drotar, Repaske, & Hood, 2009;
75
McKellar et al., 2004). The mediation can be directly tested by a methodological scheme
proposed by Baron & Kenny (1986). One study found that a composite index, which
demonstrates the extent of adherence to recommended diet and general self-care behaviors, was a
significant mediator in the relationship between depressive symptoms and self-reported diabetes
symptoms measured after one year (McKellar et al., 2004). The mediating role of self-care
behaviors was also found in a study that investigated the relationship between depressive
symptoms and glycemic control, measured by A1C (Chiu et al., 2010). When SBGM was
measured with an objective measure, the effect size for the mediator was larger than the one in
studies with self-reported self-care behaviors (McGrady et al., 2009).
However, it is unclear whether mediating roles of self-care behaviors can be observed in
clinical studies that involve an intervention addressing depression (Lin et al., 2006; Markowitz et
al., 2011). Although many depression care interventions have assumed the linear path proposed
in the conceptual model in Katon (2003), few empirical results are available (Ell et al., 2010;
Katon et al., 2004a; Katon et al., 2010; Williams et al., 2004). For instance, collaborative
depression care is seen as promising because this treatment provides an opportunity to treat
depression with either psychological or pharmacological treatments, or both, based on the
preference of patients in a stepped-care approach (Katon et al., 2010). The depression care is
designed to facilitate collaboration among providers with different roles in a primary care clinic.
The care model demonstrated improvements in self-reported diabetes symptoms and daily
functioning, as well as depression measures (Ell et al., 2010; Katon et al., 2004a; Katon et al.,
2010; Williams et al., 2004). However, it is not clear whether the improved diabetes symptoms
and daily functioning were mediated by more frequent self-care behaviors (Lin et al., 2006). For
instance, there was no difference in adherence to exercise and diet between the INT and EUC
76
groups when two-fold patients in the INT group experienced more than a 50% decrease in
depression severity and better self-rated quality of life and global functioning (Katon et al.,
2010). Another study also showed no improvement in a healthy diet frequency and exercise
frequency in the INT group, despite showing significantly improved mood (Lin et al., 2006).
Some theoretical explanations can be offered for the difference in findings from observational
and intervention studies. The assumed linear model path depression to diabetes-related outcome
by virtue of change in self-care behavior is too simplistic and underestimates the complex nature
of human motivation related to healthy behaviors. For instance, the linear relationship might be
tarnished by conflicting effects of improved depression on disease management (Detweiler-
Bedell et al., 2008). Detweiler-Bedell et al. (2008) proposed underlying mechanisms by which
this insignificant or counterintuitive effect is observed in clinical studies examining depression
care. They proposed a m is- or under-regulation of diabetes management by depression treatment
(Detweiler-Bedell et al., 2008). The mis-regulation may happen when reduction in depressed
affect enables patients to engage in social activities, such as family feasts involving high-
carbohydrate intakes (Detweiler-Bedell et al., 2008). Patients who experience improved
depression are more likely to hang out with friends and family members (Perrino et al., 2011). In
a qualitative study, Hispanic patients reported increased temptation to unhealthy food when they
met friends and family members (Early et al., 2009). Also, lower negative affect undermines
diabetes patients’ abilit ies to notice change in blood glucose level (Ryan, Dulay, Suprasongsin,
& Becker, 2002). As a result, some patients may engage less in diabetes management. The
under-regulation of depression care is found when patients put more priority on depression
treatment and focus less on diabetes management. Adherence to prescribed self-care behaviors
requires a great amount of efforts and sacrifice of alternative pleasant behaviors (American
77
Diabetes Association, 2013; Detweiler-Bedell et al., 2008). This effect was noticed in a clinical
trial that examined the effectiveness of cognitive behavioral therapy and showed undermined
SBGM, as well as significant decline of depressive symptoms (Lustman et al., 1998). The
authors argued that the finding might be attributable to a situation where patients’ adherence to
daily self-care behaviors were distracted by new assignments given during cognitive behavioral
therapy for depression (Lustman et al., 1998).
However, existing literature does not have evidence to answer to the mediating role of
self-care behaviors because of methodological issues. There exists only circumstantial evidence
without direct tests for the mediation relationship. Observed insignificant mean differences in
self-care behaviors between the INT and control groups was proposed as an evidence to reject
the hypothesized mediation relationship (e.g., Lin et al., 2006). Therefore, a mediation analysis
proposed by Baron & Kenny (1986) seems to be a more appropriate method to examine the
mediation. Specifically, we need to examine whether the change of depressive symptoms
following the implementation of depression care is predictive of change of self-care behaviors
and results in self-rated diabetes symptoms and functional impairment. To test the mediating role
of self-care behavior change in depression care, two methodological conditions need to be met.
First, a change in depression symptoms needs to be a predictor, rather than study group
membership. This allows for the determination of whether reduced depression severity has
sufficient power to change the extent of self-care behavior. Second, temporal order among the
three variables should be set up. Theoretically, depression decline should proceed the change of
self-care behavior adherence, and diabetes-related outcomes should follow the self-care behavior
measuring point. These two methodological conditions are important for validating the linear
path proposed.
78
Figure 4. Path model of mediation analysis
This dissertation demonstrated empirical tests to determine the hypothesized mediating
role of self-care behaviors within depression care, as shown in Figure 4. For this goal, secondary
analyses were performed with data collected in MDDP (Ell et al., 2010). In the previous study, a
group with collaborative depression care demonstrated more improvement in self-reported
diabetes symptoms and daily functioning, as well as depression, compared with the EUC. To
address methodological issues found, data collected at four follow-ups were examined: at
baseline, 6, 12, and 18 months. Two analyses were conducted. First, unadjusted path analysis
with structural equation modeling tested the mediational role when the baseline values for the
three variables were controlled. Also, adjusted mediation analyses were conducted by controlling
for more potential confounders in demographic and clinical domains.
79
METHODS
Sample and Procedure
This study conducted a secondary analysis of data collected in the MDDP clinical trial,
which enrolled 387 patients from two public safety-net clinics whose primary patients are low-
income, Hispanics, and recent immigrants (Ell et al., 2010; Ell et al., 2012). The total number of
diabetes patients who were identified in the study sites were 1,803, and 30.2% (n = 523) of them
were screened for probable major depressive disorder, resulting in 492 patients eligible for the
study (Ell et al., 2010). The MDDP recruited both patients with type 1 (n = 8) and type 2 (n =
379) diabetes mellitus (T2DM). Details on inclusion and exclusion criteria are available in Ell et
al. (2010; 2012).
Patients in the INT group were offered the structured stepped-care algorithm 12-month
intervention, which systematically administered either AM or PST, depending on the patient’s
preference, responsiveness to dose, and type of previous treatment (Ell et al., 2010; Katon et al.,
2010). The MDDP had tailored the manualized intervention to be culturally relevant to low-
income Hispanic patients and to address socioeconomic stress believed to contribute to
depressive symptoms (Cabassa et al., 2008; Ell et al., 2010). The EUC group was given
educational pamphlets describing depression and self-care for diabetes patients and a list for
community resources addressing nonmedical needs of the patients. Also, primary care physicians
(PCPs) were notified of the probable major depressive disorder diagnosis. These components in
the EUC seem to contribute to more than a two-fold increase of antidepressant receipt over 12
months (26.8%) since enrollment when compared with that of baseline survey (12.7%) (Ell et al.,
2010).
80
In this particular analysis, we used data from both the INT and the EUC groups. Data on
depression, self-care behaviors, disability, and diabetic symptoms were measured four times,
including at baseline, 6-, 12-, and 18-month follow-ups. According to Ell et al. (2011), the INT
group retained 78.2%, 73.6%, and 74.6% of its participants at 6-, 12-, and 18-months follow-up,
respectively, whereas the EUC group retained 78.4%, 71.6%, and 70.6% of its participants,
showing higher attrition rates in the EUC. As patients participated in assessments at these four
time points, 226 patients (58.4%) were analyzed for this particular study.
Table 10 lists the demographic and clinical variables for analyzed and excluded patients.
Patients with less than a high school education were more likely to remain in the study (p < .05),
and patients with more than a high school education were more likely to drop out (p < .05). Also,
the proportions of patients speaking only Spanish was higher in the analyzed group (p < .05). No
significant differences between the two groups were found.
81
Table 10
Description for demographic and clinical variables for all patients, analyzed patients, and
excluded patients
Analyzed
N = 226
(58.4%)
Drop-out
N = 161
(41.6%)
Group
test
P value
Age ≥50 166 (73.5%) 113 (70.2%) .48
Intervention group
117 (51.8%) 76 (47.2%) .38
Lower than High school
a
194 (85.8%) 123 (76.4%) .02
Some college or graduates
a
8 (3.5%) 14 (8.7%) .03
Female
188 (83.2%) 130 (30.7%) .54
Separated
b
82 (36.3%) 67 (41.6%) .29
Never married
b
25 (11.1%) 22 (13.7%) .44
Dysthymia
126 (55.8%) 88 (54.7%) .83
Only Spanish
c
201 (88.9%) 126 (78.3%) .00
Insulin use 66 (29.2%) 41 (25.5%) .42
Mean (SD), PHQ-9 at baseline 14.54 ± 2.90 14.98 ± 3.01 .15
Mean (SD), Diet at baseline 4.28 ± 1.69 4.26 ± 1.92 .89
Mean (SD), Exercise at baseline 2.09 ± 2.39 1.83 ± 2.32 .29
Mean (SD), SBGM at baseline 2.82 ± 2.81 2.41 ± 2.79 .15
Mean (SD), Foot care at baseline 5.07 ± 2.77 4.77 ± 2.86 .31
Mean (SD), BMI at baseline 32.74 ± 7.03 33.14 ± 8.25 .61
Mean (SD), Diabetes symptoms at baseline 17.78 ± 5.95 18.16 ± 6.15 .55
Mean (SD), PCS at baseline 35.94 ± 9.00 35.30 ± 9.33 .50
Note.
a
reference group: high school diploma;
b
reference group: currently married;
c
reference
group: English only speaking or bi-lingual patients. SD, standard deviation; PHQ-9, Patient
Health Question-9.
82
Measures
The PHQ-9 was selected to measure the extent of change during 6 months after patient’s
enrollment in the MDDP (Kroenke & Spitzer, 2002; Löwe et al., 2004; Spitzer et al., 1999). The
PHQ-9 was administered as a screening tool in the MDDP with cutoff score at least 10, which
demonstrated adequate sensitivity and specificity for diagnosis (Kroenke & Spitzer, 2002;
Wittkampf et al., 2007). The PHQ-9 asked nine items from a structural interview for diagnosis of
major depressive disorder in the DSM-4, and respondents were asked to select one of four
options ranging from “ not at all” to “ nearly every day” (Löwe et al., 2004). To obtain the
depression severity, answers from the nine items were summed and their combined value ranged
from 0 to 27. Reliability of the PHQ-9, including test-retest correlation and internal consistency,
was found to be adequate for primary care clinic patients (Spitzer et al., 1999). Concurrent
validity was tested with results from the structural interview for depression diagnosis (Spitzer et
al., 1999).
Self-care behaviors included five types: the frequency of 1) healthy diet, 2) exercise, 3)
SBGM, 4) foot care, and 5) BMI change. To measure the frequency of the four behaviors, the
SDSCA was used (Toobert & Glasgow, 1994; Toobert et al., 2000). Patients were asked to
remember the number of days in which healthy behavior was implemented during the previous
week. We averaged answers from items within each subdomain and used the mean number of
days. The SDSCA was initially developed in 1994 and was revised in 2000. Toobert et al. (2000)
proposed four subdomains: the four behaviors described above and smoking. In this study,
smoking was excluded from the analysis because only a few patients were current smokers.
Toobert et al. (2000) suggested not to aggregate scores from each subdomains because these
behaviors may not be adequately correlated with one another. Too low correlations among
83
subdomains may result in increased random error. Thus, individual scores from each subdomain
were analyzed.
The Self-Completion Outcome Measure (SCOM) was used to measure the extent of
diabetic symptoms that were perceived by patients (Whitty et al., 1997). This measure has 9
items that asked the extent of perception on nine cardinal symptoms among diabetes patients,
including abnormal thirst, blurred vision, excessive urine, unusual hunger, shakiness, cold hands
and feet, sleepiness, feeling pins and needles, and faintness. This instrument is useful to measure
the diabetes-specific conditions rather than general health (Whitty et al., 1997). Respondents
were asked to retrospectively count the number of days when each symptom was perceived.
Answering options included 1 (never), 2 (on one or a few days), 3 (on several days), 4 (on most
days), and 5 (every day). To summarize the level of self-reported diabetes symptoms, answers
were summed for a total score ranging from 9 to 45 (Whitty et al., 1997). The internal
consistency of this measure was .78 (McColl et al., 1995). With respect to validity, this measure
had significant correlations with the SF-36, which asked about generic health conditions, ranging
from .55 to .63 (McColl et al., 1995). Also, patients experienced a decline in SCOM score and
the level of A1C and non-fasting serum cholesterol when these patients enrolled in insulin
therapy (Whitty et al., 1997). Finally, in natural settings, the level of diabetes symptoms was
predictive of higher odds of having poorer glycemic control, higher risk of complications, self-
rated disability, and functional limitations (Ludman et al., 2004; Von Korff et al., 2005b).
