University of Southern California Dissertations and Theses

Older adults with visual impairments: the role of health dimensions in predicting falls

(USC Thesis Other)

Older adults with visual impairments: the role of health dimensions in predicting falls

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
OLDER ADULTS WITH VISUAL IMPAIRMENTS: THE ROLE OF HEALTH
DIMENSIONS IN PREDICTING FALLS

by

Bernard Alex Steinman
________________________________________________________

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(GERONTOLOGY)

December 2010

Copyright 2010 Bernard Alex Steinman
ii
DEDICATION
This dissertation is dedicated to my grandparents, William and Viola
Badgley, and Emil and Elizabeth Steinman; and to my parents, Bernard and Judith
Steinman, each of whom, I suspect, must share part of the credit for why I care
about these topics.
iii
ACKNOWLEDGEMENTS
I would like to extend my sincere thanks to members of my dissertation
committee for their time and efforts in helping me to complete this project: Dr. Jon
Pynoos (chair); Dr. Susan Enguidanos whose interest in the growth of students is
inspiring; and Dr. Jim Gauderman, whose teaching has provided me with valuable
tools I will build on and use for the rest of my career.
I would especially like to thank Dr. Pynoos for his guidance, support and
friendship as I plodded, sometimes quite messily and ungracefully, through the
completion of this research project. I have mentioned many times to people how
fortunate I feel to have a mentor who is so knowledgeable about the practical
aspects of gerontology, and I am inspired to be a similar type of person one day.
Without the kind support and mentorship of each member of my committee, this
entire process might have been even messier and more graceless than it already
was.
Faculty members not on my committee, but who deserve acknowledgement
of my appreciation include: Dr. Eileen Crimmins, who has challenged me and
taught me so much about research in this field; Dr. Caleb Finch, who kept me busy
and out of trouble during my first few years at USC; Dr. Greg Hise, who was good
enough to take his time away from history and policy to come across campus and
listen to “things-gerontological”; Dr. Phoebe Liebig, whose teaching will continue to
inspire my teaching, and whose wise approach to life and retirement will also be a
good model, if I’m so lucky as she is; Dr. Merril Silverstein, whose depth of
interests, methodological knowledge, and generosity with his time have always
iv
been truly appreciated; Dr. Liz Zelinski, who was kind enough to provide financial
support during my first year, and intellectual support along the way; and Assistant
Dean, Maria Henke who has helped to create a positive professional image for me,
which has been and will be very difficult for me to live up to (but I shall try). Each
member of the Davis School faculty has taught me important lessons not only
about how to treat students when I acquire a professorial teaching role; but more
importantly, how to treat people in general, in day-to-day life. I am nearly certain
that without each of their contributions to the academic infrastructure that was
afforded to me, as well as their patience and friendship that they have shown me
personally, I may not have completed even the first page of this document.
I am also indebted to my fellow USC gerontology students, who over the
years have provided me with great support, camaraderie, and (sometimes) the
inspiration to finish, and to move on. This list is long, as I am grateful to all the
students with whom I have come into contact, and with whom I have shared this
experience. However some particularly notable friends have included Caroline
Cicero; Alexis Denton; Zachary Gassoumous; Jessica Lendon; Joohong Min;
Emily Nabors; Adria Navarro; Anna Nguyen; Petrice Oyama; Sarinnapha
Vasunilashorn; and Amber Watts. Thanks guys for your patience and friendship.
Outside the hallowed halls of the Andrus Gerontology center, I am also
indebted to my good friends, who in times of uncertainty and malaise have given
me things to think about beyond vision loss and falling down. These include, but
are not limited to my buddies from the old days, Stephen T. Paul, a mentor and
friend from Mississippi State, who taught me how to read student papers
v
critically—my friendship with Dr. Paul showed me that hard efforts to write are not
always wasted; my friends at the RRTC on Blindness and Low Vision at
Mississippi State, Brenda Cavenaugh, Marty Giesen, and B.J. Lejeune, who
introduced me to the topic of vision loss and aging; Robert H. Sprague, a buddy
from Maine, and keeper of trivial, but interesting knowledge; and Richard G. Wood,
a man I’ve traveled with since our high school days, who taught me how to twist a
short story so that it is good, and makes people believe fiction.
And of course, I am indebted to my parents for their love and support:
thanks Bernard and Judith; and to Del Rae Head, who is a model sister, and good
friend.
Finally, I would like to recognize the Archstone Foundation whose
generous support made my education at USC and completion of this dissertation
possible; and the National Institute on Aging, Multidisciplinary Research Grant in
Gerontology (5T32AG000037), which provided integral support to me in my final
year, while I was finishing this manuscript. To all of the above: I extend a sincere
thank you.
vi
TABLE OF CONTENTS

Dedication

ii
Acknowledgements

iii
List of Tables

vii
List of Figures

x
Abstract

xi
Chapter I: Aging, Visual Impairment, and Falls

1
Chapter II: Vision Loss and Health Dimensions

35
Chapter III: Moderating Effects of Vision Loss on Limb Functioning and
Falls

76
Chapter IV: Mediating Effects of Disability and Functional Status in
Predicting Falls in Older People with Poor Vision

105
Chapter V: Fall Prevention Interventions for Older Adults with Vision
Impairment

128
Bibliography

152

vii
LIST OF TABLES

Table 1.1: Recent and Projected Prevalence Estimates of Clinically
Measured Blindness, Low Vision, and Total Visual
Impairment, Extrapolated from 2000 U.S. Census

9
Table 1.2: Recent and Projected Prevalence Estimates of
Self-Reported Blindness, Low Vision, and Total Visual
Impairment, Based on National Health Interview Survey
(2002), Extrapolated from 2000 U.S. Census

12
Table 1.3: Percentage of Growth in Each Age Group in 10-Year
Increments

14
Table 2.1: Unadjusted Descriptive Statistics for Sociodemographic
Variables by Self-Reported Vision Status (NHANES IV, 1999
– 2008, Weighted)

54
Table 2.2: Unadjusted Descriptive Statistics for Biological Risk
Indicators by Self-Reported Vision Status (NHANES IV,
1999 – 2008, Weighted)

55
Table 2.3: Binary Logistic Regression Models Testing the Effect of
Self-Reported Vision Status on Biological Risk Indicators
(NHANES IV, 1999 – 2008, Weighted)

57
Table 2.4: Unadjusted Descriptive Statistics for Self-Reported
Pathology Indicators by Self-Reported Vision Status
(NHANES IV, 1999 – 2008, Weighted)

58
Table 2.5: Binary Logistic Regression Models Testing the Effect of
Self-Reported Vision Status on Self-Reported Pathological
Conditions (NHANES IV, 1999 – 2008, Weighted)

59
Table 2.6: Unadjusted Descriptive Statistics for Self-Reported
Functional Indicators by Self-Reported Vision Status
(NHANES IV, 1999 – 2008, Weighted)

61
Table 2.7: Multinomial Logistic Regression Models Testing the Effect of
Self-Reported Fair Vision Status on Self-Reported Difficulty
in Functioning Indicators (NHANES IV, 1999 – 2008,
Weighted)

63
viii
Table 2.8: Multinomial Logistic Regression Models Testing the Effect of
Poor Self-Reported Vision Status on Self-Reported Difficulty
in Functioning Indicators (NHANES IV, 1999 – 2008,
Weighted)

64
Table 2.9: Unadjusted Descriptive Statistics for Self-Reported Daily
Living Activity Indicators by Self-Reported Vision Status
(NHANES IV, 1999 – 2008, Weighted)

66
Table 2.10: Multinomial Logistic Regression Models Testing the Effect of
Self-Reported Fair Vision Status on Self-Reported Difficulty
in Daily Living Activities (NHANES IV, 1999 – 2008,
Weighted)

68
Table 2.11: Multinomial Logistic Regression Models Testing the Effect of
Self-Reported Poor Vision Status on Self-Reported Difficulty
with Daily Living Activities (NHANES IV, 1999 – 2008,
Weighted)

69
Table 3.1: Demographic Descriptive Statistics by Gender (HRS, 2004,
Weighted)

91
Table 3.2: Percent Falls, Vision Status, Home Modifications, Health
Status, and Mean Number of Upper- and Lower-Limb
Disabilities by Gender (HRS, 2004 & 2006, Weighted)

93
Table 3.3: Women: Odds-Ratios Representing Independent and
Moderating Effects of Key Explanatory Variables and Control
Variables (HRS, 2004 & 2006, Weighted; N = 4,950)

96
Table 3.4: Men: Odds-Ratios Representing Independent and
Moderating Effects of Key Explanatory Variables and Control
Variables (HRS, 2004 & 2006, Weighted; N = 3,500)

97
Table 4.1: Descriptive Sociodemographic Characteristics, Number of
Chronic Diseases, and Falls by Vision Status; Test of
Significance and Pair-Wise Comparisons between Vision
Groups (HRS, 2006 & 2008, Weighted)

118
Table 4.2: Descriptive Statistics for Indices of Functioning and Disability
by Vision Status; Test of Significance and Pair-Wise
Comparisons between Vision Groups (HRS, 2006,
Weighted)

119

ix
Table 4.3: Partial Correlation Matrix for Key Explanatory Variables and
Covariates, Controlling for Age, Sex, Race, Education and
Marital Status (HRS, 2006 & 2008, Weighted)

121
Table 4.4: Women: Comparable Regression Coefficients and Tests of
Indirect Effects of Mediators, Controlling for
Sociodemographic Variables, Chronic Disease, Functioning,
and Disability (HRS, 2006 & 2008, Weighted)

122
Table 4.5: Men: Comparable Regression Coefficients and Tests of
Indirect Effects of Mediators, Controlling for
Sociodemographic Variables, Chronic Disease, Functioning,
and Disability (HRS, 2006 & 2008, Weighted)
123
x
LIST OF FIGURES

Figure 1.1: Cross-Section of the Human Eye, Divided into Optical or
Neural Functions by Anatomical Components

18

Figure 2.1: The Disability Process: Theoretical Pathway Leading from
Risk Factors to Disability

36
Figure 3.1: Independent Systems Effects of Multiple Systems Can
Influence Fall Risk

79
Figure 3.2: Integrated Moderating Effects of Vision Loss on Other
Systems Can Influence Fall Risk

81
Figure 3.3: Feedback Loops within the Disability Process Model Can
Influence Fall Risk

82
Figure 4.1: Total Effect of Vision on Fall Risk (panel A); and Total Direct
Effect Mediated by Muscle Decline (panel B)

106
Figure 4.2: Theoretical Tests of Mediated Relationships between
Self-Reported Vision Status, and Falls

111
Figure 5.1: Potential Intervention Points Exist between Health
Dimensions in the Disability Process

132

xi
ABSTRACT
The purpose of this dissertation was to examine relationships between self
reported vision impairment, health dimensions, and falls among older people; and
to describe an indirect pathway through which vision loss may increase risk for falls
by way of poor health outcomes in dimensions of health that are believed to lead to
disability. Whereas previous research has tended to focus on direct effects of
vision loss on fall risk, the primary goal of this research was to examine integrated
effects between systems that might lead to increased fall risk via moderated or
mediated relationships.
First, in order to establish health disparities by self reported vision status,
binary and multinomial logistic regression models were performed using data from
the National Health and Nutrition Examination Survey, 1999-2008
(N=6,693) to estimate the probability of negative health outcomes across four
dimensions of health, including biological risk, pathological conditions, functional
difficulties and disability in daily living activities, by vision status. Results suggest
that older adults with poor vision have greater likelihood of experiencing negative
health outcomes across the four dimensions.
Next, two waves (2004 & 2006) from the Health and Retirement Study
(HRS) (N=8,449) were analyzed using binary logistic regression to investigate
whether a moderating relationship exists between vision status and upper and
lower limb functioning. This relationship was hypothesized to derive, in part, from
decreased physical activity that often follows vision loss in late life, to result in
poorer functional ability. Little evidence was found for a moderating effect of
xii
self-reported vision status on musculoskeletal health and functioning; however,
results suggest that declines and/or gains in functioning across short periods of
time may supercede self-reported vision as a predictor of falls. Thus, poor
self-reported vision status may not be as good an indicator of fall risk in older
adults as might otherwise be assumed.
In a third study, two waves (2006 & 2008) from HRS (N = 9,143) were used
to test whether functional difficulties and disability with daily living activities
mediate self-reported vision loss to increase fall risk among older adults. Binary
logistic and Poisson regression analyses were conducted to test indirect paths
leading from self-reported vision to falls, through declines in indices of functioning
and disability. No evidence was found for a mediating effect among women;
however, for men, large muscle groups were implicated as partially mediating risk
factors for falls among participants with poor vision.
Finally, implications of the three studies are discussed, including the need
for prioritizing improved muscle strength of older persons with vision impairments
as preventive measure against falls. It is acknowledged that the most effective fall
prevention interventions are likely multifactorial in structure; therefore,
interventions pertaining to education, medical assessment, exercise, and home
assessment and modification are discussed with respect to older persons who
have vision impairments.
1
CHAPTER I: AGING, VISUAL IMPAIRMENT, AND FALLS
A. Introduction
An increasingly common concern for those who are interested in public
health and public policy in the United States is population aging and the
accompanying health outcomes which are often associated with living until an old
age. One of the more common health outcomes associated with living a long life is
the decline and/or loss of vision, which has the potential to drastically influence the
health and well-being of older adults. It is estimated that up to a quarter of
individuals older than age 75 report some form of vision loss (Horowitz, Brennan, &
Reinhardt, 2005); and a disproportionate percentage of people who are blind or
visually impaired are older than age 60 (Lighthouse International, 2008). By itself,
visual impairment that occurs in late life can reduce wellbeing of older people by
making it more difficult to perform everyday activities, which may be taken for
granted earlier in life, such as driving, working at a job, reading, enjoying crafts
and hobbies, or readily recognizing family and friends from across a crowded
room.
Nevertheless, the impact of vision loss is seldom experienced as a solitary
problem that is limited just to the inability to see. Older people with visual
impairments often experience dynamic losses across multiple systems in
conjunction with their sensory decline. Vision loss is often the result of, or
co-occurs with other health problems such as chronic diseases and secondary
health outcomes that are commonly associated with frailty, such as hip fractures
and recurrent falls. In comparison to other health issues related to aging, vision
2
impairment in late life is relatively unique due to its associations with a range of
health-related outcomes. Most directly, age-related vision loss is known to
interfere with the ability of older individuals to perform tasks that are necessary for
living independently (Crews & Campbell, 2001; Crews & Campbell, 2004;
Haymes, Johnston, & Heyes, 2002; Jacobs, Hammerman-Rozenberg, Maaravi,
Cohen, & Stessman, 2005; Lamoureux, Hassell, & Keeffe, 2004; Salive et al.,
1994). In addition, vision loss has also been linked to psychological (Brody et al.,
2001; Chia, Mitchell, Rochtchina, Foran, & Wang, 2003; Varma, Wu, Chong, Azen,
& Hays, 2006) and social wellbeing (Crews & Campbell, 2004), as well as to
self-rated health (Wang, Mitchell, & Smith, 2000), biological risk factors
(Schalkwijk et al., 1999; Steinman & Vasunilashorn, in press; Vine, Stader,
Branham, Musch, & Swaroop, 2005;), clinical co-morbidity (Crews & Campbell,
2004), and even mortality (Laforge, Spector, & Sternberg, 1992). Negative
consequences of visual impairment have also been studied with respect to
secondary health outcomes such as hospital/emergency room utilization (Jacobs
et al., 2005), hip fractures (Abdelhafiz & Austin, 2003; Anastasopoulos, Yu, &
Coleman, 2006; Felson et al., 1989), and, most germane to this dissertation, falls
(Coleman et al., 2004; Lord & Dayhew, 2001; Steinman, 2008; Steinman, Pynoos,
& Nguyen, 2009).
Although the majority of older adults enjoy good health and functioning well
into late life, vision loss and other often disabling health conditions, which may
precede falls, become more likely as individuals progress into old age. As a result,
the incidence of falling is greater in old age. It is estimated that a third of individuals
3
older than age 65 will fall in any given year and that, of those who fall, up to 20%
may experience injuries related to their falls (Masud & Morris, 2001). Falls are the
number one cause of injuries to older adults, and the number one cause of
injury-related death for this age group (Centers for Disease Control and Prevention
[CDC], 2008). When coupled with forecasts of unprecedented demographic
changes, age-related increases in the incidence of vision loss and falls portend a
potentially dire public health problem if no action is taken to support and/or
rehabilitate the effects of vision loss as it relates to falls. By the year 2030, the
number of Americans who are 65 years of age and over is expected to double to
about 71 million people, representing roughly 20% of the total population
(compared to 12% in 2000) (U.S. Census Bureau, 2010a)—many of these
individuals will likely experience declines in their vision that could reduce their
mobility, increase their dependence, and result in other negative health outcomes
(especially falls), that are expensive to address ex post facto. Fortunately, it may
be possible to reduce some of the negative health outcomes that are associated
with visual impairment through lifestyle and environmental changes, and
rehabilitation training in skills that can improve the capacity of older adults to live
independently and safely. This looming public health crisis will demand resources,
and influence health policy decisions, as institutions and programs seek to reduce
the burden of vision-related falls during the first three decades of the 21
st
century.
The pages that follow document a series of studies conducted to assess the
health status of adults who report vision impairment in late life, as well as the role
of self-reported vision in falls among older people. As a whole, the purpose of this
4
dissertation is to describe health disparities experienced by older adults with vision
impairments, and to test hypotheses about how self-reported vision impairment
could lead directly and indirectly to increased risk of falling. The remainder of
Chapter 1 is divided into sections that describe the incidence and demography of
vision loss, and describes how different ways of measuring vision loss can result in
vastly different estimates of prevalence. Finally, the chapter concludes with a
discussion about normal and pathological age-related vision impairments with
particular attention to their relationship to falls in older people. Chapter 2 describes
the first of three studies that was designed to be a survey of the health of older
adults with visual impairments. An epidemiological model of the disability process
is introduced, that functions as an overarching theoretical basis for the two
remaining studies. In Chapter 3, a distinction is introduced between independent
effects of self-reported vision loss, and integrated effects, in which vision loss may
increase fall risk through its interaction with systems outside of the visual system.
A second study is described, that was designed to test whether self-reported vision
loss moderates losses in other systems, by way of an epidemiological model of
disability. Moderating effects are found when one independent variable affects the
magnitude of another independent variable—thus, analyses in the second study
were conducted to determine whether older persons with visual impairments
experienced greater loss of limb functioning than persons with normal vision, due
to their visual impairments. By contrast, Chapter 4 describes analyses that were
designed to assess whether vision is mediated by losses in other systems.
Mediating variables intervene between independent variables and dependent
5
variables to produce an outcome that is specifically related to the mediator.
Analyses in the third study tested the hypothesis that vision loss influences fall risk,
contingent on disability and functional loss, in addition to vision’s direct effect.
Regression modeling was employed to examine relationships between these
health dimensions within the context of the epidemiological model of disability.
Finally, in Chapter 5, the relevance of the epidemiological model is summarized in
relation to interventions and programs directed at improving functioning and
reducing the negative impact of vision loss, with respect to falls. Given that the
disability process model describes a progression leading from risk factors to
disablement, the model has the potential to serve as an effective structure for
describing prospective health and fall-prevention interventions that could be
tailored specifically to older adults who have vision impairments. Interventions are
recommended at specific stages of the disablement process, as preventive
measures against negative health outcomes, progression toward disablement and
secondary health outcomes, namely falls that often accompany functional decline
and disability.
B. Prevalence and Demographics of Visual Impairment
Prevalence of Vision Impairment
The incidence of visual impairment in the U.S. is expected to rise
dramatically during the first 30 years of the 21
st
century, due to an unprecedented
growth in the absolute number and proportions of older people in the population.
Nevertheless, vision is a multidimensional sensory function, and it is therefore
difficult to reach consensus about how to measure, and therefore project the
6
prevalence of vision loss (Orr, 1992). Reporting the current prevalence and
projected rates of vision impairment is challenging due to the great amount of
variability between studies in defining what is meant by “visual impairment”. One
means of categorizing prevalence estimates is by distinguishing between
assessment methods that are based on clinical measures of visual function and
vision problems that are self-reported (Horowitz et al., 2005).
Relatively conservative estimates are derived when vision status is
measured objectively in a clinical setting. By the common U.S. definition,
individuals are considered to be legally blind when their best corrected visual
acuity is 20/200 or worse in the better-seeing eye (i.e., they are able to see from a
distance of 20 feet objects and details that someone with normal vision is able to
see from 200 feet); or a visual field of 20 degrees or less (i.e., a 20-degree angle
can be seen when each eye is directed forward). Low vision acuity is defined as
having 20/40 vision or worse. Distance visual acuity is often tested using wall
charts with high-contrast black lettering on white backgrounds. Under such optimal
testing conditions, the effects of certain types of visual impairment, which have the
potential to impact daily living functioning, and increase fall risk, could be
underestimated (West et al., 2002). A significant percentage of falls, for example,
are attributed to hazards in the environments of many older adults (Pynoos,
Steinman, & Nguyen, in press), including hallways that are inadequately
illuminated or poorly maintained sidewalks (Josephson, Fabacher, & Rubenstein,
1991). Simple measures of visual acuity in ideal and constant settings, such as a
laboratory or clinical setting, may be poor predictors of falls in actual environments
7
where older adults live and spend their time. Constant and subtle changes in
lighting, as well as dynamic changes in the positioning of objects in the
environment make the homes of older adults incomparable to the laboratory
setting. Nevertheless, clinical thresholds, such as those derived from wall charts
are commonly used in studies of vision loss, as well as for legal criteria to
determine eligibility for state and federal programs (Kirchner & Schmeidler, 1997),
and driving eligibility (Jacobs et al., 2005). In addition to standard tests for visual
acuity, researchers who are interested in prevalence estimates based on objective
clinical measures of visual function often utilize standard tests for contrast
sensitivity; glare sensitivity; stereo acuity; and width of visual fields.
Despite their tendency to underestimate the negative consequences of
visual impairments, clinical measures provide an objective standard on which to
base research. Unfortunately, costs associated with clinical assessments may be
prohibitive for use in studies with large samples. Larger studies that collect clinical
measures are rare, but include the National Health and Nutrition Examination
Survey (NHANES), a nationally representative study that collects measures of
visual acuity and refraction readings via physical examination of study participants
(CDC, 2010). Generally, only relatively small community or population-based
studies have used objective measures to estimate prevalence of visual
impairment, due to the expense and training requirements associated with
conducting examinations. Examples of prominent smaller studies that have used
clinical data to estimate the prevalence of vision impairment and eye diseases
include the Framingham Eye Study (FES) (e.g., Kahn et al., 1977), the Salisbury
8
Eye Evaluation Project (SEE) (e.g., Rubin et al., 1997), and the Baltimore Eye
Survey (BES) (e.g., Tielsch, Sommer, Witt, Katz, & Royall, 1990). Relatively small
studies like these can be advantageous because they are able to assess entire
communities with specific demographic traits. By the same token, a major
drawback of these studies is that they are often based on idiosyncratic
populations, making it difficult to generalize results to the population of older adults
at large. For example, results reported from the FES were derived from eye
examinations of survivors from the Framingham Heart Study conducted in 1948 in
Framingham, Massachusetts. Thus, Kahn et al. (1977) explicitly acknowledged
that response biases related to survival and previous participation by examinees in
the earlier study made it difficult to make valid inferences about the visual health of
the overall population in general. Nevertheless, FES has made valuable
contributions, including improving knowledge about specific age-related vision
diseases, such as cataract, diabetic retinopathy, macular degeneration and
glaucoma, which were emphasized during the study’s clinical examinations.
In a study conducted by The Eye Diseases Prevalence Research Group
(2004a), clinical prevalence of blindness and low vision was estimated based on
combining data from eight population-based studies recently conducted in the
United States, Australia, and Europe. According to this study, slightly more than
8% of individuals age 65 years and older experienced vision impairment, including
6% with clinically-defined low vision and slightly more than 2% who were blind (see
table 1.1). In 2000, this percentage represented nearly 3 million older Americans
with visual impairments, as extrapolated from the 2000 U.S. Census (U.S. Census
9
Bureau, 2010b). Of those with visual impairments, roughly 28% were legally blind,
with the remaining having low vision. Based on rates of visual impairment in 2000,
the absolute number of older individuals with clinically determined vision loss is
projected to more than double to nearly 6 million by the year 2030. Of these, nearly
1.7 million are projected to be legally blind.

Prevalence estimates of blindness and low vision often differ from clinical
measures when subjective or self-reported measures are used to assess vision
Table 1.1: Recent and Projected Prevalence Estimates of Clinically Measured
Blindness, Low Vision, and Total Visual Impairment, Extrapolated from 2000 U.S.
Census
Age % 2000 2000 2010 2020 2030
Blind
65 to 74 0.47% 86,000 99,461 148,606 177,452
>=75 4.40% 730,000 834,369 1,004,923 1,473,368
Total 65+ 2.33% 816,000 933,830 1,153,529 1,650,821
Low Vision
65 to 74 1.55% 285,000 329,608 492,473 588,069
>=75 10.99% 1,824,000 2,084,780 2,510,931 3,681,402
Total 65+ 6.03% 2,109,000 2,414,388 3,003,404 4,269,471
Total Visual Impairment
65 to 74 2.02% 371,000 429,068 641,079 765,521
>=75 15.38% 2,554,000 2,919,149 3,515,854 5,154,771
Total 65+ 8.36% 2,925,000 3,348,218 4,156,933 5,920,291
10
status. Researchers who are concerned about visual functional capacity of
participants may rely on measures that emphasize self- or proxy-reports. Such
measures are designed to determine whether a person sees well enough to
perform specific tasks, such as reading, recognizing a face, or driving a car
(Kirchner & Schmeidler, 1997). Examples of surveys that assess self-reported
visual functional limitations include the Census Bureau’s Survey of Income and
Program Participation (SIPP), the Medicare Current Beneficiary Survey (MCBS),
the National Health Interview Survey (NHIS), and the University of Michigan’s
Health and Retirement Study (HRS). The questionnaire section of NHANES also
requests that participants self-report their vision status.
In large surveys that quantify vision status via subjective measures,
participants are often classified as having visual impairments when they answer
“yes” to questions such as “Do you have any trouble seeing even when wearing
glasses or contact lenses?” Participants are categorized as blind when they
answer “yes” to questions such as “Are you blind or unable to see at all?”
(Ryskulova et al., 2008). Large studies, such as the HRS ask participants to
self-rate their vision status, with potential responses ranging from “excellent” to
“poor” or “legally blind” (University of Michigan Health and Retirement Study,
2010). Subjective responses to items like “Are you blind?” can sometimes result in
data that are misleading or difficult to interpret— especially data pertaining to older
adults who may only recently have lost their vision (Orr, 1992). Although these
individuals might be categorized as “blind” by clinical standards, their subjective
sense may be that they are not “blind”, but instead “have troubles seeing”.
11
According to Orr, the refusal of some older adults to identify themselves as being
blind, after a lifetime of normal vision status is not uncommon and may lead to
deceptively small prevalence figures with respect to blindness. Nevertheless, most
studies that rely on self-reported measures of vision impairment in general arrive at
larger prevalence estimates than studies that assess clinical measures.
Table 1.2 presents recent and projected prevalence estimates of
self-reported blindness, low vision, and total self-reported visual impairments of
persons 65 years and older. Percentages are based on data from the 2002 wave
of NHIS (Ryskulova et al., 2008). The NHIS categorized participants as visually
impaired if they reported having “any difficulty seeing even when wearing glasses”.
Participants were categorized as blind when they answered “yes” to an item that
asked whether they were “blind or unable to see at all”. Prevalence figures and
percentages are extrapolated from census data (U.S. Census Bureau, 2010b).
According to these estimates, more than 6.5 million Americans over the age of 65
(representing about 19% of the older population) experienced self-reported visual
impairments in 2000. Of these, the vast majority (95%) reported low vision, rather
than blindness. Based on these percentages, by 2030, the number of older
Americans with self-reported visual impairments is expected to double to more
than 13 million.
Although vision loss can occur at any age, the prevalence of visual
impairment increases dramatically in the last third of the life-span (Ryskulova et al.,
2008). Furthermore, as life expectancies continue to increase and a larger portion
of the population lives longer, the largest percentage of older people experiencing
12
significant vision loss will continue to shift toward the oldest-old age group. In
1950, when life expectancy was 68.2 years, it was estimated that about 860,000
people over age 65 (roughly 7%) experienced severe vision impairment (Crews,
1994). Among these, the vast majority (68.5%) was composed of individuals
between ages 65 and 74 years old (U.S. Census Bureau, 2010c).

Table 1.2: Recent and Projected Prevalence Estimates of Self-Reported
Blindness, Low Vision, and Total Visual Impairment, Based on National Health
Interview Survey (2002), Extrapolated from 2000 U.S. Census
Self-reported Blind
Age % 2002 NHIS 2000 2010 2020 2030
65 to 74 0.5% 91,955 106,348 158,896 189,740
>=75 1.5% 249,012 284,613 342,791 502,583
Total 65+ 0.97% 340,967 390,961 501,687 692,323
Self-reported Low Vision
65 to 74 14.5% 2,666,693 3,084,079 4,607,978 5,502,450
>=75 21.1% 3,502,762 4,003,557 4,821,926 7,069,669
Total 65+ 17.63% 6,169,455 7,087,636 9,429,904 12,572,119
Total Self-reported Visual Impairment
65 to 74 15.00% 2,758,648 3,190,427 4,766,874 5,692,190
>=75 22.60% 3,751,774 4,288,170 5,164,717 7,572,252
Total 65+ 18.61% 6,510,422 7,478,597 9,931,591 13,264,442

In contrast, by 1990, life expectancy had increased to slightly more than 75
years, and the number of older persons with visual impairments had grown to 2.7
13
million (roughly 8.6%; Crews, 1994). Compared to 1950, the proportion of persons
75 years and older had increased from 31.5% to about 42.1% (U.S. Census
Bureau, 2010). Currently, the largest proportion of vision loss is experienced by
those who are over age 75 (The Eye Diseases Prevalence Research Group,
2004a; U.S. Census Bureau, 2010b) by both clinical and self-reported standards
(see tables 1.1 and 1.2). This trend toward greater numbers of older people
experiencing visual impairments is projected to continue for decades to come, as
the relative proportion of persons over 75 increases, and continues to be the
fastest growing segment of the population.
Table 1.3 displays projections of population growth in each age category,
by ten-year increments. Between 2010 and 2020, the total population of
individuals over age 65 is expected to increase by about 36% (U.S. Census
Bureau, 2010a). This period represents the time interval in which the oldest
baby-boomers will begin to enter retirement age. As such, the age group that will
experience the greatest amount of growth (almost 50%) during this period will be
between 65 and 74 years old. However, between 2020 and 2030, as baby
boomers begin to move into the 75 and older age group, this older group will
experience the greatest amount of growth (almost 47%). By 2030, the entire cohort
of baby boomers will be older than 65 (representing about 71.5 million people),
and more than 33.5 million Americans will be over age 75.
Viewed in conjunction with population growth estimates that place those
age 75 and above in the fastest growing age category (U.S. Census Bureau,
2010a), it is clear that much of the expected increase in the numbers of older
14
adults who are blind or visually impaired will stem from growth within the oldest old
group. Among persons age 75 and older, the total incidence of clinically measured
visual impairment is about 1 in 6 (15%; see table 1.1). In 2030, this will equate to
more than 5.1 million people. The total rate of self-reported visual impairment
among this age group is slightly more than 1 in 5 (23%; see table 1.2), which in
2030, will equal nearly 7.6 million older persons.

Table 1.3: Percentage of Growth in Each Age Group in 10-Year Increments
Age 2000 2010 2020
2030
65 to 74

18,390,986

21,269,509
(15.65%)

31,779,159
(49.41%)

37,947,933
(19.41%)

>=75

16,600,767

18,974,204
(14.30%)

22,852,732
(20.44%)

33,505,538
(46.62%)

Total 65+

34,991,753

40,243,713
(15.01%)

54,631,891
(35.75%)

71,453,471
(30.79%)

Demographics of Visual Impairment
According to the CDC (2006), the absolute number of women at any age
who have vision impairments exceeds the number of men by nearly 45%. This is
not surprising given differences in life expectancies of men and women. In 2006,
life expectancy at age 65 of women was an additional 19.7 years; whereas the life
expectancy of men at age 65 was only 17.0 years (Arias, 2010). Most studies that
have examined differences in prevalence of vision impairment based on sex have
reported few statistical differences when analyses were adjusted for age. In
meta-analyses conducted by The Eye Diseases Prevalence Research Group
(2004a; 2004b), none of the six studies described, nor the pooled total of all six
15
studies found statistical differences between white women and men, with respect
to visual acuity of 20/200 or worse (blindness) when age was controlled. With
respect to low vision, only three out of the six studies and the pooled total reflected
statistical differences, with women experiencing slightly greater prevalence in
each case. Additionally, few clinically significant differences between men and
women have been reported with respect to specific clinical vision tests (such as
tests for contrast sensitivity, glare sensitivity, and stereo acuity) (Rubin et al.,
1997). Some small differences between sexes have been reported with respect to
the types of vision pathologies that lead to vision loss. For instance, findings
reported by Leibowitz et al. (1980) based on FES data, concluded that a slightly
larger proportion of the women (13.2%) than men (11.2%) were affected by single
eye diseases, and a much larger proportion of women than men had multiple eye
diseases (34.6% and 22.9%, respectively). With respect to specific diseases,
women over the age of 75 experienced greater prevalence of senile cataracts,
diabetic retinopathy, and macular degeneration; whereas men had greater
prevalence of open angle glaucoma.
Thus, a good portion of the crude differences observed between women
and men, with respect to the prevalence of vision impairment can be attributed to
greater age and the longer life expectancies of women. Many of the differences by
race can also be attributed to interactions between race and life expectancy, as
well as health disparities, and socioeconomic variables. According to the CDC
(2006), the prevalence of “vision troubles” among African Americans is 9.4%, but
only 8.9% for their white counterparts. Another analysis of self-reported measure
16
of vision found that nearly 12% of non-Hispanic blacks reported some form of
visual impairment, compared to 9.6% among non-Hispanic Whites (Ryskulova et
al., 2008). As with Whites, the incidence of blindness and visual impairment among
minority groups increases dramatically with age. Unfortunately, meta-analyses
such as that conducted by The Eye Diseases Prevalence Research Group
(2004a) have not provided robust prevalence information specific to age and
racial/ethnic groups because most study samples are too small to provide enough
power to detect differences among race-based subgroups within their samples. By
contrast, the Baltimore Eye Survey was specifically designed to determine
prevalence for a multiracial urban population with a large minority population. Of
the 5,341 people who received ophthalmological screenings as part of the BES,
nearly half (2,395 participants) were black, providing a strong bases for making
inferences about prevalence and causes of vision loss in this subgroup of the
population. According to results reported by Tielsch et al. (1990), the age-adjusted
rates of blindness and visual impairment among Blacks were twice those of Whites
in the BES. Furthermore, the difference in prevalence between Blacks and Whites
continued to grow from age 40 until age 70, when the two curves converged.
According to The Eye Diseases Prevalence Research Group (2004a), the specific
pathological causes of blindness and visual impairment differ by race. For
example, the leading cause of blindness among older white people is age-related
macular degeneration, causing more than half (54%) of the cases. By contrast, the
leading causes of blindness for blacks are cataracts (50.9%) followed by glaucoma
(26.0%). These differences likely result from various interactions between race,
17
sex, age, and socioecomonic differences, including reduced access to eye care
that can influence health outcomes of individuals regardless of race. For example,
data from the FES were used by Leibowitz et al. (1980) to characterize macular
degeneration as a retinal abnormality that affects a disproportionate number of
persons over the of 75. Thus, the disease affects a greater number of whites,
because white women are more likely than any other group to survive to this age
(Arias, 2010). Similarly, the likelihood of acquiring diseases such as cataracts and
glaucoma, which often occur in relatively young people, can be influenced by
lifestyle factors, such as diet and smoking habits, and may be more prevalent
among persons with shorter life-expectancies and lower socioeconomic status.
C. Age-Related Vision Loss and Falls in Older Adults
This section provides an introduction to specific normal and pathological
age-related changes that influence clinical and self reported measures of vision
function, and their relation to falls in older adults. Changes are discussed in
relation to a two-tiered typology, which distinguishes factors based on the
structures in the eye and mechanisms by which the structures operate.
Figure 1.1 depicts a cross-section of the human eye and divides its
anatomical components into optical or neural functions, adapted from Curcio,
Millican, Allen, and Kalina (1993). Anatomical structures of the eye that have
optical functions are those that operate physically or mechanically on light, by
refraction, aperture control, or by focusing light for processing at later stages in the
system. Most of these structures, with the exception of the vitreous humor, are
situated in the anterior section of the eyeball (left, in figure 1.1), where light enters,
18
including the cornea, the iris, the lens, and ciliary bodies which shape the lens
during accommodation.

Figure 1.1: Cross-Section of the Human Eye, Divided into Optical or Neural
Functions by Anatomical Components (Conceptually Adapted from Curio et
al.,1993; Drawing by Bernard Steinman)

In contrast, eye structures that serve neural functions process visual
information by means of electrochemical reactions—similar to neurons that deliver
signals in the central and peripheral nervous systems. Retinal structures, which
line the posterior wall of the eyeball (right, in figure 1.1), and advance through the
optic disk to the brain operate primarily via neural mechanisms. Although vision
19
impairments originating in either section of the eyeball can result in vision loss that
increases fall risk; impairments that are based on optical structures are often more
amenable to repair. For example, cataracts are among the most common causes
of vision impairment in older adults. They reduce visual acuity by way of clouding
formations on the lens. Although surgery for cataracts was once a complex, and
somewhat risky inpatient procedure, advances in opthalmogical medicine have
made removal and replacement of lenses affected by cataracts into a relatively
routine outpatient procedure, that has increased at least four-fold since 1990
(Taylor, 2000).
Normal Age-related Changes to Vision
In discussing vision loss in older adults, normal age-related vision loss is
commonly distinguished from non-normal losses that result from pathology (Stuen
& Faye, 2003). Normal age-related vision loss is an unavoidable result of time on
physical mechanisms in the eye, which most people will experience if they live long
enough. These changes occur in the absence of any immediately apparent
structural or pathological defects that could otherwise explain loss of visual acuity
in old age (Haegerstrom-Portnoy & Morgan, 2007). The most commonly affected
structures of the eye are described below, along with the normal age-related
changes that affect them. Changes are discussed in terms of the optical or neural
factors in which they originate, as well as their influence on fall risk.
Optical Factors. The processing of visual information begins in the anterior
chamber of the eye, when light from the environment enters through two concave
transparent layers. Together, the functioning of these initial structures is important,
20
as declines in their integrity have the potential to disrupt the amount and quality of
light that is initially admitted into the eye. The first layer is called the conjunctiva—it
consists of a thin transparent mucous membrane that connects to the upper and
lower eyelids. In addition to providing a protective layer to the eye, the conjunctiva
contains cells that regulate the secretion of tears that lubricate parts of the eye that
are exposed to air (Millodot, 2004). Effects of aging on the conjunctiva include a
decrease in the number and functioning of these tear-secreting cells, resulting in
progressive reduction of in the amount of fluid and dry eyes (Bennett, Weissman, &
Remba, 2007). Aging may also cause distortions in the conjunctiva, which may
scatter and/or reduce the amount of light that enters the eye (Digiovanna, 2000).
Directly beneath the conjunctiva lies the cornea, the second transparent
structure that functions to bend, or refract light into the eyeball. Corneal refraction
of visual stimuli occurs along two meridians— one oriented horizontally, and the
other vertically across the front of the eye. In normal vision, curvatures of each
meridian work in conjunction to direct unblurred light deeper into the eyeball. With
age, however, corneal curvature may change, and astigmatism (refractive error
caused by irregularities in the shape of the cornea) may increase, especially along
the horizontal meridian (Haegerstrom-Portnoy & Morgan, 2007). Refractive errors
in the cornea may cause individuals to perceive visual fields less clearly. Age also
results in the gradual decrease in the transparency of the cornea, which can
contribute to color-blindness and loss of visual acuity in older people (Kahan &
Goodstadt, 2001).
21
The next visible layer in the anterior chamber of the eye, situated behind the
cornea is the iris— a circular membrane, made up of pigments (which determine
eye color), vessels (which nourish the structure), and tiny muscles that constrict
and dilate the pupil, through a process called miosis. The function of miosis in the
iris is to regulate the amount of light that is allowed into the posterior chamber of
the eye (Millodot, 2004). Muscles in the iris are controlled involuntarily by
parasympathetic fibers that respond to varying degrees of light in the environment.
As an involuntary system, the iris/pupil operates similarly to a thermostat, by
making automatic adjustments that allow optimal amounts of available light to
reach the retina (Gregory, 1997). As individuals age, the strength of the muscles in
the iris diminish, and blood vessels in the structure may become more rigid. As a
result of this change, the pupil becomes smaller at all levels of illumination
(Haegerstrom-Portnoy & Morgan, 2007), and adjustment to different lighting
conditions takes longer for older people. With respect to fall-risk, these changes
have important implications. First, older adults may take longer and have greater
difficulty adjusting, when changes in illumination occur rapidly (as when going from
a dark room to a lighter room in their homes). Furthermore, older adults may
sometimes require greater amounts of light to function adequately and safely in
their environments, compared to younger adults (Orr & Rogers, 2006).
The crystalline lens of the eye is situated directly behind the iris and
pupillary aperture, but in front of the vitreous chamber and the retina at the back of
the eye. The lens is structured in thin concentric layers like an onion (Gregory,
1997); it is composed of soft yellowish cortical material, a firm nucleus, and is
22
partially covered at the outermost anterior layer by a transparent epithelium
(Millodot, 2004). The purpose of the lens is to bring visual images from varying
distances into focus on photoreceptors in the retina at the rear of the eye. To
accomplish this, ciliary bodies composed of tiny muscles act on and regulate the
curvature of the lens by changing tension in the ligaments that suspend the lens
over the retina, called zonula. The involuntary process of changing focus is called
accommodation.
Normal changes to the lens and its supporting structures are called
presbyopia. In younger individuals with normal vision, the lens may remain
relatively elastic and its shape may be easily adjusted to bring near and far objects
into focus on the retina. By comparison, age-related changes to the lens and ciliary
bodies decrease the ability of the lens to accommodate different focal distances.
Most notably, the lens nucleus becomes thicker, resulting in sharper anterior and
posterior curvatures that reduce refractive power (Koretz, Cook, & Kaufman,
2002). Lens growth reduces the depth of the anterior chamber, affecting refraction
by the cornea, and making the eye slightly more myopic, or near sighted
(Haegerstrom-Portnoy & Morgan, 2007). Growth of the lens also makes it less
malleable so that it loses its ability to change shape. At the same time, small
declines in the strength of ciliary muscles decrease their ability to contract on the
lens, further reducing accommodative response (Strenk et al., 1999). Each of
these factors reduces the range of distances that older adults are able to see
clearly, and make it more difficult to change focal distances quickly, as may be
necessary to function safely in the environment.
23
Whereas presbyopia is not “curable”, difficulty with accommodating is
commonly addressed by way of corrective bifocal or trifocal lenses, or with
over-the-counter reading glasses. Multifocal lenses allow older adults to change
refractive powers quickly by looking through different regions of the lens. However,
in tradeoff, multifocal glasses may also increase fall-risk by interfering with the
wearer’s ability to accurately perceive environmental hazards, particularly while
walking outside their homes, or on steps (Lord, Dayhew & Howland, 2002).
Nevertheless, older persons with multifocal lenses can learn techniques for
reducing fall-risks that are related to eyewear. Those with outdated prescriptions
or no glasses at all can reduce fall-risk by wearing corrective lenses with the most
up-to-date prescriptions available to address their vision loss (Lord, 2006).
Similarly, research has shown that a large percentage of vision impairments could
be reduced or eliminated by updating vision prescriptions to correct refractive
errors (Resnikoff, Pacolini, Mariotti, & Pokharel, 2008).
Neural Factors. Normal age-related changes in the retina are also germane
to a discussion of vision and falls in older adults. The retina, located in the back of
the eye (right, in figure 1.1), is composed of thin layers of ganglion cells, integrating
neurons, and photoreceptor cells. To function effectively, the retina also depends
on other anatomical structures that provide support, including the retinal pigment
epithelium, a complex vascular system, and a transparent vitreous liquid that
maintains the shape of the eyeball. Retinal cell layers are oriented such that
photoreceptors are located at the back of the eye; thus, incoming light must pass
through layers of ganglion and integrating cells before photoreceptors are
24
activated. The layer of photoreceptors is made up of two kinds of cells. In the
macula, the area where best visual acuity is attained, cones are the most
prominent type of cell. Cone cells are specialized to perceive color, and in high
concentrations (as in the fovea of the macula), provide good visual acuity and color
perception, especially during daytime or in well-lit environments. Outside of the
macula, rod cells are more common. Rods are highly sensitive to light, but do not
perceive color; therefore, they are most important in situations where luminance is
low, as in a dark room, or at nighttime. The pigment epithelium is situated behind
photoreceptor cells—it functions to supply metabolites and other nutritional
support to photoreceptors via a complex vascular network (Millodot, 2004). Due to
its dark color, the pigment epithelium also reduces the amount of scattered light
that could blur the visual image (Coren & Ward, 1984). As a whole, the retina
processes light through chemical reactions in photoreceptors, which are
transmitted via integrating neurons, and ganglion cells, the axons of which
compose the optic nerve leading to the brain (Millodot, 2004). Indeed, structures of
the eye that are neuronal in function are often considered to be part of the brain
and central nervous system (Marmor, 1982).
With age, changes to the retina occur that are similar to changes in the
central nervous system—most notably, is the loss of cells. Within the visual cortex
of the brain, it is estimated that nearly 50% of neurons are lost by age 70 (Devaney
& Johnson, 1980). Similarly, the overall thickness of the retina decreases as
individuals age and the number of cells is reduced, particularly in the retinal
periphery where rods are most common. In the cells that remain, a degenerative
25
pigment called lipofuscin can accumulate and interfere with the functioning of
photoreceptors. Over time, rods and cones may become disorganized in structure
and arrangement. In addition, cells in the pigment epithelium may become
irregular in size and shape, and a gradual loss of capillary cells may interfere with
the capacity to maintain metabolic equilibrium in the retina. Although these normal
changes may have only limited direct effect on the functioning of the retina
(Marmor, 1982), age related reduction in the number of cells, and the declining
vasculature, along with a lifetime of exposure to light may make the system more
vulnerable to pathological conditions that can cause loss of vision.
Pathological Age-related Changes to Vision
To some degree, each of us can expect to experience some normal
age-related losses to vision. Aging is often accompanied by normal changes in the
eyes that can increase the risk of falls. In contrast, some vision loss is associated
with diseases that are not a normal part of aging, but that are more likely to affect
older adults. The most common pathological eye conditions, to which older adults
in particular are vulnerable, include age related macular degeneration (AMD),
diabetic retinopathy (DR), glaucoma, and cataracts (Leibowitz et al., 1980; Stuen
& Faye, 2003).
Age-related macular degeneration occurs as a result of histological
changes in the macula— the region of the retina which perceives fine details.
Macular degeneration is the leading cause of vision loss in the United States and
other industrialized countries, and is associated with age-related changes,
precursor lesions, cataracts, family history, alcohol consumption, smoking, and the
26
apolipoproteins A1 and B. Among Whites, AMD is the leading cause of visual
impairment, accounting for more than half (54%) of the cases of blindness, and
23% of cases of low vision (The Eye Diseases Prevalence Research Group,
2004a). Because AMD is more likely to occur in persons over age 75 (Ryskulova et
al., 2008), it is most common in white women, who tend to live longer than other
subgroups of the population. Persons with AMD experience a progressive loss
and/or distortion of vision in the central visual field. As a result, persons with AMD
may often be forced to use their peripheral vision when they move about in their
environments. Vision loss in the central field greatly increases the risk of falling, by
reducing the amount of information available about the environment or the area
immediately in front of the individual. Age-related macular degeneration has been
shown to interfere with mobility (Hassan, Lovie-Kitchin, & Woods, 2002), and to be
associated with increased fall-related injuries (Haronian, Wheeler, & Lee, 1993).
Diabetic retinopathy, another leading cause of low vision and blindness in
older adults, is a secondary condition related to systemic diabetes. It is
characterized by retinal vascular leakage, capillary nonperfusion, and endothelial
cell damage that results from autoimmune responses of white blood cells.
Stagnated white blood cells attach to the lining of vessels, causing endothelial cell
injury and premature cell death in vessels that nourish the retina (Adamis, 2002).
Diabetic retinopathy may become progressively worse and eventually lead to
blindness if diabetes is not controlled. Retinal scarring is responsible for the
unpredictable occurrence of distorted, “splotchy” vision, and blind spots across the
visual field, that characterize vision losses associated with DR. The incidence of
27
DR is greatest in minority populations, who may have the least access to treatment
and resources to aid in management of diabetes. Among Blacks, DR causes 7% of
blindness and 15% of low vision cases; among Hispanics 14% of blindness and
13% of low vision cases are attributed to the disease. For whites, slightly more
than 5% of blindness, and 3% of low vision cases are related to DR (The Eye
Diseases Prevalence Research Group, 2004a). Like AMD, DR contributes to the
incidence of falling by distorting or eliminating portions of the central and
peripheral visual fields. In addition, fall risk is also increased by diabetic
neuropathy in other sensory receptors, such as receptors in the skin on the
bottoms of the feet, which provide proprioceptive information about the body’s
movement and position in space and in environments (Ducic, Short, & Dellon,
2004).
Glaucoma (open angle) occurs when the ducts that regulate fluid in the eye
become blocked, resulting in a build up of intraocular pressure that damages the
optic nerve. Glaucoma results in permanent loss of vision in the peripheral fields,
night-blindness, and blind spots (scotoma). The destructive effects of glaucoma
occur very gradually, and may go unnoticed until the disease is in an advanced
state. For this reason, Stuen and Faye (2003) referred to glaucoma as the “sneak
thief of sight”, because its symptoms may not appear until after irreversible
damage has occurred. Risk factors for glaucoma include age (glaucoma is
significantly more common after age 45), family history, diabetes, and most
notably, race/ethnicity (Ryskulova et al., 2008). According to The Eye Diseases
Prevalence Research Group (2004a), the incidence of glaucoma among Blacks
28
and Hispanics as a cause of blindness are each at least four times that of Whites
(26%, 29%, and 6%, respectively). On average about 10% of older adults who are
visually impaired report glaucoma as their primary cause of visual impairment,
across races. With respect to falls, the presence of glaucoma is a key risk factor
because of losses in the peripheral fields. Peripheral losses may create mobility
difficulties for older adults who may not see environmental cues and hazards,
resulting in increased bumps and trips as they maneuver through the environment.
Usually, the central fields of persons with glaucoma remain in tact (Stuen & Faye,
2003), but progression of the disease may leave many older adults with “pinhole”
vision, and occlusions across the majority of the visual field. Additionally, the
destruction of peripheral retinal cells reduces the ability of the individual to
perceive details in dark environments, which could also increase fall risk.
Cataracts, by far, constitute the most likely age-related vision condition to
be experienced by individuals after age 65 (Stuen & Faye, 2003). A cataract is the
abnormal clouding of the lens, which causes vision to be blurred or foggy, and
decreases the amount of light that reaches the retina for processing. Risk factors
for cataracts include sex (women are significantly more likely than men to acquire
cataracts; Ryskulova et al., 2008), smoking status, heavy drinking, trauma,
exposure to light, and genetic factors. Confusion about whether cataracts
represent normal aging or whether the condition occurs due to pathology is
common because of the high prevalence of cataracts among older people. It
appears, however, that cataracts are a pathological manifestation of normal
age-related oxidative processes that occur rapidly in predisposed individuals
29
(Williams, 2006). Just as the purpose of the lens is to focus light onto the retina, it
is, therefore, exposed to great amounts of photo-oxidation, in which relatively
harmless oxygen molecules are converted into highly reactive free radicals, which
can damage the lens. Frequent exposure to photo-oxidation may lead to the build
up of proteins on the lens, and may cause other age-related damage to tissues in
the central nervous system, which (notably) may also be associated with
Alzheimer’s disease as well as stroke (Harding, 2002). Intraocular inflammation,
whether stemming from local trauma or systemic inflammatory conditions, also
appears to play a direct role in the formation of cataracts (Klein, Klein, Lee,
Knudtson, & Tsai, 2006).
Cataracts contribute to fall risk by interfering with vision acuity, and
increasing glare in bright settings. The treatment for cataracts has become a
common and relatively risk-free out-patient surgical procedure, involving the
extraction of the clouded lens and replacement with prosthetics. Individuals with
prosthetic lenses typically notice great improvement in their vision; however,
studies that have evaluated the effects of surgically corrected vision (such as in
cataract extraction) on fall risk have produced mixed results, with some indicating
reduced risk (Harwood et al., 2005) and others finding no difference or even
increased fall risk, as a result of cataract removal (Cumming et al., 2007).
Age-related Changes to Vision and Falls
People may fall at any age, but falls among older people especially are of
concern because of age-related changes in the body that make older individuals
more likely to experience a serious injury as a result of a fall. For example, declines
30
in the musculoskeletal system with age can interfere with mobility, and increase
the risk of falls and fall-related injuries (Evans, 1995; Frost, 1997; Lindle et al.,
1997). As described above, aging is also associated with changes in vision that
increase risk for falls (Hassan et al., 2002; Menz, Lord, St George, & Fitzpatrick,
2004), multiple falls (Coleman et al., 2004) and being injured in falls
(Anastasopoulos et al., 2006). Age-related changes in vision, whether normal or
the result of disease, have the capacity to affect at least four clinical measures of
vision status, which are typically used in research studies to quantify and
operationalize the concept of vision impairment. Decrements in visual acuity
(Coleman et al., 2004; Ivers, Cumming, Mitchell, & Attebo, 1998; Ivers, Norton,
Cumming, Butler, & Campbell, 2000; Klein, Moss, Klein, Lee, & Cruickshanks,
2003; Lord & Dayhew, 2001), contrast sensitivity (De Boer et al., 2004; Ivers et al.,
1998; Klein et al., 2003; Lord & Dayhew, 2001; Lord, Clark, & Webster, 1991),
depth perception (Felson et al., 1989; Lord & Dayhew, 2001; Lord et al., 2002),
and visual field (Ivers et al., 1998; Lord & Dayhew, 2001; Lord, 2006; Turano et al.,
2004) have each been identified and tested to determine their influence on fall risk.
These four factors represent mutually exclusive components of visual functioning,
which can affect the mobility and stability of older adults who experience
decrements in them.
In clinical exams, the acuteness or clearness of vision is referred to as
visual acuity. This measure of vision function is dependent in large part on the
quality of retinal focus within the eye (Millodot, 2004). Optical structures in the
anterior section of the eye admit and focus light onto the retina in the posterior
31
chamber (see figure 1.1). Therefore, decreased visual acuity could result from
decrements in functioning at multiple locations in the eye. Under normal
circumstances, visual acuity varies according to where on the retina light is
focused. The greatest visual acuity occurs in the macula, a small oval-shaped area
in the retina where photoreceptors are densely packed. Tests of central visual
acuity are based on the functioning of cells in the macular area. Decrements in
central visual acuity, whether from normal loss of photoreceptive cells or disease,
could cause blurriness, distortion, or occlusion of fall hazards (for example, throw
rugs, pets, furniture, or people) that may exist in the central field. By contrast,
peripheral visual acuity is a measure of clearness in regions of the retina, outside
the macula, where the dispersion of photoreceptors favors vision that is less
detailed, but more sensitive to light. Peripheral regions of the retina are important
for vision when illumination is low. Decrements in peripheral visual acuity could
lead to difficulties and increased fall risk in environments where lighting is poor (for
example, in dimly lit rooms, or outside at nighttime), or when fall hazards exist in
the peripheral eccentricities of the visual field. Research findings suggest that poor
visual acuity approximately doubles the risk of falls (Harwood, 2001). Thus, eye
diseases such as cataracts and other age-associated conditions which affect
vision acuity may increase fall risk by reducing the clarity of detail in a visual image,
and reducing the individual’s ability to discern fine details in the environment.
Individuals with good contrast sensitivity are able to detect fine luminance
differences in a visual image. In clinical settings, contrast sensitivity is usually
measured using gratings (repeated patterns) that vary by spatial frequency, or with
32
letter charts that have varying contrast (Harwood, 2001). In studies and clinical
evaluations, contrast sensitivity is operationalized as the minimum contrast level
required for detection of stimuli of varied frequencies. Low spatial frequencies are
optically sparse; whereas high spatial frequency stimuli are optically dense (Lord
et al., 1991). Age-related changes to contrast sensitivity seem to affect high spatial
frequency the most, and low-spatial frequency to a lesser degree. In persons with
no optical pathology, contrast sensitivity remains constant into middle age,
followed by declines around age 50 (Berbaken & Johnston, 1986). By age 80,
persons with normal vision require almost twice as much contrast as their
middle-aged counterparts to detect differences.
Many studies have shown an association between contrast sensitivity and
fall risk (Lord & Dayhew, 2001), and risk for multiple falls (Ivers et al., 1998). In
practical terms, older adults who have impaired contrast sensitivity may have
difficulty seeing edges that mark important changes in surfaces. In real-world
terms, poor contrast sensitivity can negatively influence the individual’s ability to
detect ground-level hazards, such as steps, curbs, pavement cracks or sidewalk
misalignments that normally may be evidenced by differences in light dispersal
(Lord, 2006). Reduced contrast sensitivity can make the edges of curbs nearly
invisible, and may make clutter or obstacles very difficult to detect by their blending
into the background. As such, intact contrast sensitivity is important for older adults
in avoiding ground-level trip, slip, and misstep hazards that are commonly found in
homes and public settings.
33
Depth perception refers to the ability to judge the distance of an object from
the viewer, or the distance between two objects. Accurate depth perception is
made possible by the disparity of images cast on the two retinas (called binocular
disparity). Distances are also inferred by viewers from environmental cues (for
example, the relative sizes of objects in the environment, and/or patterns of light
and shade). Older people who have good vision in one eye and moderate or poor
vision in the other are said to have poor stereo-acuity, which may cause greater
difficulty with accurately perceiving depth. Stereo-acuity is important for judging
distances from hazardous obstacles and for making the visual world seem more
3-dimensional. According to Lord (2006), impaired depth perception is a significant
risk factor for falls, multiple falls, and fractures. Findings by Felson et al. (1989),
and Lord and Dayhew (2001) suggested that fall rates of persons with poor vision
in just one eye are equal to those of persons with poor vision in both eyes, whereas
persons with good vision in both eyes had the lowest rates of falls. These findings
have important implications, with respect to measurement definitions of visual
impairment—if impairment is defined as acuity in the best eye, then the increased
risk associated with loss in one eye could be ignored (Felson et al., 1989), resulting
in under-estimates about the effects of vision loss.
Finally, visual field refers to the area that can be seen without moving the
head or eyes, and is measured either binocularly or monocularly. For persons with
normal vision, the visual field in each eye extends 60 degrees nasally (toward the
nose) and 100 degrees temporally (away from the nose), as well as 60 degrees
above and 75 degrees below the horizontal meridian. Diseases and disorders of
34
the retina, the optic nerve, and/or the brain may result in losses in vision
above/below the horizon, in the periphery, or in the central field. Whereas some
eye diseases such as glaucoma may affect peripheral visual fields, other eye
diseases like AMD or DR may cause blind spots in the central field or throughout
the visual field. Not surprisingly, decrements in visual fields have been found to be
an independent risk factor for falls (Ivers et al., 1998).
The purpose of this chapter was to describe important concepts related to
aging, age-related vision loss, and falls. In the studies that follow, self-reported
vision loss is used as a key explanatory variable, with respect to health conditions
and falls among older people. As noted above, self-reported vision has the
potential to be a powerful predictor of falls, because self-reports are often a good
reflection of how older individuals live and function with their recently acquired
visual impairments.
35
CHAPTER II: VISION LOSS AND HEALTH DIMENSIONS
A. Introduction
Visual impairments in late life have been widely associated with outcomes
pertaining to a range of health-dimensions that influence and reflect the well-being
of older adults. In addition to the association between vision loss and falls, vision
impairment has been linked to reduced functional capacity and physical activity
levels (Crews & Campbell, 2004; Haymes, Johnston, & Heyes, 2002; West et al.,
2002), reduced psychosocial wellbeing (Chia, Mitchell, Rochtchina, Foran, &
Wang, 2003), morbidity (Crews & Campbell, 2004), self-rated health (Wang,
Mitchell, & Smith, 2000), and even mortality (Laforge, Spector, & Sternberg, 1992;
McCarty, Nanjan, & Taylor, 2001; Wang, Mitchell, Simpson, Cumming, & Smith,
2001). A large number of studies have examined relationships between vision loss
and various risk-factors, such as obesity (Capella-McDonnall, 2007), inflammation
(Adamis, 2002; de Maat, Pietersma, Kofflard, Sluiter, & Kluft, 1996; Joussen et al.,
2004; Seddon, Gensler, Milton, Klein, & Rifai, 2004), and cholesterol levels
(Curcio, Millican, Bailey, & Kruth, 2001). In addition, the negative consequences of
visual impairment have also been studied with respect to secondary health
outcomes such as hospital and emergency room utilization (Jacobs,
Hammerman-Rozenberg, Maaravi, Cohen, & Stessman, 2005), and hip fractures
(Felson et al., 1989). Thus, associations between vision loss and sundry health
outcomes are not uncommon, and seem to reflect the importance of vision as an
indicator of the general health status of adults as they age.

36
B. The Disability Process
One potential framework for thinking about how vision loss is associated
with health is the disability process model, originally provided as a sociological
model of disability by Nagi (1965; 1976) and later was conceptualized further by
Verbrugge and Jette (1994), Crimmins (2004), and others. The model consists of
four main stages that progress from (1) risk factors, through (2) pathology and
impairments to result in (3) loss of functioning, and eventually (4) disability (see
figure 2.1).

Figure 2.1 The Disability Process: Theoretical Pathway Leading from Risk Factors
to Disability