To measure daily functioning, the PCS in the Medical Outcomes Study Short-Form
Health Survey (MOS SF-12) was used (Ware, Kosinski, & Keller, 1996). The PCS in MOS SF-
12 measures the extent of limitations that respondents experienced daily (Ware et al., 1996). The
PCS assessed the extent of physical functioning, roles related to physical health, bodily pain, and
84
general health (Ware et al., 1996). The SF-12 is a brief version of the SF-36 (Ware &
Sherbourne, 1992). The MOS SF-12 is a norm-based measure that has a mean of 50 with 10 as a
standard deviation. The mean of 50 represents the health-related quality of life for the general
U.S. population (Ware et al., 1996). Scoring follows predetermined algorithms and provides
scores ranging between 0 (high functional impairment) and 100 (minimal functional
impairment). With respect to psychometric characteristics, the test-retest correlation of .89 and
.86 were reported in studies with samples from the U.S. and the U.K. (Ware et al., 1996). The
construct validity was also evaluated by comparisons with scores measured with the SF-36
(Ware et al., 1996). The PCS in the SF-12 was used in a study that investigated the correlation
between depression and the functional impairment (Ciechanowski et al., 2003).
Analysis
A descriptive analysis was conducted to find the mean and standard deviations for
depression, diet, exercise, SBGM, foot care, BMI, self-reported diabetes symptoms, and daily
functioning. Data on the variables measured at baseline, 6-, 12-, and 18-month follow-up were
recorded. Mean comparisons were conducted by the level of change in depressive symptoms
between baseline and 6-month follow-up to explore possible correlations. Specifically, we
ranked the extent of depression change during the 6-month period and dichotomized the change
score of patients by two groups: lower 50% vs. upper 50%. Independent t-tests were conducted
to determine whether the mean of the focal variables for the two groups were significantly
different. A two-tailed significance level of 95% was used.
To test the existence of the mediation relationship, both unadjusted and adjusted analyses
were conducted. As an unadjusted analysis, path analyses were conducted. The conceptual model
depicted in Figure 3 was directly tested and controlled for baseline values for a predictor, a
85
mediator, and an outcome. M-plus version 7.11 (Muthen & Muthen, Los Angeles, CA) was used
to find effect sizes and test for significance for two relationships. Also, the bootstrapping method
with 1,000 times sampling was used to examine the significance of an indirect effect mediated by
each self-care behavior, resulting in a 95% CI of the mediation effect. The bootstrapping method
is seen as an appropriate analysis to overcome the violation of statistical assumption in the Sobel
test (Shrout & Bolger, 2002). The Sobel test inherently violates the normality assumption
because this test assumes the multiplied ab is normally distributed, which is not viable. Second,
the bootstrapping method requires a substantially small sample size compared with the three-
stepped method (Fritz & MacKinnon, 2007). Therefore, this method is more appropriate for
studies using data collected in clinical trials. Second, the adjusted mediation analysis was
conducted by controlling for confounders, including age, study group, education, gender, marital
status, dysthymia, the severity of depression at baseline, primary language use, insulin treatment
at baseline, baseline values of depression, focal self-care behavior, and the outcome. For this
adjusted analysis, MEDIATE macro (Hayes, n.d.) installed in SPSS 21.0 (Chicago, IL, USA)
was used. The macro also provided a 95% CI of a mediation effect with the bootstrapping
method. Finally, sensitivity analysis was conducted with different sets of data (data not shown).
Because the MDDP followed patients up to 24-months post-baseline, sensitivity analysis with
different data points was possible. Thus, models with self-care behaviors measured at the 18-
month follow-up and diabetes outcomes measured at the 24-month follow-up were examined to
determine whether relationships with different lagged times between a predictor and mediator
changes the results from significant tests.
86
RESULTS
Table 11 shows the mean and standard deviations of depression and self-care behaviors at
baseline, 6-, 12-, and 18-month follow-up for the analyzed sample (N = 226). The mean of
depression declined between baseline and 6 months post-baseline and remained stable afterward.
Notably, the standard deviation of depression was amplified since the 6-month follow-up,
suggesting an increased variation of depressive symptoms across patients subsequent to
treatment. In other words, a symptom difference among patients possibly became larger because
of different levels of adherence and responsiveness to available depression care regimens (e.g.,
antidepressants, PST, or other counseling). In comparisons between groups selected by the extent
of depression change during six months, significant mean differences at every assessment were
found (p < .001). Interestingly, the group with the largest decline during 6 months since baseline
had a higher baseline PHQ-9 score but lower depressive symptoms at the 6-, 12-, and 18-month
follow-up. Self-care behavior frequency showed less increase between baseline and 6-month
assessment compared with the findings in depression. Exercise and foot care frequencies even
declined slightly at the 6-month follow-up. Few notable changes were found in the mean
comparisons by depressive symptoms change during 6 months post-baseline with the exception
of diet frequency. Specifically, the upper 50% change group had significantly more frequent
healthy diet weekly than the counterpart at the 6- (p < .05), 12- (p < .05), and 18-month follow-
up (p < .05). Finally, self-reported diabetes symptoms and daily functioning had sudden changes
at the 6-month assessment, similar to depressive symptoms. Also, in the mean comparison
analysis, diabetes symptoms declined more in the group with the upper 50% change in
depressive symptoms at the 6- (p < .001), 12- (p < .01), and 18-month follow-up (p < .01).
Similar results were observed in the PCS during follow-up.
87
Table 11
Mean comparisons of self-care behaviors and self-reported diabetes symptoms and daily
functioning depending on the extent of change in depressive symptoms during the first 6 months
since enrollment in the MDDP (lower 50% vs. upper 50%) (N = 226)
Depression
change
Baseline
Mean (SD)
6 month
Mean (SD)
12 month
Mean (SD)
18 month
Mean (SD)
Depression All 14.54 (2.90) 7.09 (5.24) 7.00 (5.27) 7.94 (5.83)
Lower 50% 13.95 (2.61)** 10.91 (4.36)*** 8.71 (5.13)*** 9.67 (5.84)***
Upper 50% 15.10 (3.07)** 3.40 (2.80)*** 5.36 (4.88)*** 6.27 (5.32)***
Diet All 4.28 (1.69) 4.45 (1.45) 4.47 (1.56) 4.43 (1.74)
Lower 50% 4.08 (1.70) 4.21 (1.55)* 4.24 (1.60)* 4.16 (1.79)*
Upper 50% 4.48 (1.67) 4.67 (1.30)* 4.69 (1.50)* 4.64 (1.66)*
Exercise All 2.09 (2.39) 1.98 (1.91) 2.37 (2.22) 2.79 (2.55)
Lower 50% 1.97 (2.32) 2.00 (1.98) 2.29 (2.23) 3.00 (2.67)
Upper 50% 2.20 (2.46) 1.97 (1.85) 2.45 (2.23) 2.59 (2.43)
SBGM All 2.82 (2.81) 3.06 (2.52) 2.70 (2.53) 3.21 (2.83)
Lower 50% 2.86 (2.79) 3.07 (2.66) 2.69 (2.62) 3.13 (2.79)
Upper 50% 2.79 (2.83) 3.06 (2.38) 2.70 (2.45) 3.29 (2.88)
Foot care All 5.07 (2.77) 4.61 (2.02) 4.72 (1.99) 5.19 (2.08)
Lower 50% 5.07 (2.60) 4.44 (2.08) 4.63 (2.08) 5.14 (2.15)
Upper 50% 5.07 (2.93) 4.77 (1.95) 4.80 (1.91) 5.24 (2.03)
BMI
a
All 32.74 (7.03) 32.34 (6.96) 32.60 (6.92) 32.42 (7.14)
Lower 50% 32.45 (7.01) 32.49 (7.42) 32.55 (6.92) 32.52 (7.11)
Upper 50% 33.01 (7.08) 32.19 (6.51) 32.66 (6.96) 32.31 (7.19)
Diabetes All 17.78 (5.95) 14.89 (4.96) 14.78 (4.81) 15.54 (5.99)
88
symptoms
Lower 50% 18.06 (6.01) 16.63 (5.35)*** 15.87 (4.90)** 16.72 (6.23)**
Upper 50% 17.51 (5.90) 13.21 (3.90)*** 13.72 (4.51)** 14.39 (5.55)**
PCS All 35.94 (9.00) 40.61 (11.17) 40.10 (11.58) 40.63 (11.21)
Lower 50% 36.53 (9.36) 38.21 (10.76)** 37.86 (11.51)** 38.90 (10.46)*
Upper 50% 35.38 (8.64) 42.93 (11.11)** 42.25 (11.28)** 42.30 (11.70)*
*p < .05; **p < .01; ***p < .001.
Figure 4 demonstrates parameters, standard errors, and significance test results from the
10 models. These models were unadjusted and examined with path analysis. As opposed to
Katon (2003), no mediation relationship was found to be significant (p > .05). In other words,
change in frequency of each self-care behavior did not significantly explain the improvements in
self-reported diabetes symptoms and daily functioning, which were significantly predicted by the
reduction of depressive symptoms gained during 6 months since the enrollment. Other than that,
only the change of diet was associated with change in diabetes symptoms (p < .01) and daily
functioning (p < .05).
89
Figure 5. Unadjusted results from path analyses with each self-care behavior as a mediator
Tables 12, 13, 14, 15, and 16 demonstrate results from adjusted mediation analyses. First,
in Path c, total effect of change in depressive symptoms on both diabetes symptoms and daily
functioning were significant (p < .001). The effect sizes in these two variables seem to be
clinically meaningful. However, Paths a and b demonstrated insignificant associations (p < .05)
in most mediation analyses. For instance, changes in depressive symptoms were not associated
with a change in the frequency of healthy diet (p > .05), SBGM (p > .05), foot care (p > .05),
and BMI change (p > .05). We found that decline of depressive symptoms predicted more
frequent exercise (p < .05). Because the average decline during 6 months was 7.5 in the
depression measurement, this extent of change in depression severity estimated a 0.6-day more
frequent exercise in a week on average in the analysis. With respect to Path b, increased exercise
90
frequency predicted increased daily functioning (p < .05), increased SBGM frequency predicted
decreased diabetes symptoms (p < .05), and increased BMI change predicted lower daily
functioning (p < .05). Other than one anomaly, evidence for a significant mediation relationship
was not found (p > .05). Only the change in exercise frequency showed a significant mediation
effect for the PCS (p < .05), demonstrating that 13.6% of the association between depression and
daily functioning was explained by this mediator. According to sensitivity analysis with different
lagged time between depression change and self-care behavior frequency change, no significant
result supporting mediation relationship was found.
91
Table 12
Evaluating the mediation of diet change (baseline - 12 months) in the association between
change of depression (baseline - 6 months) and self-reported diabetes symptoms and functional
impairment (baseline - 18 months)
Diabetes Symptoms Daily Functioning
Mediation test with bootstrapping method
a
B SE LL UL B SE LL UL
Indirect effects .02 .02 -.00 .07 -.00 .02 -.05 .04
total effect model
B SE t P B SE t P
Path c (total effect) .28 .08 3.55 .00 -.47 .15 -3.16 .00
R
2
.28 .22
causal steps approach
Path a (DEP → Diet) -.04 .02 -1.67 .10 -.03 .02 -1.50 .14
Path b (Diet → DV) -.46 .25 -1.85 .07 .08 .49 .17 .86
Path c` (DEP → DV) .26 .08 3.35 .00 -.47 .15 -3.12 .00
R
2
.29 .22
Note. LL, lower limit; UL, upper limit; DEP, depression; DV, dependent variable;
Linear
regression adjusted for age (<50 vs. 50+), study groups (intervention vs. enhanced usual care),
education (lower vs. high school vs. above), gender, marital status (separated vs. marriage vs.
never married), dysthymia at baseline, PHQ-9 score at baseline, primary language (Spanish vs.
English only or bilingual), insulin treatment at baseline, and baseline values of predictors.
Sample size: n = 226;
a
Sampling was repeated 1,000 times.
92
Table 13
Evaluating the mediation of exercise change (baseline - 12 months) in the association between
change of depression (baseline - 6 months) and self-reported diabetes symptoms and functional
impairment (baseline - 18 months)
Diabetes Symptoms Daily Functioning
Mediation test with bootstrapping method
a
B SE LL UL B SE LL UL
Indirect effects .02 .01 -.00 .05 -.06 .03 -.14 -.01
total effect model
B SE t P B SE t P
Path c (total effect) .24 .08 3.09 .00 -.44 .15 -3.02 .00
R
2
.26 .23
causal steps approach
Path a (DEP →
EXER)
-.08 .03 -2.60 .01 -.08 .03 -2.68 .01
Path b (EXER→ DV) -.21 .17 -1.23 .22 .67 .32 2.05 .04
Path c` (DEP → DV) .22 .08 2.83 .01 -.39 .15 -2.62 .01
R
2
.26 .24
Note. LL, lower limit; UL, upper limit; DEP, depression; DV, dependent variable;
Linear
regression adjusted for age (<50 vs. 50+), study groups (intervention vs. enhanced usual care),
education (lower vs. high school vs. above), gender, marital status (separated vs. marriage vs.
never married), dysthymia at baseline, PHQ-9 score at baseline, primary language (Spanish vs.