Risk Factors
In developing Nagi’s (1965; 1976) model, Verbrugge and Jette (1994)
defined risk factors as characteristics of an individual (including biological
characteristics) that can affect the presence and severity of pathology and
impairment. Biological indicators include cholesterol levels, body mass index
(BMI), markers of inflammation (such as C-reactive protein; CRP), and indicators
37
of insulin regulation (such as glycated hemoglobin). Accordingly, these biological
factors often precede diseases and conditions that lead to functional impairment
(including visual impairment), disability, and mortality. For example, glycated
hemoglobin— a measure of the average amount of hemoglobin bound to glucose
over a prolonged period— could serve as a preclinical indicator of diabetes.
Patients with diabetic retinopathy (DR), one of the most common causes of vision
loss in older adults, may experience visual impairments which could inhibit their
ability to perform activities of daily living (ADLs). With consideration to this
trajectory—in which high-risk levels of certain biomarkers precede disease states
and poor physical functioning— older adults, especially those with greater genetic
or behavioral risk, could monitor levels as a preventative measure against disease
and related disability.
Several biomarkers have been associated with higher rates of mortality
(Harris et al., 1999), reduced cognitive (Wilson, Finch, & Cohen, 2002) and
physical (Cohen, Harris, & Pieper, 2003) functioning, and heart disease (Cesari et
al., 2003), and are generally viewed as good indicators of the health status of older
people (Crimmins et al., 2005), without regard to vision status. In a study that
directly compared biological risks of older persons with clinical blindness and low
vision with older adults with normal vision, Steinman and Vasunilashorn (in press)
compared at-risk levels for nine biological markers with underlying clinical
relevance in predicting clinical manifestation of conditions commonly experienced
by older adults and associated with poorer physiological functioning. In this study,
older adults who were blind were more likely to have high-risk levels of low density
38
lipoprotein (LDL) cholesterol, homocysteine, and to be underweight. Thus it was
noted that differences between vision groups were likely in part, diet related, and
could potentially be addressed by way of nutrition and diet programs aimed toward
education of older people who are visually impaired.
Pathology/Impairment
Pathology is described by Verbrugge and Jette (1994) as physiological
abnormalities that are labeled as disease, and presumably arise from biological as
well as behavioral risk factors; whereas impairments represent dysfunctions in
body systems. In the current study, these categories have been combined. The
incidence of chronic conditions increases dramatically with age for older adults in
general. Previous studies have reported that by age 65, the vast majority of
individuals have multiple chronic conditions (Wolff, Starfield, & Anderson, 2002),
many of which are associated with vision loss. Furthermore, older individuals who
are visually impaired have more health problems than their peers with normal
vision, including lower bone mineral density, higher rates of osteoporosis,
depression, and diabetes (Ray & Wolf, 2008). However, the extent and direction of
the association between vision loss and disease is not always clear. Of the studies
that have examined relationships between biomarkers and vision loss in older
people, most have linked the two via the shared risk profiles of specific eye
diseases that occur most frequently in old age and age-related health conditions
and systemic diseases outside of the vision system. For example, some types of
age-related macular degeneration (AMD) have been positively associated with
high-risk levels of diastolic blood pressure, LDL cholesterol, and total cholesterol
39
(Hyman, Schachat, He, & Leske, 2000). Furthermore, the hypothesis that AMD is
mediated by inflammatory reactions similar to those found in atherosclerosis is
supported by studies that show the presence of similar inflammatory markers
within retinas affected by AMD. Several studies have reported elevated levels of
biomarkers that predict cardiovascular disease (such as fibrinogen, homocysteine,
and CRP) in individuals who develop AMD (Smith, Mitchell, Leeder, & Wang,
1998; Vine, Stader, Branham, Musch, & Swaroop, 2005). Similarly, levels of CRP
have been found to positively correlate with vessel damage in individuals with
Type 1 diabetes (Schalkwijk et al., 1999). Inflammatory markers are present in the
retinal vasculature of individuals with DR, as vessel dilation decreases the flow of
plasma proteins, resulting in inactive/acellular capillaries. The accumulation of
damaging white blood cells in the vessels and neural retina can result in damage to
endothelial cells and premature cell death that could result in blindness (Adamis,
2002). Thus, vision impairment may often be caused by systemic diseases that
have origins in other body systems.
Functional Limitations
Limitations of function are difficulties performing fundamental physical and
mental activities used in daily life, and may include overall mobility impairment,
losses of discrete motions and strengths (such as the ability to climb stairs or grip
strength), trouble hearing, trouble communicating, and trouble seeing (Verbrugge
& Jette, 1994). Functional activities may be thought of as component tasks of most
daily living activities which are more complex. For example, in order to successfully
shop for groceries, one must have enough strength and agility in his/her limbs to
40
walk a quarter mile to and from one’s car in the parking lot; to stoop to reach low
items on the shelf; to lift heavier items into the cart; and to stand long periods while
waiting in the checkout line. Each physical functioning task, when combined with
others make up more complex tasks, within specific domains of day to day living.
Vision impairments have been widely associated with declines in physical
functioning. In research by Salive et al. (1994), older adults from several
communities were assessed to reveal cross sectional and longitudinal
associations between visual function and physical function. Compared to
participants with low levels of visual impairment, those with severe and moderate
impairments had more self-reported difficulties with mobility. With respect to
observed physical performance measures, participants with severe and moderate
visual impairments had the greatest difficulty performing functional tasks. Results
of longitudinal analyses suggested that individuals with severe visual impairments
were three and a half times more likely to develop new mobility limitations; and
about 60% less likely to show improvement in mobility across multiple waves, than
individuals with low levels of visual impairment.
In a study by West et al. (2002), four factors representing different aspects
of visual functioning (including spatial vision, binocularity, field integrity, and
adaptation) were used to predict dependent measures of physical functioning
(including tests of mobility, walking, standing from a chair, and tandem balance).
When each vision factor was included individually in models, spatial vision
predicted mobility, walking, and chair stand limitations; binocularity predicted
failure on the chair stand and tandem balance tests; field integrity predicted
41
limitations in all four functional tests; and adaptation predicted mobility limitations
and failure in the tandem balance test. When all four factors were included in a
model together, binocularity predicted only failed chair stand; field integrity
predicted mobility limitations, failed walking test, and failed tandem test; and
adaptation predicted only mobility limitations. Spatial vision was not a significant
predictor of any functional test when the other aspects of vision were controlled.
Thus, functional limitations that result from vision impairments would seem to be
influenced by specific losses within the vision system.
Disability
Finally, disability is defined as having difficulty performing activities in
specific age-appropriate domains of life, ranging from personal care activities to
household management activities. In the case of older adults, domains have
traditionally consisted of ADLs, such as eating, toileting, dressing and bathing; and
instrumental activities of daily living (IADLs), including managing money, preparing
meals, doing household chores, and shopping for meals (Verbrugge & Jette,
1994). Perhaps not surprisingly, limitations in daily living activities are more
common among older adults with visual impairments. Due to the nature of daily
living activities, one of the more troublesome effects of vision loss in old age is its
negative influence on the ability to perform activities that are necessary to live
independently. For example, research by Cavenaugh and Steinman (2003) used
data from the Medicare Current Beneficiaries Survey to show that nearly half
(46%) of Medicare beneficiaries who are blind or visually impaired had difficulty
performing ADLs and IADLs, in comparison to 21% of beneficiaries without vision
42
impairments. Similarly, Salive et al. (1994) demonstrated that persons with
“severe” visual impairments were nearly five times as likely as participants with
“low levels” of impairment to report difficulties with ADLs. Furthermore, participants
with severe impairments were about 3 times more likely to develop new ADL
limitations later, compared to those with low-level impairments. Branch, Horowitz,
and Carr (1989) analyzed self-reported measures of visual decline to compare
daily living outcomes of community dwelling older adults. Those with visual
impairments showed decreased ability to perform daily living activities, and to
participate in social functions, compared to older adults with normal vision. In
addition, Branch et al. demonstrated that some visually impaired older adults may
not seek appropriate services to ameliorate their vision-related problems, despite
experiencing greater difficulties with daily living activities. Finally, a study of
institutionalized older adults conducted by Horowitz (1994) demonstrated that both
moderately and severely visually impaired nursing home residents were
significantly more dependent while performing daily living tasks related to
transferring, bathing, and eating. Even after controlling for multiple chronic
impairments, vision loss was implicated as a sizable contributor to the losses
experienced by institutionalized older adults.
Taken together, limitations with daily living activities such as those
described above may often precede general declines in overall physical activity
that could influence the health and well-being of older persons with visual
impairments (Ray & Wolf, 2008). Furthermore, older individuals who cannot
adequately perform specific daily-living activities such as preparing meals for
43
themselves, or taking medications appropriately could be at greater risk of
maintaining poorer health practices, such as eating inadequate or imbalanced
diets (Crews & Campbell, 2001), or mismanaging their single or co-occurring
medical conditions (Claesson, Morrison, Wertheimer, & Berger, 1999). Such
practices have the potential to negatively affect biological and other health
indicators that underlie further disability (Crimmins et al., 2005; Steinman &
Vasunilashorn, in press).
Feedback Loop
Under disease conditions, high-risk levels of some biomarkers would be
expected, in part, due to underlying systemic pathologies that result in late-life
visual impairment (such as vascular disease and diabetes). However, it is also
possible that high-risk levels of some biomarkers, disease, and decreased
functioning could result secondarily from behavioral changes that often occur after
older individuals become visually impaired, due to vision-related disability.
Verbrugge and Jette (1994) described a system by which stages of the disability
process could feed back to result in interrelated declines in multiple systems. For
example, consider how obesity (a biological risk factor) may lead to disregulated
insulin production or diabetes (pathology/impairment). Diabetes is a systemic
disease which often causes damage to cells that innervate systems throughout the
body, including the retinas (visual impairment). Individuals with DR have reduced
vision sporadically throughout their visual fields, making it difficult to complete
visual tasks (functional loss). Finally, due to their vision loss, individuals may lose
confidence in their ability to cook and clean for themselves and to perform other
44
daily living activities (disability). Because individuals with vision impairments are
likely to reduce their daily activity levels, new biological risks could arise that cause
impairment and pathology in different systems, such as the cardiovascular or
musculoskeletal systems. After multiple iterations, feedback loops in the disability
process could result in rapid declines, which correspond to the common
“downward spiral” that is often associated with frailty in older people. Therefore,
one advantage to considering vision impairment within the context of the disability
process model is that it provides a framework by which interventions could be
applied to prevent or reduce the effects of vision loss in the development of frailty.
C. Purpose of Study
The overarching hypothesis of this study is that visually impaired older
adults have more negative health outcomes due, in part, to the relationship
between visual functional capacity and disability. In general, visually impaired
people are more likely to limit activity-levels, eat poorer diets, and have increased
stress due to their declining visual functioning (Crews & Campbell, 2004). These
lifestyle differences would likely lead to health outcomes that are more negative
compared to peers with normal vision. Thus, the purpose of this initial study was to
compare the health of older persons with and without self-reported visual
impairments along the health dimensions described above. There are no known
studies that have taken a comprehensive view of how vision loss relates to health;
or that have attempted to provided a structure by which vision impairment could
influence (and be influenced by) other health outcomes. Such comparisons are
important, because of the potential to identify points in the disability process where
45
older persons with visual impairments differ from those with normal vision, and to
develop interventions at those points to prevent feedback to earlier stages and/or
progression to later stages in the disability process. In addition, understanding how
older people with visual impairments differ from older adults with normal vision is
an important starting point for understanding how health disparities between vision
groups could precipitate falls.
D. Methods
Data
Variables corresponding to each health dimension in Nagi’s (1965; 1976)
disability process were selected for analyses from five cycles (1999 – 2008) of the
National Health and Nutrition Examination Survey (NHANES IV). Within NHANES,
cross sectional data are collected biennially to form a continuous data set, which
when weighted is representative of the noninstitutionalized American population of
persons ages 2 and up. The study uses a complex sampling design in which
counties are the primary sampling unit; NHANES over-samples racial minorities
and persons who are 60 years and older, in order to provide statistical power to
studies of persons in these subgroups. When cases from five cycles were
combined, a total of 51,623 cases were available to be analyzed. An exclusion
criterion of age greater than or equal to 65 was applied, resulting in a total
weighted sample size of 6,693 participants.
The majority of analyses conducted for this study were based on the five
most recent cycles of NHANES; however, when variables within the survey were
discontinued, had their wording changed, or were included in cycles any time after
46
the 1999 cycle, variables were either appropriately combined, or available
subsamples of the entire sample were used. For example, some variables, such as
biomarkers, when taken following a fasting period, were based on subsamples of
the entire sample. Therefore, sample sizes often varied between analyses,
depending on the availability of data and whether items had changed from year to
year.
The NHANES is composed of four main sections from which variables for
the current study were selected. In the demographics section, information was
provided about age, sex, race, educational attainment, and marital status at
screening. An examination section provided information collected through physical
exams and dietary interview components—this section included body
measurements such as BMI, and blood pressure measures. A series of laboratory
files contained results from analyses of blood, and urine specimens—and provided
information about health risk factors such as total cholesterol levels, glycated
hemoglobin, fasting triglycerides, CRP, and serum homocysteine. Finally, a
questionnaire section gathered by face-to-face interviews, provided information
pertaining to self-reported health and functioning, including visual functioning,
physical health conditions, mobility and physical functioning, and disability with
respect to various daily-living activities.
Independent variables
Sociodemographic Measures. In each analysis, five sociodemographic
characteristics of participants were controlled including age, gender, race,
marital/partner status, and education. Age was analyzed as a continuous variable.
47
Gender, race, and marital status were coded into indicator variables with “male,”
“white” and “not married/partnered” as reference categories, respectively.
Education was coded as an ordered categorical variable comprised of three
levels—less than a high school diploma, graduated from high school or equivalent
(e.g., completion of GED requirements), and greater than a high school education.
Self-Reported Vision Measures. In addition to the five sociodemographic
variables, a self-reported measure of visual functioning was included as the key
explanatory independent variable. During the interview section of NHANES,
respondents were asked to rate their present eyesight, with glasses or contact
lenses, if she/he regularly wore them. Respondents could rate their vision as
excellent, good, fair, poor, or very poor. Participants could also refuse to answer
the item, or say that they did not know about the general condition of her/his
eyesight. This item was recoded into three indicator variables—respondents who
said that their vision was poor or very poor made up one group; whereas those
who rated their eyesight as fair made up a second group— participants who rated
their vision as good or excellent made up a third group, which served as a
reference category in analyses.
Model Covariates. In addition to sociodemographic and self-reported vision
variables, additional covariates were included in models, based on the position of
dependent variables within Nagi’s theoretical pathway to disability. Variables
contained in health dimensions to the left of dependent variables (see figure 2.1)
were controlled to assess the relative role of preceding health dimensions, in
addition to the variables being tested. For example, in analyses of disability
48
variables, two models were tested. First, each disability variable was assessed
controlling only for sociodemographic control variables. The second model
controlled all previous variables including biological risk variables, health
conditions, and mobility/functioning. By contrast, in analyses of biological risk
factors, only sociodemographic variables and self-reported vision were controlled
in analyses, since risk factors are the first health dimension in the model of the
disability process.
Dependent variables
Dependent variables were chosen from the four sections of NHANES
based on whether they fit into the categories of health dimensions described by
Nagi (1965; 1976). Dependent variables were analyzed to assess whether
participants with fair or poor self-reported visual functioning experienced greater
odds of specific high-risk levels of biological indicators, pathology/impairments,
reduced mobility/functioning, and disability compared to persons with good or
better self-reported visual functioning.
Biological Indicators. In addition to controlling sociodemographic risk
variables in all models, dichotomous variables were computed based on clinically
defined and well-established at-risk levels for each biological marker. The ten
biomarkers included systolic and diastolic blood pressure, high-density lipoprotein
(HDL) cholesterol, LDL cholesterol, total cholesterol, glycated hemoglobin, two
extremes of BMI (underweight and obese), fasting triglycerides, CRP, and plasma
homocysteine. Cut points for some biomarkers remain debated relative to their use
in the older adult population (e.g., proposed BMI cut-points used for older adults).
49
The U-shaped relationship of BMI to mortality across age, which has been
reported (Adams et al., 2006) informed the cutoffs for BMI used in the current
study. The high and low biomarker cut-points used here have also been used in
other studies of older adults (Crimmins et al., 2005) and have been associated with
multiple outcomes, including disability and mortality (Adams et al., 2006; Alley &
Chang, 2007). Defined cutoffs for at-risk levels are shown in Table 2.2.
Pathology/Impairments. A combination of NHANES items from various
sections of the questionnaire were used to derive 7 dependent variables
representing pathological conditions commonly experienced by older people.
Within the medical conditions section of the NHANES questionnaire, participants
were asked whether a doctor had ever told them that they had arthritis, congestive
heart failure, coronary heart disease, angina pectoris, heart attack, stroke,
emphysema, chronic bronchitis, or cancer. In response to these items, participants
could state that they had or had not been told by a doctor that they had each
condition. A variable representing heart problems was created, and coded in the
positive direction (has heart problem = 1), if respondents confirmed that a doctor
had told them they had any one of the heart problems probed. Similarly, a variable
representing respiratory problems was created, and coded positively if the
respondent said that he/she had been told by a doctor that they had emphysema
or chronic bronchitis.
Hearing was assessed in the interview section of NHANES with an item that
asked participants to pick the statement that best reflected their ability to hear.
Respondents could state that their hearing was good, that they had a little trouble
50
hearing, a lot of trouble, or that they were deaf. This variable was recoded into a
dichotomous variable that compared participants who reported a lot of trouble
hearing or deafness against a reference group of participants who reported little
trouble, or good hearing.
Finally, within the diabetes section of the questionnaire, participants were
asked whether a doctor had ever told them that they have diabetes or sugar
diabetes. Participants could state that they had or had not been told by a doctor
that they had diabetes, or that their status with respect to blood sugar was
bordering on high-risk levels for diabetes. Those who stated that they were
bordering on high-risk levels for diabetes were categorized with those who stated
that they had been told that they had diabetes. Participants who refused to answer,
or who stated that they did not know were coded as missing data.
Functional limitations. Dependent variables representing functional
limitations were quantified using measures of physical mobility and functioning
found in NHANES. Participants were asked whether, due to a health problem, they
had difficulty performing 10 functional activities including walking a quarter mile;
walking up ten steps; stooping, crouching or kneeling; lifting or carrying; walking
between rooms on the same floor; standing up from an armless chair; standing for
about 2 hours; sitting for long periods; reaching up over head; and grasping small
objects. Respondents could report having no difficulty, some difficulty, much
difficulty, that they were unable to do the activity, or that they did not do the activity.
These functional items were recoded into polytomous ordered categorical
variables, with respondents reporting some or much difficulty grouped together,
51
and those reporting inability grouped together. Participants who reported no
difficulty were included as a third group to serve as a reference category in
multinomial analyses. Those who said that they did not do the activities were
coded as missing data.
Disability. Nine dependent variables representing disability were selected
from the physical functioning section of the NHANES questionnaire. Items were
classified within this particular dimension of health if they assessed difficulty
performing specific activities in age-appropriate domains of life. In the case of older
adults, age-appropriate activities typically include traditional ADLs and IADLs;
however, in this study, other activities that reflected the individual’s ability to
participate in social activities, and to go out into social situations were also
analyzed. Participants were asked whether, due to a health problem, they had
difficulty managing money; doing house chores; preparing meals; getting in and
out of bed; using fork, knife, or drinking from a cup; dressing themselves; going out
to movies; attending a social event; and performing leisure activities at home. To
each of these items, respondents could report having no difficulty, some difficulty,
much difficulty, that they were unable to do the activity, or that they did not do the
activity. Disability items were recoded into polytomous ordered categorical
variables, with respondents who reported some or much difficulty grouped
together, and those reporting inability grouped together. Participants who reported
no difficulty were included to serve as a reference category in multinomial analyses.
Those who said that they did not do the activities were coded as missing data.

52
Analyses
Analyses for this study were conducted using Statistical Analysis Software
(SAS), version 9.1 for Windows. All analyses were modified in order to account for
the survey design and selection effects of the NHANES complex sampling
method. For example, in main analyses, the proc surveylogistic command was
used in lieu of proc logistic, because the former command employs stratification
and clustering variables available in NHANES to account for unequal probabilities
of being selected to participate in the study, based on sampling design.
Basic descriptive statistics were calculated for selected sociodemographic
variables, vision acuity, and outcome variables chosen to represent each of the
four health dimensions along the disability process. For all variables, including
outcome variables, percentages representing the proportion of individuals who
were at or above high-risk thresholds, who experienced disease outcomes, or who
experienced difficulty or dependence with functional and daily living activities were
computed for each vision category. A Wald Χ
2
statistic was computed, and
probabilities were reported to determine statistically significant differences in
proportions between vision groups.
Main analyses consisted of a series of binary and multinomial logistic
regressions. The purpose of these analyses was to estimate odds ratios (ORs) and
95% confidence intervals (95% CI) to reflect the likelihood of negative health
outcomes among persons with self-reported poor and fair vision relative to
persons with good or better self-reported vision. Covariates of each model were
selected based on the position of dependent variables in Nagi’s (1965; 1976)
53
model of the disability process. For each dependent variable, two models were
assessed. The first model controlled for sociodemographics and self-reported
vision status; whereas subsequent models also controlled covariates considered
to be within health dimensions to the left of the health dimension category being
tested. For example, the odds of self-reporting a diagnosed health condition was
tested with control for covariates related to biological risk factors. This hierarchical
covariate control structure was used to assess the relative impact of preceding
health dimensions on the dependent variables, compared to crude analyses,
controlling only for sociodemographic covariates.
E. Results
Table 2.1 displays unadjusted percentages and means of selected
sociodemographic variables by self-reported vision status. Overall, more than
three-quarters of the sample reported having good or better vision; whereas 18%
reported fair vision, and 7% self-reported their vision as poor. The mean age of the
total sample was 74.7 years (Standard Error of the Mean; SE = 0.13). On average,
participants with who reported poor vision were older (mean; x = 77.3; SE = 0.33)
than those with fair vision ( x = 75.8; SE = 0.21), and those with good vision ( x =
74.2; SE = 0.14). The majority (58%) of the sample was female, and proportions of
female participants were greater in poor (66%) and fair (62%) vision categories,
compared to participants with good self-reported vision (58%). The vast majority of
the sample was White (83%), whereas 8% were Black, 6% were non-white
Hispanic, and 3% reported some other race. Disproportionately low percentages
of Blacks and Hispanics reported having good vision (7% and 5%, respectively).
54

Table 2.1: Unadjusted Descriptive Statistics for Sociodemographic Variables by
Self-Reported Vision Status (NHANES IV, 1999 – 2008, Weighted)
Good Fair Poor Total Wald χ
2
χ
2
prob
(n = 4916) (n = 1169) (n= 465) (N = 6550)
% total 75.1% 17.8% 7.1% 100.0%
Age (mean yrs.) 74.2 75.8 77.3 74.7 435.3 <0.0001
Female 57.7% 61.7% 66.3% 58.2% 7.1 0.04
Race
White 85.6% 76.4% 75.6% 83.0% 49.5 <0.0001
Hispanic 5.0% 8.9% 9.7% 6.1% 15.5 0.0009
Black 6.8% 11.3% 10.4% 7.9% 33.9 <0.0001
Other 2.7% 3.4% 4.3% 3.0% 2.06 0.36
Education
< High school 25.8% 43.4% 55.5% 31.0% 145.8 <0.0001
High school 30.7% 27.9% 22.5% 29.6%
> High school 43.5% 28.7% 22.0% 39.4%
Married/Partnered 59.0% 50.3% 39.0% 55.8% 41.7 <0.0001

The vast majority (69%) of the sample reported having attained a high
school diploma or more education, although only 45% of those with poor vision,
and 57% with fair vision had attained as much education, compared to 74% with
good vision. Finally, the majority of participants (56%) reported being married or
partnered. Among the vision categories, those with poor vision were the least likely
to report being married or partnered (39%); only 50% of participants with fair vision
reported being married or partnered, compared to 59% with good vision.

55

Table 2.2 displays results of tests that were conducted to assess whether
proportions of participants above biological risk cut points differed by self-reported
vision status group. With few exceptions, the proportions of persons with fair vision
or poor vision who had at-risk levels of biomarkers were greater than those with
self-reported good or better vision. Wald χ
2
probability tests indicated significant
differences in four biomarkers in the proportions that were at high risk between
vision groups. Whereas 41.4% of participants with good or better vision had
high-risk levels of systolic blood pressure, 47.3% of persons with fair vision, and
50.7% of participants with poor vision had high risk levels. Larger proportions of
Table 2.2: Unadjusted Descriptive Statistics for Biological Risk Indicators by
Self-Reported Vision Status (NHANES IV, 1999 – 2008, Weighted)
Good Fair Poor Total Wald
χ
2

χ
2
prob
(n = 4916) (n = 1169) (n= 465) (N = 6550)
Indicator At Risk Cut Point Percent At-Risk
Diastolic blood pressure >90 mmHg 4.0% 4.8% 4.1% 4.1% 0.70 0.7108
Systolic blood pressure >140 mmHg 41.4% 47.3% 50.7% 43.3% 13.89 0.0017
HDL cholesterol <40 mg/dl 16.9% 21.0% 21.5% 18.0% 11.05 0.0058
Fasting LDL cholesterol >160 mg/dl 12.6% 9.7% 11.0% 12.3% 2.64 0.2808
High total cholesterol <240 mg/dl 19.5% 17.1% 19.8% 19.2% 1.53 0.4767
Low total cholesterol <160 mg/dl 12.8% 14.3% 14.5% 13.1% 1.05 0.5984
Glycated hemoglobin >6.4% 13.2% 17.6% 24.7% 14.8% 23.08 <0.0001
Obese (BMI) >30 kg/m
2
29.3% 29.7% 29.3% 29.3% 0.04 0.9783
Under weight (BMI) <18.5 kg/m
2
1.7% 1.4% 1.7% 1.6% 0.31 0.8573
Fasting triglycerides >200 mg/dl 18.9% 19.4% 15.6% 18..9% 1.29 0.5321
C-reactive protein >4.0 mg/l 32.8% 35.7% 37.3% 33.7% 4.00 0.1465
Plasma homocysteine >15 μmol/l 10.5% 18.2% 24.8% 12.8% 48.69
<0.0001
LV= low vision; BMI = body mass index; HDL = high-density lipoproteins; LDL = low-density lipoproteins
56
participants with fair and poor vision were also at higher risk for HDL cholesterol
than those with good vision (21.0%, 21.5% and 16.9%, respectively); Glycated
hemoglobin (17.6%, 24.7% and 13.2%, respectively); and plasma homoscysteine
(18.2%, 24.8% and 10.5%, respectively).
Table 2.3 presents results of binary logistic regressions conducted to
determine the relative odds that persons in each self-reported vision category
would be at high-risk for each of the 10 biological risk indicators. Having fair
self-reported vision was statistically associated with HDL cholesterol, glycated
hemoglobin, and plasma homocysteine. After controlling for sociodemographic
variables, persons with fair vision were about 33% more likely to have high-risk
levels of HDL cholesterol (OR = 1.33; 95% CI 1.11-1.60), and about 30% more
likely to have high-risk levels of glycated hemoglobin (OR = 1.30; 95% CI
1.03-1.64) compared to people with good or better vision; older persons with fair
vision were about 56% more likely to have high-risk levels of plasma homocysteine
(OR = 1.56; 95% CI 1.20 – 2.04). Similarly, self-reported poor vision among older
persons was statistically associated with HDL cholesterol, glycated hemoglobin,
and plasma homocysteine. After controlling for sociodemographic variables,
persons with poor vision were about 42% more likely to have high-risk levels of
HDL cholesterol (OR = 1.42; 95% CI 1.02-1.99); 91% more likely to have high-risk
levels of glycated hemoglobin (OR = 1.91; 95% CI 1.35-2.70); and more than twice
as likely to have high-risk levels of plasma homocysteine (OR = 2.06; 95% CI 1.43
– 2.96) compared to people with good or better vision.

57

Table 2.4 displays unadjusted percentages of persons in each self-reported
vision category who reported being diagnosed with any of seven pathological
conditions commonly experienced by older adults. With the exception of cancer, a
statistically greater proportion of older persons with fair and poor vision
experienced diseases compared to persons with good or better vision. For
instance, whereas 51% of older adults with good vision reported a diagnosis of
arthritis, more than 58% of participants with fair vision, and more than 63% of
Table 2.3: Binary Logistic Regression Models Testing the Effect of
Self-Reported Vision Status on Biological Risk Indicators (NHANES IV, 1999 –
2008, Weighted)
Fair Vision

Poor Vision
Indicator

At-Risk Cut
Point
N OR 95% CI OR 95% CI
Diastolic blood pressure >90 mmHg 5491 1.27 (0.81 – 1.99) 1.16 (0.63 – 2.15)
Systolic blood pressure >140 mmHg 5491 1.13 (0.96 – 1.34) 1.17 (0.88 – 1.55)
HDL cholesterol <40 mg/dl 5422 1.33 (1.11 – 1.60) 1.42 (1.02 – 1.99)
Fasting LDL cholesterol >160 mg/dl 1761 0.71 (0.46 – 1.09) 0.91 (0.45- 1.83)
High total cholesterol >240 mg/dl 4138 0.83 (0.62 – 1.10) 0.99 (0.70 – 1.41)
Low total cholesterol <160 mg/dl 4138 1.14 (0.87 - 1.50) 1.13 (0.64 – 2.00)
Glycated hemoglobin >6.4% 5511 1.30 (1.03 – 1.64) 1.91 (1.35 – 2.70)
Obese (BMI) >30 kg/m
2
5499 1.05 (0.88 – 1.25) 1.07 (0.83 – 1.38)
Underweight (BMI) <18.5 kg/m
2
5499 0.62 (0.33 – 1.20) 0.79 (0.33 – 1.90)
Fasting triglycerides >200 mg/dl 2375 1.11 (0.76 – 4.61) 0.87 (0.52 – 1.46)
C-reactive protein >4.0 mg/l 5457 1.09 (0.90 – 1.31) 1.11 (0.87 – 1.41)
Plasma homocysteine >15 μmol/l 4208 1.56 (1.20 – 2.04) 2.06 (1.43 – 2.96)
Note: Models controlled for age, gender, race, education, marital status
Odds Ratio (OR) with good vision as the reference group; CI = confidence interval; HDL = high-density lipoprotein; LDL =
low-density lipoprotein; BMI = body mass index
Bold items indicate significance at the p<.05 level
58
participants with poor vision reported that diagnosis. Of all the pathological
conditions, the greatest differences in proportions between poor or worse and
good or better vision groups occurred with respect to stroke (20.9% and 7.6%,
respectively), and diabetes (29.4% and 16.9%, respectively).

Table 2.4: Unadjusted Descriptive Statistics for Self-Reported Pathology
Indicators by Self-Reported Vision Status (NHANES IV, 1999 – 2008, Weighted)
Good Fair Poor Total Wald χ
2
χ
2
prob
(n = 4594) (n = 1411) (n= 582) (N = 6693)
Indicator Percent With Disease
Cancer
24.6% 24.5% 24.3% 24.4% 0.03 0.98
Arthritis
51.0% 58.4% 63.3% 53.1% 40.7 <.0001
Heart Problem
19.6% 27.7% 31.0% 21.9% 33.6 <.0001
Stroke
7.6% 13.5% 20.9% 9.7% 50.3 <.0001
Respiratory
11.2% 15.1% 20.3% 12.5% 25.4 <.0001
Diabetes
16.9% 25.9% 29.4% 19.4% 49.3 <.0001
Hearing
3.0% 4.8% 6.6% 3.6% 18.4 0.0003

In Table 2.5, results are reported from seven binary logistic regression
analyses conducted to determine the odds that older persons in fair or poor
self-reported vision categories report a diagnosis of each pathological condition,
relative to older persons with good or better vision. Results from two models are
shown. The first model displays odds ratios controlling for sociodemographic
covariates for each vision group relative to good or better vision. In model 2, the
odds of self-reported disease are shown relative to the good or better vision
category, controlling for sociodemographic and biological risk covariates.

59

Table 2.5: Binary Logistic Regression Models Testing the Effect of Self-Reported Vision Status on Self-Reported
Pathological Conditions (NHANES IV, 1999 – 2008, Weighted)

Fair Vision

Poor Vision
Model 1* Model 2† Model 1* Model 2†
Pathological Condition

N
(models 1/2)
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Cancer
6420/3731 1.07 (0.90 – 1.28) 1.06 (0.87 – 1.30) 1.10 (0.89 – 1.35) 1.20 (0.86 – 1.67)
Arthritis
6414/3725 1.30 (1.12 – 1.51) 1.20 (0.93 – 1.54) 1.56 (1.24 – 1.97) 1.52 (1.11 – 2.08)
Heart Problem
6429/3735 1.57 (1.30 – 1.90) 1.51 (1.18 – 1.95) 1.71 (1.31- 2.22) 1.82 (1.24 – 2.66)
Stroke
6399/3722 1.72 (1.35 – 2.20) 1.68 (1.14 – 2.49) 2.48 (1.90 – 3.31) 2.00 (1.30 – 3.06)
Respiratory
6381/3705 1.47 (1.16 - 1.88) 1.68 (1.20 – 2.37) 1.97 (1.53 – 2.53) 2.46 (1.70 – 3.55)
Diabetes
6423/3705 1.66 (1.41 – 1.96) 1.63 (1.24 – 2.13) 1.92 (1.52 – 2.44) 1.71 (1.09 – 2.69)
Hearing
6429/3735 1.47 (0.94 – 2.28) 0.84 (0.55 – 1.26) 1.83 (1.16 – 2.91) 1.68 (0.63 – 4.49)
Note:* Model 1 controlled for age, gender, race, education, marital status
†Model 2 controlled sociodemographic and biological risk factors.
Odds Ratio (OR) with good vision as the reference group; CI = confidence interval; Bold items indicates significant at the p<.05 level

60
With the exception of cancer and hearing impairment, older persons with
poor vision experienced greater odds of reporting each pathological condition,
after controlling for sociodemographic and biological risk covariates. Compared to
those with good vision, older people with poor vision were about 52% more likely to
report arthritis (OR = 1.52; 95% CI 1.11-2.08); 82% more likely to report some
heart problem (OR = 1.82; 95% CI 1.24-2.66); twice as likely to report having
experienced a stroke (OR = 2.00; 95% CI 1.30-3.06); and about 71% more likely to
report diabetes (OR = 1.71; 95% CI 1.09 -2.69).
Results were similar for older persons with fair vision. With the exception of
cancer, arthritis, and hearing impairment, persons with fair vision on average were
more likely to report having all other impairments than persons with good vision.
Older persons reporting fair vision were about 51% more likely to report a heart
problem (OR = 1.51; 95% CI 1.18 -1.95); 68% more likely to report a stroke (OR =
1.68; 95% CI 1.14 -2.49); 68% more likely to report a respiratory problem (OR =
1.68; 95% CI 1.20 -2.37); and 63% more likely to report diabetes (OR = 1.63; 95%
CI 1.24 -2.13), after controlling for sociodemographic and biological risk
covariates.
Table 2.6 shows unadjusted percentages of persons in each self-reported
vision category who reported any difficulty with, or being unable to perform ten
functional activities. The largest difference in proportions of older persons with
poor vision compared to those with good vision, who had difficulties occurred in
activities related to lower-limb functioning. Whereas 41.4% of persons with poor
vision reported difficulty standing from an armless chair, only 18.7% of people with
61
good vision, and 23.1% of the total sample reported difficulty in this activity. Other
areas with remarkably large differences in proportions with difficulties included
reaching up over head, and grasping/holding small objects.
Table 2.6: Unadjusted Descriptive Statistics for Self-Reported Functional
Indicators by Self-Reported Vision Status (NHANES IV, 1999 – 2008, Weighted)
Indicator Good Fair Poor Total Wald χ
2
χ
2
prob
(n = 4594) (n = 1411) (n= 582) (N = 6587)
Percent At-Risk
Walk ¼ Mile
Difficulty 16.7% 22.1% 19.8% 17.8% 240.8 <0.0001

Unable 16.6% 33.1% 51.5% 22.0%
Walking 10 Steps
Difficulty 14.0% 20.0% 22.1% 15.6% 207.6 <0.0001

Unable 14.8% 27.9% 44.6% 19.2%
Stoop, Crouch, Kneel
Difficulty 39.4% 47.3% 46.2% 41.2% 202.2 <0.0001

Unable 8.6% 17.1% 32.0% 11.8%
Lifting or Carrying
Difficulty 16.6% 26.3% 30.7% 19.3% 151.1 <0.0001

Unable 5.7% 13.0% 21.7% 8.4%
Walk Between Rooms
Difficulty 6.9% 15.4% 19.6% 9.8% 113.0 <0.0001

Unable 1.8% 4.3% 20.9% 2.7%
Stand from Armless Chair
Difficulty 18.7% 34.1% 41.4% 23.1% 187.2 <0.0001

Unable 2.7% 4.7% 12.8% 3.8%
Stand Long Periods
Difficulty 26.6% 35.6% 32.2% 30.1% 241.7 <0.0001

Unable 12.3% 23.9% 43.4% 16.6%
Sit Long Periods
Difficulty 15.6% 23.2% 29.3% 17.9% 87.3 <0.0001

Unable 0.9% 1.6% 6.7% 1.4%
Reach Up Over Head
Difficulty 12.4% 23.0% 27.6% 15.4% 128.0 <0.0001

Unable 1.6% 3.5% 7.7% 2.4%
Grasp/Hold Small Object
Difficulty 13.9% 23.7% 29.2% 16.7% 73.1 <0.0001

Unable 0.4% 1.53% 4.11% 0.9%

62
With respect to dependence, the largest differences in proportions of
persons reporting inability were also in activities related to lower-limb functioning.
More than half (51.5%) of persons with poor vision reported being unable to walk a
quarter mile, whereas only 16.7% of persons with good vision, and 17.8% of the
sample over all reported this inability; 44.6% of persons with poor vision reported
being unable to walk 10 steps compared to 14.0% of persons with good vision and
15.6% overall; and 43% of persons with poor vision reported being unable to stand
for long periods, compared to 12.3% of persons with good vision, and 16.6%
overall.
Odds ratios and 95% confidence intervals representing difficulty and
dependence in each functional activity, by fair and poor vision groups are
displayed in Tables 2.7 and 2.8, respectively. Two models are depicted in each
table. In model 1, sociodemographic covariates are controlled; model 2 shows
odds ratios after controlling sociodemographic, biological risk covariates, and
pathological conditions. Without exception, persons with fair and poor vision on
average had greater odds of reporting difficulty with all functional activities,
compared to older persons with good or better vision, even after accounting for
covariates from each of the other health dimension groups. For older persons with
fair vision, the greatest difficulties were observed in walking between rooms (OR =
2.09; 95% CI 1.40 -3.11); grasping/holding small objects (OR = 2.02; 95% CI 1.43
-2.85); and standing from an armless chair (OR = 1.90; 95% CI 1.41 -2.55). Older
persons with fair vision experienced the greatest dependence with walking a
quarter mile (OR = 2.01; 95% CI 1.46 -2.77); stooping, crouching and kneeling (OR
63
= 1.92; 95% CI 1.27 -2.92); and standing from an armless chair (OR = 1.92; 95%
CI 1.06 -3.49), compared to respondents with good or better vision.
Table 2.7: Multinomial Logistic Regression Models Testing the Effect of
Self-Reported Fair Vision Status on Self-Reported Difficulty in Functioning
Indicators (NHANES IV, 1999 – 2008, Weighted)
Fair Vision
Model 1* Model 2†
Activity

N
(models 1 & 2)
OR 95% CI OR 95% CI
Walk ¼ Mile
6352/3640 Difficulty 1.83 (1.49 – 2.25) 1.72 (1.30 – 2.28)

Unable 2.57 (2.01 – 3.29)

2.01 (1.46 – 2.77)
Walking 10 Steps
6379/3649 Difficulty 1.81 (1.52 – 2.15) 1.52 (1.16 –2.01)

Unable 2.18 (1.74 – 2.74) 1.53 (1.08 – 2.18)
Stoop, Crouch, Kneel
6323/3632 Difficulty 1.71 (1.43 – 2.05) 1.46 (1.46 – 3.24)

Unable 2.48 (1.99 – 3.10) 1.92 (1.27 – 2.92)
Lifting or Carrying
6299/3626 Difficulty 1.80 (1.54 – 2.10) 1.61 (1.25 – 2.08)

Unable 2.28 (1.67 – 3.11) 1.81 (1.09 – 3.02)
Walk Between Rooms
6410/3673 Difficulty 2.10 (1.61 – 2.73) 2.09 (1.40 – 3.11)

Unable 2.40 (1.55 – 3.73) 1.29 (0.49 – 3.42)
Stand from Chair
6414/3673 Difficulty 2.03 (1.68 – 2.45) 1.90 (1.41 – 2.55)

Unable 1.81 (1.24 – 2.64) 1.92 (1.06 – 3.49)
Stand Long Periods
6242/3594 Difficulty 1.71 (1.40 – 2.10) 1.51 (1.19 – 1.91)

Unable 2.36 (1.90 – 2.93) 1.71 (1.24 – 2.36)
Sit Long Periods
6406/3664 Difficulty 1.65 (1.36 – 2.01) 1.72 (1.35 – 2.20)

Unable 1.78 (1.04 – 3.05) 1.49 (0.84 – 2.65)
Reach Up Over Head
6408/3670 Difficulty 1.91 (1.58 – 2.31) 1.54 (1.14 – 2.06)

Unable 2.01 (1.36 – 2.95) 1.64 (0.81 – 3.35)
Grasp/Hold Small Object
6421/3675 Difficulty 1.86 (1.45 – 2.39) 2.02 (1.43 – 2.85)

Unable 2.75 (1.30 – 5.79) 2.60 (0.55 – 12.33)
* Model 1 controlled for age, gender, race, education, marital status
†Model 2 controlled for sociodemographic variables, biological risk factors, and pathological conditions.
Odds Ratio (OR) with good vision as the reference group; CI = confidence interval
Bold items indicate significance at the p<.05 level
64

Table 2.8: Multinomial Logistic Regression Models Testing the Effect of
Self-Reported Poor Vision Status on Self-Reported Difficulty in Functioning
Indicators (NHANES IV, 1999 – 2008, Weighted)
Poor Vision
Model 1* Model 2†
Activity

N
(models 1 & 2)
OR 95% CI OR 95% CI
Walk ¼ Mile
6352/3640 Difficulty 2.24 (1.67 – 3.02) 1.73 (1.09 – 2.75)

Unable 5.23 (4.01 – 6.82) 3.93 (2.50 – 6.16)
Walking 10 Steps
6379/3649 Difficulty 2.79 (2.00 – 3.89) 1.69 (1.08 – 2.41)

Unable 4.64 (3.59 – 5.99) 2.81 (1.82 – 4.35)
Stoop, Crouch, Kneel
6323/3632 Difficulty 2.54 (1.91 – 3.39) 2.17 (1.46 – 3.24)

Unable 6.52 (4.80 – 8.87) 4.24 (2.47 – 7.28)
Lifting or Carrying
6299/3626 Difficulty 2.64 (1.99 – 3.51) 1.84 (1.08 – 3.13)

Unable 5.61 (4.06 – 7.74) 4.43 (2.65 – 7.42)
Walk Between Rooms
6410/3673 Difficulty 3.98 (2.99 – 5.30) 3.36 (2.22 – 5.09)

Unable 4.78 (2.93 – 7.80) 2.91 (0.94 – 9.01)
Stand from Chair
6414/3673 Difficulty 2.99 (2.28 – 3.92) 2.27 (1.55 – 3.33)

Unable 5.46 (3.65 – 8.17) 3.87 (2.10 – 7.14)
Stand Long Periods
6242/3594 Difficulty 2.38 (1.74 – 3.26) 2.18 (1.47 – 3.25)

Unable 6.27 (4.53 – 8.68) 4.09 (2.53 – 6.63)
Sit Long Periods
6406/3664 Difficulty 2.39 (1.79 – 3.21) 2.22 (1.52 – 3.25)

Unable 8.59 (4.76 – 15.51) 4.89 (2.52 – 9.49)
Reach Up Over Head
6408/3670 Difficulty 2.32 (1.85 – 2.90) 1.51 (1.03 – 2.22)

Unable 4.03 (2.38 – 6.82) 2.07 (0.74 – 5.78)
Grasp/Hold Small Object
6421/3675 Difficulty 2.26 (1.74 – 2.95) 1.50 (1.05 – 2.16)

Unable 9.23 (4.00 – 21.28) 5.50 (1.36 – 22.24)
* Model 1 controlled for age, gender, race, education, marital status
†Model 2 controlled for sociodemographic variables, biological risk factors, and pathological conditions.
Odds Ratio (OR) with good vision as the reference group; CI = confidence interval
Bold items indicate significance at the p<.05 level

65
Compared to older persons with good vision, those with poor vision
reported the greatest difficulty with walking between rooms (OR = 3.36; 95% CI
2.22 -5.09); standing from an armless chair (OR = 2.27; 95% CI 1.55 -3.33); and
sitting for long periods (OR = 2.22; 95% CI 1.52 -3.25). With respect to
dependence, older persons with poor vision were more likely than their
counterparts with good or better vision to report an inability to grasp/hold small
objects (OR = 5.50; 95% CI 1.36 -22.24); sit long periods (OR = 4.89; 95% CI 2.52
-9.49); and lifting or carrying (OR = 4.43; 95% CI 2.65 -7.42).
Table 2.9 shows unadjusted percentages of persons in each self-reported
vision category who reported any difficulty with, or being unable to perform nine
daily-living activities representing the disability dimension of health. Significant
differences in proportions of persons with difficulty were found for each activity,
between persons with self-reported poor vision and those with good vision;
however, the largest difference in proportions were found for dressing (31.7% and
9.6%, respectively); getting in and out of bed (32.1% and 10.4%, respectively); and
activities related to doing chores around the house (38.2% and 17.1%,
respectively). Other areas with large differences in proportions reporting difficulty
included going to movies/events, attending social events, and doing leisure
activities at home. With respect to dependence, the largest differences in
proportions of persons reporting inability were for attending social events.
Whereas 30.8% of older persons with poor vision reported this disability, only 2.8%
of persons with good vision, and 4.3% overall reported being unable to attend
66
social events. Other activities with relatively large differences between groups
included doing chores around the house, and going to movies/events.

Table 2.9: Unadjusted Descriptive Statistics for Self-Reported Daily Living
Activity Indicators by Self-Reported Vision Status (NHANES IV, 1999 – 2008,
Weighted)
Indicator Good Fair Poor Total Wald χ
2
χ
2
prob
(n = 4594) (n = 1411) (n= 582) (N = 6587)
Percent At-Risk
Manage Money
Difficulty 5.1% 11.8% 20.0% 7.2% 133.5 <0.0001

Unable 2.0% 4.3% 15.6% 3.3%
House Chores
Difficulty 17.1% 31.2% 38.2% 21.1% 206.4 <0.0001

Unable 4.2% 7.7% 23.5% 6.2%
Preparing Meals
Difficulty 5.6% 12.8% 22.3% 8.0% 128.3 <0.0001

Unable 2.4% 4.5% 15.1% 3.6%
Get In/Out Bed
Difficulty 10.4% 23.0% 32.1% 14.2% 130.4 <0.0001

Unable 0.9% 1.8% 4.8% 1.4%
Using Fork, Knife, Cup
Difficulty 3.9% 9.9% 16.7% 5.9% 70.3 <0.0001

Unable 0.2% 0.4% 2.5% 0.4%
Dressing
Difficulty 9.6% 18.1% 31.7% 12.6% 135.8 <0.0001

Unable 0.7% 1.6% 5.9% 1.2%
Go to Movies/Events
Difficulty 12.7% 23.9% 33.6% 16.1% 218.5 <0.0001

Unable 3.2% 6.4% 21.9% 5.0%
Attend Social Event
Difficulty 8.6% 20.0% 28.6% 11.9% 194.2 <0.0001

Unable 2.8% 5.0% 30.8% 4.3%
Leisure At Home
Difficulty 3.2% 10.1% 24.0% 5.9% 132.0 <0.0001

Unable 0.3% 0.8% 3.6% 0.6%
67
In tables 2.10 and 2.11, odds ratios and 95% confidence intervals are
displayed for each daily-living activity by fair and poor vision groups, respectively.
Two models are depicted in each table. Model 1 shows odds ratios with controls for
sociodemographic covariates; model 2 shows odds ratios after controlling
sociodemographic, biological risk, pathological conditions, and functioning
covariates — variables from all four previous health dimension categories. After
controlling for other health dimensions, older persons who reported fair vision were
significantly more likely to have difficulty with few daily living activities. Significant
differences between persons with fair vision, and those with good vision were
found for difficulty managing money (OR = 1.65; 95% CI 1.08 -2.53) and doing
leisure activities at home (OR = 1.86; 95% CI 1.17 -2.96). Older persons with fair
vision were significantly less likely to report being unable to managing money (OR
= 0.38; 95% CI 0.17 -0.85).
In contrast, persons with poor vision were more likely to experience difficulties with
most daily living activities. For example, older persons with poor vision were more
than twice as likely as their counterparts with good or better vision to report
difficulty preparing meals (OR = 2.20; 95% CI 1.15 -4.19), using a knife, fork, and
cup (OR = 2.18; 95% CI 1.21 -9.95), and dressing (OR = 2.31; 95% CI 1.42 -3.77).
Older persons with poor vision were also more likely to report being unable to
perform most daily living activities. Relatively large odds ratios were reported for
activities that are usually performed outside of the home. For example, persons
with poor vision were nearly 3 times more likely to report being unable to attend
social events (OR = 2.64; 95% CI 1.08 -6.45), and more than six times more likely
68
to report an inability to go to the movies or to events (OR = 6.12; 95% CI 2.83
-13.22). In addition, older persons with poor vision were nearly five times more
likely to report being unable to do leisure activities at home (OR = 4.99; 95% CI
1.48 -16.81).
Table 2.10: Multinomial Logistic Regression Models Testing the Effect of
Self-Reported Fair Vision Status on Self-Reported Difficulty in Daily Living
Activities (NHANES IV, 1999 – 2008, Weighted)
Fair Vision
Model 1* Model 2†
Activity

N
(models 1 & 2)
OR 95% CI OR 95% CI
Manage Money
6110/3350 Difficulty 2.17 (1.66 – 2.84) 1.65 (1.08 – 2.53)

Unable 1.75 (1.19 – 2.56) 0.38 (0.17 – 0.85)
House Chores
6124/3360 Difficulty 2.20 (1.79 – 2.72) 1.11 (0.74 – 1.67)

Unable 1.85 (1.32 – 2.61) 0.59 (0.28 – 1.25)
Preparing Meals
6088/3358 Difficulty 2.19 (1.71 – 2.81) 1.02 (0.66 – 1.59)

Unable 1.48 (1.01 – 2.17) 0.23 (0.10 – 0.55)
Get In/Out Bed
6416/3454 Difficulty 2.19 (1.80 – 2.66) 1.10 (0.79 – 1.53)

Unable 2.07 (1.06 – 4.03) 2.08 (0.20 – 22.12)
Using Fork, Knife, Cup
6428/3457 Difficulty 2.30 (1.71 – 3.09) 1.33 (0.78 – 2.27)

Unable 1.83 (0.46 – 7.24) -- --
Dressing
6422/3455 Difficulty 1.84 (1.44 – 2.35) 0.84 (0.55 – 1.28)

Unable 2.33 (1.26 – 4.32) 0.86 (0.21 – 3.56)
Go to Movies/Events
6187/3390 Difficulty 2.11 (1.74 – 2.55) 1.02 (0.69 – 1.50)

Unable 1.91 (1.27 – 2.89) 0.70 (0.32 – 1.53)
Attend Social Event
6185/3390 Difficulty 2.44 (1.99 – 3.00) 1.27 (0.80 – 2.03)

Unable 1.59 (1.04 – 2.41) 0.59 (0.27 – 1.30)
Leisure At Home
6416/3453 Difficulty 3.11 (2.27 – 4.25) 1.86 (1.17 – 2.96)

Unable 2.34 (0.91 – 5.99) 0.60 (0.19 – 1.90)
* Model 1 controlled for age, gender, race, education, marital status
† Model 2 controlled sociodemographic variables, biological risk factors, pathological conditions, and functional difficulty
Odds Ratio (OR) with good vision as the reference group; CI = confidence interval
Bold items indicate significance at the p<.05 level
69
Table 2.11: Multinomial Logistic Regression Models Testing the Effect of
Self-Reported Poor Vision Status on Self-Reported Difficulty with Daily Living
Activities (NHANES IV, 1999 – 2008, Weighted)
Poor Vision
Model 1* Model 2†
Activity

N
(models 1 & 2)
OR** 95% CI OR** 95% CI
Manage Money
6110/3350 Difficulty 4.28 (3.16 – 5.79) 2.81 (1.72 – 4.59)

Unable 7.01 (4.69 – 10.44) 3.07 (1.29 – 7.29)
House Chores
6124/3360 Difficulty 4.05 (3.27 – 5.02) 1.78 (1.08 – 2.94)

Unable 8.06 (5.57 – 11.68) 2.95 (1.27 – 6.87)
Preparing Meals
6088/3358 Difficulty 4.40 (3.13 – 6.19) 2.20 (1.15 – 4.19)

Unable 6.35 (4.29 –9.38) 2.29 (0.91 – 5.76)
Get In/Out Bed
6416/3454 Difficulty 3.27 (2.56 – 4.17) 1.05 (0.59 – 1.86)

Unable 6.23 (3.10 – 12.54) 9.55 (1.17 – 77.59)
Using Fork, Knife, Cup
6428/3457 Difficulty 3.68 (2.70 – 5.02) 2.18 (1.21 – 9.95)

Unable 13.34 (3.45 – 51.77) -- --
Dressing
6422/3455 Difficulty 3.80 (2.92 – 4.93) 2.31 (1.42 – 3.77)

Unable 9.54 (4.58 – 19.92) 1.50 (0.13 – 17.68)
Go to Movies/Events
6187/3390 Difficulty 4.40 (3.42 – 5.67) 2.77 (1.63 – 4.69)

Unable 8.72 (5.91 – 12.87) 6.12 (2.83 – 13.22)
Attend Social Event
6185/3390 Difficulty 4.67 (3.74 – 5.82) 3.01 (1.78 – 5.08)

Unable 7.72 (5.33 – 11.19) 2.64 (1.08 – 6.45)
Leisure At Home
6416/3453 Difficulty 8.24 (6.21 – 10.93) 5.91 (3.24 – 10.77)

Unable 11.47 (5.25 – 25.09) 4.99 (1.48 – 16.81)
* Model 1 controlled for age, gender, race, education, marital status
† Model 2 controlled sociodemographic variables, biological risk factors, pathological conditions, and functional difficulty
Odds Ratio (OR) with good vision as the reference group; CI = confidence interval
Bold items indicate significance at the p<.05 level
70
F. Discussion
The purpose of this study was to compare older adults with self-reported
fair and poor vision status with those who reported good vision on the basis of four
health dimensions that make up a theoretical pathway to disability in later life.
According to Verbrugge and Jette (1994), biological factors are among many
personal risk characteristics that represent the first stage leading to disability, via
disease/impairment and functional decline. Therefore, understanding how older
people with visual impairments differ from older adults with normal vision, with
respect to these health dimensions, is an important starting point for
understanding how health disparities between vision groups could precipitate
different health and functioning outcomes, as well as secondary health outcomes
such as falls.
In this study, comparisons were made based on high-risk levels of
biomarkers of physiological functioning, self-reported pathological conditions,
difficulty and/or inability to perform functional activities, and difficulty/inability with
daily living activities experienced by older persons participating in NHANES.
Results of this study suggested that less-than good vision status was associated
with negative health outcomes across the four health dimensions. Even with
statistical control for covariates in preceding health dimensions, older adults who
reported fair or poor vision status consistently indicated worse biological health,
greater comorbidity, and more difficulty performing functional and daily living
tasks than their counterparts who reported good or better vision.
71
Observed disparities across health dimensions could derive from at least
two potential explanations. First, as hypothesized, disparities in health
dimensions among the three vision status groups may originate as a result of
disability—that is to say, limitations in daily activities that could lead to changes in
diet, health maintenance, and activity levels that often follow the onset of vision
loss in late life (Crews & Campbell, 2004). Functional losses in general, inherently
suggest that older adults who experience them will be less physically active. Older
persons with varying degrees of vision loss have been reported to experience
greater declines in their ability to perform functional and daily-living activities
(Berger & Porell, 2008; Hassell, Lamoureux, & Keeffe, 2006; Horowitz, 2004;
Travis, Boerner, Reinhardt, & Horowitz, 2004; West et al., 2002). Some specific
measures of ADLs and IADLs that are typically used to quantify disability in older
people place direct emphasis on activities that would influence the dietary choices
of persons who could not perform them independently (such as the ability to eat
independently, prepare meals for themselves, or to shop for groceries without
assistance). Consistent with this view, results of this study indicated that older
adults with vision impairments were more likely to report difficulties in an array of
daily living activities including preparing meals, or using utensils to eat. Similar
difficulties over a long enough period of time would likely affect other aspects of
an individual’s health, starting by reducing physical activity and dietary quality,
and consequently exerting negative influence on specific biomarkers that
reflected the individual’s dietary practices and activity levels.
72
Second, disparities among health dimensions reported here may be a
reflection of primary pathologies and conditions that often precede vision loss and
that are indirectly related to poor vision status. For example, with respect to
biomarkers, results of this study suggested that less-than good self-reported
vision status was associated high-risk levels of HDL cholesterol, glycated
hemoglobin, and plasma homocysteine, each of which could be associated with
systemic diseases such as atherosclerosis or diabetes, which often result in loss
of vision. In this case, disease may precede or even cause vision loss. Disease
states may occur along the continuum of the biomarker-to-disability pathway, and
adjusting for these conditions modifies relationships between health dimensions
and vision loss, as demonstrated by decreased effects of poor vision when
previous health conditions were controlled in the models. Ultimately, it is likely that
disparities in health dimensions can be explained by both differences in the
processes underlying vision status (and any potentially related comorbidities), as
well as by behavioral factors that likely place older adults who are visually
impaired at greater risk with respect to their disability status, and subsequent
levels of biological indicators.
With respect to the biomarkers tested in this study, differences in risk
among the blood lipid parameter for persons with poor vision is consistent with
differences in cholesterol findings that have been reported among older adults
participating in the MacArther Study of Successful Aging and for adults age 40
and older in the NHANES III (Crimmins, Vasunilashorn, Kim, & Alley, 2008;
Seeman, Crimmins, Huange, Singer, Bucur et al., 2004). Elevated levels of
73
homocysteine were also exhibited among persons with fair or poor vision. The
mechanisms by which vision loss and high-risk levels of homocysteine interact
could be explained in part by diet and lifestyle circumstances. One important
factor that is known to determine homocysteine concentration is nutrition, and it
as been suggested that elevated levels of plasma homocysteine are indicative of
poor nutrition (Danesh & Lewington, 1998; Selhub et al., 1999). For older adults
with vision impairment, this finding is especially important given the associations
reported between elevated plasma homocysteine and increased risk of other
negative health outcomes, including cardiovascular disease (Boushey, Beresford,
Omenn, & Motulsky, 1995; Refsum, Ueland, Nygard, & Vollset, 1998; Wilcken,
1998), depression (Tiemeier et al., 2002), Alzheimer’s disease (Clarke et al.,
1998; Seshadri et al., 2002), and colorectal cancer (Kato et al., 1999). Given that
a significant portion of homocysteine levels can be attributed to nutrition, diet
modifications that include adequate amounts of folate and vitamins B6 and B12
could potentially improve the health of individuals with visual impairments or any
individuals with inadequate nutrient intake. Folate, vitamins B6 and B12 play
important roles in homocysteine metabolism, such that deficiencies in any or in all
have been associated with elevated levels of total homocysteine (Bates et al.,
1997; Nygard, Refsum, Ueland, & Vollset, 1998; Selhub, Jacques, Wilson, Rush,
& Rosenberg, 1993; Selhub et al., 1999). However, vitamin use or levels of
circulating vitamins (e.g., B-vitamins) alone may not be associated with
cardiovascular outcomes (Baigent & Clarke, 2007). It is likely that a broader
74
combination of nutrition and lifestyle changes, aside from greater intake of
B-vitamins, is important in determining the onset of adverse health outcomes.
The primary goal of this study was to demonstrate relationships between
self-reported visual impairments and specific health dimensions that could
influence the overall health and disability status of older adults. Although the
findings of this study are believed to reflect these associations accurately, it is
recognized that the study has at least one important limitation. The relationship
between vision and health is complicated because it is often difficult, if not
impossible to determine directionality of the relationship—that is, to sort out the
extent to which vision loss causes changes in biomarkers that often precede
chronic diseases, functional loss, and disability (as hypothesized); versus the
extent to which chronic diseases may lead to vision loss. Unfortunately, the use of
cross-sectional data was limiting in the current study, because it did not provide
the opportunity to assess the temporal order of these important events. It is
possible that at-risk levels of some biomarkers could exist prior to vision loss or as
a result of chronic diseases (such as diabetes or vascular disease), which often
result in vision loss. Therefore, future studies should employ a more rigorous
longitudinal analysis to take into consideration the order of key events, and to
make a more powerful assertion of causality. Results of this study, nevertheless,
point to important differences between older persons with normal vision and those
with vision impairment. Fortunately, poor nutrition and decreased activity levels
that often accompany vision impairment in late life, and may affect many of the
biomarkers, pathological conditions and activities examined in this study, can be
75
addressed through relatively simple and inexpensive interventions within
programs that serve this population. Nutrition and education programs that serve
older adults in general, as well as senior activity centers and rehabilitation
programs that target services to the population of older persons with visual
impairments can benefit from this study by increasing awareness of how certain
health dimensions may be experienced differently by older persons who are
visually impaired.
76
CHAPTER III: MODERATING EFFECTS OF VISION LOSS ON LIMB
FUNCTIONING AND FALLS
A. Introduction
Vision loss can be experienced at any age; however, loss of vision in late
life is often accompanied by comorbidities, and functional impairments that make
adjustment and rehabilitation more difficult than when vision loss occurs earlier in
the life course. Similarly, people may fall at any time during their lives, but falling is
of greater concern to older people because of age-related changes to the body
that make older individuals more likely to experience a fall and serious injury as a
result of a fall. A large number of studies have identified risk factors for falling,
including muscle weakness, a history of falls, gait and balance deficits, impaired
activities of daily living (ADLs), and visual deficit (American Geriatrics Society,
2010). Considerable evidence has shown that attention to these risk factors can
significantly reduce rates of falling (Abdelhafiz & Austin, 2003; Feldman &
Chaudhury, 2008; Lord, Ward, Williams, & Strudwick, 1995), and that the most
effective fall-reduction programs involve systemic fall-risk assessment with
targeted interventions that include physical activity, medical management, and
environmental inspection and hazard removal (Campbell & Robertson, 2006; Day
et al., 2002; Pynoos, Rose, Rubenstein, Choi, & Sabata, 2006; Rubenstein, 2006).
Bearing in mind findings that implicate interactions between multiple factors in
causing falls, an important early step in creating effective fall-prevention programs
for older adults with vision impairments is to understand relationships that exist
between vision and other common factors that predict falls.
77
Fall Risk Factors
Vision impairments are among more than 400 fall-risk factors that have
been identified in the literature (Masud & Morris, 2001). In general, fall-risk factors
are described as being extrinsic, intrinsic, or behavioral in origin (Bath & Morgan,
1999; Bueno-Cavanillas, Padilla-Ruiz, Jimenez-Moleon, Peinado-Alonso, &
Galvez-Vargas, 2000; Cesari et al., 2002; Graafmans et al., 1996). Extrinsic
factors tend to be environmentally oriented, and are shared among persons who
inhabit a common environment. Many extrinsic factors have been identified such
as slippery surfaces; inadequate lighting; loose, deep pile, or worn carpets;
staircases without railings; unsupportive and/or badly arranged furniture; poorly
designed tubs, toilets and fixtures in the bathroom; clutter; and pets underfoot
(Clemson, Cumming, & Roland, 1996; Pynoos et al., 2006; Rogers, Rogers,
Takeshima, & Islam, 2004). Extrinsic factors often result in trips, slips, or missteps,
posing increased fall risk, especially for community dwelling older adults whose
homes may contain many hazards. The prevalence of environmental hazards in
the homes of older adults is high, with approximately 80% of homes containing at
least one identifiable hazard, and 39% containing five or more hazards (Carter,
Campbell, Sanson-Fisher, & Gillespie, 2000). Therefore, home assessments and
modifications designed to reduce or eliminate hazards in the environment would
seem to be integral in programs aimed at reducing fall risk, especially since, as
extrinsic in origin, they are amenable to correction.
In contrast, intrinsic factors are individually oriented, as in health conditions
(e.g., chronic diseases), degrees of functional loss (e.g., impaired mobility), or
78
states of being (e.g., advanced age). Other intrinsic fall-risk factors include muscle
weakness, gait/balance disorders, reduced mental status, drug interactions,
having a history of falls, and sensory loss including vision loss (American
Geriatrics Society, 2010; Pynoos et al., 2006). Intrinsic risk factors, including vision
impairment, are dynamic insofar as they may change over time, resulting in health
and disability status that is in constant flux between losses and gains. To address
intrinsic factors, multifactorial fall-prevention programs may employ medical risk
assessments to identify health and functioning problems that can lead to falls.
Exercise programs designed to build strength and improve balance may be
recommended to reduce the negative impact of intrinsic fall-risk factors on the
wellbeing of older people who have visual impairments.
In addition to intrinsic and extrinsic factors leading to increased fall risk,
behavioral fall-risk factors are those that reflect choices of older individuals with
respect to how they interact within their environments. Risk-taking behaviors are
those that would increase the likelihood of adverse physical consequences, such
as a fall (Feldman & Chaudhury, 2008). Examples of behavioral risk factors include
performing behaviors that could increase fall risk (e.g., standing on unstable
objects to reach items that are stored on high shelves); failing to perform behaviors
that could reduce fall risk (e.g., not turning on lights when using the bathroom at
night, or not using grab bars or handrails when they are present); or selecting
unsafe clothing, footwear, or inappropriate/outdated eyewear prescriptions. In
research by Hornbrook et al. (1994), behavioral fall-risk factors such as not being
79
careful or alert, not looking where one is going, or being in a hurry while performing
daily activities were cited as preceding circumstances in up to 63% of falls.
Whereas the majority of falls occur as a result of interactions between
extrinsic, intrinsic, and behavioral risk factors, the relationship between falls and
vision factors are often complex and specific to the individual (Leslie & Pierre,
1999). Thus, the effects of vision loss and other systems on fall risk may be further
conceptualized in either one of two ways (Steinman, 2008). The first perspective is
an independent systems view, which posits that age-related changes to
independent systems contribute separately to fall risk by reducing perceptual and
physical functioning of the individual (see figure 3.1).

Figure 3.1: Independent Systems Effects of Multiple Systems Can Influence Fall
Risk

The independent role of vision loss in falls has been described from many
perspectives. Various clinical measures of vision status have been associated
independently with increased fall risk, including visual acuity (Ivers, Cumming,
Mitchell, & Attebo, 1998; Lord & Dayhew, 2001), contrast sensitivity (De Boer, et
80
al., 2004; Ivers et al., 1998; Lord & Dayhew, 2001), depth perception (Lord &
Dayhew, 2001), and loss in visual fields (Ivers et al., 1998; Lord, 2006).
Researchers have also looked at the independent effects of vision loss in directly
reducing mobility (Crews & Campbell, 2004) and functional capacity (Berger &
Porell, 2008; West et al., 2002). Numerous studies have demonstrated that
perceptual degradation caused by vision loss may directly interfere with an
individual’s ability to maneuver safely through the environment (Hassan,
Lovie-Kitchin, & Woods, 2002; Patel et al., 2006; Soong, Lovie-Kitchin, & Brown,
2004).
A second way to conceptualize the effects of vision loss and declines in
other systems on fall risk is an integrated systems view, which acknowledges the
potential for systems to interact with each other (Steinman, 2008). Under this view,
losses in the visual system could moderate losses in limb functioning. According to
Muller, Judd, and Yzerbyt (2005), moderating effects are found when one
independent variable affects the magnitude of another, with respect to a
dependent variable. In this scenario, losses in visual acuity could lead to
synergistic changes in musculoskeletal functioning by way of reduced physical
activity— a combined effect that is greater than effects that occur due to changes
in either single system independently (see figure 3.2).
81

Figure 3.2: Integrated Moderating Effects of Vision Loss on Other Systems Can
Influence Fall Risk

According to Steinman (2008), changes in the musculoskeletal system may
be moderated by declines in vision such that losses are magnified, resulting in less
stability and a greater likelihood of falling in old age. Figure 3.3 displays the
disability process described by Nagi (1965; 1976) and shows how impairments
associated with vision could relate to new functional losses. According to this
model, feedback loops are a potential mechanism by which limits in visual
functioning could result in functional limitations in other systems that could
increase the risk of falls (Verbrugge & Jette, 1994). For example, older persons
with vision impairment are more likely to have disabilities related to their vision that
could limit their physical activity levels (see study 1, above), resulting in new
biological risk factors and pathological conditions/impairments that could lead to
declines in the physical functioning of upper and lower limbs. Steinman (2008)
tested a moderating relationship between vision loss and declines in the
musculoskeletal system; however, methodological difficulties made interpretation
of results difficult, with respect to integrated effects. Specifically, Steinman
82
considered both parameters (vision loss and limb functioning) at a common point
in time. Nevertheless, if vision loss moderates upper- and lower-limb functioning
as hypothesized, then vision loss would precede declines in physical functioning,
with its moderating effects. Consequently, previous results made it difficult to infer,
with certainty, the role of moderating effects of vision loss on musculoskeletal
decline in predicting falls.