English only or bilingual), insulin treatment at baseline, and baseline values of predictors.
Sample size: n = 226;
a
Sampling was repeated 1,000 times.
93
Table 14
Evaluating the mediation of SBGM change (baseline - 12 months) in the association between
change of depression (baseline - 6 months) and self-reported diabetes symptoms and functional
impairment (baseline - 18 months)
Diabetes Symptoms Daily Functioning
Mediation test with bootstrapping method
a
B SE LL UL B SE LL UL
Indirect effects .01 .02 -.02 .04 .00 .01 -.02 .05
total effect model
B SE t P B SE t P
Path c (total effect) .24 .08 3.11 .00 -.46 .15 -3.16 .00
R
2
.26 .22
causal steps approach
Path a (DEP → Blood) -.02 .03 -.45 .65 -.01 .03 -.27 .79
Path b (Blood→ DV) -.38 .16 -2.38 .02 -.24 .31 -.77 .44
Path c` (DEP → DV) .24 .08 3.07 .00 -.46 .15 -3.17 .00
R
2
.28 .22
Note. LL, lower limit; UL, upper limit; DEP, depression; DV, dependent variable;
Linear
regression adjusted for age (<50 vs. 50+), study groups (intervention vs. enhanced usual care),
education (lower vs. high school vs. above), gender, marital status (separated vs. marriage vs.
never married), dysthymia at baseline, PHQ-9 score at baseline, primary language (Spanish vs.
English only or bilingual), insulin treatment at baseline, and baseline values of predictors.
Sample size: n = 226;
a
Sampling was repeated 1,000 times.
94
Table 15
Evaluating the mediation of foot care change (baseline - 12 months) in the association between
change of depression (baseline - 6 months) and self-reported diabetes symptoms and functional
impairment (baseline - 18 months)
Diabetes Symptoms Daily Functioning
Mediation test with bootstrapping method
a
B SE LL UL B SE LL UL
Indirect effects .00 .01 -.01 .04 .00 .01 -.02 .04
total effect model
B SE t P B SE t P
Path c (total effect) .24 .08 3.03 .00 -.47 .15 -3.22 .00
R
2
.26 .22
causal steps approach
Path a (DEP → Foot) -.02 .03 -.53 .60 -.02 .03 -.66 .51
Path b (Foot→ DV) -.28 .18 -1.52 .13 -.07 .35 -.19 .85
Path c` (DEP → DV) .23 .08 2.98 .00 -.47 .15 -3.22 .00
R
2
.27 .22
Note. LL, lower limit; UL, upper limit; DEP, depression; DV, dependent variable;
Linear
regression adjusted for age (<50 vs. 50+), study groups (intervention vs. enhanced usual care),
education (lower vs. high school vs. above), gender, marital status (separated vs. marriage vs.
never married), dysthymia at baseline, PHQ-9 score at baseline, primary language (Spanish vs.
English only or bilingual), insulin treatment at baseline, and baseline values of predictors.
Sample size: n = 226;
a
Sampling was repeated 1,000 times.
95
Table 16
Evaluating the mediation of BMI change (baseline - 12 months) in the association between
change of depression (baseline - 6 months) and self-reported diabetes symptoms and functional
impairment (baseline - 18 months)
Diabetes Symptoms Daily Functioning
Mediation test with bootstrapping method
a
B SE LL UL B SE LL UL
Indirect effects -.00 .01 -.01 .01 .00 .02 -.04 .05
total effect model
B SE t P B SE t P
Path c (total effect) .24 .08 3.08 .00 -.45 .15 -3.07 .00
R
2
.26 .22
causal steps approach
Path a (DEP → BMI) -.00 .04 -.10 .92 -.00 .04 -.09 .93
Path b (BMI→ DV) .02 .13 .12 .91 -.56 .24 -2.31 .02
Path c` (DEP → DV) .24 .08 3.08 .00 -.45 .15 -3.11 .00
R
2
.26 .24
Note. LL, lower limit; UL, upper limit; DEP, depression; DV, dependent variable;
Linear
regression adjusted for age (<50 vs. 50+), study groups (intervention vs. enhanced usual care),
education (lower vs. high school vs. above), gender, marital status (separated vs. marriage vs.
never married), dysthymia at baseline, PHQ-9 score at baseline, primary language (Spanish vs.
English only or bilingual), insulin treatment at baseline, and baseline values of predictors.
Sample size: n = 225;
a
Sampling was repeated 1,000 times.
96
DISCUSSION
In our knowledge, this study is the first effort to validate hypothesized mediating role of
self-care behaviors in the observed significant association between depression severity and
diabetes-related outcomes by conducting empirical analysis of clinical data. It has been unclear
whether reduced depression severity may lead to more frequent self-care behaviors and result in
subsequent improvement in diabetes-related outcomes. We tested the hypothesis with data that
demonstrated sudden depressive symptoms decreases between baseline and a 6-month follow-up.
In the results, no significant mediation relationship was found in unadjusted models. In adjusted
analyses, only exercise frequency was found to have a significant indirect effect.
Examining the findings with previous studies on related topics (e.g., Lin et al., 2006), we
tentatively conclude that Katon’s (2003) proposed path model is not observed other than for
exercise. Thus, the observed depressive symptoms decrease gave a direct effect on the
significantly improved self-reported diabetes symptoms and daily functioning, rather than the
indirect effect posited in Katon (2003). Lacking power for significance test does not appear to
explain the null findings because the effect sizes for indirect effect were very trivial in most of
the models. In most mediational models, Path a was insignificant, suggesting that most self-care
behaviors other than exercise were not affected by the depression change during 6 months since
enrollment. The findings were contrary to the existing data in observational studies (Gonzalez et
al., 2007; Gonzalez et al., 2008a; Gonzalez et al., 2008b; Katon et al., 2004b; Wagner et al.,
2010). The discrepancy between observational studies and clinical studies may be explained by
more than one theory. First, it is possible that depression change is necessary but not sufficient to
cause actual behavioral changes. Although given the efficacious depression intervention,
environmental factors, which had strong impact on patient’s self-care behavior (King et al.,
97
2010), might not be changed. Community is an important social outlet that either facilitates or
hinders positive behaviors for diabetes (Shaw, Gallant, Riley-Jacome & Spokane, 2006). Thus, it
is important to examine the roles of community resources for the effectiveness of depression care
on diabetes management.
Also, the significant associations found in the observational studies are between-person
differences in depressive symptoms and self-care behavior frequency, whereas clinical trials
examined the correlation between within-person changes in depressive symptoms and in self-
care behaviors. Thus, it is possible that the effect size of correlation between within-person
changes in the two variables is reduced because other factors contributing to self-care behavior
frequency have not changed. Some examples may include illness representation (Searle, Norman,
Thompson, & Vedhara, 2007), diabetes-related burden (Fisher et al., 2010), and self-efficacy
(King et al., 2010). Although these variables are correlated with depressive symptoms (Fisher et
al., 2010), the effect of depression care on these exogenous variables is subdued because a
product term between depressive symptoms and the outside variable shrink the effect size as an
indirect effect (Baron & Kenny, 1986).
Third, as proposed in Detweiler-Bedell et al. (2008), depression care may have
conflicting effects on diabetes management. The significant association of depression with self-
care behaviors in observational studies is subject to noise, which would increase random errors.
As Lustman et al. (1998) showed, depression care might work as a barrier for diabetes patients to
focus on burdensome diabetes management. Depression may be a product of commitment to
diabetes management, which requires a serious workload to meet the recommended guidelines
(Russell et al., 2005; Safford et al., 2005). It is also possible that patients underestimate diabetes-
related symptoms as their depressive symptoms decline (Detweiler-Bedell et al., 2008). In an
98
experimental study that assessed whether anxiety symptoms are associated with accuracy of self-
perceived hypoglycemia among type 1 diabetes patients, patients with increased anxiety
symptoms demonstrated increased sensitivity to physical symptoms (Ryan et al., 2002). Because
the level of symptoms perceived by a patient is a stronger predictor for care seeking and
engaging in self-care behaviors (Leventhal, Brissette, & Leventhal, 2003), decreased depression
could work in the opposite direction from which Katon (2003) believed.
Finally, social connection theory may explain the null mediation relationship. Depressed
patients are often asked to engage in pleasurable activities, and these activities might cut off time
for SBGM and foot care compared with a situation when patients solitarily stay at home due to
depressive symptoms. Also, Latino patients reported increased temptation to high-fat and
carbohydrate foods when they join family feasts (Early et al., 2009). Thus, the positive effect of
alleviated depression may be cancelled out by increased consumption of restricted menus by
health providers.
Exercise is an important activity that simultaneously affects depression and diabetes-
related outcomes (Detweiler-Bedell et al., 2008). For behavioral activation, many depression
care models involve exercise (e.g., Piette et al., 2011). Also, exercise is proposed as one of major
self-care behaviors for diabetes management because many life modification interventions
promoting the exercise frequency demonstrated weight loss and reduced diabetes related
consequences (American Diabetes Association, 2013; Fabricatore et al., 2011). In addition,
increased exercise frequency may have positive benefits on subsequent depressive symptoms
(Oh & Ell, unpublished). A one-day increase in exercise frequency was associated with 17% and
15% lower risk of clinical depression at assessments conducted 6 and 12 months after the
measurement of depression (Oh & Ell, unpublished).
99
This study has strengths in the design by using prospective analysis, the ability to conduct
sensitivity analysis, well-validated measurements for variables, and the data from depression
care trial, which showed notable decline in depressive symptoms. However, several limitations
should be considered for more accurate interpretation. First, the findings were observed in a
sample recruited from safety-net clinics located in Southern California. Latino patients with
lower cultural acculturation were a predominant ethnic group in the sample. Also, the geographic
areas covered by participating clinics suggest higher structural barriers for successful self-care
behaviors. Many patients were from high-crime areas that possibly hindered patients from
exercising or walking in the evening, resulting in more subdued increases of exercise frequency
than it would be in safer communities at the same level of decrease in depression. The low-
income community has lower access to fresh fruits and vegetables, leaving few options beyond
the lower-priced fast food restaurants in the community. This environmental barrier would
prevent patients from eating a healthy diet, which was measured with four items in the SDSCA.
Also, many patients had low-income jobs that are often characterized with insufficient job
control, and this property may stifle the effect of patient’s changed mot ivation to influence
SBGM at workplace. Therefore, further studies need to examine the effect of various barriers
that might be unequally distributed by income level. Second, we posited that the reduced
depressive symptoms were triggered by having depression care. However, many patients had not
had any depression treatment during the 6-month period, when a notable depressive symptoms
decline was observed (Ell et al., 2010; 2011). This does not deviate from results of other
depression care studies that used clinical depression screening at recruitment (Chou, Chi,
Weisner, Pentz, & Hser, 2010; Katon et al., 2004a). Some patients’ symptoms reduction may
relate to situations where scores to the PHQ-9 move to the average level, resulting in a decrease
100
in the scores on the PHQ-9. Also, trainings for primary care physicians and nurses on depression
treatment seemingly developed spillover effects on patients in the EUC group, resulting in
significant symptoms reduction in the EUC as well. Lastly, the MDDP reported to primary care
physicians of depression diagnosis for all patients who were recruited. This notification helped
physicians and nurses provide brief versions of depression treatment, which may not to be
perceived as a professional depression care by patients. Another concern for removing cases
without report of receiving depression treatment was adequate sample size for mediation analysis.
Mediation analysis requires a larger sample size than multivariate depression treatment (Fritz &
MacKinnon, 2007). In clinical trials, increasing sample size is very costly. Therefore, because of
feasibility, we analyzed all patients who enrolled in the MDDP.
Ten sets of unadjusted and adjusted mediational analytic models demonstrated the
complicated nature for self-care behavior as a mediator. To develop a depression care model,
which is also able to trigger better diabetes management, as well as mental health condition,
deeper understandings on this complexity should be necessary. Environmental barriers hindering
the implementation of self-care behaviors should be considered. Also, Detweiler-Bedell et al.
(2008) provided a decent theoretical framework for studies contemplating the complex
relationship between mood disorder and chronic illness management. We suggest further efforts
to understand the mechanisms by which null results from mediation analyses were produced.
Also, it is important to find demographic and clinical characteristics of patients that influence the
benefit of depression remission on self-care behaviors and diabetes-related outcomes.
101
Chapter 5. CONCLUSION
It is the well-established understanding that depression is a debilitating condition that has
an association with undermined adherence to health behaviors among the general population
(Leventhal, Weinman, Leventhal, & Phillips, 2008), patients with chronic illness (DiMatteo et al.,
2000), and diabetes patients (Gonzalez et al., 2008a). Based on this understanding, it is logical to
hypothesize that diabetes patients would show improved self-care behaviors if their depressive
symptoms were reduced through depression care (Egede & Hernández-Tejada, 2013; Katon,
2003). However, previous clinical trials on the effectiveness of depression treatment do not
appear to bolster the function of depressive symptoms change related to self-care behavior
adherence (Gask et al., 2011; Markowitz et al., 2011). In other words, although the significant
association between depression and self-care behaviors was obviously observed in observational
studies, reduced depressive symptoms may not be sufficient to trigger self-care behavior
increases in depression care (e.g., Katon et al., 2010). According to a meta-analysis, no
significant association between depression remission and glycemic control was found under the
collaborative care model for diabetes patients (Atlantis, Fahey, & Foster, 2014).