Figure 3.3: Feedback Loops within the Disability Process Model Can Influence
Fall Risk

A second limitation in research by Steinman (2008) was that it did not
control potentially important factors that could influence fall risk, such as the
presence of home modifications. For the majority of older adults who state that
they prefer to age in place within their homes and communities, the environment
may afford a setting which is accommodative to age-related changes in health and
functioning. In environments that are well-designed, or which are improved with
home modifications to support changes in physical abilities, the functioning of
older adults can be maximized to facilitate physical health, a sense of security, and
continued social engagement with others in the community (Pynoos, Caraviello, &
Cicero, 2009). However, home and community environments may also be rife with
83
potential problems, which can increase risk of negative outcomes including falls
and increased rates of injury and mortality related to falls.
B. Purpose of study
The aim of the current study was to address short-comings in previous
research by employing independent variables that accounted for weaknesses
described above. In addition to assessing the direct effects of self-reported vision
status and upper- and lower-limb functioning, the current study controlled change
in functioning of upper- and lower-limbs across a two-year period. Interaction
terms computed from change variables were used to test whether a significant
moderating relationship exists between self-reported vision and limb-function
status. The current study also controlled main effects of home modification to
determine the role of this variable in falls and fall risk. To control for fundamental
physiological differences between men and women, analyses were conducted
separately for each sex.
C. Method

Data
Data from the 2004 and 2006 panels of the University of Michigan’s Health
and Retirement Study (HRS) were analyzed. The HRS is a national longitudinal
steady state survey that contacts participants every two years. In 2004, HRS
continued data collection on four cohorts, made up of 18,491 Americans over the
age of 55. A fifth replenishing cohort was added to the sample that year, making a
total of 21,831. By comparison, in 1998 when four original cohorts were merged,
the sample was made up of 22,641 participants (over age 51), and 25,725 at the
84
study’s inception (attrition of roughly 4% and 15%, respectively). In 2006, the study
had 19,740 participants who were known to be living (attrition of roughly 13%
compared to 1998, and 23% overall). Baseline responses of each cohort were
taken via face-to-face interviews in the respondent’s home. Follow-up interviews,
including those in 2004 and 2006, were conducted via telephone.
The sample for this study was composed of participants who were 65 years
old or older. Participants were excluded if they were known to have died between
the 2004 and 2006 panels, or if they were institutionalized. Participants were also
excluded from analyses if their interviews were conducted by proxy, due to the
possibility that proxy respondents may not be knowledgeable about whether or not
individuals had fallen. Finally, participants were excluded if they scored a 6 or
lower out of 10 on the Telephone Interview of Cognitive Status (TICS). The
aggregate TICS item measured general orientation to time and place, and was
assumed, for this study, to function as an indicator of the respondent’s ability to
remember and report falls accurately. After all exclusion criteria were applied to the
data, an initial unweighted sample of 8,449 participants remained for analyses.
Sociodemographic measures
Four sociodemographic characteristics of participants, including age,
education, race, and presence of a spouse or partner in the household were
statistically controlled. Age and education were coded and analyzed as continuous
variables. Race, and spouse/partner status were coded as indicator variables, with
“White”, and “no spouse/partner present” as reference categories. Sex was
controlled on the basis of doing separate analyses for men and women.
85
Independent Variables
All independent variables were taken from the 2004 panel of HRS, with the
exception of change in functioning measures, which were computed using both
2004 and 2006 data (described below). The dependent variable, a dichotomous
measure of whether or not participants had fallen in the last two years was taken
from the 2006 panel, and merged to 2004 data for subsequent analyses. Key
explanatory independent variables included measures of self-reported vision
status, home modifications, change in upper- and lower-limb functioning, and four
interaction terms that were computed by multiplying vision status variables by
change in upper- and lower-limb functioning status. Other control variables
included baseline upper- and lower-limb functioning, self-reported pathology,
depression, hearing status, total number of medications, and number of alcoholic
drinks per day.
Vision. In the HRS, vision status was assessed using an item that asked
respondents to rate whether their eyesight was excellent, very good, good, fair, or
poor when using glasses or corrective lenses. Respondents could also state that
they were legally blind. This variable was recoded into three indicator variables—
participants who were legally blind or who reported having poor eyesight were
categorized together, whereas those who rated their eyesight as fair made up a
second group. A third group of participants with good or better eyesight served as
the reference category.
Home Modifications. Respondents representing each household were
asked in 2004 whether within the last two years (since 2002) their homes or
86
apartments had been modified to make it easier or safer for an older person or a
disabled person to live there; or whether their homes were already handicap
accessible prior to the 2002 interview. Responses were recoded into an indicator
variable in which homes that were previously or recently modified were coded
together, and compared to a reference group, in which homes had never been
modified. Household-level responses were applied to person-level participants
who indicated that they resided in households which had home modifications.
Upper- /Lower-Limb Functioning. Independent variables representing
baseline limb functioning in 2004, and change in functioning from 2004 to 2006
were quantified using measures of physical functioning found in the HRS.
Participants were asked whether, due to health or physical problems, they had any
difficulty performing 11 functional activities. For the baseline measures, six
activities associated with upper-limb functioning (ULF) (i.e., sitting 2 hours; getting
up from a chair; reaching arms; pulling/pushing large objects; lifting weights; and
picking up a dime) were grouped together. Likewise, six activities associated with
lower-limb functioning (LLF) (i.e., walking several blocks; walking 1 block; climbing
stairs; climbing 1 flight of stairs; and stooping) were grouped together. Continuous
variables representing baseline functioning were computed by summing the
number of difficulties experienced within each (ULF and LLF) category in 2004.
Change in functioning variables (∆ULF and ∆LLF) were computed in two
steps. First, dichotomous responses for each activity in 2006 were subtracted from
responses of respective activities in 2004. Each activity, then, could have a value
of 1 (indicating improvement of functioning between 2004 and 2006), 0 (indicating
87
no change), or -1 (indicating decline of functioning). Next, these scores were
summed, according to their ULF/LLF groupings; thus, scores for ∆ULF could range
from 6 (if the participant reported difficulty in all six activities in 2004, but no
difficulty in 2006) to -6 (if the participant reported no difficulties in 2004, but
difficulties in all activities in 2006). Likewise, scores for ∆LLF ranged from 5 to -5. In
both groups, scores close to 0 indicated no or little change in the number of
activities with which participants reported difficulties between the two waves.
Interaction Terms. Four interaction terms were computed by multiplying
indicators of each vision status (poor, fair) by ∆ULD and ∆LLD status. These new
variables were used in analyses testing the moderating effects of vision on limb
functioning.
Health and Medical Conditions. Self-ratings of the physical health of
respondents were assessed in HRS, with respect to 8 medical conditions that are
commonly associated with old age. Participants were asked whether a doctor had
ever told them that they had hypertension, diabetes, cancer, respiratory problems,
heart conditions, stoke, psychiatric impairment, or arthritis. Each condition was
coded into an indicator variable (yes = 1).
Within HRS, depression was measured using a shortened (9-item) version
of the Center for Epidemiologic Studies Depression Scale (CES-D). Responses to
items were reversed when appropriate, and recoded into dichotomous variables.
Items were then summed into a continuous variable, such that scores represented
the number of items that were answered in the direction indicating depression—
that is, higher scores indicated more depressive symptoms (range 0 to 9).
88
Hearing status was assessed in the HRS with an item that asked
respondents to rate whether their hearing was excellent, very good, good, fair, or
poor when using prescribed hearing aids as usual. This variable was recoded into
a single indicator variable that compared participants with poor and fair hearing
against a reference group of participants with good or better hearing.
Medications and Alcohol. Participants who reported experiencing any of the
8 medical conditions were subsequently asked whether they were currently taking
any medications to address those health problems. Participants who reported
having heart problems were also asked whether they were taking medications for
heart attack, angina, or congestive heart failure. As a measure of the number of
medications used by participants, items were summed into a continuous variable.
A total possible of 11 medications could be reported.
Alcohol use was quantified by dichotomizing an item that captured the
number of drinks per day that participants consumed. If participants reported
having 2 or fewer drinks per day, then they were categorized as non/rare drinkers,
versus a reference group who had 3 or more drinks per day (drinker = 1).
Dependent variable
Participants in the 2006 wave of the HRS were asked whether they had
fallen in the last two years. Responses to this item were recoded and analyzed as
a dichotomous variable (did fall = 1).
Analyses
Analyses for study 2 were conducted using Statistical Package for Social
Sciences (SPSS), version 15.0 for Windows. Basic descriptive statistics, including
89
means, standard deviations, and percentages were calculated for selected
sociodemographic variables, vision status, home modification, measures of
physical function, health status, and falls. Independent sample t-tests were
conducted to determine statistical differences between women and men. Eta
square (η
2
) was calculated as an index of effect size to determine the practical
significance of observed statistical differences. In accordance with
recommendations by Green, Salkind, and Akey (2000), η
2
cutoffs of 0.01, 0.06,
and 0.14 were used to denote small, medium, and large effect sizes, respectively.
Main analyses consisted of two binary logistic regressions, differentiated by
gender. Separate analyses were performed for women and men based on
differences found in demographic, and health variables (reported in tables 3.1 and
3.2, below). Further justification for separate analyses is provided by literature that
reports fundamental differences between sexes, with respect to the types of falls
experienced (O'Neill et al., 1994), incidence of falling (Bath & Morgan, 1999), and
differences in the muscle and bone structures of men and women (Frontera,
Hughes, Lutz, & Evans, 1991; Hannan et al., 2000). Physiological differences
between sexes portend different mechanisms that could be responsible for falls in
women and men.
Originally, regression models contained variables controlling for number of
medications and alcohol consumption (not shown); however, when these models
were run, the variables did not statistically predict falling, and the likelihood ratio
tests for models including these variables were not significantly different from
90
models in which the variables were not included. Consequently, the variables were
dropped from analyses, and the regressions were rerun.
Independent variables were entered hierarchically in 4 blocks. In model 1,
sociodemographic variables, vision status, and home modifications variables were
entered, followed by measures of physical functioning (model 2), health and
medical conditions (model 3), and finally, interaction terms (model 4). Although the
likelihood ratio tests indicated that models including interaction terms were not
significant, they were retained in tables for the sake of discussion. Because tests of
significance were likely to be biased by large sample sizes, the Nagelkerke
pseudo-R
2
— a logistic analogy to R
2
in ordinary least squares (OLS) regression—
was reported for each model as an index of effect size.
D. Results
Descriptive statistics for selected sociodemographic variables and results of
independent t-tests by gender are presented in table 3.1. The mean age of
participants in the sample was 74.1 years (SD = 6.6). Women were statistically
older than men (74.4 vs. 73.7 years, respectively). Men had more education than
women (12.7 vs. 12.3 years); and men were statistically more likely than women to
live with a spouse or partner (74.8% vs. 46.3%). Whereas 90.6% of the sample
was White, 6.8% were Black, and 2.5% identified themselves as some other race.
According to standards cited by Green et al. (2000), effect sizes leading to
observed statistical differences of sociodemographic variables were very small for
age (η
2
= 0.002), small for education (η
2
= 0.01), and medium for coresidence
status (η
2
= 0.081).
91

Table 3.2 displays descriptive statistics, and results of independent sample
t-tests conducted to compare genders on the dependent variable (falls), as well as
key explanatory variables and control variables. In total, about a third of the
participants reported falling in 2006, since their last interviews in 2004. Women
were statistically more likely than men to fall, t = 4.26, p < .0005, although this
result was based on a very small effect size (η
2
= 0.002). With respect to vision
status, women were statistically more likely than men to rate their vision as poor, t
= 4.01, p < .0005; whereas men were more likely to rate their vision as good or
better, t = -2.30, p <= .05. There was no statistical difference between percentages
of men and women who rated their vision as fair. Effect sizes leading to differences
in vision status were very small. Only 13.9% of the sample reported that they had
modified their homes to accommodate disabilities. A statistically greater
percentage of women than men had made home modifications (15.0% vs. 12.3%),
Table 3.1: Demographic Descriptive Statistics by Gender (HRS, 2004,
Weighted)
Women
(n = 4,950)
Men
(n = 3,500)
Total
(N = 8,450)
t (df) p
η
2

Age 74.4 (SD = 6.7) 73.7 (SD = 6.5) 74.1 (SD = 6.6) 4.44 (8448) *** 0.002
Education 12.3 (SD = 2.7) 12.7 (SD = 3.2) 12.5 (SD = 2.9) -6.49 (8431) *** 0.005
Race
White 89.9% 91.7% 90.6% -2.71 (8448)
Black 7.5% 5.8% 6.8% 2.97 (8448)
Other 2.6% 2.5% 2.5% 0.26 (8448)
Spouse/partner 46.3% 74.8% 58.1% -27.33 (8442) *** 0.081
*** p < .0005
92
t = 3.61, p < .0005; however the effect size leading to this difference was very small
(η
2
= 0.002).
Differences in functioning between women and men were found in baseline
(2004) measures. Compared to men, women reported statistically more difficulties
with ULF, t = 17.56, p < .0005; and LLF, t = 15.62, p < .0005. Eta square values
(0.036 and 0.029, respectively) indicated that these differences were based on
small effect sizes. By comparison, no statistical differences were found in ∆ULF or
∆LLF between genders.
There were significant differences between women and men with respect to
health and medical conditions. Women were statistically more likely to report the
presence of hypertension, t = 4.05, p < .0005; psychiatric disorders, t = 10.88, p <
.0005; and arthritis, t = 12.87, p < .0005. Women also reported experiencing
greater depressive symptoms, per the modified CES-D instrument than men, t =
10.82, p < .0005. In contrast, men were statistically more likely to report having
diabetes, t = -4.50, p < .0005; cancer, t = -3.80, p < .0005; heart problems, t =
-9.74, p < .0005; stroke, t = -2.47, p < .05; and hearing impairments, t = -12.81, p <
.0005. All of the statistical differences between women and men, with respect to
health and medical conditions were based on small or very small effect sizes.
93

Table 3.2: Percent Falls, Vision Status, Home Modifications, Health Status, and
Mean Number of Upper- and Lower-Limb Disabilities by Gender (HRS, 2004 &
2006, Weighted)
Women
(n = 4,950)
Men
(n = 3,500)
Total
(N = 8,450)
t (df) p
η
2

Fall 34.7% 30.2% 32.9% 4.26 (8192) *** 0.002
Vision
Poor 6.1% 4.1% 5.2% 4.01 (8442) *** 0.002
Fair 15.5% 15.6% 15.6% 0.11 (8442) ns
Good 78.3% 80.4% 79.2% -2.30 (8442) * 0.001
Home Mod. 15.0% 12.3% 13.9% 3.61 (8410) *** 0.002
Baseline ULF 1.53 (SD = 1.56) 0.96 (SD = 1.26) 1.29 (SD = 1.28) 17.56 (8207) *** 0.036
Baseline LLF 2.19 (SD = 1.79) 1.57 (SD = 1.71) 1.93 (SD = 1.93) 15.62 (8082) *** 0.029
∆ ULF -0.18 (SD = 1.25) -0.15 (SD = 1.10) -0.17 (SD = 1.19) -1.41 (8207) ns
∆ LLF -0.23 (SD = 1.11) -0.20 (SD = 1.12) -0.22 (SD = 1.12) -1.29 (8067) ns
Hypertension 60.7% 56.3% 58.8% 4.05 (8439) *** 0.002
Diabetes 16.3% 20.1% 17.9% -4.50 (8434) *** 0.002
Cancer 16.4% 19.6% 17.7% -3.80 (8427) *** 0.002
Respiratory 9.9% 10.7% 10.3% -1.21 (8439) ns
Heart 25.3% 35.0% 29.3% -9.74 (8436) *** 0.011
Stroke 8.0% 9.5% 8.6% -2.47 (8440) * 0.001
Psychiatric 15.1% 7.4% 11.9% 10.88 (8438) *** 0.014
Arthritis 72.5% 59.2% 67.0% 12.87 (8436) *** 0.019
Hearing 19.4% 31.4% 24.3% -12.81 (8442) *** 0.019
Depression 2.0 1.5 1.77 10.82 (8339) *** 0.014
*** p < .0005 * p <= .05

Table 3.3 displays the main effects and interactions of vision status, home
modifications, and functional status for women. In model 1, which contained
demographic, vision, and home modification variables, poor vision (compared to
94
good or better vision) was a significant predictor of falls, OR = 1.737, p < .0005, as
was fair vision, OR = 1.299, p < .005. In this first model, home modifications were
also a significant predictor of falls in women; those who lived in homes which had
been modified were about 25% more likely to fall than those whose homes had not
been modified.
In model 2, when baseline and change measures of functioning were
included in the equation, the effects of vision status and home modifications were
reduced; however, baseline measures of both ULF and LLF were strong predictors
of falls. For each additional upper-limb activity with which women reported having
difficulties at baseline, the likelihood of a fall increased by about 21%, p < .0005.
Each additional lower-limb activity increased the likelihood of a fall by about 16%, p
< .0005. Change in functioning between 2004 and 2006 also was a strong
predictor of falls. For each upper-limb functional activity that participants lost the
ability to perform during the 2-year time period, the likelihood of falling increased
by about 10%, p < .0005. Similarly, each additional lower-limb functional activity
lost was associated with about a 12% increase in fall likelihood, p < .0005. Upper-
and lower-limb functioning remained strong predictors in models 3 and 4, when
health status variables and interaction terms were added to equations,
respectively. In the final model, baseline measures of functioning were highly
significant. Whereas the effect of ∆ULF became somewhat weaker (though it
remained significant), the effects ∆LLF became stronger after health status, and
interaction terms were controlled. Other significant predictors of falls in women in
the final model were diabetes, OR = 1.224, p <= .05; psychiatric impairments, OR
95
= 1.297, p <= .05; and depression, OR = 1.055, p < .005. The interaction between
poor eyesight and ∆ULF was also statistically significant, OR = 0.816, p <= .05;
however, likelihood ratio test indicated that this model was not significantly
different from model 3. From models 1 and 4, Nagelkerke’s R
2
increased from
0.031 to 0.109, suggesting good predictive efficacy of the key explanatory
variables.
Table 3.4 displays the main effects and interactions of vision status, home
modifications, and functional status for men. The pattern of results is similar to that
of women with a few notable exceptions. As in women, poor vision and fair vision
were associated with statistically greater likelihood of experiencing a fall for men,
in model 1. Compared to men who reported good or better vision, those with poor
vision were almost twice as likely to experience a fall, OR = 1.980, p < .0005;
whereas men with fair vision were about 36% more likely to experience a fall, p <
.005. However, the effects of vision disappeared in model 2, after upper- and
lower-limb functional measures were controlled. Home modifications were also a
significant predictor of falls for men in model 1. Those who reported living in homes
with home modifications were about 51% more likely to report a fall than those
without home modifications. This result carried through to the final model, when all
other variables were added to the equation. After controlling for ULF and LLF,
health status, and interaction terms, home modifications were still a significant
predictor of falls for men, OR = 1.319, p <= .05.
96

Table 3.3: Women: Odds-Ratios Representing Independent and Moderating
Effects of Key Explanatory Variables and Control Variables (HRS, 2004 & 2006,
Weighted; N = 4,950)
Independent Variable Model 1 Model 2 Model 3 Model 4
Demographics
Age 1.031*** 1.022*** 1.024*** 1.024***
Education (years) 1.010 1.035* 1.043** 1.043**
Race (White, ref.)
Black 0.776 0.700* 0.708* 0.714*
Other 0.845 0.852 0.881 0.882
Spouse/partner (no, ref.) 0.028 0.898 0.934 0.933
Vision (good vision, ref)
Poor Eyesight 1.737*** 1.219 1.089 1.014
Fair Eyesight 1.299** 1.057 0.989 1.004
Home Modifications 1.252* 1.068 1.043 1.043
Baseline ULF 1.214*** 1.165*** 1.162***
Baseline LLF 1.164*** 1.122*** 1.123***
∆ ULF 0.899*** 0.911** 0.930*
∆ LLF 0.878*** 0.891*** 0.876***
Health (no, ref)
Hypertension 1.032 1.030
Diabetes 1.231* 1.224*
Cancer 1.025 1.029
Respiratory 1.013 1.019
Heart Condition 1.155 1.158
Stroke 1.110 1.103
Psychiatric impairment 1.297* 1.297*
Arthritis 1.175 1.175
Depression 1.053** 1.055**
Hearing impairment 1.004 1.001
Interaction Terms
Poor Eyesight by ∆ ULD 0.816*
Fair Eyesight by ∆ ULD 0.984
Poor Eyesight by ∆ LLD 0.967
Fair Eyesight by ∆ LLD 1.100
-2log-likelihood 5403.358 5193.866 5156.899 5150.818
Nagelkerke R-square 0.031 0.096 0.107 0.109
*** p < .0005 ** p < .005 * p <= .05

97
Table 3.4: Men: Odds-Ratios Representing Independent and Moderating
Effects of Key Explanatory Variables and Control Variables (HRS, 2004 & 2006,
Weighted; N = 3,500)
Independent Variable Model 1 Model 2 Model 3 Model 4
Demographics
Age 1.030*** 1.015* 1.020** 1.020**
Education (years) 0.980 1.009 1.018 1.018
Race (White, ref.)
Black 0.821 0.755 0.778 0.786
Other 0.944 0.909 0.937 0.938
Spouse/partner (no, ref.) 0.948 0.984 1.062 1.063
Vision (good vision, ref)
Poor Eyesight 1.980*** 1.352 1.133 1.193
Fair Eyesight 1.362** 1.027 0.929 0.929
Home Modifications 1.511*** 1.332* 1.316* 1.319*
Baseline ULF 1.311*** 1.239*** 1.243***
Baseline LLF 1.204*** 1.144*** 1.143***
∆ ULF 0.799*** 0.808*** 0.796***
∆ LLF 0.830*** 0.841*** 0.839***
Health (no, ref)
Hypertension 0.892 0.894
Diabetes 1.342** 1.339**
Cancer 0.803* 0.812*
Respiratory 1.028 1.030
Heart Condition 1.039 1.033
Stroke 1.175 1.167
Psychiatric impairment 1.607** 1.608**
Arthritis 1.123 1.119
Depression 1.114*** 1.114***
Hearing impairment 1.059 1.061
Interaction Terms
Poor Eyesight by ∆ ULD 1.131
Fair Eyesight by ∆ ULD 1.036
Poor Eyesight by ∆ LLD 1.176
Fair Eyesight by ∆ LLD 0.977
-2log-likelihood 3851.187 3623.004 3567.579 3565.365
Nagelkerke R-square 0.033 0.128 0.150 0.151
*** p < .0005 ** p < .005 * p <= .05

98
The effects of ULF and LLF were strong predictors of falls across models.
In model 4, baseline measures of functioning were highly significant, as were the
effects of ∆ULF, OR = 0.796, p < .0005; and ∆LLF, OR = 0.839, p < .0005. Other
significant predictors of falls in men in the final model were diabetes, OR = 1.339,
p < .005; psychiatric impairments, OR = 1.608, p <= .005; and depression, OR =
1.114, p < .0005. None of the interaction terms were statistically significant.
Between models 1 and 4, Nagelkerke’s R
2
increased from 0.033 to 0.151,
suggesting good predictive efficacy of the key explanatory variables.
E. Discussion
A key finding of this study replicates important previous findings by
Steinman (2008) — namely that effects of self-reported vision status in predicting
falls are reduced as a result of controlling baseline functioning of upper and lower
extremities. Furthermore, the current study extends earlier findings by showing
that declines and/or gains in functioning across short periods of time may also
supercede self-reported vision in predicting falls. However, consistent with
previous findings, little evidence was found for a moderating effect of self-reported
vision status on musculoskeletal health and functioning. Nevertheless, results
reported here are important because they suggest that the relative strength of
poor self-reported vision status may not be as good an indicator of fall risk in older
adults as might otherwise be assumed. According to these results, independent
measures of baseline functioning, as well as recent declines in functional status in
the upper- and lower-limbs are relatively more important predictors of falls. Thus,
preserving residual limb-functioning through exercise has the important potential
99
to improve muscle function (McCartney, Hicks, Martin, & Webber, 1995) and bone
strength (Wolff, van Croonenborg, Kemper, Kostense, & Twisk, 1999), while
improving balance and gait (Daley & Spinks, 2000), and reducing fall risk (Carter,
Kannus, & Khan, 2001).
A second key finding of these analyses involves the effects of home
modifications in predicting falls in older people. The presence of home
modifications was included in analyses with the expectation that older individuals
living in modified homes would experience decreased risk of falling. However,
results suggested that in the HRS sample, the opposite was true. For both women
and men whose homes were previously modified to accommodate disabilities, the
risk of falling was greater when no other functional or health measures were
controlled. This effect disappeared for women, but remained for men, when all
other predictors were included in the regression equations. Overall, this pattern of
results suggested that older individuals may primarily install home modifications
in response to health conditions or events which would otherwise reduce their
ability to live independently. Despite modifications to their homes, fall risk for
these individuals could be higher due to physical impairments and functional
difficulties, which initially prompted the installation of their home modifications. In
other words, with respect to findings reported here and elsewhere in the literature,
it is possible that presence of home modifications often predicts increased falling
because persons who have them are often older adults with disabling conditions
and impairments, or those who need home modifications the most.
100
As an alternative, a preventive approach to modifying homes of older
adults could focus on recommendation and provision of modifications universally,
and without regard to the perceived or objective needs of the consumer, based on
their physical health and functioning. Indeed, in some studies when home
modifications have been found to reduce falls, the effect of the home environment
is stronger for more vigorous older adults, compared to more frail persons, who
are less likely to have exposure to hazards when they exist (Lord et al., 2006).
While there are only a limited number of randomized intervention studies focusing
on the environment, there is evidence that home modifications can have an
important impact in reducing falls; however, benefits of home modifications are
most likely to be seen in combination with other fall-prevention strategies, and in a
multifactorial interactive manner. Professional assessment, education, and
installation of home modifications, when part of an integrated risk-management
intervention, can be effective for improving functioning, decreasing fear of falling,
and reducing the incidence of falls among older adults in the community (Close et
al., 1999; Cumming et al., 1999; Day et al., 2002; Gillespie et al., 2003; Gitlin et
al., 2006). In conjunction with other proven fall-prevention interventions, specific
supportive features, such as grab bars in bathrooms, handrails on both sides of
staircases, non-slip flooring, and adequate lighting (which provide universal
benefits for all users regardless of age, vision status, or physical functioning),
could potentially help to prevent falls and promote safety, even in homes of the
most vigorous and healthy older adults.

101
Limitations
Although the results of this study are thought to accurately reflect important
relationships between falls and various risk factors that could lead to falls, some
limitations of the study should be acknowledged. First, there are several
difficulties with the measurement of the dependent variable. To assess falls, HRS
asked participants if they had fallen in the last 2 years. This extended period is
troublesome because of the difficulties that participants may have accurately
recalling falls that occurred in that duration (Ganz, Higashi, & Rubenstein, 2005).
In an attempt to address issues of memory, participants with cognitive
impairments were excluded, who may be more likely to experience a fall, but less
likely to report falls accurately due to forgetting. Thus, results potentially represent
a conservative bias, which may be less generalizable with respect to the true
number of fallers in the older population at large. In addition, the HRS item does
not define what is meant by a fall. One commonly accepted definition is that set
forth in the International Classification of Diseases (ICD-10) that calls a fall “an
unexpected event where a person falls to the ground from an upper level or the
same level” (Masud & Morris, 2001). Nevertheless, no definition was given to
HRS participants, so it can not be assured that all participants interpreted the item
in the same way. Therefore, problems with item validity could have resulted in
inconsistent response patterns between participants that would make it difficult to
compare this study with findings from similar studies where consistent definitions
were provided. Other fall-related questions that were asked in the HRS study
pertained to the number of falls, and whether participants were injured in their
102
falls. These items were not analyzed because low response rates and missing
data would have resulted in reduced statistical power.
Similarly, a second limitation of the study relates to the subjective nature of
the key explanatory variables. Participants were asked to self report the status of
their vision and their functional capacity, thus inviting subjectivity into responses
that could influence the validity of items, and introduce error into analyses.
Although results based on self-reported measures are telling about the
participant’s perceptions of their situation, it is likely that a similar study that used
clinical measures of vision and musculoskeletal functioning would report stronger
or weaker relationships between variables relative to the results reported here.
Another problem concerning key independent variables is the
nonspecificity of the item regarding home modifications. Whereas the HRS
question used in this study asked whether modifications to homes had been made
in the previous two years, a more important query involves the types of
modifications that were made. The range of home modifications is large, and
varies with respect to cost and scale. It is possible that participants of HRS biased
their responses to include only major (or more expensive) home modifications,
such as ramps, widening doorways, or remodeling bathrooms, while other less
expensive home modifications such as grab bars and improving lighting may be
overlooked. Although it is not certain whether participants biased their responses
to exclude smaller, but important home modifications, such a bias could result in a
weaker effect, with respect to preventing falls.
103
Finally, a third limitation of this study relates to the choice of independent
control variables. Although models introduced here revealed important
statistically significant relationships between variables that predicted falls, it is
noteworthy that the amount of variance accounted for by the models was
relatively small. The modest explanatory power of even the best models (between
10% and 15%) suggests that some important variables that predict falls were
absent from analyses. In comparison to previous research that employed similar
control variables (Steinman, 2008), greater predictive power was found in the
current study when the presence of home modifications and change in functioning
were controlled. However, an even more rigorous analysis would identify and
include other potentially important predictor variables, such as activity-levels,
participation in exercise programs, and severity of health problems experienced
by participants, that would improve results and could add support to the
hypotheses of this line of study. Nevertheless, results are telling about the relative
impact of self-reported vision and its relationship with other explanatory variables
in causing falls. Further research should continue to explore interactive
relationships between variables that are assumed to have important roles in
determining whether or not older adults experience falls.
Conclusion
The practical implications of this study merit some discussion. While
certainly, older adults with visual impairments can and do benefit from vision
rehabilitation techniques that train consumers to use their residual vision
effectively to respond to fall hazards, and by creating home environments that are
104
safer with respect to fall prevention— results reported here also suggest the
importance of maintaining good physical fitness as a high priority for older people
with visual impairments. Physical fitness, especially in the lower limbs is important
for recovering and avoiding a fall after a misstep. Similarly, fitness in upper limbs
is important for maintaining balance, and for avoiding injury after a fall.
Nevertheless, getting adequate exercise— whether through daily functional
activities, leisure and/or social activities; or through participation in exercise
programs— is likely to be more difficult for older adults with visual impairments
due to the independent effects of vision loss in day-to-day life. The independent
effects of vision loss, which may make mobility more difficult, are also likely to
reduce the practicality of attending rehabilitation/exercise programs, unless
support mechanisms are provided to meet special needs of older adults who are
visually impaired. Furthermore, exercise programs that are developed for older
people with visual impairments can be enhanced by taking into consideration
differences between sexes, with respect to the types of chronic diseases and
other factors that affect people differently as they age. Thus, this study has the
potential to inform developers of health/functioning programs for older adults with
visual impairments, insofar as programs will take into consideration the unique
circumstances and needs of individuals within the older population.

105
CHAPTER IV: MEDIATING EFFECTS OF DISABILITY AND FUNCTIONAL
STATUS IN PREDICTING FALLS IN OLDER PEOPLE WITH POOR VISION
A. Introduction
Age-related vision loss is widely acknowledged to be an important risk
factor for falls in older adults. Many studies that have examined the association
between vision impairment and fall risk have assessed total effects— most
commonly those associated with decrements in specific measures of vision, such
as acuity, visual fields, binocularity, and contrast sensitivity. A comparatively
modest number of studies have examined the contributions of integrated (i.e., the
moderated and/or mediated) effects of vision loss on and by other systems as they
relate to falls in late life. As suggested in Steinman (2008), it is possible that vision
loss could moderate declines in other systems, as when age-related losses in
limb-function are magnified synergistically due to decreased physical activity that
may accompany disability associated with vision impairment. Losses in limb
function due to reduced activity levels have the potential over time to result in less
stability and a greater likelihood of falling in old age.
A second means by which vision could be integrated with other systems is
by way of a mediated effect, in which effects of vision loss on falls are contingent
on declines in other systems. Mediating variables are said to intervene between
independent variables and dependent variables to produce an outcome that is
specifically related to the mediator (Muller, Judd, & Yzerbyt, 2005). For example,
an older individual with reduced lower limb functioning may be less likely to recover
when concomitant losses in his central visual field cause him to trip over objects in

106
the environment. In this scenario, the total effect of vision can be partialed into
effects related to vision loss, and visual effects that are contingent on losses in
functioning. Figure 4.1 depicts a potentially mediated relationship between vision
and falls by way of losses in the musculoskeletal system.

Figure 4.1: Total Effect of Vision on Fall Risk (panel A); and Total Direct Effect
Mediated by Muscle Decline (panel B) (Adapted from Baron & Kenny, 1986).

The usual way for determining whether a mediating relationship exists is to
quantify the relationships in figure 4.1 using unstandardized regression
coefficients. According to Baron and Kenny (1986), in order for a mediating
relationship to be established, four steps must be followed. First, it must be shown
that there is an effect that can be mediated—this is done by establishing a direct
effect between the independent variable (vision loss) and the dependent variable

107
(falls). Panel A (figure 4.1) shows an unmediated model, in which an initial variable
(vision) is associated directly with an outcome (a fall). The total effect of vision on
fall risk, when mediating variables are not considered is represented by line c.
Second, it must be shown that vision is correlated with the mediator (in this case,
muscle loss). This step involves treating muscle loss as an outcome variable that is
predicted by vision loss (depicted as line a in panel B). Next, the mediator (muscle
loss) must be shown to affect the outcome variable (falls), independent of vision
(depicted as line b in panel B). In this step, if vision is not controlled, then the
relationship between muscle loss and falls could be attributed to the correlation
between vision loss and muscle loss. Panel B of figure 4.1 shows the original
relationship between vision and falls when mediated by the intervening variable
(muscle loss). Finally, to establish whether the relationship is fully, partially or
unmediated, the direct effect of vision loss on falls controlling for the potential
mediating effects of muscle loss (shown as line c’ in panel B), is compared to the
total direct effect (line c, in panel A). Line c’ also corresponds to what are referred
to as independent effects above (see figure 3.1). If all relationships are significant,
then the mediated effect is the difference between lines c and c’, or the product of
lines a and b.
B. Purpose of Study
The purpose of this study was to examine mediating effects of six measures
of functioning and disability on vision, as it relates to falls in older adults. To
address this issue, the study relied on the steps outlined by Baron and Kenny
(1986). In contrast to study 2 which tested whether vision loss moderates change

108
in ULF and LLF, the current study tested the degree to which limb functioning
contributes to the total effect of vision loss in predicting falls. As a theoretical
framework, variable selection and analyses for this study were guided by the
epidemiological model of the disability process that was originally discussed by
Nagi (1965; 1976) (described extensively above). The model consists of four main
stages that progress from (1) risk factors, through (2) pathology/impairments to
result in (3) losses in mobility and functioning that eventually lead to (4) disability in
domain specific age-appropriate daily-living activities (see figure 2.1). Disparities
in these dimensions between older persons with vision impairment and those with
good vision are described in study 1. The current study employed longitudinal data
to address whether physical functioning and disability mediate the effects of vision
in predicting whether or not older adults experience a fall, with controls for
covariates related to sociodemographic risk factors, pathological conditions, as
well as functional ability, and disability. In addition, the goal of this study was to
begin to describe a series of prospective pathways by which vision could
interrelate with disease, functional limitations and disability to increase fall risk.
C. Methods
Data
Data for this study were derived from two waves (2006 and 2008) of the
University of Michigan Health and Retirement Study (HRS), including variables
that were cleaned and compiled by the RAND Corporation (2010). Variables
representing sociodemographic traits, chronic conditions, functional and disability
status, and vision were taken from the 2006 wave of HRS. Subsequent measures

109
representing fall status were acquired from the 2008 wave of HRS. In 2006, the
study had 19,740 participants who were known to be living (attrition of roughly 13%
compared to HRS baseline in 1998, and 23% overall). Participants of HRS were
included for analyses in the current study if they were 65 years or older; if their
interviews were not conducted by proxy; and if they were alive during both waves
considered. With this selection criteria applied, a weighted sample of 11,080
participants remained for analyses.
Analyses
Predictive Analytics Software (PASW; formerly SPSS), version 18.0 for
Windows was used for all analyses. Basic descriptive statistics were computed,
including means, standard deviations, and percentages for selected
sociodemographic and health variables, including age, race, educational
attainment, marital status, aggregate health conditions, self-reported vision status,
six indices of functioning and disability (including mobility, large muscle, gross and
fine motor functioning, activities of daily living (ADL), and instrumental activities of
daily living (IADL)), and the main outcome variable (whether participants reported
a fall in the 2008 wave). Univariate analyses of variance (ANOVA) were conducted
to evaluate relationships between vision status and these variables. The
independent variable in these ANOVAs, included three levels of self-reported
vision: good, fair, and poor. When appropriate, follow-up tests were conducted to
evaluate pair-wise differences among the means. In each analysis, variances
among the three vision status groups were not homogeneous; therefore, Dunnett’s
C assessment of comparisons, a test that does not assume equal variance among

110
groups, was used in post-hoc tests (Green, Salkind, & Akey, 2000). Partial eta
square (η
2
), an index of effect size, was calculated to determine the practical
significance of statistical differences found. According to Green et al., η
2
values of
.01, .06, and .14 represent small, medium, and large effect sizes, respectively.
Main analyses consisted of a series of regressions designed to test the
steps outlined by Baron and Kenny (1986) to show mediation between vision and
fall variables. Because scaling and distributions of the key dependent variables
differed, both Poisson regression and logistic regression were used. Specifically,
models were designed to test whether the proposed mediators (indices of
functioning) exercised their influence on the independent variable (self-rated vision
in 2006) with respect to its effect on the dependent variable (falls in 2008). Thus,
Poisson regression was used to test the relationship between vision and mediating
count variables; whereas logistic regression was used to test the effects of vision
and the mediators on the key dichotomous outcome variable, whether or not
individuals had fallen. Because analyses were based on different scales,
regression coefficients were made comparable using formulas described by
MacKinnon and Dwyer (1993). These comparable coefficients were then used to
compute Sobel’s test statistic which compares the strength of the indirect effect to
the null hypothesis that the effect equals zero (Preacher & Hayes, 2004).
Figure 4.2 displays theoretical relationships between key variables and the
waves from which variables were derived for six different analyses conducted for
men and women. According to the proposed theoretical model, a total effect
(depicted by line c) is made up of a direct effect of vision on falls (line c’), plus the

111
total indirect effects via the functioning and/or disability variables. Indirect effects
were calculated separately for mobility, large muscle, gross motor, and fine motor
functioning, as well as ADLs and IADLs. The total effect of prospective mediators
was assessed by computing the products of comparable nonstandardized
regression coefficients representing the effects of vision on functioning/disability
(paths a
1…
a
6
) and comparable regression coefficients representing the effects of
the functioning/disability on falls (paths b
1
… b
6
).
Figure 4.2: Theoretical Tests of Mediated Relationships between Self-Reported
Vision Status, and Falls (Dashed lines indicate separate analyses)

112

Statistically significant indirect effects, calculated based on the products of
a
i
and b
i
, represent the total mediating effect, and should be equal to c – c’ in each
analysis. In addition to the functional/disability indices, each model also controlled
covariates that have theoretical significance with respect to Nagi’s (1965; 1976)
model— namely sociodemographic risk variables, and number of chronic health
conditions. The theoretical basis for conducting separate analyses for men and
women is provided by the literature that shows important differences between men
and women, with respect to the types of falls experienced (O'Neill et al., 1994),
incidence of falling (Bath & Morgan, 1999), and differences in the muscle and bone
structures of women and men (Frontera, Hughes, Lutz, & Evans, 1991; Hannan et
al., 2000). In addition, results from study 2 (above) support the hypothesis that fall
risk is experienced differently by the sexes, and that different mechanisms could
be responsible for falls in women and men.
Variables in Models
Covariates. Five sociodemographic risk variables, including age, race,
presence of a spouse or partner in the household, years of education attainment
and number of health conditions at baseline, were statistically controlled. Age and
education attainment were coded as continuous variables; whereas race, and
spouse/partner status were coded as indicator variables with “White” and
“spouse/partner present” as reference categories, respectively. Gender was
controlled on the basis of performing separate analyses for men and women.

113
A variable representing the number of chronic health conditions was
computed by summing the conditions that participants reported having, as
diagnosed by their doctor. Participants were asked whether they had ever been
told by a doctor that they had high blood pressure, diabetes, cancer, lung
problems, heart problems, stroke, psychiatric problems, or arthritis. When
summed, values for this aggregate continuous variable could range from 0 to 8.
Functional/disability indices (described below) that were not being tested as
mediators in each equation were also included as covariates.
Mediators. Summary indices of physical functioning and disability were
taken from the RAND (2010) HRS data set, and are based on factors compiled by
Wallace and Herzog (1995). Six indices represent major areas of functioning,
including upper and lower body mobility, strength, and daily living activities.
Specifically, the mobility index included variables assessing the participant’s
difficulty walking one block; walking several blocks; walking across a room;
climbing one flight of stairs; and climbing several flights of stairs. The large muscle
index included items that assessed the participant’s difficulty sitting for 2 hrs;
getting up from a chair; stooping, kneeling or crouching; and pushing or pulling
large objects. The gross motor index assessed the participant’s difficulty with
walking one block; walking across a room; climbing one flight of stairs; getting in or
out of bed; and bathing. Finally, the fine motor index assessed difficulty with
picking up a dime; eating; and dressing. With respect to disability, the ADL index
was a summation of three items proposed by Wallace and Herzog, including
bathing; dressing; and eating. The IADL index was composed of items measuring

114
participants’ difficulty using the phone; managing money; managing medications;
shopping and preparing meals. In all waves the "some difficulty" versions of
individual measures were used in constructing indices, such that participants were
coded in the positive direction if they indicated any difficulty, or being unable to
perform each task. Each limitation added one to the summary measure, such that
mobility could range from 0 to 5; large muscle index could range from 0 to 4; gross
motor index could range from 0 to 5; fine motor index could range from 0 to 3; ADL
ranged from 0 to 3; and IADL ranged from 0 to 5.
Vision. Self-reported vision status was assessed in the HRS by asking
participants to state whether their vision was excellent, very good, good, fair, or
poor when wearing glasses or corrective lenses. Participants could also state that
they were legally blind. This variable was recoded into a ordered-categorical
variable with 3 levels— one combining persons with self-reported good or better
vision, a second composed of persons with fair vision, and a third composed of
persons with poor vision and legal blindness.
Dependent variable. The key dependent variable of this study was whether
or not participants had fallen in the two years prior to the 2008 wave of the HRS.
Participants were asked whether they had fallen in the past 2 years (since the last
wave of the study). Responses to this item were recoded and analyzed as a
dichotomous indicator variable (yes = 1).
D. Results
Descriptive statistics and results from one way univariate ANOVAs
comparing sociodemographic variables, number of diseases, and falls by vision

115
status are displayed in table 4.1. In addition, results of follow-up tests, which
evaluated pair-wise differences between vision groups are also displayed.
Dunnett’s C test was used for post-hoc tests, because it allows comparisons when
variances among the three vision status groups are heterogenous. Overall, the
average age of participants was 73.7 years (SD = 6.8). The mean age of
participants who had poor vision ( x = 77.5; SD = 7.9) and fair vision ( x = 74.5; SD
= 7.1) were significantly different from the mean of participants who reported good
or better vision ( x = 73.3; SD = 6.6); F (2, 11,077) = 125.3, p < .0005. The strength
of the relationship between vision status and age was assessed by partial η
2
.

About 2% of the variance in the dependent variable age was related to vision
status—a small effect size, according to criteria cited by Green et al. (2000).
Around 58% of the total sample was composed of women, with statistically more
women reporting poor vision than reporting fair or good vision; F (2, 11,077) = 4.9,
p < .05. Eta Square (η
2
=

.001) indicated that less than 1% of variance in gender
was accounted for by vision status. The vast majority of the sample (90%) was
white; 7% were black, and 3% reported some other race. Significantly greater
percentages of Blacks reported fair or poor vision; F (2, 11,077) = 38.9, p < .0005.
Conversely, a significantly greater number of Whites reported good vision F (2,
11,077) = 45.0, p < .0005. The strength of relationships between vision status and
white and black race, as measured by partial η
2
(0.008 and 0.007, respectively)
was small. Fifty-seven percent of the sample reported that they were either
married or partnered during the 2006 wave of HRS. A larger percentage of persons
with good or better vision (59%) reported being married or partnered, compared to

116
persons with poor vision (42%); F (2, 11,077) = 49.6, p < .0005; however this
difference was based on a small effect size (partial η
2
= .009). Finally, the average
years of education obtained among the sample was 12.4 (SD = 3.1) years.
Compared to participants with good vision ( x = 12.8; SD = 2.9), those who
reported fair vision ( x = 11.3; SD = 3.5) and poor vision ( x = 10.6; SD = 3.7) had
significantly fewer years of education; F (2, 11,051) = 280.7, p < .0005. The effect
size, as assessed by η
2
, indicated that around 4.8% of the variance in education
was associated with vision status. On average, respondents reported 2.4 (SD =
1.4) chronic diseases; persons with poor vision reported statistically more chronic
health conditions ( x = 3.3; SD = 1.7) than those with fair vision ( x = 2.7; SD = 1.5)
and good vision ( x = 2.3; SD = 1.3). Partial η
2
indicated that around 4% of
variance in health conditions was accounted for by vision status. With respect to
the study’s dependent variable, whether participants had fallen between the 2006
and 2008 waves, older persons with poor vision reported statistically more falls
than persons with fair or good vision (F (2, 11,052) = 50.1, p < .0005). On average
more than half (54%) of those with poor vision reported a fall, compared to 36% of
those with good vision and 43% with fair vision, though this finding was based on a
small effect size (partial η
2
= .009).
Table 4.2 displays descriptive statistics for indices of functioning/disability
by vision status, as well as results of univariate ANOVAs conducted to test overall
significance and pair-wise differences between vision groups. On average,
participants reported 1.2 (SD = 1.5) mobility-related difficulties; 1.4 (SD = 1.3) large

117
muscle-related difficulties; 0.5 (SD = 1.0) gross motor related difficulties; and 0.2
(SD = 0.4) fine motor-related difficulties. With respect to disability indices, on
average participants reported difficulty with 0.2 (SD = 0.5) ADLs; and 0.2 (SD =
0.6) IADLs. Older adults with poor vision had the most difficulty with all five indices
of functioning and disability. On average, those with poor vision reported 2.2 (SD =
1.7) difficulties with mobility, compared to 1.6 (SD = 1.56) for persons with fair
vision, and 1.0 (SD = 1.4) for older persons with good vision (F (2, 11,077) = 313.8,
p < .0005; partial η
2
=

.054). Older adults with poor vision also reported a greater
number of difficulties with large muscle activities ( x = 2.1; SD = 1.3) compared to
those with fair vision ( x = 1.7; SD = 1.4) and good vision ( x = 1.2; SD = 1.2) (F (2,
11,047) = 215.9, p < .0005; partial η
2
=

.038). Participants with poor vision had
statistically more difficulty with gross motor activities ( x = 1.3; SD = 1.5) than
participants with fair ( x = 0.7; SD = 1.1) and good vision ( x = 0.4; SD = 0.9) (F (2,
11,077) = 325.4, p < .0005; partial η
2
=

.055). Finally, older persons with poor
vision on average were more likely to report difficulties with fine motor activities (F
(2, 11,045) = 205.6, p < .0005; partial η
2
=

.036). Compared to a mean of 0.2 (SD =
0.4) for older persons with good vision, and 0.3 (SD = 0.6) for persons with fair
vision; those with poor vision on average experienced difficulties with 0.5 (SD =
0.8) fine motor activities.

118
Subscripts represent Dunnett C pair-wise differences between vision categories, Significant at <=.05
Table 4.1: Descriptive Sociodemographic Characteristics, Number of Chronic Diseases, and Falls by Vision Status; Test
of Significance and Pair-Wise Comparisons between Vision Groups (HRS, 2006 & 2008, Weighted)
Variables All Test of Significant Differences Between Vision Groups
X SD Range n F df P Partial η2
Age 73.7 6.8 65-92 11,080 125.3 2, 11,077 <.0005 0.022
Female 58% 0.5 0-1 11,080 4.9 2, 11,077 .007 0.001
White 89% 0.3 0-1 11,080 45.0 2, 11,077 <.0005 0.008
Black 8% 0.3 0-1 11,080 38.9 2, 11,077 <.0005 0.007
Other 3% 0.2 0-1 11,080 5.7 2, 11,077 .003 0.001
Mar./Part. 57% 0.5 0-1 11,080 49.6 2, 11,077 <.0005 0.009
Education 12.4 3.1 0-17 11,054 280.7 2, 11,051 <.0005 0.048
Diseases
2.4 1.4 0 to 8 11,078 222.4 2, 11,047 <.0005 0.039
Fall
38% 0.5 0 to 1 11,055 50.1 2, 11,052 <.0005 0.009
Variables Good Vision Fair Vision Poor Vision

x SD Range n x SD Range n x SD Range n
Age 73.3 a 6.6 65-92 8747 74.5 b 7.1 65-92 1,711 77.5 c 7.9 65-91 622
Female 58% a 0.5 0-1 8,747 59% a 0.5 0-1 1,711 64% b 0.5 0-1 622
White 90% a 0.3 0-1 8,747 84% b 0.4 0-1 1,711 82% b 0.4 0-1 622
Black 7% a 0.3 0-1 8,747 12% b 0.3 0-1 1,711 14% b 0.4 0-1 622
Other 3% 0.2 0-1 8,747 4% 0.2 0-1 1,711 5% 0.2 0-1 622
Mar./Part. 59% a 0.5 0-1 8,747 51% b 0.5 0-1 1,711 42% c 0.5 0-1 622
Education 12.8 a 2.9 0-17 8,728 11.3 b 3.5 0-17 1,705 10.6 c 3.7 0-17 621
Diseases
2.3 a 1.3 0 to 8 8,745 2.7 b 1.5 0 to 8 1,711 3.3 c 1.7 0 to 7 622
Fall
36% a 0.5 0 to 1 8726 43% b 0.5 0 to 1 1,709 54% c 0.5 0 to 1 620

119

Subscripts represent Dunnett C pair-wise differences between vision categories, Significant at <=.05.
Table 4.2: Descriptive Statistics for Indices of Functioning and Disability by Vision Status; Test of Significance and
Pair-Wise Comparisons between Vision Groups (HRS, 2006, Weighted)
Variables All Test of Significant Differences Between Vision Groups
X SD Range n F Df P Partial η2
Mobility
1.2 1.5 0 to 5 11,080 313.8 2, 11,077 <.0005 0.054
Large Muscle
1.4 1.3 0 to 4 11,077 215.9 2, 11,074 <.0005 0.038
Gross Motor
0.5 1.0 0 to 5 11,080 325.4 2, 11,077 <.0005 0.055
Fine Motor
0.2 0.5 0 to 3 11,078 205.6 2, 11,045 <.0005 0.036
ADL
0.2 0.5 0 to 3 11,078 234.5 2, 11,075 <.0005 0.041
IADL
0.2 0.6 0 to 5 11,078 646.1 2, 11,075 <.0005 0.104
Variables Good Vision Fair Vision Poor Vision

x
SD Range n
x
SD Range n
x
SD Range n
Mobility
1.0 a 1.4 0 to 5 8,747 1.6 b 1.6 0 to 5 1,711 2.2 c 1.7 0 to 5 622
Large Muscle
1.2 a 1.2 0 to 4 8,745 1.7 b 1.4 0 to 4 1,710 2.1 c 1.3 0 to 4 622
Gross Motor
0.4 a 0.9 0 to 5 8,747 0.7 b 1.1 0 to 5 1,711 1.3 c 1.5 0 to 5 622
Fine motor
0.2 a 0.4 0 to 3 8,745 0.3 b 0.6 0 to 3 1,711 0.5 c 0.8 0 to 3 622
ADL
0.1 a 0.4 0 to 3 8,745 0.2 b 0.6 0 to 3 1,711 0.5 c 0.8 0 to 3 622
IADL
0.1 a 0.5 0 to 5 8,745 0.3 b 0.7 0 to 5 1,711 1.0 c 1.4 0 to 5 622

120
Participants with poor vision also reported statistically more difficulties on disability
indices. With respect to ADLs, participants with poor vision reported difficulties with
0.5 (SD = 0.8) activities, compared to 0.2 (SD = 0.6) for those with fair vision, and
0.1 (SD = 0.4) for those with good vision (F (2, 11,075) = 234.5, p <
.0005; partial η
2
=

.041). Similarly, participants with poor vision reported
statistically more difficulty with IADLs (F (2, 11,075) = 646.1, p < .0005; partial η
2
=

.104); on average, persons with poor vision experienced difficulty with 1.0 (SD =
1.4) IADLs, compared to 0.3 (SD = 0.7) for those with fair vision, and 0.1 (SD = 0.5)
for those with good vision.
Table 4.3 displays partial correlation coefficients of key explanatory
variables and covariates, holding constant sociodemographic variables. A p-value
of less than .005 (.05/9) was required for significance to control for type 1 error
across six partial correlations. All coefficients were in the correct direction, and
statistically significant at p = .0005 or less. The strongest correlation with vision
was found with variables pertaining to IADLs (0.24), gross motor activities (0.18)
and mobility (0.17). In general, these partial correlations suggest that greater
degree of vision loss is associated with greater degrees of difficulty across indices.
Recall that statistically significant correlations are necessary between vision and
prospective mediating variables as a preliminary step in establishing a mediation
relationship (Baron & Kenny, 1986). These results indicated that there was
adequate justification to test for the possible role of mediators.

121

Table 4.3: Partial Correlation Matrix for Key Explanatory Variables and Covariates, Controlling for Age, Sex, Race,
Education and Marital Status (HRS, 2006 & 2008, Weighted)
Variables Vision Disease Mobility Large
Muscle
Gross
Motor
Fine
Motor
ADL IADL Fall
Vision 1.00
Diseases 0.15* 1.00
Mobility 0.17* 0.38* 1.00
Large
Muscle
0.14* 0.34* 0.56* 1.00
Gross Motor 0.18* 0.31* 0.84* 0.47* 1.00
Fine motor 0.15* 0.19* 0.37* 0.37* 0.45* 1.00
ADL 0.15* 0.19* 0.43* 0.35* 0.61* 0.77* 1.00
IADL 0.24* 0.20* 0.35* 0.25* 0.47* 0.39* 0.48* 1.00
Fall 0.08* 0.15* 0.19* 0.17* 0.17* 0.12* 0.11* 0.11* 1.00
*p < .0005

122

Table 4.4: Women: Comparable Regression Coefficients and Tests of Indirect Effects of Mediators, Controlling for
Sociodemographic Variables, Chronic Disease, Functioning, and Disability (HRS, 2006 & 2008, Weighted)
Women Path c
Total effect
(Vision to Falls)

Path c’
Direct effect
(Vision to Falls)

Path a
(Vision to
Mediator)
Path b
(Mediator to
Falls)
Indirect effect
(a x b)
Test of
Significance

Mediator Comp. Beta
(SE); prob.
Comp. Beta
(SE); prob.
Comp. Beta
(SE); prob.
Comp. Beta
(SE); prob.
Sobel (SE); prob.
Mobility .013 (.016); n.s. .012 (.016); n.s. .005 (.034); n.s. .077 (.032); * 0.15 (.003); n.s.
Large
Muscle
.013 (.016); n.s. .013 (.016); n.s. .005 (.004); n.s. .120 (.022); ** 1.24 (.0002); n.s.
Gross Motor .013 (.016); n.s. .013 (.016); n.s. .008 (.005); n.s. .057 (.034); n.s. 1.16 (.0004); n.s.
Fine motor .015 (.016); n.s. .013 (.016); n.s. -.044 (.108); *** .062 (.023); * 0.41 (.007); n.s.
ADL .013 (.016); n.s. .013 (.016); n.s. -.024 (.015); n.s. -.13 (.100); n.s. -1.02 (.001); n.s.
IADL .017 (.015);n.s. .012 (.015); n.s. .141 (.014); *** .025 (.018); n.s. 1.34 (.003); n.s.
***p < .0005; **p < .005; *p <= .05; n.s. = not significant

123

Table 4.5: Men: Comparable Regression Coefficients and Tests of Indirect Effects of Mediators, Controlling for
Sociodemographic Variables, Chronic Disease, Functioning, and Disability (HRS, 2006 & 2008, Weighted)
Men Path c
Total effect
(Vision to Falls)

Path c’
Direct effect
(Vision to Falls)

Path a
(Vision to
Mediator)
Path b
(Mediator to
Falls)
Indirect effect
(a x b)
Test of
Significance

Mediator Comp. Beta
(SE); prob.
Comp. Beta
(SE); prob.
Comp. Beta
(SE); prob.
Comp. Beta
(SE); prob.
Sobel (SE); prob.
Mobility .065 (.019); *** .063 (.018); ** .0003(.058); n.s. .106 (.036); ** 0.01 (.006); n.s.
Large
Muscle
.067 (.019); *** .063 (.018); ** .014 (.006); * .120 (.022); *** 1.98 (.001); *
Gross Motor -.063(.019); ** .063 (.019); ** -.004 (.008); n.s. .029 (.038); n.s. -0.39 (.0002); n.s.
Fine motor .063 (.019); ** .063 (.019); ** -.011 (.015); n.s. .051 (.031); n.s. -0.67 (.001); n.s.
ADL .062 (.019); ** .063 (.019); ** -.015 (.019); n.s. -.063 (.035); n.s. 0.73 (.001); n.s.
IADL .068 (.019); *** .063 (.019); ** .070 (.024); n.s. .040 (.021); n.s. 1.61 (.002); n.s.
***p < .0005; **p < .005; *p <= .05; n.s. = not significant

124
Finally, tables 4.4 and 4.5 shows results of regression analyses designed to
test mediating relationships, per the steps outlined by Baron and Kenny (1986) for
women and men, respectively. With respect to the women, no statistically
significant indirect effects (path a x b) between vision and falls were found for any
of the functioning/disability indices, suggesting that mediation by these variables
did not occur for women. In addition, no direct effects of vision on falls were found
when respective mediators and all covariates were in the models (path c’). This
finding is consistent with previous research, which showed that controlling for
covariates related to functioning reduced or eliminated the effects of self-reported
vision as it relates to falls (Steinman, 2008; Steinman et al., 2009). However, vision
was significantly associated with fine motor functioning, and IADL functioning
(path a); and mobility, large muscle, and fine motor activities were statistically
associated with falls (path b). Results for men suggested that large muscle
functioning was a statistically significant mediator between vision and falls (Sobel
statistic = 1.98; p < .05).
In addition, the direct effect (path c’) of vision on falls, after controlling
sociodemographic and functioning covariates, was statistically significant in all
analyses for men; thus, mediation of large muscle activities was partial in effect.
Finally, it is noteworthy that self reported vision was not a statistically significant
predictor of functional indices (path a) in five out of six of the tests; and that only
mobility and large muscle activities were statistically associated with falls (path b).

125
G. Discussion
A key premise of the research reported here, was that the disability process
described by Nagi (1965; 1976) and others may serve as a model to explain how
functional losses and disability that often follow vision loss may mediate some of
the effects of vision, with respect to its influence on fall risk. The results of this
study suggest that mediation effects of various measures of functioning and
disability are limited, and may differ by gender. For example, no mediation effects
were found for women, suggesting that no portion of the effects of self-reported
vision on falls is contingent on losses in the musculoskeletal system. In addition,
direct effects of self-reported vision were not significant, which suggests that some
other system provides the primary mechanism by which falls occur. Since all of the
models controlled for functional indices not being tested as mediators, it is
reasonable to assume that functional and disability measures, which operate as
proxy variables for musculoskeletal health, are important predictors of falls. For
women, this hypothesis was supported by significant paths leading from several
functional/disability indices—specifically, mobility, large muscle and fine motor
activities— to falls. Implications of these results point to the importance of
maintaining functioning across domains. This important goal can be achieved
through interventions that focus on improving confidence and strength of older
individuals after they experience vision loss. Currently, rehabilitation programs that
train older adults with visual impairments to move safely through their
environments; or exercise programs that take into consideration the special needs

126
of older people with vision impairments, are well positioned to make fall prevention
a priority in programs designed for this high-risk segment of the older population.
Maintaining muscle strength and functioning is also important for men. As in
women, reduced muscle capacity may have a direct influence on fall risk. In
addition, results of this study suggest that declines in the large muscle groups,
including the chest, legs and back may mediate the effects of vision on falls. That is
to say, that a small portion of the effects of self-reported vision as a predictor of
falls may be through declines in large muscles. For example, older men who
experience vision loss may be less likely to avoid a full fall after tripping over an
unseen object if their muscles are not fit, and they are unable to regain stability. In
this scenario, decrements in vision cause the primary event which sets off a series
responses within multiple other systems, that are designed to prevent the fall or
protect from injury due to the fall—for example, positioning one’s arms to regain
balance, or widening the base of one’s stance, to increase stability. However,
because of declines in a secondary system, perhaps reduced strength in leg
muscles, a fall could occur which might otherwise have been avoided if strength
had been maintained.
In contrast to women, the men in this study also experienced greater fall risk
due to the direct effects of vision. After controlling for all other covariates,
self-reported vision remained a strong predictor of falls. Vision loss in late life has
the potential to increase fall risk by occluding hazards in the environment. In
addition, vision loss could influence fall risk by way of behavioral changes that
place people at greater risk. A strong causal relationship has been noted between

127
declining vision and declining physical activity (Lamoureux, Hassell, & Keeffe,
2004/2). According to Crews and Campbell (2004), the independent effects of
vision loss on mobility may result in many older adults who are blind and visually
impaired reducing their leisure and social activities, forgoing proper nutrition,
and/or neglecting exercise regimens. Vision loss may lead directly to increased
disability (Burmedi, Becker, Heyl, Wahl, & Himmelsbach, 2002; Haymes,
Johnston, & Heyes, 2002) as measured by the ability to perform ADLs and IADLs
resulting in a net reduction of physical activity, and introducing the potential for
reduced musculoskeletal functioning. Older individuals who are visually impaired
may become relatively less fit (Capella-McDonnall, 2007) and, therefore, be at
greater risk of experiencing declines in their ability to perform physical functional
tasks such as walking or rising up from a chair (West et al., 2002) that are crucial
for maintaining effective mobility and that reflect a degree of fitness that is
important for avoiding falls, either directly, or as a mediator to vision. State/federal,
as well as private vision rehabilitation services that train older adults to move
around safely and effectively in their homes and communities and to maintain skills
that foster independence play an important roll in addressing the health of older
adults with vision impairments. In addition, it is incumbent on those who provide
rehabilitation services to older adults with vision impairments to recognize the
important relationship between functioning/disability and fall prevention; and for
service providers who focus on fall-prevention to recognize the limitations, and
special needs of older persons who have vision impairments.

128
CHAPTER V: FALL PREVENTION INTERVENTIONS FOR OLDER ADULTS
WITH VISION IMPAIRMENT
A. Summary of Key Findings
The purpose of this dissertation was to address issues pertaining to
self-reported vision loss and related health outcomes that often accompany vision
impairments in late life, with a primary focus on the outcome of falls. As
demonstrated by the studies reported here, the incidence of health problems in
general is greater among older persons with vision impairments than for those with
normal vision, including the greater incidence of falls among this group. Disparities
in health that are associated with vision status have potential to reduce the quality
of life of older persons with vision impairments by reducing the range of activities in
which they are able to participate, and by the increased possibility of experiencing
secondary health outcomes (especially falls) that often accompany age-related
vision loss.
Each of the three studies described above makes a unique contribution to
knowledge about the relationship between self-reported vision status, health
dimensions, and falls in late life. Specifically, the first study demonstrated that
older adults with visual impairments had worse outcomes across an array of health
dimensions leading to disability. These findings suggest that the effects of vision
loss are dynamic in a way that could perpetuate a downward decline in multiple
independent systems that place older persons who experience them at greater risk
of falling. For instance, older adults in the NHANES sample were much more likely
to report difficulties with mobility and daily living activities if they reported poor

129
vision. This finding inherently suggests that older persons who report their vision
as poor would be less active, which could result in them being less healthy overall.
Reduced physical activity along with dependence in personal maintenance
activities, including diet and medication management could result in reduced
strength and stability which is crucial for avoiding falls in late life.
The theme of cyclical effects was central to the remaining studies, which
employed longitudinal data analyses to assess integrated effects of vision loss in
producing secondary health outcomes. In study 2, analyses were conducted to
determine whether self-reported vision loss over time could result in weaker limbs,
decreased mobility, and greater likelihood of falling, than would be expected when
individuals reported normal vision. Although no strong evidence for a moderating
effect of self-reported vision loss on limb functioning was found, the data did point
to an interesting hierarchical relationship between these factors. Results
suggested that self-reported vision may be secondary to losses in limb functioning,
such that persons who reported poor vision, but who retained strong limb
functioning would be at reduced risk of falling, regardless of their vision status.
Finally, study 3 assessed the possibility that total effects of self-reported
vision loss on falls might be divided into direct and indirect effects. In accordance
with previous research on the relationship between self-reported vision and falls,
direct effects correspond to fall risk that can be attributed to specific decrements in
vision (e.g. losses in vision acuity, visual field), which could result in trips,
stumbles, or bumps into objects in the environment. By contrast, indirect effects, or
mediated effects are those that originate in the visual system, but are contingent

130
on problems in different systems to result in the outcome falls. As noted in study 2,
older persons with vision impairments have reduced risk of falling when their upper
and lower limb functioning is maintained. Thus, it follows that some vision related
falls may be contingent on a weak musculoskeletal system. For example,
individuals who have inadequate lower body strength are less likely to recover
following a trip, than individuals who are relatively strong in the lower body. Results
of the third study reported above suggest that this may be the case, especially for
men, whose vision impairments seem to be mediated by losses in large muscle
groups, with respect to the effect on falls.
B. Interventions to Prevent Falls
Implications of these findings seem obvious in pointing to the importance of
keeping older people with vision impairments from acquiring avoidable functional
limitations by implementing programs and policies that encourage and facilitate
appropriate diets, disease management, exercise, and hazard assessments of the
home environment. Indeed, age-related vision impairments can lead to losses in
functioning by way of medical conditions that cause vision impairments (e.g.,
diabetic neuropathy; atherosclerosis); by reducing the individual’s confidence in
their ability to interact in the environment; or indirectly through secondary health
conditions acquired as a result of maladaptive behavioral modifications in
response to declining vision. In formulating fall-prevention programs for persons
with vision impairment, each of these possibilities merits some consideration.
The primary goal of most fall prevention programs is to implement
prevention strategies that decrease the likelihood of falls among older people,

131
while maintaining maximum levels of mobility and functioning. Fall risk
assessments are often utilized to identify individual factors that place people at risk
for falling. In addition to evaluating vision status, a fall risk assessment may include
inquiries about previous falls; circumstances that lead to falls; medical histories,
including health conditions and medications; mobility and functioning; and about
the environments in which older persons spend their time. Vision impairments are
often cited at the top of lists indicating the many factors that place people at risk for
falling; however, results of the studies reported in this dissertation suggest that
self-reported vision impairments rarely operate independently of other systems to
increase the likelihood of a fall. In thinking of visual functioning as a system that is
integrated with other systems, it follows that effective fall-prevention programs for
older adults with vision impairments would be composed of elements that address
factors beyond vision alone.
There is general consensus among researchers that integrated
risk-management programs that emphasize multiple interventions, including
educational programs, medical risk assessment, exercise, and home hazard
assessment and implementation of modifications are most effective for improving
function and reducing falls among community-dwelling older people (Close et al.,
1999; Day et al., 2002; Gillespie et al., 2003; Tinetti et al. 1994). Studies have
compared single interventions to interventions that take into account multiple
fall-risk factors, and have demonstrated the greater efficacy of multifactorial
programs compared to most singular interventions on their own. For instance, Day
et al. (2002) tested the effectiveness of three interventions separately and

132
together, including vision improvement, group-based exercise, and home hazard
management. Whereas statistically significant effects were reported for exercise
alone, and for the other interventions when combined with exercise, neither
management of home hazards nor treatment of poor vision alone were statistically
significant in reducing falls. Similarly, Close et al. (1999) compared previous fallers
who received comprehensive medical examination/assessments and home visits
from occupational therapists to identify and document home hazards, with those
who received usual care. Compared to the control group, significantly fewer falls
and recurrent falls were reported among the group that received multiple
interventions.