Few have examined the role of changed self-care behavior frequency on negative affect.
For this question, two competing arguments seem to exist. On one hand, increased self-care
behavior frequency would reduce depressive symptoms because better diabetes management
decreases the risk of disability and quality of life, which are believed to contribute to
psychological health. It is suggested that depression among diabetes patients is associated with a
change in biomarkers, including insulin, cortisol, and catecholamines (Musselman, Betan, Larsen,
& Phillips, 2003). On the other hand, more energy and time are necessary for additional
investment in diabetes management, and the changes would increase emotional distress related to
102
diabetes management. Diabetes-related distress coming from self-care behaviors was found to be
a significant contributor for depression risk (Fisher et al., 2007; Golden et al., 2008; Golden et al.,
2007). Also, limited cognitive resources for self-regulation would tax emotional stress if diabetes
patients were asked to increase self-care behaviors too much (Hagger, Wood, Stiff, &
Chatzisarantis, 2010; Vohs et al., 2005). Thus, more frequent self-care behavior, which is
facilitated by depression care activities or self-management training, would ask to use increased
self-regulation. If the increase is not well-supported by a patient’s internal and external resources,
patients would experience burn-out and fatigue from the burden of self-care behaviors. However,
this question is left untested. Because many self-management training or depression care
activities successfully increase self-care behaviors (American Diabetes Association, 2013; Katon
et al., 2010), investigating the question seems meaningful.
This dissertation proposed three individual questions that may help document empirical
results that illuminate changes in depression, self-care behavior, and diabetes outcomes in
depression care. Furthermore, these questions lead to an understanding of the mechanisms by
which artificially reduced depression relates to behavioral change. For this goal, data in the
MDDP were used for secondary analyses. Study 1 examined the effect of reduced depressive
symptoms, actual receipt of PST, and the interaction term of these two variables on the change in
frequency of self-care behaviors. Study 2 investigated the effect of change in self-care behavior
frequency on current and future risk of depression. Finally, Study 3 empirically validated the
proposed indirect effect of self-care behavior frequency change in the significant association
between depressive symptoms change and self-reported diabetes symptoms and daily functioning
(Egede & Hernández-Tejada, 2013; Katon, 2003). In findings, Study 1 suggested that depression
alleviation is not sufficient to increase the number of weekly self-care behaviors other than
103
healthy diet frequency. Also, PST does not appear to have an effect on the self-care behavior
frequency change. Study 2’s demonstrated change in self-care behaviors frequency barely
influenced the risk of depression. Exercise was a notable exception and decreased the risk of
depression by 15% to 17% if patients increased by another day of implementing recommended
exercise compared with the start of depression care. Finally, in Study 3, unadjusted and adjusted
models for testing the mediation effect of each self-care behavior did not bolster the
hypothesized relationship other than one exception. Only exercise frequency significantly
mediated the significant total effect of depressive symptoms change on daily functioning,
measured with the SF-12 PCS.
It seems that the effect of declined depressive symptoms on self-care behavior frequency
was tarnished by other environmental factors while patients were receiving depression care. This
is consistent with studies that examined the association between behavioral change intention and
actual behavioral change. In a meta-analysis, the effect of the behavioral intention and actual
change became halves in clinical studies compared with observational studies (Webb & Sheeran,
2006). Without changing contextual factors affecting self-care behavior frequency, increased
energy and motivation co-occurring with depressive symptoms alleviation seem to be
insufficient to create behavioral change. Interestingly, diet was an exception. We speculate that a
more frequent healthy diet was observed because this behavior particularly involves social
interactions with caregivers and family members, and depression is a critical predictor for the
quality of interpersonal relationships (Hammen, 2006). Hispanic patients had difficulties when
their family members were not supportive of their changing diet (Early et al., 2009; Vincent,
Clark, Zimmer, & Sanchez, 2006). Another study found that the extent that a partner understands
the depth of diabetes’ impact significantly explained more fruit and vegetable eating (Searle et
104
al., 2007). Thus, alleviated depression could enhance a patient’s ability to communicate wit h
others who may have an influence on his or her daily diet. In this way, reduced depressive
symptoms might exhibit a positive impact on healthy diet frequency, after all. Yet, exercise,
SBGM, and BMI are often practiced or determined by patients themselves. Although inter-
individual difference in depressive symptoms often demonstrates significant differences in these
behaviors (Gonzalez et al., 2008a; Gonzalez et al., 2008b), within-person changes in self-care
behavior does not appear to be influenced by situational changes in depressive symptoms.
Self-care behaviors might receive counterproductive effects from activities-targeted
depressive symptoms and declined depression symptoms (Detweiler-Bedell et al., 2008).
Facilitating pleasurable activities in cognitive behavioral therapy for depression may reduce time
availability for following exercise recommendations, which requires three times the level of
aerobic activities and two times that of muscular physical activities in a week (American
Diabetes Association, 2013). Also, declined depression sometimes undermines a patient’s ability
to assess symptoms from diabetes that might result in reduced motivation for continuous SBGM
(Ryan et al., 2002). As a result, within-person change in depressive symptoms might have both
positive and negative effects, resulting in null findings observed. Changing self-care behavior is
a very difficult and laborious job in evidence-based self-care management training (Lorig et al.,
2009). Interventions showing efficacy on self-care behaviors require highly intensive activities
requiring significant cost, time, and sacrificing pleasurable behaviors (Leventhal et al., 2008). To
reduce 58% of diabetes risk, 16 individual sessions plus monthly behavioral activation and group
sessions were provided in Diabetes Prevention Program Research Group intervention (Knowler
et al., 2002). It seems that PST, which primarily focused on depression treatment, was
insufficient to trigger observable behavioral change. Thus, the findings advocate additional self-
105
management skills training for life modification in depression care for diabetes patients in order
to expect increased self-care behavior changes.
In Study 2, contrary to the theory of limited resource for self-regulation (Hagger et al.,
2010), we found that change in self-care behavior frequency in depression care for diabetes
patients had a positive or no effect on the risk of depression. Evidence for the deteriorating effect
of more frequent self-care behavior on depression was not found in our empirical analysis. The
findings bolstered the importance of an integrated approach that mixes diabetes-management
skills training and depression treatment. There are an increasing number of studies that
demonstrate notable conceptual overlap between depressive symptoms and distress from diabetes
management (Fisher et al., 2007; Gonzalez et al., 2011; Reddy, Philpot, Ford, & Dunbar, 2010).
By addressing diabetes-related distress, particularly from maintaining self-care behaviors, we
may prevent relapse or prolonged depression. Observing the protective effect of more frequent
exercise, it is important to add behavioral activation of exercise in depression care. This change
seemingly improved effectiveness of depression care in a previous clinical trial (Piette et al.,
2011). Piette et al. (2011) examined the effectiveness of psychotherapy plus walking for diabetes
patients with clinical depression, demonstrating positive outcomes in depression, biomarkers
related to diabetes outcomes, and self-reported quality-of-life indicators. In focused group
interviews, Latino patients called for the creation of a support that can teach or show examples
for better self-care behaviors to reduce distress from diabetes management (Vincent et al., 2006).
Thus, for depressed Latino patients, it seems important to collaboratively introduce activities for
depression and self-management improvement. For instance, cognitive-behavioral therapy
addressing irrational beliefs and norms is supplemented with skills that remove barriers for self-
106
care behavior implementation and with training for development of feasible action plan for more
frequent self-care behaviors.
Although the significant mediating role of self-care behavior was barely observed in
Study 3, the importance of comorbid depression on diabetes management and quality of life
should not be overlooked because numerous evidence from both observational and clinical
studies have accumulated. Rather, the findings are demonstrating the complicated nature of the
relationship between depression care and diabetes management. It appears that depression care
has a puzzling relationship with self-care behaviors that was hypothesized in Detweiler-Bedell et
al. (2008). In this theoretical argument, depression care might unwittingly undermine disease
management. Also, declined depression achieved by depression care may help patients “mis-
regulate” their disease. In addition to Lustman et al.’s (1998) finding, which showed worse
glycemic control in a group offered cognitive behavioral therapy for depression, the null results
in mediation analyses bolster the possible existence of unintended outcomes of depression care.
It is possible that some patients who received PST might become less attentive to self-care
behaviors that they had previously followed daily. PST might hinder some patients in
implementing self-care behaviors that were scheduled because PST assignments and activities
within the psychotherapy used time and energy that had been used for self-care behaviors. Also,
it is possible that declined depressive symptoms may reduce time for self-care behaviors because
patients with depression remission would have increased energy for more frequent social
meetings and outdoor activities. By leaving home, patients could be more distracted by
interactions with others, followed by less attention to diabetes symptoms and lower motivation to
conduct self-care behaviors. Therefore, the null findings recommend that depression care
107
providers and developers understand conflicting relationship with diabetes management and look
for measures that help achieve the dual objectives together.
This dissertation left intriguing questions after extensive investigation on the complex
association between depression and self-care behaviors under depression care. First, conflicting
effects led by depression care on self-care behaviors need more scholarly attention.
Understanding the ambivalent influences of depression care is in its infancy. Thus, it is important
to conduct qualitative studies that explore the experience of patients enrolled in depression care
related to self-care behaviors. Interesting questions can be asked, including whether
psychotherapy may attenuate interest and attention that were originally directed at self-care
behaviors, whether declined depression helps patients have more social activities, possibly
resulting in limited time and resources for diabetes management, and whether patients have
increased discrepancy between actual self-care behavior frequency and intention to implement
these behaviors after recovering from dire depression. Findings from qualitative studies may
support quantitative studies with rigorous conceptualization and adequate measures for inference
tests on the effect of depression care. Second, it would be intriguing if studies examined
environmental factors that might enable declined depression to increase self-care behavior
frequency. Based on findings and previous literature, depressive symptoms decline appears to be
a necessary condition for more frequent self-care behaviors, not a sufficient condition. In other
words, depression decline can activate healthy behaviors only when other contextual conditions
are met for the facilitation. For instance, availability of nearby park and safe community are
important environmental factors that allow motivated patients to engage in regular exercise.
Because predominate patients in the studies resided in low-income urban communities that
lacked recreational facilities and suffered from a high crime rate, patient’s improved motivation
108
may not be linked to actual behavioral changes. Also, the extent of weight control triggered by
depression remission might be influenced by community-level healthy food accessibility and
cultural norms on food sizes. Thus, further studies are recommended to investigate possible
contextual factors that help depression remission have a link to behavioral changes.
The findings have significant implications to social work education during a time when
unprecedented changes in primary care clinics are occurring in the United States because of the
Affordable Care Act (ACA). As the ACA is implemented, we expect social workers to find
increasing employment opportunities within primary care clinics that treat chronic illness
patients, including diabetes, because clinics would need new staff that provide behavioral health
services. In addition to high-profile individual mandates, the ACA is driving several experiments
testing different types of health care organizations, which is expected to increase care
coordination for better diabetes self-management and psychological health (Darnell, 2013). The
Accountable Care Organization (ACO) was proposed to create incentives for health
organizations to develop care procedures that can strengthen patient’s engagement and
participation for chronic illness management. Also, the Patient-Centered Medical Home (PCMH)
promotes collaborative approaches across providers with different backgrounds to address unmet
needs observed under traditional health care delivery with fee-for-service financing. For diabetes
management, the ACO and the PCMH are asked to devise formulated health care delivery
systems that can motivate patients to engage in healthy behaviors and that promptly address
barriers compromising the quality of self-care behaviors. Comorbid depression is selected as a
major blockade that undermines the adherence to recommended diabetes management. Social
workers may assume a position treating depression and helping patients follow the recommended
self-care behaviors. Even though the change is promising for social workers, some conditions
109
need to be met for successful positioning within primary care clinics, given the ACA’s policy
changes. First, social workers are recommended to understand the different nature of depressive
symptoms found among diabetes patients compared with physically healthy patients with clinical
depression. Depressive symptoms among diabetes patients closely interact with emotional
distress derived from diabetes management (Fisher et al., 2007). It was found that undiagnosed
diabetes patients have an even lower risk of depression, suggesting that distress from uncertainty
related to disease progression and disease management impose enormous stress and anxiety
(Golden et al., 2008; Knol et al., 2007). Thus, social workers should understand the nature of
diabetes and mechanisms by which depression and diabetes-related distress interact and lead to a
negative spiral of physical and mental health simultaneously. For this goal, social work
education requires courses that teach about chronic illnesses’ biological, psychological, and
social impacts on patient quality of life. For example, it is important to understand processes by
which patient’s cognitive perception o f chronic disease and related symptoms facilitate help-
seeking behaviors from health professionals, which was theorized in the common sense model
(Leventhal, Forster, & Leventhal, 2007). Also, social workers need to understand shared
biological pathways connecting the physical symptoms and depression via circadian rhythm and
overlapped neurotransmitters and hormones. It is also important to expose students to cutting-
edge evidence-based interventions that were developed through interdisciplinary approaches.
Second, students are asked to work in a multi-disciplinary environment because the ACA
requires more frequent communication among health providers within the primary care setting.
This collaborative work setting may expect social workers to have a skill set that facilitates
effective problem solving with traditional medical providers. Therefore, students are asked to
understand medical terminology and medical service procedures. For this goal, a primary care
110
clinic is a good placement for the field practicums for master-level students. From the field
practicum, students can observe medical service delivery, work culture, and attitudes of
physicians and nurses in primary care clinics. Yet, many primary care clinics don’ t have a social
worker because this big transition is in its infancy. Thus, these clinics could not serve as an
organization accepting a social work student for the field practicum. This barrier needs to be
addressed by social work education leadership. One option may be allowing an exception for a
primary clinic that has an experienced clinical psychologist, and the clinical psychologist may
assume a field instructor. In addition to the field practicum, dual-degree programs provide
students with a good opportunity to expand their employment in a primary care clinic. In this
way, social workers would maximize their ability to provide care for needs either physical or
psychosocial or both.
111
REFERENCES
Adewuya, A. O., Ola, B. A., & Afolabi, O. O. (2006). Validity of the patient health questionnaire
(PHQ-9) as a screening tool for depression amongst Nigerian university students. Journal of
Affective Disorders, 96(1), 89-93.
Aikens, J. E. (2012). Prospective associations between emotional distress and poor outcomes in
type 2 diabetes. Diabetes Care, 35(12), 2472-2478. doi:10.2337/dc12-0181; 10.2337/dc12-
0181
Ali, S., Stone, M., Peters, J., Davies, M., & Khunti, K. (2006). The prevalence of co ‐morbid
depression in adults with type 2 diabetes: A systematic review and meta ‐analysis. Diabetic
Medicine, 23(11), 1165-1173.
American College of Sports Medicine. (2000). Exercise and type 2 diabetes. Medicine & Science
in Sports & Exercise, 32(7), 1345-1360.
American Diabetes Association. (2013). Standards of medical care in Diabetes—2013. Diabetes
Care, 36(Supplement 1), S11-S66. doi:10.2337/dc13-S011
Antero Kesaniemi, Y., Danforth, E. J., Jensen, M. D., Kopelman, P. G., Lefebvre, P., & Reeder,
B. A. (2001). Dose-response issues concerning physical activity and health: An evidence-
based symposium. Medicine & Science in Sports & Exercise, 33(6) (Supplement), S351-
S358.
Appelhans, B. M., Whited, M. C., Schneider, K. L., Ma, Y., Oleski, J. L., Merriam, P. A., …
Pagoto, S. L. (2012). Depression severity, diet quality, and physical activity in women with
obesity and depression. Journal of the Academy of Nutrition and Dietetics, 112(5), 693-698.
112
Areán, P. A., Hegel, M. T., & Reynolds III, C. F. (2001). Treating depression in older medical
patients with psychotherapy. Journal of Clinical Geropsychology, 7(2), 93-104.
Asimakopoulou, K. G., & Hampson, S. E. (2005). Biases in self-reports of self-care behaviours
in type 2 diabetes. Psychology, Health & Medicine, 10(3), 305-315.
Atlantis, E., Fahey, P., & Foster, J. (2014). Collaborative care for comorbid depression and
diabetes: A systematic review and meta-analysis. BMJ Open, 4(4) doi:10.1136/bmjopen-
2013-004706
Bao, Y., Alexopoulos, G. S., Casalino, L. P., Ten Have, T. R., Donohue, J. M., Post, E. P., et al.
(2011). Collaborative depression care management and disparities in depression treatment
and outcomes. Archives of General Psychiatry, 68(6), 627.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51(6), 1173-1182.
Bauer, I. M., & Baumeister, R. F. (2011). Self-regulatory strength. In K. D. Vohs, & R. F.
Baumeister (Eds.), Handbook of self-regulation: Research, theory, and applications (pp. 64-
82). New York: Guilford.
Baumeister, H., Hutter, N., & Bengel, J. (2012). Psychological and pharmacological
interventions for depression in patients with diabetes mellitus and depression. Cochrane
Database of Systemic Reviews, 12, CD008381.
113
Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998). Ego depletion: Is the
active self a limited resource? Journal of Personality and Social Psychology, 74(5), 1252.
Beatty, L., & Lambert, S. (2013). A systematic review of internet-based self-help therapeutic
interventions to improve distress and disease-control among adults with chronic health
conditions. Clinical Psychology Review, 33(4), 609-622.
Black, S. A. (1999). Increased health burden associated with comorbid depression in older
diabetic Mexican Americans: Results from the Hispanic Established Population for the
Epidemiologic Study of the Elderly survey. Diabetes Care, 22(1), 56-64.
Black, S. A., Markides, K. S., & Ray, L. A. (2003). Depression predicts increased incidence of
adverse health outcomes in older Mexican Americans with type 2 diabetes. Diabetes Care,
26(10), 2822-2828.
Bogner, H. R., Morales, K. H., de Vries, H. F., & Cappola, A. R. (2012). Integrated management
of type 2 diabetes mellitus and depression treatment to improve medication adherence: A
randomized controlled trial. Annals of Family Medicine, 10(1), 15-22.
Bolger, N., Foster, M., Vinokur, A. D., & Ng, R. (1996). Close relationships and adjustment to a
life crisis: The case of breast cancer. Journal of Personality and Social Psychology, 70, 283-
294.
Bowser, D. M., Utz, S., Glick, D., & Harmon, R. (2010). A systematic review of the relationship
of diabetes mellitus, depression, and missed appointments in a low-income uninsured
population. Archives of Psychiatric Nursing, 24(5), 317-329.
114
Braveman, P. A., Cubbin, C., Egerter, S., Williams, D. R., & Pamuk, E. (2010). Socioeconomic
disparities in health in the United States: What the patterns tell us. American Journal of
Public Health, 100(S1), S186-S196.
Cabassa, L. J., Hansen, M. C., Palinkas, L. A., & Ell, K. (2008). Azucar y nervios: Explanatory
models and treatment experiences of Hispanics with diabetes and depression. Social Science
& Medicine, 66(12), 2413-2424.
Castaneda, C., Layne, J. E., Munoz-Orians, L., Gordon, P. L., Walsmith, J., Foldvari, M., …
Nelson, M. E. (2002). A randomized controlled trial of resistance exercise training to
improve glycemic control in older adults with type 2 diabetes. Diabetes Care, 25(12), 2335-
2341.
Center for Disease Control and Prevention. (2012). 2011 national diabetes fact sheet. Retrieved
October 1, 2013, from http://www.cdc.gov/diabetes/pubs/pdf/ndfs_2011.pdf
Chen, S., Chiu, H., Xu, B., Ma, Y., Jin, T., Wu, M., & Conwell, Y. (2010). Reliability and
validity of the PHQ ‐9 for screening late ‐life depression in Chinese primary care.
International Journal of Geriatric Psychiatry, 25(11), 1127-1133.
Chiu, C., Wray, L. A., Beverly, E. A., & Dominic, O. G. (2010). The role of health behaviors in
mediating the relationship between depressive symptoms and glycemic control in type 2
diabetes: A structural equation modeling approach. Social Psychiatry and Psychiatric
Epidemiology, 45(1), 67-76.
Chou, C., Chi, F., Weisner, C., Pentz, M., & Hser, Y. (2010). Initial status in growth curve
modeling for randomized trials. Journal of Drug Issues, 40(1), 155-172.
115
Ciechanowski, P. S., Katon, W. J., & Russo, J. E. (2000). Depression and diabetes: Impact of
depressive symptoms on adherence, function, and costs. Archives of Internal Medicine,
160(21), 3278-3285.
Ciechanowski, P. S., Katon, W. J., Russo, J. E., & Hirsch, I. B. (2003). The relationship of
depressive symptoms to symptom reporting, self-care and glucose control in diabetes.
General Hospital Psychiatry, 25(4), 246-252.
Coleman, S. M., Katon, W., Lin, E., & Von Korff, M. (2013). Depression and death in diabetes;
10-year follow-up of all-cause and cause-specific mortality in a diabetic cohort.
Psychosomatics, 54(5), 428-436. doi:10.1016/j.psym.2013.02.015
Cornil, Y., & Chandon, P. (2013). From fan to fat? Vicarious losing increases unhealthy eating,
but self-affirmation is an effective remedy. Psychological Science, 24(10), 1936-1946.
doi:10.1177/0956797613481232
Cusi, K., & Ocampo, G. L. (2011). Unmet needs in Hispanic/Latino patients with type 2 diabetes
mellitus. American Journal of Medicine, 124(10), S2-S9.
Darnell, J. S. (2013). Navigators and assisters: Two case management roles for social workers in
the affordable care act. Health & Social Work, 38(2), 123-126.
Delahanty, L., Grant, R., Wittenberg, E., Bosch, J., Wexler, D., Cagliero, E., & Meigs, J. B.
(2007). Association of diabetes ‐related emotional distress with diabetes treatment in primary
care patients with type 2 diabetes. Diabetic Medicine, 24(1), 48-54.
116
Derogatis, L. R., Lipman, R. S., Rickels, K., Uhlenhuth, E. H., & Covi, L. (1974). The Hopkins
Symptom Checklist (HSCL): A self ‐report symptom inventory. Behavioral Science, 19(1),
1-15.
Detweiler-Bedell, J. B., Friedman, M. A., Leventhal, H., Miller, I. W., & Leventhal, E. A.
(2008). Integrating co-morbid depression and chronic physical disease management:
Identifying and resolving failures in self-regulation. Clinical Psychology Review, 28(8),
1426-1446.
DiMatteo, M. R., Lepper, H. S., & Croghan, T. W. (2000). Depression is a risk factor for
noncompliance with medical treatment: A meta-analysis of the effects of anxiety and
depression on patient adherence. Archives of Internal Medicine, 160(14), 2101-2107.
Dunn, A. L., Trivedi, M. H., & O'Neal, H. A. (2001). Physical activity dose-response effects on
outcomes of depression and anxiety. Medicine & Science in Sports & Exercise, 33(6)
(Supplement), S587-S597.
Early, K. B., Shultz, J. A., & Corbett, C. (2009). Assessing diabetes dietary goals and self-
management based on in-depth interviews with Latino and Caucasian clients with type 2
diabetes. Journal of Transcultural Nursing, 20(4), 371-381.
doi:10.1177/1043659609334928
Egede, L. E. (2005). Effect of depression on self-management behaviors and health outcomes in
adults with type 2 diabetes. Current Diabetes Reviews, 1(3), 235-243.
117
Egede, L. E., Ellis, C., & Grubaugh, A. L. (2009). The effect of depression on self-care
behaviors and quality of care in a national sample of adults with diabetes. General Hospital
Psychiatry, 31(5), 422-427.
Egede, L. E., Grubaugh, A. L., & Ellis, C. (2010). The effect of major depression on preventive
care and quality of life among adults with diabetes. General Hospital Psychiatry, 32(6),
563-569.
Egede, L. E., & Hernández-Tejada, M. A. (2013). Effect of comorbid depression on quality of
life in adults with type 2 diabetes. Expert Review of Pharmacoeconomics & Outcomes
Research, 13(1), 83-91.
Egede, L. E., Nietert, P. J., & Zheng, D. (2005). Depression and all-cause and coronary heart
disease mortality among adults with and without diabetes. Diabetes Care, 28(6), 1339-1345.
Ell, K., Xie, B., Quon, B., Quinn, D. I., Dwight-Johnson, M., & Lee, P. (2008). Randomized
controlled trial of collaborative care management of depression among low-income patients
with cancer. Journal of Clinical Oncology, 26(27), 4488-4496.
doi:10.1200/JCO.2008.16.6371
Ell, K., Katon, W., Cabassa, L. J., Xie, B., Lee, P., Kapetanovic, S., & Guterman, J. (2009).
Depression and diabetes among low-income Hispanics: Design elements of a socio-
culturally adapted collaborative care model randomized controlled trial. International
Journal of Psychiatry in Medicine, 39(2), 113-132.
Ell, K., Katon, W., Xie, B., Lee, P., Kapetanovic, S., Guterman, J., & Chou, C. P. (2010).
Collaborative care management of major depression among low-income, predominantly
118
Hispanic subjects with diabetes: A randomized controlled trial. Diabetes Care, 33(4), 706-
713.
Ell, K., Katon, W., Xie, B., Lee, P., Kapetanovic, S., Guterman, J., & Chou, C. P. (2011). One-
year postcollaborative depression care trial outcomes among predominantly Hispanic
diabetes safety net patients. General Hospital Psychiatry, 33(5), 436-442.
Fabricatore, A. N., Wadden, T. A., Higginbotham, A. J., Faulconbridge, L. F., Nguyen, A. M.,
Heymsfield, S. B., & Faith, M. S. (2011). Intentional weight loss and changes in symptoms
of depression: A systematic review and meta-analysis. International Journal of Obesity,
35(11), 1363-1376.
Fisher, L., Skaff, M., Mullan, J., Arean, P., Glasgow, R., & Masharani, U. (2008). A longitudinal
study of affective and anxiety disorders, depressive affect and diabetes distress in adults
with type 2 diabetes. Diabetic Medicine, 25(9), 1096-1101.
Fisher, L., Mullan, J. T., Arean, P., Glasgow, R. E., Hessler, D., & Masharani, U. (2010).
Diabetes distress but not clinical depression or depressive symptoms is associated with
glycemic control in both cross-sectional and longitudinal analyses. Diabetes Care, 33(1),
23-28.
Fisher, L., Skaff, M. M., Mullan, J. T., Arean, P., Mohr, D., Masharani, U., … Laurencin, G.
(2007). Clinical depression versus distress among patients with type 2 diabetes: Not just a
question of semantics. Diabetes Care, 30(3), 542-548.
119
Fitzpatrick, S. L., Schumann, K. P., & Hill-Briggs, F. (2013). Problem solving interventions for
diabetes self-management and control: A systematic review of the literature. Diabetes
Research and Clinical Practice, 100(2), 145-161.
Fritz, M. S., & MacKinnon, D. P. (2007). Required sample size to detect the mediated effect.
Psychological Science, 18(3), 233-239.
Gask, L., Macdonald, W., & Bower, P. (2011). What is the relationship between diabetes and
depression? A qualitative meta-synthesis of patient experience of co-morbidity. Chronic
Illness, 7(3), 239-252. doi:10.1177/1742395311403636
Glasgow, R. E., Strycker, L. A., Toobert, D. J., & Eakin, E. (2000). A social–ecologic approach
to assessing support for disease self-management: The chronic illness resources survey.
Journal of Behavioral Medicine, 23(6), 559-583.
Glasgow, R. E., Wagner, E. H., Schaefer, J., Mahoney, L. D., Reid, R. J., & Greene, S. M.
(2005). Development and validation of the patient assessment of chronic illness care
(PACIC). Medical Care, 43(5), 436-444.
Glazier, R. H., Bajcar, J., Kennie, N. R., & Willson, K. (2006). A systematic review of
interventions to improve diabetes care in socially disadvantaged populations. Diabetes Care,
29(7), 1675-1688.
Golden, S. H., Lazo, M., Carnethon, M., Bertoni, A. G., Schreiner, P. J., Diez Roux, A. V., …
Lyketsos, C. (2008). Examining a bidirectional association between depressive symptoms
and diabetes. Journal of the American Medical Association, 299(23), 2751-2759.
120
Golden, S. H., Lee, H. B., Schreiner, P. J., Diez Roux, A., Fitzpatrick, A. L., Szklo, M., &
Lyketsos, C. (2007). Depression and type 2 diabetes mellitus: The multiethnic study of
atherosclerosis. Psychosomatic Medicine, 69(6), 529-536.
doi:10.1097/PSY.0b013e3180f61c5c
Goldhaber-Fiebert, J. D., Goldhaber-Fiebert, S. N., Tristan, M. L., & Nathan, D. M. (2003).
Randomized controlled community-based nutrition and exercise intervention improves
glycemia and cardiovascular risk factors in type 2 diabetic patients in rural Costa Rica.
Diabetes Care, 26(1), 24-29.
Goldney, R. D., Phillips, P. J., Fisher, L. J., & Wilson, D. H. (2004). Diabetes, depression, and
quality of life: A population study. Diabetes Care, 27(5), 1066-1070.
Gonzalez, J. S., Peyrot, M., McCarl, L. A., Collins, E. M., Serpa, L., Mimiaga, M. J., & Safren,
S. A. (2008a). Depression and diabetes treatment nonadherence: A meta-analysis. Diabetes
Care, 31(12), 2398-2403.
Gonzalez, J. S., Safren, S. A., Cagliero, E., Wexler, D. J., Delahanty, L., Wittenberg, E., …
Grant, R. W. (2007). Depression, self-care, and medication adherence in type 2 diabetes:
Relationships across the full range of symptom severity. Diabetes Care, 30(9), 2222-2227.
Gonzalez, J., Safren, S., Delahanty, L., Cagliero, E., Wexler, D., Meigs, J. B., Grant, R. W.
(2008b). Symptoms of depression prospectively predict poorer self ‐care in patients with type
2 diabetes. Diabetic Medicine, 25(9), 1102-1107.
Gonzalez, J. S., Fisher, L., & Polonsky, W. H. (2011). Depression in diabetes: Have we been
missing something important? Diabetes Care, 34(1), 236-239. doi:10.2337/dc10-1970
121
Gonzalez, J. S., Schneider, H. E., Wexler, D. J., Psaros, C., Delahanty, L. M., Cagliero, E.,
Safren, S. A. (2013). Validity of medication adherence self-reports in adults with type 2
diabetes. Diabetes Care, 36(4), 831-837.
Gregg, J. A., Callaghan, G. M., Hayes, S. C., & Glenn-Lawson, J. L. (2007). Improving diabetes
self-management through acceptance, mindfulness, and values: A randomized controlled
trial. Journal of Consulting and Clinical Psychology, 75(2), 336.
Hagger, M. S., Wood, C., Stiff, C., & Chatzisarantis, N. L. (2010). Ego depletion and the
strength model of self-control: A meta-analysis. Psychological Bulletin, 136(4), 495.
Hamer, M., & Stamatakis, E. (2014). Prospective study of sedentary behavior, risk of depression,
and cognitive impairment. Medicine & Science in Sports & Exercise, 46(4), 718-723.
Hammen, C. (2006). Stress generation in depression: Reflections on origins, research, and future
directions. Journal of Clinical Psychology, 62(9), 1065-1082. doi:10.1002/jclp.20293
Harkness, E., Macdonald, W., Valderas, J., Coventry, P., Gask, L., & Bower, P. (2010).
Identifying psychosocial interventions that improve both physical and mental health in
patients with diabetes: A systematic review and meta-analysis. Diabetes Care, 33(4), 926-
930.
Harris, M. I., Klein, R., Cowie, C. C., Rowland, M., & Byrd-Holt, D. D. (1998). Is the risk of
diabetic retinopathy greater in non-Hispanic blacks and Mexican Americans than in non-
Hispanic whites with type 2 diabetes? A US population study. Diabetes Care, 21(8), 1230-
1235.
122
Hay, J. W., Katon, W. J., Ell, K., Lee, P., & Guterman, J. J. (2012). Cost-effectiveness analysis
of collaborative care management of major depression among low-income, predominantly
Hispanics with diabetes. Value in Health, 15(2), 249-254.
Heckbert, S. R., Oliver, M., Williams, L. H., Ciechanowski, P., M., Lin, E. H., & Katon, W. J.
(2010). Depression in relation to long-term control of glycemia, blood pressure, and lipids in
patients with diabetes. Journal of General Internal Medicine, 25(6), 524-529.
Hegel, M. T., Barrett, J. E., & Oxman, T. E. (2000). Training therapists in problem-solving
treatment of depressive disorders in primary care: Lessons learned from the" treatment
effectiveness project". Families, Systems, & Health, 18(4), 423.
Hegel, M. T., Imming, J., Cyr-Provost, M., Noel, P. H., Arean, P. A., & Unutzer, J. (2002). Role
of behavioral health professionals in a collaborative stepped care treatment model for
depression in primary care: Project IMPACT. Families, Systems & Health, 20(3), 265-277.
Heisler, M., Smith, D. M., Hayward, R. A., Krein, S. L., & Kerr, E. A. (2003). How well do
patients' assessments of their diabetes self-management correlate with actual glycemic
control and receipt of recommended diabetes services? Diabetes Care, 26(3), 738-743.
Hibbard, J. H., Greene, J., & Tusler, M. (2009). Improving the outcomes of disease management
by tailoring care to the patient's level of activation. American Journal of Managed Care,
15(6), 353-360.
Hickey, D., Carr, A., Dooley, B., Guerin, S., Butler, E., & Fitzpatrick, L. (2005). Family and
marital profiles of couples in which one partner has depression or anxiety. Journal of
Marital and Family Therapy, 31(2), 171-182. doi:10.1111/j.1752-0606.2005.tb01554.x
123
Hill-Briggs, F. (2003). Problem solving in diabetes self-management: A model of chronic illness
self-management behavior. Annals of Behavioral Medicine, 25(3), 182-193.
Hill-Briggs, F., & Gemmell, L. (2007). Problem solving in diabetes self-management and
control: A systematic review of the literature. Diabetes Educator, 33(6), 1032-1050;
discussion 1051-1052. doi:10.1177/0145721707308412
Huang, F. Y., Chung, H., Kroenke, K., Delucchi, K. L., & Spitzer, R. L. (2006). Using the
patient health Questionnaire ‐9 to measure depression among racially and ethnically diverse
primary care patients. Journal of General Internal Medicine, 21(6), 547-552.
Huang, X., Song, L., Li, T., Li, J., Wu, S., & Li, N. (2002). Effect of health education and
psychosocial intervention on depression in patients with type 2 diabetes. China Mental
Health Journal, 16, 149-151.
Huang, H., Russo, J., Von Korff, M., Ciechanowski, P., Lin, E., Ludman, E., & Katon, W.
(2012). The effect of changes in depressive symptoms on disability status in patients with
diabetes. Psychosomatics, 53(1), 21-29.
Katon, W. J. (2003). Clinical and health services relationships between major depression,
depressive symptoms, and general medical illness. Biological Psychiatry, 54(3), 216-226.
Katon, W. J., Lin, E. H., Von Korff, M., Ciechanowski, P., Ludman, E. J., Young, B., …
McCulloch, D. (2010). Collaborative care for patients with depression and chronic illnesses.
New England Journal of Medicine, 363(27), 2611-2620.
124
Katon, W. J., Von Korff, M., Lin, E. H., Simon, G., Ludman, E., Russo, J., … Bush, T . (2004a).
The pathways study: A randomized trial of collaborative care in patients with diabetes and
depression. Archives of General Psychiatry, 61(10), 1042.
Katon, W., Lin, E. H., & Kroenke, K. (2007). The association of depression and anxiety with
medical symptom burden in patients with chronic medical illness. General Hospital
Psychiatry, 29(2), 147-155.
Katon, W., Russo, J., Lin, E. H., Heckbert, S. R., Ciechanowski, P., Ludman, E. J., & Von Korff,
M. (2009a). Depression and diabetes: Factors associated with major depression at five-year
follow-up. Psychosomatics, 50(6), 570-579.
Katon, W., Russo, J., Lin, E. H., Heckbert, S. R., Karter, A. J., Williams, L. H., … Von Korff,
M. (2009b). Diabetes and poor disease control: Is comorbid depression associated with poor
medication adherence or lack of treatment intensification? Psychosomatic Medicine, 71(9),
965-972.
Katon, W., Unützer, J., Wells, K., & Jones, L. (2010). Collaborative depression care: History,
evolution and ways to enhance dissemination and sustainability. General Hospital
Psychiatry, 32(5), 456-464.
Katon, W., von Korff, M., Ciechanowski, P., Russo, J., Lin, E., Simon, G., … Young, B .
(2004b). Behavioral and clinical factors associated with depression among individuals with
diabetes. Diabetes Care, 27(4), 914-920.
Katon, W. J. (2008). The comorbidity of diabetes mellitus and depression. American Journal of
Medicine, 121(11, Supplement 2), S8-S15. doi:10.1016/j.amjmed.2008.09.008
125
King, D. K., Glasgow, R. E., Toobert, D. J., Strycker, L. A., Estabrooks, P. A., Osuna, D., &
Faber, A. J. (2010). Self-efficacy, problem solving, and social-environmental support are
associated with diabetes self-management behaviors. Diabetes Care, 33(4), 751-753.
doi:10.2337/dc09-1746; 10.2337/dc09-1746
Kirk, J. K., Passmore, L. V., Bell, R. A., Narayan, K. V., D'Agostino, R. B., Arcury, T. A., &
Quandt, S. A. (2008). Disparities in A1C levels between Hispanic and non-Hispanic white
adults with diabetes: A meta-analysis. Diabetes Care, 31(2), 240-246.
Knol, M. J., Heerdink, E. R., Egberts, A. C., Geerlings, M. I., Gorter, K. J., Numans, M. E., …
Burger, H. (2007). Depressive symptoms in subjects with diagnosed and undiagnosed type 2
diabetes. Psychosomatic Medicine, 69(4), 300-305.
Knowler, W. C., Barrett-Connor, E., Fowler, S. E., Hamman, R. F., Lachin, J. M., Walker, E. A.,
… Diabetes Prevention Program Research Group . (2002). Reduction in the incidence of type
2 diabetes with lifestyle intervention or metformin. New England Journal of Medicine,
346(6), 393-403. doi:10.1056/NEJMoa012512
Kroenke, K., Spitzer, R. L., & Williams, J. B. (2001). The PHQ ‐9. Journal of General Internal
Medicine, 16(9), 606-613.
Lamers, F., Jonkers, C. C., Bosma, H., Knottnerus, J. A., & van Eijk, J. T. (2011). Treating
depression in diabetes patients: Does a nurse ‐administered minimal psychological
intervention affect diabetes ‐specific quality of life and glycaemic control? A randomized
controlled trial. Journal of Advanced Nursing, 67(4), 788-799.
126
Leventhal, H., Brissette, I., & Leventhal, E. A. (2003). The common-sense model of self-
regulation of health and illness. In K. J. Petrie & J. A. Weinman (Eds.), The self-regulation
of health and illness behaviour (pp. 42-65). London: Routledge.
Leventhal, H., Weinman, J., Leventhal, E. A., & Phillips, L. A. (2008). Health psychology: The
search for pathways between behavior and health. Annual Review of Psychology, 59, 477-
505.
Leventhal, H., Forster, R., & Leventhal, E. (2007). Self-regulation of health threats, affect, and
the self. In C. M. Aldwin, C. L. Park, & A. Spiro (Eds.), Handbook of health psychology and
aging (pp. 341-366). New York: Guilford.
Lin, E. H., Katon, W., Rutter, C., Simon, G. E., Ludman, E. J., Von Korff, M., … Walker, E .
(2006). Effects of enhanced depression treatment on diabetes self-care. Annals of Family
Medicine, 4(1), 46-53.
Lin, E. H., Katon, W., Von Korff, M., Rutter, C., Simon, G. E., Oliver, M., … Young, B . (2004).
Relationship of depression and diabetes self-care, medication adherence, and preventive
care. Diabetes Care, 27(9), 2154-2160.
Lloyd, C., Pambianco, G., & Orchard, T. (2010). Does diabetes ‐related distress explain the
presence of depressive symptoms and/or poor self ‐care in individuals with type 1 diabetes?
Diabetic Medicine, 27(2), 234-237.
Look AHEAD Research Group. (2013). Cardiovascular effects of intensive lifestyle intervention
in type 2 diabetes. New England Journal of Medicine, 369(2), 145-154.
127
Lorig, K. (1996). Outcome measures for health education and other health care interventions.
Thousand Oaks, CA: Sage.
Lorig, K., Ritter, P. L., Villa, F., & Piette, J. D. (2008). Spanish diabetes self-management with
and without automated telephone reinforcement: Two randomized trials. Diabetes Care,
31(3), 408-414. doi:10.2337/dc07-1313
Lorig, K., Ritter, P. L., Villa, F. J., & Armas, J. (2009). Community-based peer-led diabetes self-
management: A randomized trial. Diabetes Educator, 35(4), 641-651.
doi:10.1177/0145721709335006; 10.1177/0145721709335006
Löwe, B., Kroenke, K., Herzog, W., & Gräfe, K. (2004). Measuring depression outcome with a
brief self-report instrument: Sensitivity to change of the patient health questionnaire (PHQ-
9). Journal of Affective Disorders, 81(1), 61-66.
Ludman, E. J., Katon, W., Russo, J., Von Korff, M., Simon, G., Ciechanowski, P., … Young, B .
(2004). Depression and diabetes symptom burden. General Hospital Psychiatry, 26(6), 430-
436.
Lustman, P. J., Griffith, L. S., Freedland, K. E., Kissel, S. S., & Clouse, R. E. (1998). Cognitive
behavior therapy for depression in type 2 diabetes mellitus: A randomized, controlled trial.
Annals of Internal Medicine, 129(8), 613-621.
Markowitz, S. M., Gonzalez, J. S., Wilkinson, J. L., & Safren, S. A. (2011). A review of treating
depression in diabetes: Emerging findings. Psychosomatics, 52(1), 1-18.
128
Margaret, K. Y., Weiss, N. S., Ding, X., Katon, W. J., Zhou, X., & Young, B. A. (2014).
Associations between depressive symptoms and incident ESRD in a diabetic cohort. Clinical
Journal of the American Society of Nephrology, 9(5), 920-928. doi:10.2215/CJN.08670813
Martin, A., Rief, W., Klaiberg, A., & Braehler, E. (2006). Validity of the brief patient health
questionnaire mood scale (PHQ-9) in the general population. General Hospital Psychiatry,
28, 71-77.
McColl, E., Steen, I., Meadows, K., Hutchinson, A., Eccles, M., Hewison, J., … Blades, S. M .
(1995). Developing outcome measures for ambulatory care: An application to asthma and
diabetes. Social Science & Medicine, 41(10), 1339-1348.
McGrady, M. E., Laffel, L., Drotar, D., Repaske, D., & Hood, K. K. (2009). Depressive
symptoms and glycemic control in adolescents with type 1 diabetes mediational role of
blood glucose monitoring. Diabetes Care, 32(5), 804-806.
McKellar, J. D., Humphreys, K., & Piette, J. D. (2004). Depression increases diabetes symptoms
by complicating patients' self-care adherence. Diabetes Educator, 30, 485-492.
Minet, L., Møller, S., Vach, W., Wagner, L., & Henriksen, J. E. (2010). Mediating the effect of
self-care management intervention in type 2 diabetes: A meta-analysis of 47 randomised
controlled trials. Patient Education and Counseling, 80(1), 29-41.
Miranda, J., Chung, J. Y., Green, B. L., Krupnick, J., Siddique, J., Revicki, D. A., … Belin, T .
(2003). Treating depression in predominantly low-income young minority women: A
randomized controlled trial. Journal of the American Medical Association, 290(1), 57-65.
129
Morgan, M. A., Coates, M. J., Dunbar, J. A., Reddy, P., Schlicht, K., & Fuller, J. (2013). The
TrueBlue model of collaborative care using practice nurses as case managers for depression
alongside diabetes or heart disease: A randomised trial. BMJ Open, 3(1), pii: e002171.
doi:10.1136/bmjopen-2012-002171.
Muraven, M., & Baumeister, R. F. (2000). Self-regulation and depletion of limited resources:
Does self-control resemble a muscle? Psychological Bulletin, 126(2), 247.
Musselman, D. L., Betan, E., Larsen, H., & Phillips, L. S. (2003). Relationship of depression to
diabetes types 1 and 2: Epidemiology, biology, and treatment. Biological Psychiatry, 54(3),
317-329. doi:10.1016/S0006-3223(03)00569-9
Narayan, K. V., Boyle, J. P., Thompson, T. J., Sorensen, S. W., & Williamson, D. F. (2003).
Lifetime risk for diabetes mellitus in the United States. Journal of the American Medical
Association, 290(14), 1884-1890.
Nezu, A. M., Nezu, C. M., & D'Zurilla, T. (2013). Problem-solving therapy: A treatment
manual. New York, NY: Springer Publishing Company.
Nezu, A. M., Nezu, C. M., Felgoise, S. H., McClure, K. S., & Houts, P. S. (2003). Project
genesis: Assessing the efficacy of problem-solving therapy for distressed adult cancer
patients. Journal of Consulting and Clinical Psychology, 71(6), 1036-1048.
doi:10.1037/0022-006X.71.6.1036
Oh, H., Ell, K., & Subica, A. (2014). Depression and family interaction among low-income,
predominantly Hispanic cancer patients: A longitudinal analysis. Supportive Care in
Cancer, 22(2), 427-434.
130
Palinkas, L. A., Wingard, D. L., & Barrett-Connor, E. (1996). Depressive symptoms in
overweight and obese older adults: A test of the “jolly fat” hypothesis. Journal of
Psychosomatic Research, 40(1), 59-66.
Pan, A., Lucas, M., Sun, Q., van Dam, R. M., Franco, O. H., Manson, J. E., … Hu, F. B . (2010).
Bidirectional association between depression and type 2 diabetes mellitus in women.
Archives of Internal Medicine, 170(21), 1884-1891.
Park, M., Katon, W. J., & Wolf, F. M. (2013). Depression and risk of mortality in individuals
with diabetes: A meta-analysis and systematic review. General Hospital Psychiatry, 35(3),
217-225.
Paschalides, C., Wearden, A., Dunkerley, R., Bundy, C., Davies, R., & Dickens, C. (2004). The
associations of anxiety, depression and personal illness representations with glycaemic
control and health-related quality of life in patients with type 2 diabetes mellitus. Journal of
Psychosomatic Research, 57(6), 557-564.
Perrino, T., Brown, S. C., Huang, S., Brown, C. H., Gómez, G. P., Pantin, H., … Szapocznik, J .
(2011). Depressive symptoms, social support, and walking among Hispanic older adults.
Journal of Aging and Health, 23(6), 974-993. doi:10.1177/0898264311404235
Piette, J. D., Richardson, C., Himle, J., Duffy, S., Torres, T., Vogel, M., … Valenstein, M .
(2011). A randomized trial of telephone counseling plus walking for depressed diabetes
patients. Medical Care, 49(7), 641.
Pollard, S. L., Zachary, D. A., Wingert, K., Booker, S. S., & Surkan, P. J. (2014). Family and
community influences on diabetes-related dietary change in a low-income urban
131
neighborhood. Diabetes Educator. Advanced online publication.
doi:10.1177/0145721714527520
Polonsky, W. H., Fisher, L., Schikman, C. H., Hinnen, D. A., Parkin, C. G., Jelsovsky, Z., …
Wagner, R. S. (2011). Structured self-monitoring of blood glucose significantly reduces
A1C levels in poorly controlled, noninsulin-treated type 2 diabetes: Results from the
structured testing program study. Diabetes Care, 34(2), 262-267. doi:10.2337/dc10-1732;
10.2337/dc10-1732
Pouwer, F. (2009). Should we screen for emotional distress in type 2 diabetes mellitus? Nature
Reviews Endocrinology, 5(12), 665-671.
Reddy, P., Philpot, B., Ford, D., & Dunbar, J. A. (2010). Identification of depression in diabetes:
The efficacy of PHQ-9 and HADS-D. British Journal of General Practice, 60(575), e239-
e245.
Renn, B. N., Feliciano, L., & Segal, D. L. (2011). The bidirectional relationship of depression
and diabetes: A systematic review. Clinical Psychology Review, 31(8), 1239-1246.
doi:10.1016/j.cpr.2011.08.001
Rieck, T., Jackson, A., Martin, S. B., Petrie, T., & Greenleaf, C. (2013). Health-related fitness,
body mass index, and risk of depression among adolescents. Medicine & Science in Sports
and Exercise, 45(6), 1083-1088.
Russell, L. B., Suh, D., & Safford, M. (2005). Time requirements for diabetes self-management:
Too much for many. Journal of Family Practice, 54(1), 52-56.
132
Ryan, C. M., Dulay, D., Suprasongsin, C., & Becker, D. J. (2002). Detection of symptoms by
adolescents and young adults with type 1 diabetes during experimental induction of mild
hypoglycemia role of hormonal and psychological variables. Diabetes Care, 25(5), 852-858.
Ryba, M. M., Lejuez, C. W., & Hopko, D. R. (2014). Behavioral activation for depressed breast
cancer patients: The impact of therapeutic compliance and quantity of activities completed
on symptom reduction. Journal of Consulting and Clinical Psychology, 82(2), 325-335.
doi:http://dx.doi.org/10.1037/a0035363
Sacco, W. P., Wells, K. J., Friedman, A., Matthew, R., Perez, S., & Vaughan, C. A. (2007).
Adherence, body mass index, and depression in adults with type 2 diabetes: The mediational
role of diabetes symptoms and self-efficacy. Health Psychology, 26(6), 693.
Sacco, W. P., & Yanover, T. (2006). Diabetes and depression: The role of social support and
medical symptoms. Journal of Behavioral Medicine, 29(6), 523-531.
Safford, M. M., Russell, L., Suh, D. C., Roman, S., & Pogach, L. (2005). How much time do
patients with diabetes spend on self-care? Journal of the American Board of Family Practice,
18(4), 262-270.
Schillinger, D., Handley, M., Wang, F., & Hammer, H. (2009). Effects of self-management
support on structure, process, and outcomes among vulnerable patients with diabetes: A
three-arm practical clinical trial. Diabetes Care, 32(4), 559-566. doi:10.2337/dc08-0787;
10.2337/dc08-0787
133
Schneider, K. L., Andrews, C., Hovey, K. M., Seguin, R. A., Manini, T., LaMonte, M. J., …
Pagoto, S. L. (2014). Change in physical activity after a diabetes diagnosis: Opportunity for
intervention. Medicine & Science in Sports & Exercise, 46(1), 84-91.
Schoenthaler, A., Ogedegbe, G., & Allegrante, J. P. (2009). Self-efficacy mediates the
relationship between depressive symptoms and medication adherence among hypertensive
African Americans. Health Education & Behavior, 36(1), 127-137.
Searle, A., Norman, P., Thompson, R., & Vedhara, K. (2007). Illness representations among
patients with type 2 diabetes and their partners: Relationships with self-management
behaviors. Journal of Psychosomatic Research, 63(2), 175-184.
doi:10.1016/j.jpsychores.2007.02.006
Shaw, B. A., Gallant, M. P., Riley-Jacome, M., & Spokane, L. S. (2006). Assessing sources of
support for diabetes self-care in urban and rural underserved communities. Journal of
Community Health, 31(5), 393-412.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New
procedures and recommendations. Psychological Methods, 7(4), 422-445.
Simon, G. E., Katon, W. J., Lin, E. H., Rutter, C., Manning, W. G., Von Korff, M., … Young, B.
A. (2007). Cost-effectiveness of systematic depression treatment among people with
diabetes mellitus. Archives of General Psychiatry, 64(1), 65-72.
Skaff, M. M., Mullan, J. T., Fisher, L., & Chesla, C. A. (2003). A contextual model of control
beliefs, behavior, and health: Latino and European Americans with type 2 diabetes.
Psychology and Health, 18(3), 295-312.
134
Spitzer, R. L., Kroenke, K., & Williams, J. B. W. (1999). Validation and utility of a self-report
version of PRIME-MD. Journal of the American Medical Association, 282(18), 1737-1744.
Stuart, M. J., & Baune, B. T. (2012). Depression and type 2 diabetes: Inflammatory mechanisms
of a psychoneuroendocrine co-morbidity. Neuroscience & Biobehavioral Reviews, 36(1),
658-676.
Toobert, D. J., & Glasgow, R. E. (1994). Assessing diabetes self-management: The summary of
diabetes self-care activities questionnaire. In C. Bradley (Ed.), Handbook of psychology and
diabetes (pp. 351-375). Berkshire, UK: Harwood Academic.
Toobert, D. J., Hampson, S. E., & Glasgow, A. R. (2000). The summary of diabetes self-care
activities measure: Results from 7 studies and a revised scale. Diabetes Care, 23(7), 943-
950.
U.S. Census Bureau. (2011). Statistical abstract of the United States: 2012 (131st ed.).
Washington, DC: U.S. Census Bureau.
Umpierrez, G. E., Gonzalez, A., Umpierrez, D., & Pimentel, D. (2007). Diabetes mellitus in the
Hispanic/Latino population: An increasing health care challenge in the United States.
American Journal of the Medical Sciences, 334(4), 274-282.
Unützer, J., Katon, W., Callahan, C. M., Williams, J. W., Jr., Hunkeler, E., Harpole, L., …
IMPACT Investigators: Improving Mood-Promoting Access to Collaborative Treatment.
(2002). Collaborative care management of late-life depression in the primary care setting: A
randomized controlled trial. Journal of the American Medical Association, 288(22), 2836-
2845. doi:10.1001/jama.288.22.2836
135
van der Feltz-Cornelis, C. M, Nuyen, J., Stoop, C., Chan, J., Jacobson, A. M., Katon, W., …
Sartorius, N. (2010). Effect of interventions for major depressive disorder and significant
depressive symptoms in patients with diabetes mellitus: A systematic review and meta-
analysis. General Hospital Psychiatry, 32(4), 380-395.
Vincent, D., Clark, L., Zimmer, L. M., & Sanchez, J. (2006). Using focus groups to develop a
culturally competent diabetes self-management program for Mexican Americans. Diabetes
Educator, 32(1), 89-97.
Vincent, D., McEwen, M. M., & Pasvogel, A. (2008). The validity and reliability of a Spanish
version of the summary of diabetes self-care activities questionnaire. Nursing Research,
57(2), 101-106.
Vohs, K. D., Baumeister, R. F., & Ciarocco, N. J. (2005). Self-regulation and self-presentation:
Regulatory resource depletion impairs impression management and effortful self-
presentation depletes regulatory resources. Journal of Personality and Social Psychology,
88(4), 632-657.
Von Korff, M., Katon, W., Lin, E. H., Simon, G., Ciechanowski, P., Ludman, E., … Young, B .
(2005a). Work disability among individuals with diabetes. Diabetes Care, 28(6), 1326-1332.
Von Korff, M., Katon, W., Lin, E. H., Simon, G., Ludman, E., Oliver, M., … Bush, T . (2005b).
Potentially modifiable factors associated with disability among people with diabetes.
Psychosomatic Medicine, 67(2), 233-240. doi:10.1097/01.psy.0000155662.82621.50
136
Wagner, J. A., Tennen, H., & Osborn, C. Y. (2010). Lifetime depression and diabetes
self ‐management in women with type 2 diabetes: A case–control study. Diabetic Medicine,
27(6), 713-717.
Wagner, E. H., Austin, B. T., Davis, C., Hindmarsh, M., Schaefer, J., & Bonomi, A. (2001).
Improving chronic illness care: Translating evidence into action. Health Affairs (Project
Hope), 20(6), 64-78.
Ware, J. E., Jr., Kosinski, M., & Keller, S. D. (1996). A 12-item short-form health survey:
Construction of scales and preliminary tests of reliability and validity. Medical Care, 34(3),
220-233.
Ware, J. E., Jr., & Sherbourne, C. D. (1992). The MOS 36-item short-form health survey (SF-
36): I. Conceptual framework and item selection. Medical Care, 30(6), 473-483.
Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender behavior
change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132(2), 249-
268.
Weinger, K., Butler, H. A., Welch, G. W., & La Greca, A. M. (2005). Measuring diabetes self-
care: A psychometric analysis of the self-care inventory-revised with adults. Diabetes Care,
28(6), 1346-1352.
Wen, L. K., Shepherd, M. D., & Parchman, M. L. (2004). Family support, diet, and exercise
among older Mexican Americans with type 2 diabetes. Diabetes Educator, 30(6), 980-993.
137
Whitty, P., Steen, N., Eccles, M., McColl, E., Hewison, J., Meadows, K., … Hutchinson, A .
(1997). A new self-completion outcome measure for diabetes: Is it responsive to change?
Quality of Life Research, 6(5), 407-413.
Williams, J. W., Katon, W., Lin, E. H., el, P. H., Worchel, J., Corne ll, J., … IMPACT
Investigators. (2004). The effectiveness of depression care management on diabetes-related
outcomes in older patients. Annals of Internal Medicine, 140(12), 1015-1024.
Wittkampf, K. A., Naeije, L., Schene, A. H., Huyser, J., & van Weert, H. C. (2007). Diagnostic
accuracy of the mood module of the patient health questionnaire: A systematic review.
General Hospital Psychiatry, 29(5), 388-395.
Woltmann, E., Grogan-Kaylor, A., Perron, B., Georges, H., Kilbourne, A. M., & Bauer, M. S.
(2012). Comparative effectiveness of collaborative chronic care models for mental health
conditions across primary, specialty, and behavioral health care settings: Systematic review
and meta-analysis. American Journal of Psychiatry, 169(8), 790-804.
Abstract (if available)
Abstract
This dissertation examined three research questions focused on relationships between depression and self‐care behaviors, self‐reported diabetes symptoms, and daily functioning among diabetes patients receiving depression care. Depression among diabetes patients has been believed to influence disease management. Many researchers have developed and tested depression care models for diabetes patients that have different formats and contents. However, many of the previous clinical trials found that patients assigned to innovative depression care groups did not report significantly increased self‐care behaviors. Yet, this group had significantly improved depressive symptoms, self‐reported diabetes symptoms, and daily functioning. Despite the contradicting evidence, few studies have examined potential underlying mechanisms and relationships across these variables. ❧ To bridge this literature gap, three individual studies were conducted to better understand equivocal findings in previous clinical trials. These studies involved secondary data analyses of data (N = 387) collected from the Multifaceted Diabetes and Depression Program (MDDP) randomized clinical trial (RCT) that tested the effectiveness of socio‐culturally adapted collaborative depression care for low‐income, predominantly Hispanic diabetes patients in safety‐net clinics. In this trial, data on the frequency of healthy diet, exercise, self‐blood glucose monitoring (SBGM) and foot care, and body mass index (BMI) were obtained. The value of depression care for diabetes patients was assessed. Depression care for diabetes patients is often believed to improve self‐care behaviors and diabetes symptoms by reducing apathy, a common cardinal symptom. However, belief is rarely based on empirical evidence, and findings from recent studies provide contradicting implications. The dissertation is aimed to provide the groundwork for future debates and rigorous empirical tests. ❧ Study 1 assessed whether problem‐solving therapy (PST) receipt, depressive symptoms change, and the interaction of these two independent variables predict patient diabetes self‐care improvement. Depression is a prevalent comorbidity associated with lower adherence to recommended self‐care behaviors and suboptimal diabetes outcomes. This association is attributable to feelings of hopelessness, lack of interest in daily jobs, and less optimism observed among depressed patients. Unfortunately, few RCTs demonstrated significantly increased self‐care behaviors in an intervention group that showed significant decline in depressive symptoms. However, this lack of evidence for this association lends to further questioning about whether reduced depressive symptoms observed during depression care are associated with self‐care behaviors. To examine this question, regression‐based analyses of data in a RCT are necessary. Also, PST for diabetes patients was aimed at identifying emotional distress associated with diabetes self‐care management behaviors in an effort to reduce emotional distress and improve self‐care management. We also examined the effect of PST on self‐care behaviors. The last question asked whether PST amplified the effect of declined depression on self‐care behaviors. For this question, the moderating effect of PST was examined. Secondary data analysis was conducted with data (N = 387) collected from a RCT that tested collaborative depression care for diabetes patients in safety‐net clinics. Because both the intervention group and enhanced usual care group showed a notable decrease in depressive symptoms, we analyzed these groups together with statistical control of group membership. PST receipt, depression symptoms change during 12 months after baseline, and the interaction of these two variables were regressed on each self‐care behavior change at 12 (N = 281), 18 (N = 249), and 24 (N = 235) months. A bilingual social worker provided PST sessions scheduled between 8 and 12 times. Depression was measured with the Patient Health Questionnaire-9 (PHQ-9). For the self‐care behavior frequency, the Summary of Diabetes Self‐Care Activities (SDSCA) was used to measure weekly frequency of the following behaviors: healthy diet, exercise, SBGM, and foot care. In addition, BMI was included as a self‐care behavior. Because each self‐care behavior had lower inter‐correlation, each behavior was regressed by depressive symptom change and demographic and clinical confounders. For analysis, multivariate regression analysis was conducted with SPSS 21.0. Three notable findings were found. First, PST receipt was not associated with concurrent and prospective increased self‐care behaviors. Second, decreased depression was associated with more frequent healthy diet at the 12‐ (p < .01), 18- (p < .05), and 24‐month follow-up (p < .05) and increased foot care at the 12‐ (p < .05) and 24‐month follow-up (p < .01). Finally, the interaction between PST receipt and depressive symptoms change was significantly associated with decreased foot care at the 12‐month follow‐up (p < .05) and decreased SBGM at the 18‐month follow‐up (p < .05), suggesting that patients receiving PST had decreased depressive symptoms, lower frequency of foot care, and SBGM. Future depression care should incorporate standard self‐care management education programs that have demonstrated positive impacts on health behaviors, health status, and healthcare use. ❧ Study 2 aimed to investigate whether improved self‐care behaviors, which are often observed among patients in depression care, predict better or worse depressive symptoms concurrently and prospectively. In several clinical trials testing the effectiveness of depression care, patients increased self‐care behavior frequency
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
A series of longitudinal analyses of patient reported outcomes to further the understanding of care-management of comorbid diabetes and depression in a safety-net healthcare system
PDF
Using a human factors engineering perspective to design and evaluate communication and information technology tools to support depression care and physical activity behavior change among low-inco...
PDF
Low-income, minority cancer patients who drop out of depression treatment
PDF
Clinical prediction models to forecast depression in patients with diabetes and applications in depression screening policymaking
PDF
Getting to end-of-life discussions in advanced cancer care: barriers and attitudes that limit end-of-life communication for disadvantaged Latinos
PDF
A study of depression disclosure among Latino older adults in Los Angeles
PDF
Social network influences on depressive symptoms among Chinese adolescents
PDF
Cultural risk and protective factors for tobacco use behaviors and depressive symptoms among American Indian adolescents in California
PDF
Mindfulness-based self-regulation for psychotic disorders: a feasibility study
PDF
The role of depression symptoms on social information processing and tobacco use among adolescents
PDF
Factors and correlates of sexual behaviors among young adults from continuation high schools
PDF
The role of protective factors on outcomes for Latinos with schizophrenia
PDF
Using observed peer discussions to understand adolescent depressive symptoms and interpersonal interactions
PDF
The role of social control in promoting healthy eating behavior among Chinese immigrants: an ecological approach
PDF
Untangling the developmental relations between depression and externalizing behavior among maltreated adolescents
PDF
The link between maternal depression and adolescent daughters' risk behavior: the mediating and moderating role of family
PDF
Addressing unmet needs and harnessing social support to improve diabetes self-care among low-income, urban emergency department patients with diabetes
PDF
Incarceration trajectories of mothers in state and federal prisons and their relation to the mother’s mental health problems and child’s risk of incarceration
PDF
Using mobile health to improve social support for low-income Latino patients with diabetes: a randomized mixed methods feasibility trial of TExT-MED FANS
PDF
Social self-control and adolescent substance use
Asset Metadata
Creator
Oh, Hyunsung
(author)
Core Title
Depression severity, self-care behaviors, and self-reported diabetes symptoms and daily functioning among low-income patients receiving depression care
School
School of Social Work
Degree
Doctor of Philosophy
Degree Program
Social Work
Publication Date
08/12/2014
Defense Date
08/12/2014
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Depression,depression care,Diabetes,Hispanic,OAI-PMH Harvest,safety‐net clinic,self‐care
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Ell, Kathleen (
committee chair
), Chou, Chih-Ping (
committee member
), Palinkas, Lawrence A. (
committee member
)
Creator Email
hyunsungoh@gmail.com,hyunsuno@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-460345
Unique identifier
UC11287116
Identifier
etd-OhHyunsung-2812.pdf (filename),usctheses-c3-460345 (legacy record id)
Legacy Identifier
etd-OhHyunsung-2812.pdf
Dmrecord
460345
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Oh, Hyunsung
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 a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
depression care
Hispanic
safety‐net clinic
self‐care