Figure 5.1: Potential Intervention Points Exist between Health Dimensions in the
Disability Process

Theoretically then, the most effective programs for reducing falls would
consider vision impairments in their context among other fall-risk factors in the
disability process. Intervention strategies would be composed of multiple

133
components that addressed relationships between vision loss and other fall risk
factors that are intrinsic, extrinsic, and behavioral in origin (Clemson, Mackenzie,
Ballinger, Close, & Cumming, 2008; Pynoos, Rose, Rubenstein, Choi, & Sabata,
2006). Figure 5.1 shows the disability process model and prospective
interventions that are thought to represent necessary components of the most
effective fall-prevention programs (Pynoos et al., 2006). Arrows indicate points at
which selected interventions could be implemented at each stage to address
problems that persons with visual impairments, in particular, are likely to
experience, as informed by the studies reported above. The advantage to
conceptualizing the relationship between vision status and fall risk in terms of the
disability process is that this structure provides a framework and rationale for
considering potential interventions across health dimensions that may seem,
otherwise, to have little relation to falls.
Education and Rehabilitation
Increasingly, the realm of healthcare policy has focused on preventive
efforts as means to reduce expenditures associated with negative health
outcomes, including falls. An important aspect of prevention is educating
individuals about behaviors that can increase risks associated with better or worse
health outcomes. In study 1 of this dissertation, disparities across health
dimensions were found between older persons who reported poor vision
compared to those with normal vision. Several biological risk factors were found to
have greater incidence among persons with vision impairments than in older
persons with normal vision. For instance, persons with vision impairments were

134
found to have greater at-risk levels of homocysteine, an amino acid that has been
linked to cardiovascular disease. It is likely that educating older adults with vision
impairments about nutritional and rehabilitation treatments that improve diet and
reduce high-risk levels of specific biomarkers could exert positive influence on an
array of health factors that may be directly related to falls, including cardiovascular
health. The efficacy of dietary changes has been borne out in several studies,
including a controlled trial that found that increased consumption of dietary folate
via vegetables and citrus fruits resulted in lowered plasma homocysteine
concentrations (Brouwer et al., 1999).
Extant nutrition programs that are administered by the Administration on
Aging (AoA) and implemented locally by area agencies on aging (AAA) and their
contracted agencies have the potential to function as effective vessels for targeting
and educating older adults with visual impairments about special diets and
disseminating dietary information related to specific biomarkers. In congregate
and home-delivered meal programs, which pursue the explicit goal of improving
the health of older adults (Gelfand, 2006), consumers with visual impairments (as
well as older adults in general) might benefit from the option of special diets that
address issues related to elevated homocysteine and cholesterol levels (similar to
low-salt diets offered to control hypertension), as well as low body weight.
Biological risks are considered to be foundational on the path to functional loss and
disability; therefore, ameliorating problems that are noted early on in the disability
process, through education about simple dietary and other behavioral changes,
can serve to reduce fall risk by preventing health problems that are known to be

135
factors that predict falls. Similarly, education of older adults with vision
impairments can also extend to better management of chronic conditions, such as
diabetes, which if not properly managed, could result in additional impairments
that increase the older individual’s risk for falling.
Older adults with vision impairments may also acquire new functional
losses because they are not comfortable interacting in their home environment.
They may avoid specific activities (such as walking up a flight of stairs) because
they no longer feel safe, due primarily to their inability to see. These types of
functional losses are potentially avoidable through the provision of training in daily
living activities that have the potential to make up a large portion of any individual’s
daily physical activity level. Rehabilitation services, such as those provided by
state/federal independent living programs for older individuals who are blind can
serve as a vital model in training older adults with visual impairments to maintain
ADL/IADL functioning that is crucial to health maintenance and for preserving
independence. In particular, public and private agencies that serve older people
who are blind or have low vision under the Older Blind Independent Living program
(Title VII, Chapter 2, of the Rehabilitation Act of 1973, as amended— hereafter,
VII-2) may have much to contribute in efforts to prevent falls. Similar to the
relationship between AoA and state units on aging, services to older persons with
visual impairments are regulated by the federal Rehabilitation Services
Administration and are administered by designated state units in each state. VII-2
independent living services that are available (but not necessarily provided) under
the program include services to help correct blindness (including visual screening

136
and surgical or therapeutic treatment); hospitalization related to such services; the
provision of visual aids (such as magnifiers or eyeglasses), and other specific
services that are designed to assist older individuals to adjust to blindness,
maintain independence, and become more mobile and more self-sufficient (Orr,
1998). These services in particular have been shown to be effective in helping
older adults complete daily functional activities independently, including preparing
meals for themselves, and continuing to participate in the lives and activities of
family, friends and their communities (Moore, Steinman, Giesen, & Frank, 2006).
Although VII-2 programs serve only a fraction of the number of older adults who
could benefit from vision rehabilitation services, the program’s access to older
persons with vision impairments make VII-2 potentially quite effective for
disseminating information about falls and fall prevention among this high-risk
population.
Older persons with visual impairments can also benefit greatly from
educational materials that are created and disseminated by private organizations
and government agencies, whose collective mission is to prevent falls among
older adults. For example, there are a range of materials available from the Fall
Prevention Center of Excellence (FPCE), located at the University of Southern
California. Documentation compiled at FPCE is disseminated to a coalition of
agencies across California that provide services to older adults. The materials are
also available on the FPCE website (http://www.stopfalls.org) for service providers
and researchers, as well as for individuals and their families. Information is
available regarding fall-risk factors (including visual impairment), assessment tools

137
for evaluating environments, and programs and services that are available locally,
aimed at preventing falls. Unfortunately (and perhaps ironically), many of the
materials that are available on this, and other websites targeting an older
audience, are not accessible to persons with vision impairments due to limited print
size options, or due to inaccessible web-design, which does not accommodate
screen readers and other technologies that are commonly used by persons with
blindness and low vision (Milne et al., 2005).
Finally, many simple and inexpensive assessment tools and checklists
designed to help older adults to identity common risk factors and environmental
hazards are readily available from multiple sources. Checklists, which are
relatively easy to administer, and require little or no training to conduct, are often
disseminated directly to older adults via facilities where older people may
congregate or seek services, such as senior centers and health clinics. Because of
their relative ease of use, checklists can provide older adults with a practical and
inexpensive basis for evaluating risk related to the presence of hazards, and the
safety of their homes. Although checklists are often relatively user-friendly for older
adults or their caregivers, they may vary greatly with respect to their
comprehensiveness and suggested solutions may be generic and may not apply in
all cases (Pynoos, Steinman, & Nguyen, in press). Due to their limited content,
checklists may overlook some important problems in home settings and may
present only a limited array of possible solutions to problems. The adoption of
recommendations presented by checklist assessments may also vary based on
the willingness of older adults to change aspects of their homes, and their beliefs

138
as to whether making changes would influence their likelihood of falling (Cumming
et al., 2001). An example of a home assessment checklist is Check for Safety: A
Home Fall Prevention Checklist for Older Adults, disseminated by the Centers for
Disease Control and Prevention (CDC; 2005), and attainable from CDC’s website
(http://www.cdc.gov).
Medical Assessment
Customarily, older people may begin to seek out redress for age-related
vision loss, or related health conditions through consultations with physicians, who
are trained to address vision loss on the basis of a medical examination.
Physicians and other health care professionals are an important source of
information for older individuals with vision impairments who are concerned about
falling. Increasingly, geriatricians have begun to integrate fall-prevention into
practices that have traditionally focused on treating acute and chronic conditions,
as well as falls, after they occur. Indeed, many of the most common chronic
conditions experienced by older people have been associated with falls, including
stroke, diabetes, osteoarthritis, cardiovascular problems and cognitive factors
(Lord, Sherrington, & Menz, 2001). Furthermore, it is well documented that among
older adults in general, the risk of mobility loss and falls increases relative to the
individual’s degree of frailty and to the number of chronic conditions that are
experienced by the individual (Guralnik et al. 1993; Tinetti, Williams, & Mayewski,
1986). The disability model used as a key theoretical construct in the studies
described above serves here to demonstrate the potential trajectory leading from
disease to functional loss, disability, and finally, to falls.

139
Many of the most common chronic conditions cited above are experienced
with greater frequency among older persons with vision impairments. In fact, there
is often a direct association between some chronic diseases/conditions and loss of
vision. For example, some systemic diseases such as diabetes are directly related
to some forms of vision loss, especially diabetic retinopathy. Atherosclerosis, an
impairment that affects the cardiovascular system and vessels throughout the
body including in the retinas, is associated with development of age-related
macular degeneration. Systemic hypertension and glaucoma often co-occur and
are believed to develop from common mechanisms. A serious stroke that affects
both hemispheres of the visual cortex has the potential to cause instant blindness
that is not amenable to rehabilitation. Thus, the results of study 1, which
demonstrated greater incidence of chronic disease among persons with vision
impairments is likely to reflect, at least in part, the causal pathway that is
hypothesized to link disease and impairment in the disability process model.
The association between disease and vision status may also be apparent
when vision loss is primary—that is, when vision impairments result in disabilities
that reduce quality of life, and affect overall health, as hypothesized in study 2. This
association, wherein vision loss precedes losses in other systems is important,
because it points to a pathway that could be moderated through interventions to
prevent further decline that could result in falls.
Regardless of the direction of their relationship, chronic disease and vision
impairment have a unique association that is directly relevant to an individual’s fall
risk. Identifying integrated relationships between vision and other health

140
dimensions during a formal assessment of fall risk has the potential to inform
effective interventions designed to prevent falls and to improve overall health of
persons with vision impairments.
A formal fall-risk assessment should include a visual exam, as well as an
assessment of the individual’s eyewear prescription for currency. Uncorrected
refractive errors have been demonstrated to be a significant cause of vision
impairments among people of all ages (Resnikoff, Pascolini, Mariotti, & Pokharel,
2008), as well as a fall risk for older people (Day et al., 2002). Additional risk
factors can be gleaned through a detailed history of the patient’s health, including
their medical conditions and medications, and information about falls episodes,
and circumstances leading up to falls (Tideiksaar, 1998). According to Rubenstein,
Robbins, Josephson, Schulman and Ostwerweil (1990), much can be learned by
conducting post-fall medical assessments that are intended to identify underlying
and possibly modifiable fall risk factors that could inform preventive and
therapeutic interventions. Even when falls do not result in injuries, reviewing the
circumstances that led up to fall episodes with physicians may be informative, as
these discussions may elucidate symptoms of underlying disease or functional
losses that, if addressed early on, could prevent more serious health problems,
functional losses, and falls in the future.
Exercise
Results of studies reported above suggest that persons with self-reported
vision impairments have greater risk of acquiring functional difficulties that, in part,
may reflect age-related losses in the musculoskeletal system. For this high-risk

141
group of individuals and for older adults in general, exercise is an important
preventive measure against many negative health outcomes commonly
experienced by older people. Even minimal doses of regular exercise by older
adults have been shown to modify many of the most prevalent chronic medical
conditions experienced in old age, including arthritis, heart disease, stroke, and
pulmonary disease (Bean, Vora, & Frontera, 2004). Studies that have assessed
the effects of exercise on the musculoskeletal system have reported positive
effects on muscle strength, neuromuscular performance, and bone mineral density
(Taaffe, Duret, Wheeler, & Marcus, 1999), in addition to improving balance and
mobility (Shumway-Cook, Gruber, Baldwin, & Liao, 1997), and increasing flexibility
and joint range of motion (Fatouros et al., 2002). Studies that have directly
assessed the impact of exercise in reducing fall-risk have generally confirmed the
efficacy of various exercise regimes alone and in conjunction with other
interventions for reducing falls (Day et al., 2002; Province et al., 1995). In the
meta-analysis conducted by Province and colleagues, which assessed effects of
various exercise interventions administered as part of the FICSIT (frailty and
injuries: cooperative studies of intervention techniques) trials, general exercise
programs were found to reduce falls by about 10%, whereas balance exercises
reduced falls by 17%. Thus, all exercise programs are not equal in their beneficial
effects with respect to fall-prevention, and it appears that the greatest benefit is
achieved when exercises are chosen to address specific decrements experienced
by older individuals—that is, when exercises are tailored to address the
individual’s specific functional limitations. For example, Lord et al. (2001)

142
described resistance or strength training as a means to increase the ability of
muscles to generate force, such as that needed to stand from a chair or to climb
stairs. Various studies have demonstrated that resistance training, including
weight training, can reduce fall-risk by increasing muscle strength and improving
functioning and performance in daily tasks (Hunter, McCarthy, & Bamman, 2004).
This may be especially true for older men with vision impairments. In study 3
above, large muscle functioning was found to mediate the effects of self-reported
vision loss in men; therefore, strengthening large muscles in legs, arms and torsos
would likely reduce indirect fall risks associated with poor vision. Lord et al. (2001)
noted, however, that strength training for isolated muscle groups that are not used
in everyday tasks may do little to improve functional capacity, because carryover to
muscles that are not directly trained is limited due to specificity of effect. Greater
improvement may be possible by practicing particular functional activities in which
individuals have deficits, with and/or without added resistance. In addition to
formal exercise classes then, older persons with vision impairments could
potentially acquire valuable physical training in programs (such as VII-2, described
above) that teach Orientation and Mobility (O&M), a protocol from the field of vision
rehabilitation, which focuses on safe and effective travel through the environment.
Although the evidence overwhelmingly supports the need for all older
people to remain physically active in order to maintain optimal health, older
persons with vision loss are at a decided disadvantage in their ability to easily
participate in formal exercise programs, even when they are available to them in
their communities. Older adults with visual impairments may have any number of

143
reasons for choosing not to participate in formal exercise programs. For example,
vision loss often results in difficulty or inability to drive safely. At the same time,
other forms of transportation may be difficult to procure, and problems reading bus
schedules due to their small print size may make public transportation undesirable,
relative to staying at home. Some agencies, such as the Braille Institute in Los
Angeles, provide limited door-to-door transportation from the homes of
participants to their classes; however, expenses associated with overhead costs,
including paying drivers, gasoline, and maintenance of a fleet, make this option
impractical and prohibitive to most agencies that serve older people.
Vision impairments may also result in motivational difficulties which could
arise due to health disparities related to vision impairments (described in detail
above). Chronic conditions, which are experienced disproportionately by people
with vision impairments, may limit the types of exercise that can be performed, and
the duration of time which they are able to participate. Older people who have
mobility difficulties due to their vision loss may feel that their safety is threatened by
venturing beyond their homes; or they may feel uncomfortable or unstable while
participating in formal exercise programs. Finally, older people with visual
impairments may have greater difficulty in following instructors, who may rely
heavily on visual cues while leading classes. Some exercise regimes that have
been adapted specifically for fall prevention, such as tai chi for older adults, may
involve a complex series of choreographed movements, which are not easily
replicated without being able to see them demonstrated by the instructor. In this

144
case, older adults with vision impairments would likely forgo attending a tai chi
class if special provisions were not available to address their concerns.
Within fall prevention programs that offer exercise components there is a
clear need for protocol development with consideration to the special needs of
persons who have vision impairments. Those who instruct older adults in physical
fitness classes should become knowledgeable about the unique circumstances
that older people with vision impairments face when attempting to participate in
groups where the majority have normal vision. Morgenthal and Shephard (2005)
provided some basic recommendations and modifications which would foster
participation by persons with low vision, including making sure that exercise areas
are well lighted, and that instructions are provided in large type or are spoken
clearly and slowly to assure comprehension. Exercise areas should be kept clear
of objects that can cause trips or stumbles; and if possible, provision of external
support devices, such as chairs, or wall bars should be provided to individuals who
have a history of falls. Exercise classes could also be specifically designed for
older people with vision impairments, to target their unique balance, strength and
coordination problems, while simultaneously providing structure, motivation, and
social support to this group.
Home Hazards and Home Modifications
The settings in which older people with vision impairments live often contain
hazards and problem areas, or lack supportive features which could ameliorate
risks associated with dangerous areas in the home and community if they were
present (Stevens, 2002/2003). To compound these problems, the oldest old, those

145
who are most likely to experience vision impairment, may become less mobile due
to their functional losses, and spend more of their time in and around their homes
(Pynoos, Sabata, & Choi, 2005), thus increasing the potential for environmental
factors to influence fall risk. Among older adults in general, the majority of
fall-related injuries (55%) occur inside the home, including falls on stairs and in
rooms throughout the house. An additional 23% of injuries experienced by older
adults due to falls occur outside but near the home, as on curbs and sidewalks,
and other familiar routes. The remaining 22% occur in the communities of older
adults (Kochera, 2002), in public and private spaces, as well as in the built
environment (Pynoos et al., 2005). According to Li et al. (2006) risk profiles for
indoor and outdoor falls are different, with higher leisure-time physical activity
being associated with outdoor falls; whereas a greater number of physical
difficulties and indicators of poor health are associated with indoor falls. In all, it is
estimated that between 35% and 40% of falls result from factors that are related to
the environment (Josephson, Fabacher, & Rubenstein, 1991).
In accordance with Lawton & Nahemow’s (1973) theory of environmental
press, well-designed environments and effective home modifications should
function to reestablish equilibrium between a person’s capabilities, which may
have declined due to their vision loss, and the demands of the environment. The
concept of universal design (UD) has been employed to create products, buildings,
and exterior spaces that reduce environmental demands for people of all ages,
sizes, and abilities to the greatest extent possible (Sanford, in press). Effective UD
minimizes barriers and increases supportive features to facilitate participation in

146
daily living activities (Mace, Hardie & Place, 1996). Among many other possibilities,
UD features that may be especially helpful for reducing falls among older persons
with visual impairments include a zero-step entrance with flush or low profile
threshold; high contrast trim and glare-free floor surfaces; a curbless or roll-in
shower in bathrooms; and motion sensor lighting that automatically turns on and
off when individuals enter or exit the room (Pynoos, 1992; Young, 2006).
In older homes, and in homes that have not been well-designed for older
persons with vision impairments and related disabilities, home modifications can
be employed to address hazardous areas that could increase fall risk. Home
modification refers to the converting or adapting of environments to make
everyday tasks easier, reduce accidents, and support independent living (Pynoos
et al., 2005). Just as visual functioning of older individuals is dynamic,
environments also can change over time, as homes become older and keeping up
with repairs may become more difficult. The range of home modifications available
is wide and their expense may vary from low-cost adaptations to more expensive
renovations (Pynoos et al., in press). Home modifications that are particularly
germane to older adults with visual impairments include improving lighting
throughout the home; removing hazards (e.g., clutter and throw rugs); adding
special features or assistive devices (e.g., grab bars wrapped with brightly colored
contrasting tape); moving furnishings to create clear pathways; placing contrasting
nonskid textured mats in showers and tubs; and painting door frames in bright solid
colors (Duffy, 2002; Pynoos et al., 2005).

147
Researchers generally agree that home modifications are important for
promoting safety and independent living; however, research findings are
inconsistent and sometimes counterintuitive in their reports of the impact of hazard
reduction and home modification on fall rates. For example, in study 2 above, it
was reported that persons whose homes had been modified to accommodate
someone living with a disability were more likely to fall than persons who lived in
homes that had not been modified. This finding was interpreted to reflect the
greater likelihood of those who need home modifications, whether due to their
chronic conditions or functional difficulties, acquiring them—that is, that those who
are at the greatest risk of falls are more likely to invest in home modifications to
prevent a fall.
When statistically significant effects of hazard reduction/home modification
have been reported, it has usually been described as a combined effect, in which
the environment interacts with other fall-risk factors. The general removal of home
hazards without consideration for interactions of physical and behavioral traits with
specific aspects of the environment may not only prove ineffectual for reducing
falls, but also has the potential to increase fall risk by interfering with idiosyncratic
relationships established over time between individuals and their home
environments. For instance, a common recommendation is to clear all walkways
and paths of obstacles that may be viewed generally as trip hazards. On the
surface, this recommendation seems sensible, since one cannot trip over hazards
that are not present. However, many older adults may have adopted personal
strategies based on environmental features to improve mobility and compensate

148
for declines in their vision. For example, older people with poor vision may use
large pieces of furniture as cues to orient themselves as to their location in a room,
or relative to other known but less salient environmental hazards. Similarly, older
people with balance impairments may use table tops, or the backs of sofas to
support themselves as they make their way across a room. By exploring unique
forms of support that environments afford and observing how persons carry out
tasks, recommendations by professionals can be better tailored to address
specific combinations of intrinsic (including vision-related) factors experienced as
a function of age, behavioral attributes, and extrinsic fall-risk factors.
C. Conclusion
Previous studies, as well as the evidence presented above suggest that
effective fall prevention programs that target older adults with vision impairments
would need to address chronic conditions, functional difficulties and disability in
daily living activities, as well as vision loss, in order to maximize the impact of
services. This more comprehensive approach to fall prevention with respect to
vision impairment might include teaching techniques for coping with vision-related
disability, and providing assistive technology and home modifications that are
specifically designed to address difficulties with daily living activities. Similarly, with
respect to mobility issues, current training that is available through the older blind
independent living program may be an appropriate medium for introducing older
adults with vision impairments to physical exercise regimes that are effective for
countering age-related losses in upper and lower limbs. In addition, physical and
occupational therapists could provide training and exercise programs, in

149
conjunction with home assessments and modifications that are specifically
adapted to meet the unique circumstances of older people with visual
impairments.
For practitioners who are interested in the impact of visual impairment on
everyday health, functioning, and falls, the findings described in this dissertation
are important and practical because they point to a relatively unacknowledged
interconnectedness between vision and the multiple other systems that influence
health, quality of life, and fall risk of older adults. Vision loss has been cited as a
major factor predicting falls among older people; however, most studies have
looked only at total and direct effects of specific aspects of vision, and have
neglected to consider how vision status may be integrated with other health
dimensions to increase fall risk. Thus, the three studies reported in this dissertation
collectively contribute to the research body by acknowledging the potential role of
self-reported vision impairment as part of a larger integrated system, which causes
falls by way of specific and predictable pathways.
Future research studies should attempt to refine and extend the results
reported here, through the development and use of data that specifically address
important questions at hand. Most notably, studies should focus on improving
measures that operationalize falls. Whereas survey-based data, such as the HRS,
are useful for making rough estimates about the incidence of falls, there remain
implicit problems with data collected over long intervals, used to make inferences
about the larger population. Future studies will benefit by the use of data derived in
the shortest possible intervals, and by assuring that respondents have a clear and

150
consistent understanding of how falls are defined. Similarly, it is possible that
clinical measures of vision loss would yield stronger evidence in support of the
hypotheses tested, than the measures of self-reported vision that were used in the
studies above. More precise clinical measures of musculoskeletal functioning, in
lieu of proxy measures of upper and lower-limb functioning, would likely produce
results that better reflect the relationship between vision loss and physical
functioning, with respect to falls.
In conclusion, programs that hold the greatest potential in fall prevention
would explore dynamic interactions between the abilities of older adults and their
surroundings. Toward this aim, Practice Guidelines for the Prevention of Falls in
Older Adults established by the American Geriatrics Society (AGS; 2010)
recommends that older adults who are at high risk of falling undergo multifactorial
fall-risk assessments. Implicit in the recommendation of AGS is the
acknowledgement that any single intervention is likely to be less effective than
when complex relationships between fall-risk factors are considered together. It is
clear, given projections of a growing number of older individuals with vision
impairments, that practitioners who are interested in preventing falls among older
adults will need to become increasingly aware of issues related to medical,
functional, and rehabilitative aspects of vision loss, especially as they relate to fall
prevention. The assertion that vision declines may be indirectly associated with
declines in muscle strength and functional capacity has implications for
professionals who may improve fall-prevention programs by acknowledging a
more complex set of physical associations with vision. Furthermore, through

151
understanding a specific mechanism by which frailty may develop—that is,
feedback loops in the disability process— there is potential to prevent disease, and
maintain residual vision and physical functional capacity to reduce dangers
associated with falling. This worthy goal can be achieved by continuing rigorous
research, dissemination of findings, and by assuring adequate funding for
rehabilitation and fall-prevention programs that address factors across health
dimensions.

152
BIBLIOGRAPHY
Abdelhafiz, A. H., & Austin, C. A. (2003). Visual factors should be assessed in
older people presenting with falls or hip fracture. Age and Ageing, 32, 26-30.

Adams, K. F., Schatzkin, A., Harris, T. B., Kipnis, V., Mouw, T., Ballard-Barbash, et
al. (2006). Overweight, obesity, and mortality in a large prospective cohort of
persons 50 to 71 years old. New England Journal of Medicine, 355, 763-778.

Adamis, A. P. (2002). Is diabetic retinopathy an inflammatory disease? British
Journal of Ophthalmology, 86, 363-365.

Alley, D. E., & Chang, V. W. (2007). The changing relationship of obesity and
disability, 1988-2004. Journal of the American Medical Association, 298,
2020-2027.

American Geriatrics Society. (2010). AGS clinical practice guideline: Preventing
falls in older persons. Retrieved on June 28, 2010 from http://www.american
geriatrics.org/health_care_professionals/clinical_practice/clinical_guidelines_
recommendations/2010/.

Anastasopoulos, E., Yu, F., & Coleman, A. L. (2006). Age-related macular
degeneration is associated with an increased risk of hip fractures in the
Medicare database. American Journal of Ophthalmology, 142, 1081-1083.

Arias, E. (2010). United States life tables, 2006. In National Vital Statistics Report,
58 (21). Hyattsville, MD: National Center for Health Statistics.

Baigent, C., & Clarke, R. (2007). B vitamins for the prevention of vascular disease:
Insufficient evidence to justify treatment. Journal of the American Medical
Association, 298, 1212-1214.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction
in social psychological research: Conceptual, strategic, and statistical
considerations. Journal of Personality & Social Psychology, 51, 1173-1182.

Bates, C. J., Mansoor, M. A., van der Pols, J., Prentice, A., Cole, T. J., & Finch, S.
(1997). Plasma total homocysteine in a representative sample of 972 British
men and women aged 65 and over. British Journal of Clinical Nutrition, 51,
691-697.

Bath, P. A., & Morgan, K. (1999). Differential risk factor profiles for indoor and
outdoor falls in older people living at home in Nottingham, UK. European
Journal of Epidemiology, 15, 65-73.

153
Bean, J. F., Vora, A., Frontera, W. R. (2004). Benefits of exercise for
community-dwelling older adults. Archives of Physical Medicine and
Rehabilitation, 85 (suppl. 3), 31-42.

Bennett, E. S., Weissman, B. A., & Remba, M. J. (2007). Contact lenses and the
older adult. In A. A. Rosenbloom (Ed.), Rosenbloom and Morgan's Aging and
Vision (pp. 215-240). St. Louis, MO: Butterworth Heinemann.

Berbaken, J. H., & Johnston, A. W. (1986). Population norms for edge contrast
sensitivity. American Journal of Optometry and Physiological Optics, 63,
724-732.

Berger, S., & Porell, F. (2008). The association between low vision and function.
Journal of Aging and Health, 20, 504-525.

Boushey, C. J., Beresford, S. A. A., Omenn, G. S., & Motulsky, A. G. (1995). A
quantitative assessment of plasma homocysteine as a risk factor for vascular
disease: Probable benefits of increasing folic acid intakes. Journal of the
American Medical Association, 274, 1049-1057.

Branch, L. G., Horowitz, A., & Carr, C. (1989). The implications for everyday life of
incident self-reported visual decline among people over age 65 living in the
community. Gerontologist, 29, 359-365.

Brody, B. L., Gamst, A. C., Williams, R. A., Smith, A. R., Lau, P. W., Dolnak, D., et
al. (2001). Depression, visual acuity, comorbidity, and disability associated
with age-related macular degeneration. Ophthalmology, 108, 1893-1900.

Brouwer, I. A., van Dusseldorp, M., West, C. E., Meyboom, S., Thomas, C. M. G.,
Duran, M., et al. (1999). Dietary folate from vegetables and citrus fruit
decreases plasma homocysteine concentrations in humans in a dietary
controlled trial. Journal of Nutrition, 129, 1135-1139.

Bueno-Cavanillas, A., Padilla-Ruiz, F., Jiménez-Moleón, J. J., Peinado-Alonso, C.
A., & Gálvez-Vargas, R. (2000). Risk factors in falls among the elderly
according to extrinsic and intrinsic precipitating causes. European Journal of
Epidemiology, 16, 849-859.

Burmedi, D., Becker, S., Heyl, V., Wahl, H. W., & Himmelsbach, I. (2002).
Behavioral consequences of age-related low vision. Visual Impairment
Research, 4, 15-45.

Campbell, A. J., & Robertson, M. C. (2006). Implementation of multifactorial
interventions for fall and fracture prevention. Age and Ageing, 35 (suppl ii),
60-64.

154
Capella-McDonnall, M. (2007). The need for health promotion for adults who are
visually impaired. Journal Visual Impairment and Blindness, 101, 133-145.

Carter, N. D., Kannus, P., & Khan, K. M. (2001). Exercise in the prevention of falls
in older people. Sports Medicine, 31, 427-438.

Carter, S. E., Campbell, E. M., Sanson-Fisher, R. W., & Gillespie, W. J. (2000).
Accidents in older people living at home: A community-based study assessing
prevalence, type, location and injuries. Australia New Zealand Journal of
Public Health, 24, 633-636.

Cavenaugh, B. S., & Steinman, B. A. (2005). Estimated number of persons eligible
for vision rehabilitation services under expanded Medicare. Journal of Visual
Impairment and Blindness, 99, 431-436.

Centers for Disease Control. (2005). Check for safety: A home fall prevention
checklist for older adults. Retrieved on June 16, 2010 from
http://www.cdc.gov/ncipc/pub-res/toolkit/Falls_ToolKit/DesktopPDF/English/b
ooklet_Eng_desktop.pdf

Centers for Disease Control and Prevention (2006). Summary Health Statistics for
U.S. Adults: National Health Interview Survey, 2004. National Center for
Health Statistics. Vital Health Statistics 10(228).

Centers for Disease Control and Prevention. (2008). Self-reported falls and
fall-related injuries among persons aged >=65 years—United States, 2006.
Journal of the American Medical Association, 299, 1658-1659.

Centers for Disease Control and Prevention (2010). National Health and Nutrition
Examination Survey. Retrieved June 30, 2010, from
http://www.cdc.gov/nchs/nhanes.htm

Cesari, M., Landi, F., Torre, S., Onder, G., Lattanzio, F., & Bernabei, R. (2002).
Prevalence and risk factors for falls in an older community-dwelling population.
Journals of Gerontology Series A: Biological Sciences and Medical Sciences,
57, M722-726.

Cesari, M., Penninx, B. W. J. H., Newman, A. B., Kritchevsky, S. B., Nicklas, B. J.,
Sutton-Tyrrell, K., et al. (2003). Inflammatory markers and cardiovascular
disease (the health, aging and body composition [health ABC] study). The
American Journal of Cardiology, 92, 522-528.

Chia, E., Mitchell, P., Rochtchina, E., Foran, S., & Wang, J. (2003). Unilateral
visual impairment and health related quality of life: The Blue Mountains eye
study. British Journal of Ophthalmology, 87, 392-395.

155

Claesson, S., Morrison, A., Wertheimer, A. I., & Berger, M. L. (1999). Compliance
with prescribed drugs: Challenges for the elderly population. Pharmacy World
& Science, 21, 256-259.

Clarke, R., Smith, A. D., Jobst, K. A., Refsum, H., Sutton, L., & Ueland, P. M.
(1998). Folate, vitamin B12, and serum total homocysteine levels in confirmed
Alzheimer disease. Archives of Neurology, 55, 1449-1455.

Clemson, L., Cumming, R. G., & Roland, M. (1996). Case-control study of hazards
in the home and risk of falls and hip fractures. Age and Ageing, 25, 97-101.

Clemson, L., Mackenzie, L., Ballinger, C., Close, J. C. T., & Cumming, R. G.
(2008). Environmental interventions to prevent falls in community-dwelling
older people: A meta-analysis of randomized trials. Journal of Aging and
Health, 20, 954-971.

Close, J., Ellis, M., Hooper, R., Glucksman, E., Jackson, S., & Swift, C. (1999).
Prevention of falls in the elderly trial (PROFET): A randomised controlled trial.
The Lancet, 353, 93-97.

Cohen, H. J., Harris, T., & Pieper, C. F. (2003). Coagulation and activation of
inflammatory pathways in the development of functional decline and mortality
in the elderly. The American Journal of Medicine, 114, 180-187.

Coleman, A. L., Stone, K., Ewing, S. K., Nevitt, M., Cummings, S., Cauley, J. A., et
al. (2004). Higher risk of multiple falls among elderly women who lose visual
acuity. Ophthalmology, 111, 857-862.

Coren, S., & Ward, L. M. (1984). Sensation and perception. San Diego: Harcourt
Brace Jovanovich.

Crews, J. E. (1994). Aging and disability: The issues for the 1990s. In S. E. Boone,
D. Watson & M. Bagley (Eds.), The Challenge to Independence: Vision and
Hearing Loss among Older Adults (pp. 47-60). University of Arkansas:
Rehabilitation Research and Training Center for Persons who are Deaf or
Hard of Hearing.

Crews, J. E., & Campbell, V. A. (2001). Health conditions, activity limitations, and
participation restrictions among older people with visual impairments. Journal
of Visual Impairment & Blindness, 95, 453-467.

Crews, J. E., & Campbell, V. A. (2004). Vision impairment and hearing loss among
community-dwelling older Americans: Implications for health and functioning.
American Journal of Public Health, 94, 823-829.

156

Crimmins, E., Vasunilashorn, S., Kim, J. K., & Alley, D. (2008). Biomarkers related
to aging in human populations. In G. S. Makowski (ed.), Advances in Clinical
Chemistry, 46 (pp. 161 – 216). Burlington: Academic Press.

Crimmins, E. M. (2004). Trends in the health of the elderly. Annual Review of
Public Health, 25, 79-98.

Crimmins, E. M., Alley, D., Reynolds, S. L., Johnston, M., Karlamangla, A., &
Seeman, T. (2005). Changes in biological markers of health: Older Americans
in the 1990s. Journal of Gerontology: Medical Sciences, 60A, 1409-1413.

Cumming, R. G., Ivers, R., Clemson, L., Cullen, J., Hayes, M. F., Tanzer, M., et al.
(2007). Improving vision to prevent falls in frail older people: A randomized
trial. Journal of the American Geriatrics Society, 55, 175-181.

Cumming, R. G., Szonyi, T. M., Salkeld, G., O'Neill, E., Westbury, C., & Frampton,
G. (1999). Home visits by an occupational therapist for assessment and
modification of environmental hazards: A randomized trial of falls prevention.
Journal of the American Geriatrics Society, 47, 1397-1402.

Cumming, R. G., Thomas, M., Szonyi, G., Frampton, G., Salkeld, G., & Clemson,
L. (2001). Adherence to occupational therapist recommendations for home
modifications for falls prevention. American Journal of Occupational Therapy,
55, 641-648.

Curcio, C. A., Millican, C. L., Bailey, T., & Kruth, H. S. (2001). Accumulation of
cholesterol with age in human Bruch's membrane. Investigative
Ophthalmology Visual Science, 42, 265-274.

Curcio, C., Millican, C., Allen, K., & Kalina, R. (1993). Aging of the human
photoreceptor mosaic: Evidence for selective vulnerability of rods in central
retina. Investigative Ophthalmology Visual Science, 34, 3278-3296.

Daley, M. J., & Spinks, W. L. (2000). Exercise, mobility, and aging. Sports
Medicine, 29, 1-12.

Danesh, J., & Lewington, S. (1998). Plasma homocysteine and coronary heart
disease: Systematic review of published epidemiological studies. Journal of
Cardiovascular Risk, 5, 229-232.

Day, L., Fildes, B., Gordon, I., Fitzharris, M., Flamer, H., & Lord, S. (2002).
Randomised factorial trial of falls prevention among older people living in their
own homes. British Medical Journal, 325 (7356), 128.

157
De Boer, M. R., Pluijm, S. M. F., Lips, P., Moll, A. C., Volker-Dieben, H. J., Deeg,
D. J. H., et al. (2004). Different aspects of visual impairment as risk factors for
falls and fractures in older men and women. Journal of Bone and Mineral
Research, 19, 1539-1547.

de Maat, M. P., Pietersma, A., Kofflard, M., Sluiter, W., & Kluft, C. (1996).
Association of plasma fibrinogen levels with coronary artery disease, smoking
and inflammatory markers. Atherosclerosis, 121, 185-191.

Devaney, K. O. & Johnson, H. A. (1980). Neuron loss in the aging visual cortex of
man. Journal of Gerontology, 15, 836 – 841.

Digiovanna, A. G. (2000). Human aging: Biological perspectives (2nd ed.). Boston,
MA: McGraw-Hill Higher Education.

Ducic, I., Short, K. W., & Dellon, A. L. (2004). Relationship between loss of pedal
sensibility, balance, and falls in patients with peripheral neuropathy. Annals of
Plastic Surgery, 52, 535-540.

Duffy, M. A. (2002). Making life more livable: Simple adaptations for living at home
after vision loss. New York: AFB Press.

Evans, W. J. (1995). What is sarcopenia? Journals of Gerontology: Series A:
Biological Sciences and Medical Sciences, 50, 5-8.

Fatouros, I. G., Taxildaris, K., Tokmakidis, S. P., Kalapotharakos, V., Aggelousis,
N. et al. (2002). The effects of strength training, cardiovascular training and
their combination on flexibility of inactive older adults. International Journal of
Sports Medicine, 23, 112 – 119.

Feldman, F., & Chaudhury, H. (2008). Falls and the physical environment: A
review and a new multifactorial falls-risk conceptual framework. Canadian
Journal of Occupational Therapy, 75, 82-95.

Felson, D. T., Anderson, J. J., Hannan, M. T., Milton, R. C., Wilson, P. W., & Kiel,
D. P. (1989). Impaired vision and hip fracture: The Framingham study. Journal
of the American Geriatrics Society, 37, 495-500.

Frontera, W. R., Hughes, V. A., Lutz, K. J., & Evans, W. J. (1991). A
cross-sectional study of muscle strength and mass in 45- to 78-yr-old men and
women. Journal of Applied Physiology, 71, 644-650.

Frost, H. M. (1997). On our age-related bone loss: Insights from a new paradigm.
Journal of Bone and Mineral Research, 12, 1539-1546.

158
Ganz, D. A., Higashi, T., & Rubenstein, L. Z. (2005). Monitoring falls in cohort
studies of community-dwelling older people: Effect of the recall interval.
Journal of the American Geriatrics Society, 53, 2190-2194.

Gelfand, D. E. (2006). Nutrition programs. The aging network: Programs and
services (pp. 129-139). New York: Springer Publishing Company.

Gillespie, L. D., Gillespie, W. J., Robertson, M. C., Lamb, S. E., Cumming, R. G., &
Rowe, B. H. (2003). Interventions for preventing falls in elderly people.
Cochrane Database of Systematic Reviews, 4, CD000340.

Gitlin, L. N., Winter, L., Dennis, M. P., Corcoran, M., Schinfeld, S., & Hauck, W. W.
(2006). A randomized trial of a multicomponent home intervention to reduce
functional difficulties in older adults. Journal of the American Geriatrics
Society, 54, 809-816.

Green, S. B., Salkind, N. J., & Akey, T. M. (2000). Using SPSS for Windows:
Analyzing and understanding data (2nd ed.). Upper Saddle River, New Jersey:
Prentice Hall.

Graafmans, W. C., Ooms, M. E., Hofstee, H. M. A., Bezemer, P. D., Bouter, L. M.,
& Lips, P. (1996). Falls in the elderly: A prospective study of risk factors and
risk profiles. American Journal of Epidemiology, 143, 1129-1136.

Gregory, R. L. (1997). Eye and brain: The psychology of seeing (5th ed.).
Princeton, NJ: Princeton University Press.

Guralnik, J. M., LaCroix, A. Z., Abbott, R. D., Berkman, L. F., Satterfield, S. Evans,
D. A. (1993). Maintaining mobility in late life. I. Demographic characteristics
and chronic conditions. American Journal of Epidemiology, 137, 845 – 857.

Haegerstrom-Portnoy, G., & Morgan, M. W. (2007). Normal age-related vision
changes. In A. A. Rosenbloom (Ed.), Rosenbloom & morgan's vision and
aging (pp. 31-48). St. Louis, MO: Butterworth Heinemann.

Hannan, M. T., Felson, D. T., Dawson-Hughes, B., Tucker, K. L., Cupples, L. A.,
Wilson, P. W. F., et al. (2000). Risk factors for longitudinal bone loss in elderly
men and women: The Framingham osteoporosis study. Journal of Bone and
Mineral Research, 15, 710-720.

Harding, J. J. (2002). Viewing molecular mechanisms of ageing through a lens.
Ageing Research Reviews, 1, 465-479.

159
Haronian, E., Wheeler, N. C., & Lee, D. A. (1993/0). Prevalence of eye disorders
among the elderly in Los Angeles. Archives of Gerontology and Geriatrics, 17,
25-36.

Harris, T. B., Ferrucci, L., Tracy, R. P., Corti, M. C., Wacholder, S., EttingerJr, W.
et al. (1999). Associations of elevated interleukin-6 and C-reactive protein
levels with mortality in the elderly. American Journal of Medicine, 106,
506-512.

Harwood, R. H. (2001). Visual problems and falls. Age and Ageing, 30-S4, 13-18.

Harwood, R. H., Foss, A. J. E., Osborn, F., Gregson, R. M., Zaman, A., & Masud,
T. (2005). Falls and health status in elderly women following first eye cataract
surgery: A randomised controlled trial. British Journal of Ophthalmology, 89,
53-59.

Hassan, S. E., Lovie-Kitchin, J. E., & Woods, R. L. (2002). Vision and mobility
performance of subjects with age-related macular degeneration. Optometry
and Vision Science, 79, 697-707.

Hassell, J. B., Lamoureux, E. L., & Keeffe, J. E. (2006). Impact of age related
macular degeneration on quality of life. British Journal of Ophthalmology, 90,
593-596.

Haymes, S. A., Johnston, A. W., & Heyes, A. D. (2002). Relationship between
vision impairment and ability to perform activities of daily living. Ophthalmic
and Physiological Optics, 22, 79-91.

Hornbrook, M. C., Stevens, V. J., Wingfield, D. J., Hollis, J. F., Greenlick, M. R., &
Ory, M. G. (1994). Preventing falls among community-dwelling older persons:
Results from a randomized trial. The Gerontologist, 34, 16-23.

Horowitz, A. (1994). Vision impairment and functional disability among nursing
home residents. The Gerontologist, 34, 316-323.

Horowitz, A. (2004). The prevalence and consequences of vision impairment in
later life. Topics in Geriatric Rehabilitation. Vision Rehabilitation, 20, 185-195.

Horowitz, A., Brennan, M., & Reinhardt, J. P. (2005). Prevalence and risk factors
for self-reported visual impairment among middle-aged and older adults.
Research on Aging, 27, 307-326.

Hunter, G. R., McCarthy, J. P., & Bamman, M. M. (2004). Effects of resistance
training on older adults. Sports Medicine, 34, 329-348.

160
Hyman, L., Schachat, A. P., He, Q., & Leske, M. C. (2000). Hypertension,
cardiovascular disease, and age-related macular degeneration. Archives of
Ophthalmology, 118, 351-358.

Ivers, R. Q., Cumming, R. G., Mitchell, P., & Attebo, K. (1998). Visual impairment
and falls in older adults: The Blue Mountains eye study. Journal of the
American Geriatrics Society, 46, 58-64.

Ivers, R. Q., Norton, R., Cumming, R. G., Butler, M., & Campbell, A. J. (2000).
Visual impairment and hip fracture. American Journal of Epidemiology, 152,
633-639.

Jacobs, J. M., Hammerman-Rozenberg, R., Maaravi, Y., Cohen, A., & Stessman,
J. (2005). The impact of visual impairment on health, function and mortality.
Aging Clinical and Experimental Research, 17, 281-286.

Joussen, A. M., Poulaki, V., Le, M. L., Koizumi, K., Esser, C., Janicki, H., et al.
(2004). A central role for inflammation in the pathogenesis of diabetic
retinopathy. The FASEB Journal, 18, 1450-1452.

Josephson, K. R., Fabacher, D. A., & Rubenstein, L. Z. (1991). Home safety and
fall prevention. Clinical Geriatric Medicine, 7, 707-731.

Kahan, B., & Goodstadt, M. (2001). The interactive domain model of best practices
in health promotion: Developing and implementing a best practices approach
to health promotion. Health Promotion Practice, 2, 43-67.

Kahn, H. A., Leibowitz, H. M., Ganley, J. P., Kini, M. M., Colton, T., Nickerson, R.
S., et al. (1977). The Framingham eye study: I. outline and major prevalence
findings. American Journal of Epidemiology, 106, 17-32.

Kato, I., Dnistrian, A. M., Schwartz, M., Toniolo, P., Koenig, K., Shore, R. E., et al.
(1999). Serum folate, homocysteine and colorectal cancer risk in women: A
nested case-control study. British Journal of Cancer, 79, 1917-1921.

Kirchner, C., & Schmeidler, E. (1997). Prevalence and employment of people in
the United States who are blind or visually impaired. Journal of Visual
Impairment and Blindness, 91, 508-511.

Klein, B. E. K., Moss, S. E., Klein, R., Lee, K. E., & Cruickshanks, K. J. (2003).
Associations of visual function with physical outcomes and limitations 5 years
later in an older population. Ophthalmology, 110, 644-650.

161
Klein, B. E. K., Klein, R., Lee, K. E., Knudtson, M. D., & Tsai, M. Y. (2006). Markers
of inflammation, vascular endothelial dysfunction, and age-related cataract.
American Journal of Ophthalmology, 141, 116-122.

Kochera, A. S. (2002). Falls among older persons and the role of the home: An
analysis of cost incidence and potential savings from home modifications.
Washington, D.C.: American Association of Retired Persons.

Koretz, J. F., Cook, C. A., & Kaufman, P. L. (2002). Aging in the human lens:
Changes in lens shape upon accommodation and with accommodative loss.
Journal of the Ophthalmalogical Society of America, 19, 144-151.

Laforge, R., Spector, W., & Sternberg, J. (1992). The relationship of vision and
hearing impairment to one-year mortality and functional decline. Journal of
Aging and Health, 4, 126-148.

Lamoureux, E. L., Hassell, J. B., & Keeffe, J. E. (2004). The determinants of
participation in activities of daily living in people with impaired vision. American
Journal of Ophthalmology, 137, 265-270.

Lawton, M. P., & Nahemow, L. (1973). Ecology and the aging process. In C.
Eisdorfer, & M. P. Lawton (Eds.), Psychology of Adult Development and
Aging. Washington D.C.: American Psychological Association.

Leibowitz, H. M., Krueger, D. E., Maunder, L. R., Milton, R. C., Kini, M. M., Kahn,
H. A., et al. (1980). The Framingham eye study monograph: An
ophthalmological and epidemiological study of cataract, glaucoma, diabetic
retinopathy, macular degeneration, and visual acuity in a general population of
2631 adults, 1973 – 1975. Survey of Ophthalmology, 24 (suppl.), 335-610.

Leslie, M., & St. Pierre, R. W. (1999). An integrated risk assessment approach to
fall prevention among community dwelling elderly. American Journal of Health
Studies, 15, 57-62.

Li, W., Keegan, T. H. M., Sternfeld, B., Sidney, S., Quesenberry, C. P.,Jr, &
Kelsey, J. L. (2006). Outdoor falls among middle-aged and older adults: A
neglected public health problem. American Journal of Public Health, 96,
1192-1200.

Lighthouse International. (2008). Prevalence of vision impairment. Retrieved June
26, 2008, from http://www.lighthouse.org/research/statistics-on-vision-
impairment/prevalence/

162
Lindle, R. S., Metter, E. J., Lynch, N. A., Fleg, J. L., Fozard, J. L., Tobin, J., et al.
(1997). Age and gender comparisons of muscle strength in 654 women and
men aged 20-93 yr. Journal of Applied Physiology, 83, 1581-1587.

Lord, S. R. (2006). Visual risk factors for falls in older people. Age and Ageing, 35
(suppl. ii), 42-45.

Lord, S. R., Clark, R. D., & Webster, I. W. (1991). Visual acuity and contrast
sensitivity in relation to falls in an elderly population. Age and Ageing, 20,
175-181.

Lord, S. R., & Dayhew, J. (2001). Visual risk factors for falls in older people.
Journal of the American Geriatrics Society, 49, 508-515.

Lord, S. R., Dayhew, J., & Howland, A. (2002). Multifocal glasses impair
edge-contrast sensitivity and depth perception and increase the risk of falls in
older people. Journal of the American Geriatrics Society, 50, 1760-1766.

Lord, S. R., Menz, H. B., & Sherrington, C. (2006). Home environment risk factors
for falls in older people and the efficacy of home modifications. Age and
Ageing, 35 (suppl. ii), 55-59.

Lord, S. R., Sherrington, C, & Menz, H. B. (2001). Falls in older people: Risk
factors and strategies for prevention. New York: Cambridge University Press.

Lord, S. R., Ward, J. A., Williams, P., & Strudwick, M. (1995). The effect of a
12-month exercise trial on balance, strength, and falls in older women: A
randomized controlled trial. Journal of the American Geriatric Society, 43,
1198-1206.

Mace, R.L., Hardie, G.J., & Place, J.P. (1996). Accessible environments: Toward
universal design. North Carolina State University: The Center for Universal
Design.

MacKinnon, D. P., & Dwyer, J. H. (1993). Estimating mediated effects in
prevention studies. Evaluation Review, 17, 144 – 158.

Marmar, M. F. (1982). Aging and the retina. In R. Sekuler, D. Kline, & K. Dismukes
(eds.) Aging and Human Visual Function (pp. 60 – 78). New York: Alan R. Liss,
Inc.

Masud, T., & Morris, R. O. (2001). Epidemiology of falls. Age and Ageing, 30
(suppl. iv), 3-7.

163
McCarty, C. A., Nanjan, M. B., & Taylor, H. R. (2001). Vision impairment predicts 5
year mortality. British Journal of Ophthalmology, 85, 322-326.

McCartney, N., Hicks, A. L., Martin, J., & Webber, C. E. (1995). Long-term
resistance training in the elderly: Effects on dynamic strength, exercise
capacity, muscle, and bone. Journals of Gerontology Series A: Biological
Sciences and Medical Sciences, 50, B97-104.

Menz, H. B., Lord, S. R., St George, R., & Fitzpatrick, R. C. (2004). Walking
stability and sensorimotor function in older people with diabetic peripheral
neuropathy. Archives of Physical Medicine and Rehabilitation, 85, 245-252.

Millodot, M. (2004). Dictionary of optometry and visual science (6th ed.). New
York: Butterworth-Heinemann.

Milne, S., Dickinson, A., Carmichael, A., Sloan, D., Eisma, R., & Gregor, P. (2005).
Are guidelines enough? An introduction to web sites accessible to older
people. IBM Systems Journal, 44, 557 – 571.

Moore, J. E., Steinman, B. A., Giesen, J. M., & Frank, J. J. (2006). Functional
outcomes and consumer satisfaction in the independent living program for
older individuals who are blind. Journal of Visual Impairment and Blindness,
100, 285-294.

Morgenthal, A. P., & Shephard, R. J. (2005). Physiological aspects of aging. In, C.
J. Jones, & D. J. Rose (Ed.s), Physical activity instruction of older adults (pp.
37-53). Champlaign, IL: Human Kinetics.

Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When moderation is mediated and
mediation is moderated. Journal of Personality and Social Psychology, 89,
852-863.

Nagi, S. Z. (1965). Some conceptual issues in disability and rehabilitation. In M. B.
Sussman (Ed.), Sociology and rehabilitation (pp. 100-113). Washington, D.C.:
American Sociological Association.

Nagi, S. Z. (1976). An epidemiology of disability among adults in the united states.
The Milbank Memorial Fund Quarterly: Health and Society, 54, 439-467.

Nygard, O., Refsum, H., Ueland, P., & Vollset, S. (1998). Major lifestyle
determinants of plasma total homocysteine distribution: The Hordaland
homocysteine study. American Journal of Clinical Nutrition, 67, 263-270.

164
O'Neill, T. W., Varlow, J., Silman, A. J., Reeve, J., Reid, D. M., Todd, C., et al.
(1994). Age and sex influences on fall characteristics. Annals of the
Rheumatic Diseases, 53, 773-775.

Orr, A. L. (1992). Aging and blindness: Toward a systems approach to service
delivery. In A. L. Orr (Ed.), Vision and aging: Crossroads for service delivery
(pp. 3-31). New York: American Foundation for the Blind.

Orr, A. L. (1998). Issues in aging and vision: A curriculum for university programs
and in-service training. New York: AFB Press.

Orr, A. L., & Rogers, P. (2006). Aging and vision loss: A handbook for families.
New York: AFB Press.

Patel, I., Turano, K. A., Broman, A. T., Bandeen-Roche, K., Munoz, B., & West, S.
K. (2006). Measures of visual function and percentage of preferred walking
speed in older adults: The Salisbury eye evaluation project. Investigative
Ophthalmology Visual Science, 47, 65-71.

Preacher, K. J., & Hayes, A. F. (2004). SPSS procedures for estimating indirect
effects in simple mediation models. Behavior Research Methods, 36, 717 –
731.

Province, M. A., Hadley, E. C., Hornbrook, M. C., Lipsitz, L. A., Miller, J. P.,
Mulrow, C. D., et al. (1995). The effects of exercise on falls in elderly patients.
A preplanned meta-analysis of the FICSIT trials. Journal of the American
Medical Association, 273, 1341-1347.

Pynoos, J. (1992). Strategies for home modification and repair. Generations,
16(2), 21-25.

Pynoos, J., Caraviello, R., & Cicero, C. (2009). Lifelong housing: The anchor in
aging-friendly communities. Generations, 33(2), 26-32.

Pynoos, J., Rose, D., Rubenstein, L., Choi, I. H., & Sabata, D. (2006).
Evidence-based interventions in fall prevention. In S. M. Enguidanos (Ed.),
Evidence-based interventions for community dwelling older adults (pp. 55-73).
New York: Hawthorn.

Pynoos, J., Sabata, D., & Choi, I. H. (2005). The role of the environment in fall
prevention at home and in the community. Falls free: Promoting a national falls
prevention action plan (pp. 41-54). Washington, DC: National Council on the
Aging.

165

Pynoos, J., Steinman, B. A., & Nguyen, A. Q. D. (in press). Environmental
assessment and modification as fall-prevention strategies for older adults.
Clinics in Geriatric Medicine.

RAND. (2010). RAND HRS Data. Version I. Produced by the Rand Center for the
Study of Aging, with funding from the National Institute on Aging and the Social
Security Administration. Santa Monica, CA (March, 2010).

Ray, C. T., & Wolf, S. L. (2008). Review of intrinsic factors related to fall risk in
individuals with visual impairments. Journal of Rehabilitation Research &
Development, 45, 1115 – 1124.

Refsum, H., Ueland, P. M., Nygard, O., & Vollset, S. E. (1998). Homocysteine and
cardiovascular disease. Annual Review of Medicine, 49, 31-62.

Resnikoff, S., Pascolini, D., Mariotti, S. P., & Pokharel, G. P. (2008). Global
magnitude of vision impairment caused by uncorrected refractive errors in
2004. Bulletin of the World Health Organization, 86, 63 – 70.

Rogers, M. E., Rogers, N. L., Takeshima, N., & Islam, M. M. (2004). Reducing the
risk of falls in the homes of older adults. Journal of Housing for the Elderly, 18,
29-39.

Rubenstein, L. Z. (2006). Falls in older people: Epidemiology, risk factors and
strategies for prevention. Age and Ageing, 35 (suppl. ii), 37-41.

Rubenstein, L. Z., Robbins, K. R., Josephson, K. R., Schulman, B. L., & Osterweil,
D. (1990). The value of assessing falls in an elderly population: A randomized
clinical trial. Annals of Internal Medicine, 113, 308 – 316.

Rubin, G. S., West, S. K., Munoz, B., Bandeen-Roche, K., Zeger, S., Schein, O.,
et al. (1997). A comprehensive assessment of visual impairment in a
population of older Americans: The Salisbury eye evaluation project.
Investigative Ophthalmology Visual Science, 38, 557-568.

Ryskulova, A., Turczyn, K., Makuc, D. M., Cotch, M. F., Klein, R. J., & Janiszewski,
R. (2008). Self-reported age-related eye diseases and visual impairment in the
United States: Results of the 2002 National Health Interview Survey. American
Journal of Public Health, 98, 454-461.

Salive, M. E., Guralnik, J., Glynn, R. J., Christen, W., Wallace, R. B., & Ostfeld, A.
M. (1994). Association of visual impairment with mobility and physical function.
Journal of the American Geriatric Society, 42, 287-292.

166
Sanford, J. A. (in press). The physical environment and home health care. In
National Research Council, The Role of Human Factors in Home Health Care:
Workshop Summary. Washington, DC: The National Academies Press.

Schalkwijk, C. G., Poland, D. C. W., van Dijk, W., Kok, A., Emeis, J. J., Dräger, A.
M., et al. (1999). Plasma concentration of C-reactive protein is increased in
type I diabetic patients without clinical macroangiopathy and correlates with
markers of endothelial dysfunction: Evidence for chronic inflammation.
Diabetologia, 42, 351-357.

Seddon, J. M., Gensler, G., Milton, R. C., Klein, M. L., & Rifai, N. (2004).
Association between C-reactive protein and age-related macular
degeneration. Journal of the American Medical Association, 291, 704-710.

Seeman, T. E., Crimmins, E., Huang, M. H., Singer, B., Bucur, A., Gruenewald, T.,
et al. (2004). Cumulative biological risk and socio-economic differences in
mortality: MacArthur studies of successful aging. Social Science and
Medicine, 58, 1985-1997.

Selhub, J., Jacques, P. F., Rosenberg, I. H., Rogers, G., Bowman, B. A., Gunter,
E. W., et al. (1999). Serum total homocysteine concentrations in the third
national health and nutrition examination survey (1991–1994): Population
reference ranges and contribution of vitamin status to high serum
concentrations. Annals of Internal Medicine, 131, 331-339.

Selhub, J., Jacques, P. F., Wilson, P. W. F., Rush, D., & Rosenberg, I. H. (1993).
Vitamin status and intake as primary determinants of homocysteinemia in an
elderly population. Journal of the American Medical Association, 270,
2693-2698.

Seshadri, S., Beiser, A., Selhub, J., Jacques, P. F., Rosenberg, I. H., D'Agostino,
R. B., et al. (2002). Plasma homocysteine as a risk factor for dementia and
Alzheimer's disease. New England Journal of Medicine, 346, 476-483.

Shumway-Cook, A., Gruber, W., Baldwin, M., & Liao, S. (1997). The effect of
multidimensional exercises on balance, mobility, and fall risk in
community-dwelling older adults. Physical Therapy, 77, 46-57.

Smith, W., Mitchell, P., Leeder, S. R., & Wang, J. J. (1998). Plasma fibrinogen
levels, other cardiovascular risk factors, and age-related maculopathy.
Archives of Ophthalmology, 116, 583-587.

Soong, G. P., Lovie-Kitchin, J. E., & Brown, B. (2004). Measurements of preferred
walking speed in subjects with central and peripheral vision loss. Ophthalmic
and Physiological Optics, 24, 291-295.

167

Steinman, B. A. (2008). Self-reported vision, upper/lower limb disability, and fall
risk in older adults. Journal of Applied Gerontology, 27, 406-423.

Steinman, B. A., Pynoos, J., & Nguyen, A. Q. D. (2009). Fall risk in older adults:
Roles of self-rated vision, home modification, and limb function. Journal of
Aging and Health, 21, 655 – 676.

Steinman, B. A., & Vasunilashorn, S. (in press). Biological risk of older adults with
visual impairment. Journal of Nutrition, Health, and Aging.

Stevens, J. A. (2002/2003). Falls among older adults: Public health impact and
prevention strategies. Generations, 26(4), 7-14.

Strenk, S., Semmlow, J., Strenk, L., Munoz, P., Gronlund-Jacob, J., & DeMarco, J.
(1999). Age-related changes in human ciliary muscle and lens: A magnetic
resonance imaging study. Investigative Ophthalmology Visual Science, 40,
1162-1169.

Stuen, C., & Faye, E. E. (2003). Vision loss: Normal and not normal changes
among older adults. Generations, 27(1), 8-14.

Taaffe, D. R., Duret, C., Wheeler, S., & Marcus, R. (1999). Once-weekly resistance
exercise improves muscle strength and neuromuscular performance in older
adults. Journal of the American Geriatric Society, 47, 1208 -1214.

Taylor, H. R. (2000). Cataract: How much surgery do we have to do? British
Journal of Ophthalmology, 84, 1-2.

The Eye Diseases Prevalence Research Group. (2004a). Causes and prevalence
of visual impairment among adults in the United States. Archives of
Ophthalmology, 122, 477-485.

The Eye Diseases Prevalence Research Group. (2004b). The prevalence of
refractive errors among adults in the United States, western Europe, and
Australia. Archives of Ophthalmology, 122, 495-505.

Tideiksaar, R. (1998). Falls in older persons: Prevention and management (2
nd

ed.) Baltimore, MD: Health Professions Press.

Tielsch, J. M., Sommer, A., Witt, K., Katz, J., & Royall, R. M. (1990). Blindness
and visual impairment in an American urban population. The Baltimore eye
survey. Archives of Ophthalmology, 108, 286-290.

168
Tiemeier, H., van Tuijl, H. R., Hofman, A., Meijer, J., Kiliaan, A. J., & Breteler, M. M.
B. (2002). Vitamin B12, folate, and homocysteine in depression: The
Rotterdam study. American Journal of Psychiatry, 159, 2099-2101.

Tinetti, M. E., Baker, D. I., McAvay, G., Claus, E. B., Garrett, P., Gottschalk, M., et
al. (1994). A multifactorial intervention to reduce the risk of falling among
elderly people living in the community. New England Journal of Medicine, 331,
821-827.

Tinetti, M. E., Williams, T. F., Mayewski, R. (1986). Fall risk index for elderly
patients based on number of chronic disabilities. American Journal of
Medicine, 80, 429 – 434.

Travis, L. A., Boerner, K., Reinhardt, J. P., & Horowitz, A. (2004). Exploring
functional disability in older adults with low vision. Journal of Visual Impairment
and Blindness, 98, 534-545.

Turano, K. A., Broman, A. T., Bandeen-Roche, K., Munoz, B., Rubin, G. S., &
West, S. K. (2004). Association of visual field loss and mobility performance in
older adults: Salisbury eye evaluation study. Optometry and Vision Science,
81, 298-307.

U.S. Census Bureau. (2010a). U.S. interim projections by age, sex, race, and
Hispanic origin: 2000-2050. Retrieved June 30, 2010, from
http://www.census.gov/ipc/www/usinterimproj/

U.S. Census Bureau. (2010b). National population estimates - characteristics.
Retrieved June 30, 2010, from http://www.census.gov/popest/national/asrh
/NC-EST2007-sa.html

U.S. Census Bureau. (2010c). International data base (IDB). Retrieved June 30,
2010, from http://www.census.gov/ipc/www/idb/pyramids.html

University of Michigan Health and Retirement Study. (2010). The Health and
Retirement Study: A longitudinal study of health, retirement, and aging
sponsored by the National Institute on Aging. Retrieved June 30, 2010, from
http://hrsonline.isr.umich.edu/

Varma, R., Wu, J., Chong, K., Azen, S., & Hays, R. (2006). Impact of severity and
bilaterality of visual impairment on health-related quality of life.
Ophthalmology, 113, 1846-1853.

Verbrugge, L. M., & Jette, A. (1994). The disablement process. Social Science
Medicine, 39, 1-14.

169
Vine, A. K., Stader, J., Branham, K., Musch, D. C., & Swaroop, A. (2005).
Biomarkers of cardiovascular disease as risk factors for age-related macular
degeneration. Ophthalmology, 112, 2076-2080.

Wallace, R. B., & Herzog, A. R. (1995). Overview of the health measures in the
Health and Retirement Study. Journal of Human Resources, 30, S84 – S107.

Wang, J. J., Mitchell, P., Simpson, J. M., Cumming, R. G., & Smith, W. (2001).
Visual impairment, age-related cataract, and mortality. Archives of
Ophthalmology, 119, 1186-1190.

Wang, J., Mitchell, P., & Smith, W. (2000). Vision and low self-rated health: The
Blue Mountains eye study. Investigative Ophthalmology Visual Science, 41,
49-54.

West, C. G., Gildengorin, G., Haegerstrom-Portnoy, G., Schneck, M. E., Lott, L., &
Brabyn, J. A. (2002). Is vision function related to physical functional ability in
older adults? Journal of the American Geriatrics Society, 50, 136-145.

Wilcken, D. E. (1998). Novel risk factors for vascular disease: The homocysteine
hypothesis of cardiovascular disease. Journal of Cardiovascular Risk, 5,
217-221.

Williams, D. L. (2006). Oxidation, antioxidants and cataract formation: A literature
review. Veterinary Ophthalmology, 9, 292-298.

Wilson, C. J., Finch, C. E., & Cohen, H. J. (2002). Cytokines and cognition-the
case for a head-to-toe inflammatory paradigm. Journal of the American
Geriatrics Society, 50, 2041-2056.

Wolff, J. L., Starfield, B., & Anderson, G. (2002). Prevalence, expenditures, and
complications of multiple chronic conditions in the elderly. Archives of Internal
Medicine, 162, 2269-2276.

Wolff, I., van Croonenborg, J. J., Kemper, H. C. G., Kostense, P. J., & Twisk, J. W.
R. (1999). The effect of exercise training programs on bone mass: A
meta-analysis of published controlled trials in pre- and postmenopausal
women. Osteoporosis International, 9, 1-12.

Young, L.C. (2006). Residential Rehabilitation, Remodeling and Universal Design.
North Carolina State University: The Center for Universal Design.

Abstract (if available)

Abstract The purpose of this dissertation was to examine relationships between self reported vision impairment, health dimensions, and falls among older people

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Falls among older adults: characteristics of fallers, predictors of falls, and the impact of falls on health care and long-term care utilization

PDF

Fall related injuries among older adults in the Los Angeles region

PDF

Palliative care: breaking the knowledge barrier among community-based older adults

PDF

Translating two physical activity programs for older adults into home- and community-based settings: "Active Start" and "Healthy Moves for Aging Well"

PDF

The effects of time and space on health status of older adults in China

PDF

Investigating the effectiveness of a social work intervention on reducing hospital readmissions among older adults

PDF

Using mixed methods to identify the characteristics of older fraud victims

PDF

Predicting cognitive decline and dementia in elderly twins from indicators of early life oral health

PDF

Is stress exposure enough? Race/ethnic differences in the exposure and appraisal of chronic stressors among older adults

PDF

Family relationships and their influence on health outcomes over time among older adults in rural China

PDF

Cognitive health and self-rated memory in later life: linkages to race/ethnicity, multimorbidity, and survival

PDF

Thinking generatively versus acting generatively: exploring the associations of generative self-concept and generative activity with cognitive function among older adults

PDF

Integrative care strategies for older adults experiencing co-occurring substance use and mental health disorders (I-CARE)…

PDF

Examination of the long-term psychosocial and functioning effects of a healthy living, life management behavior intervention for older adults

PDF

Estimating survival in the face of pain: evidence from the health and retirement study

PDF

The effects of wisdom-related personality traits on caregivers’ health: an application of the resilience model

PDF

The dollars and sense of caregiving: exploring the financial financial realities of providing informal care to an older adult

PDF

How emotional arousal influences memory and learning in younger and older adults

PDF

Self-perceptions of Aging in the Context of Neighborhood and Their Interplay in Late-life Cognitive Health

PDF

The measurement, life course patterns, and outcomes of intergenerational ambivalence among parent-adult child dyads

Asset Metadata

Creator Steinman, Bernard Alex (author)

Core Title Older adults with visual impairments: the role of health dimensions in predicting falls

School Leonard Davis School of Gerontology

Degree Doctor of Philosophy

Degree Program Gerontology

Publication Date 10/04/2010

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag aging,aging and exercise,aging and health,blindness,Disability,disability model,fall intervention,fall prevention,Falls,health,health dimensions,OAI-PMH Harvest,physical functioning,vision impairment

Place Name USA (countries)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Pynoos, Jon (committee chair), Enguidanos, Susan M. (committee member), Gauderman, W. James (committee member)

Creator Email bernard_steinman@brown.edu,bsteinma@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3491

Unique identifier UC170231

Identifier etd-Steinman-4125 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-417228 (legacy record id),usctheses-m3491 (legacy record id)

Legacy Identifier etd-Steinman-4125.pdf

Dmrecord 417228

Document Type Dissertation

Rights Steinman, Bernard Alex

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu