Full text of DOI:10.1093/gerona/56.4.M243

Journal of Gerontology: MEDICAL SCIENCES
Copyright 2001 by The Gerontological Society of America
2001, Vol. 56A, No. 4, M243–M247
Lower Body Function and Mortality in Mexican
American Elderly People
Kyriakos S. Markides,1 Sandra A. Black,1 Glenn V. Ostir,1 Ronald J. Angel,2 Jack M. Guralnik,3
and Michael Lichtenstein4
1Center on Aging and Department of Preventive Medicine and Community Health,
University of Texas Medical Branch, Galveston.
2Department of Sociology, University of Texas, Austin.
3National Institute on Aging, Bethesda, Maryland.
4Department of Medicine, University of Texas Health Science Center at San Antonio.
Background. The purpose of this analysis was to examine the differential impact of performance-based and self-
reported lower body measures on 2-year mortality in Mexican American elderly persons.
Methods. Data employed are from the Hispanic Established Population for Epidemiological Studies of the Elderly, a
probability survey of 3050 community-dwelling Mexican Americans aged 65 and older from the five Southwestern
states interviewed in 1993 and 1994. Of the baseline sample with complete data, 198 persons were confirmed deceased
2 years later. A three-task, performance-based, lower body function measure consisting of a short walk, balance, and re-
peated chair stands tests was used. Self-reported lower body function was measured by a 4-item Activities of Daily Liv-
ing (ADL) measure involving the lower body.
Results. The three-task, lower body function measure was a significant predictor of 2-year mortality. The short walk
alone was as predictive as the summary measure. The predictive ability of both measures was minimally reduced by the
inclusion of the self-reported ADL measure and life-threatening medical conditions. Finally, the ADL measure was not
a significant predictor of mortality with all the other variables in the analysis.
Conclusion. Objective measures of lower body function were significant predicators of mortality in Mexican Ameri-
can elderly persons, as found in the general population. Unlike previous studies, the ADL measure was not an indepen-
dent predictor of mortality after controlling for the objective measure and other risk factors. Additional research is
needed to address why objective measures of function are such strong predictors of death.
T is becoming increasingly clear that poor lower body
function is a major contributor to disability and loss of
independence among elderly persons. It has been found to
be associated with disability in Activities of Daily Living
(ADL), both cross-sectionally and prospectively (1–4), and
is also a major predictor of self-ratings of health, depressive
symptomatology (4), and institutionalization (5).
Despite its importance as a major predictor of the health
status of older people, few studies have examined the asso-
ciation of lower body function to mortality (6,7). In an anal-
ysis of data from three of the Established Populations for
Epidemiological Studies of the Elderly (EPESE) studies,
Brock and colleagues (6) found that self-report measures of
lower body function were stronger predictors of mortality
than were measures of ADLs. Specifically, a portion of the
Rosow-Breslau (8) scale of lower body function (ability to
walk a half mile, to walk up and down stairs, and to perform
heavy housework) was a significantly stronger predictor of
3-year mortality than was the ADL scale developed by Katz
and colleagues (9), with Branch’s additions (10). The latter
scale covers bathing, dressing, walking, grooming, transfer-
ring from bed to chair, using the toilet, and eating. Similar
results were obtained for both men and women in models
that included the measures of function separately as well as
when they were included in the same model. In the latter,
only the Rosow-Breslau predicted mortality, while the ADL
measure as well as a measure of function derived from Nagi
(11) did not.
I
In a separate analysis of the same three EPESE data sets,
Guralnik and colleagues (5) examined whether a short per-
formance-based measure of lower body function predicted
mortality over an average period of 1.39 years. They also
examined whether presence of one or more ADL limitations
that require lower body function (walking across a room,
bathing, transferring from bed to chair, and using the toilet)
predicted mortality. The performance-based measure was a
summary score of three lower body functions: standing bal-
ance, walking speed over an 8-ft distance, and repeated
chair stands. When performance and self-reported measures
were entered in the same equation, both were significant
predictors of mortality among men. However, only the per-
formance score predicted mortality among women. Thus, at
least among women, this analysis provided evidence that a
performance-based measure of physical function was more
predictive of mortality than was an ADL measure of lower
body limitations.
The purpose of this analysis was to examine the extent to
which performance-based and ADL measures of lower body
M243

M244
MARKIDES ET AL.
function differentially predict mortality in a unique cohort
of Mexican American elderly persons. We have previously
shown (12) that performance-based measures of lower body
function predict the incidence of ADL and mobility-related
disability over two years in Mexican American elderly per-
sons as has been found in the general population (5).
In attempting to predict 2-year mortality in Mexican Amer-
ican elderly persons, we hypothesized that performance-
based measures of lower body function would be stronger
predictors of mortality than self-reported lower body func-
tion. Given what little is known about self-reported physical
function (13–15) as well as more objective measures of
physical function in older Mexican Americans (16), assess-
ing their differential impact on mortality in this population
should generate much-needed knowledge. Finally, we went
beyond past research to examine the extent to which perfor-
mance measures of physical function continue to be predic-
tive of mortality after controlling for the presence of major
life-threatening diseases.
value of 0 (12). The three scores were combined to form a
total lower body function score ranging from 0 to 11. The
scale was further categorized from 0 to 4 to approximate the
timed walk categorization (Table 1).
Self-reported lower body function was assessed with four
ADLs requiring lower body action: transferring from bed to
chair, using the toilet, walking across a small room, and
bathing [same as in (5)]. In the analysis reported below,
subjects were considered ADL disabled if they were unable
to perform any of the four tasks without help.
Other variables assessed were age (75ϩ ϭ 1, 65–74 ϭ 0),
gender (male ϭ 1), and presence of a number of self-reported
potentially life-threatening conditions. These were assessed
by asking respondents if they had ever been told by a doctor
that they had any of the following conditions: heart attack,
cancer, diabetes, hip fracture, stroke, and hypertension.
Analysis
Two-year mortality was assessed through a series of mul-
tiple logistic regression models. Two sets of analyses were
performed, one involving the combined performance score
and one involving the short walk. The latter was conducted
because it has been suggested that the short walk is the most
predictive of the three performance tasks of subsequent dis-
ability (2,12,21).
METHODS
Sample
Data employed are from the Hispanic EPESE. At baseline,
3050 community-dwelling Mexican Americans aged 65 and
older were interviewed in their homes during 1993 and 1994.
The sample was drawn using area probability sampling pro-
cedures to be representative of the Mexican American elderly
population in five Southwestern states: Texas, New Mexico,
Colorado, Arizona, and California. The sample and its char-
acteristics have been described elsewhere (17,18).
The initial sample was followed up 2 years later in 1995
and 1996. Of the original 3050 persons, 2439 were reinter-
viewed, 2167 in person and 272 (8.9%) by proxy. Of the
remaining subjects, 224 (7.3%) were confirmed deceased
through the National Death Index and reports from rela-
tives, 109 (3.5%) refused to be reinterviewed, and 5 (0.2%)
were reported to be living in Mexico. The status of 273 (8.9%)
persons could not be determined (12).
RESULTS
Table 1 presents the distributions of all variables included
in the analysis. Of the total sample of 2636, 34.9% were 75
Table 1. Categories of All Variables in the Analysis (n ϭ 2636)
n
%
Age
65–74 y
75ϩ y
1715
921
65.1
34.9
Gender
Men
1106
1530
42.0
58.0
Women
Performance Score
In the analysis below we include persons with complete
data on the performance measures of physical function and
other variables of interest at baseline; the subjects were de-
termined to be living or deceased at Time 2, two years later.
Of the 2636 eligible subjects, 198 (7.5%) were deceased by
Time 2.
0 (unable)
1
2
3
334
179
689
707
727
12.7
6.8
26.1
26.8
27.6
4
Short Walk
0 (unable)
1
389
479
557
591
620
326
14.8
18.2
21.1
22.4
23.5
12.4
Measures
2
3
4
Our performance-based measure of lower body function
was identical to the one employed by Guralnik and col-
leagues (5) with other EPESE data sets [see also (19,20)].
Basically, three separate tasks were assessed, and a com-
bined score was created to represent lower body function.
The tasks were a timed 8-ft walk, timed repeated chair
stands, and standing balance. The two timed tasks were di-
vided into quartiles each and scored 1 (slowest) to 4 (fast-
est). Subjects unable to complete each task were assigned a
value of 0. The standing balance tests included tandem,
semitandem, and side-by-side stands. Individuals were scored
1 to 3, with 3 indicating the highest performance. Those
unable to perform any of the balance tests were assigned a
Any ADL†
Medical Conditions
Hypertension
Heart attack
Cancer
1132
302
149
594
96
42.9
11.5
5.7
22.5
3.6
Diabetes
Hip fracture
Stroke
175
198
6.6
7.5
Deceased
†Any ADL: unable to or needing help to perform any of the following four
tasks involving the lower body: transferring from bed to chair, using the toilet,
walking across a room, bathing.

BODY FUNCTION AND MORTALITY
M245
years and older, and 58.2% were women. On the summary
performance score, 12.7% were unable to perform any of
the three tasks and were assigned a value of 0. Similarly,
14.8% were unable to perform the short walk. Approxi-
mately 12.4% were ADL dependent at baseline, and 7.5%
were deceased by Time 2.
As in previous work (5), scores on the three performance
tasks were significantly correlated: walk and balance, 0.53;
walk and chair stands, 0.49; balance and chair stands, 0.49
(p Ͻ .0001 for all three). The internal consistency of the
3-item scale was 0.75 using Cronbach’s alpha. In addition,
the ADL item was significantly correlated with each perfor-
mance task (Ϫ0.55 with balance, Ϫ0.47 with short walk, and
Ϫ0.41 with chair stands) as well as with the combined per-
formance score (Ϫ0.57). These correlations are similar to
those previously reported in the literature (22).
Table 2 presents results of logistic regression analyses
predicting 2-year mortality using the summary performance
measure of function. Model 1 included as independent vari-
ables the summary performance score of lower body func-
tion along with age and gender serving as controls. The per-
formance score was entered as four dummy variables with
scores 9 to 11 (best function) serving as the reference cate-
gory. As can be seen, persons in lower categories were all
significantly more likely to experience 2-year mortality.
Moreover, there appeared to be a linear trend with persons
in the 0 category (unable to perform any of the tasks) con-
siderably more likely to die than persons in the highest cate-
gory (odds ratio [OR], 11.80; 95% confidence interval [CI] ϭ
6.80–20.48). Model 2 included only the self-reported lower
body ADL measure along with age and gender. Persons re-
porting disability in any one of the four ADLs were signifi-
cantly more likely to die within 2 years than persons report-
ing no disability (OR, 4.06; 95% CI ϭ 2.90–5.69). Model 3
included both the summary performance measure and the ADL
measure to examine whether both continue to significantly
predict mortality. Even though the ORs declined somewhat,
the performance measure continued to significantly affect
mortality, and a linear trend continued to be observed. How-
ever, the ADL measure was no longer a significant predictor
(OR ϭ 1.52; 95% CI ϭ 0.96–2.42). Clearly, our data show
the clinical utility of the short performance battery of lower
body function in predicting mortality in Mexican American
elderly persons as has been found in the general elderly popu-
lation (5). At the same time, the self-reported ADL measure
lost its predictive ability when the performance measure was
entered into the equation.
Previous research linking performance measures of lower
body function adverse outcomes including mortality (5,23)
typically stops at this stage. We went a step further by re-
peating our analyses controlling for common life-threaten-
ing conditions (hypertension, heart disease, cancer, diabe-
tes, hip fracture, and stroke) in model 4. Interestingly, even
though the ORs for the performance measure declined some-
what, the general pattern remained the same. As in the pre-
vious model, the ADL measure was not a significant predic-
tor of mortality.
It has been suggested that the short walk component of
the three-task summary measure of lower body function might
be as predictive or almost as predictive of subsequent dis-
ability as the summary measure and that it may be more pre-
dictive than either chair stands or balance (2,12,21). How-
ever, the predictive ability of the short walk has not been
investigated with respect to mortality. We examined this re-
lationship and report the results in Table 3.
Model 1, which includes only the short walk in addition
to age and gender, yielded results almost identical to those
yielded by the summary performance measure (Table 2).
Significant ORs were computed for each of the four catego-
ries rising to 11.52 (CI ϭ 6.35–20.90) for category 0, which
contained subjects unable to complete the task.
Model 3 included the addition of the ADL measure. The
ORs were reduced somewhat, but, again, the results were al-
most identical to those for the summary measure in Table 2,
with the ADL measure being only marginally significant. Fi-
nally, when the medical conditions were added to the equa-
Table 2. Logistic Regression Models Predicting Mortality from Summary Performance Score, ADL Disability, Chronic Conditions,
Gender, and Age (n ϭ 2636)
Model 1
Model 2
Model 3
Model 4
Age
(75ϩ y vs 65–74 y)
Gender (male)
1.43 (1.05–1.96)
1.75 (1.29–2.37)
1.62 (1.19–2.20)
1.58 (1.17–2.12)
1.38 (1.01–1.89)
1.76 (1.30–2.38)
1.42 (1.02–1.97)
1.84 (1.34–2.52)
Summary Performance Score (vs 9–11)
7–8
4–6
1–3
0
2.11 (1.18–3.84)
3.44 (1.98–5.96)
3.56 (1.74–7.30)
11.80 (6.80–20.48)
2.11 (1.18–3.80)
3.41 (1.96–5.92)
3.17 (1.52–6.60)
8.93 (4.72–16.91)
1.52 (0.96–2.42)
2.02 (1.12–3.64)
3.25 (1.86–5.67)
2.87 (1.37–6.05)
7.39 (3.86–14.13)
1.46 (0.90–2.36)
Lower Body ADL Disability (any ADL)
4.06 (2.90–5.69)
Chronic Conditions
Hypertension
Heart disease
Cancer
1.09 (0.79–1.51)
1.59 (1.07–2.38)
3.18 (2.02–5.01)
1.78 (1.29–2.51)
1.35 (0.73–2.50)
0.81 (0.48–1.36)
Diabetes
Hip fracture
Stroke
Notes: Values represent odds ratios and 95% confidence intervals. ADL ϭ Activity of Daily Living.

M246
MARKIDES ET AL.
Table 3. Logistic Regression Models Predicting Mortality From Short Walk, ADL Disability, Chronic Conditions, Gender, and Age (n ϭ 2636)
Model 1
Model 2
Model 3
Model 4
Age
(75ϩ y vs 65–74)
Gender (male)
1.51 (1.11–2.06)
1.67 (1.23–2.26)
1.62 (1.19–2.20)
1.58 (1.17–2.12)
1.43 (1.04–1.95)
1.69 (1.24–2.29)
1.46 (1.06–2.03)
1.76 (1.29–2.42)
Short Walk (vs 4)
3
2
1
0
2.34 (1.22–4.46)
2.78 (1.45–5.32)
3.79 (2.02–7.09)
11.52 (6.35–20.90)
2.34 (1.23–4.47)
2.76 (1.44–5.30)
3.69 (1.97–6.93)
8.62 (4.56–16.66)
1.62 (1.04–2.53)
2.16 (1.12–4.14)
2.57 (1.33–4.94)
3.64 (1.93–6.85)
7.47 (3.83–14.55)
1.48 (0.93–2.36)
Lower Body ADL Disability (any ADL)
4.06 (2.90–5.69)
Chronic Conditions
Hypertension
Heart disease
Cancer
1.11 (0.80–1.53)
1.60 (1.08–2.39)
3.37 (2.14–5.30)
1.80 (1.29–2.52)
1.35 (0.73–2.48)
0.82 (0.49–1.38)
Diabetes
Hip fracture
Stroke
Notes: Values represent odds ratios and 95% confidence intervals. ADL ϭ activity of daily living.
tion (model 4), for the short walk we again obtained ORs al-
most identical to those obtained for the summary measure
and reported in Table 2. Similarly, with the scores on the
short walk and medical conditions entered into the equation,
the self-reported ADL measure was not a significant predic-
tor of mortality.
Thus it appears that the short walk is as predictive of 2-year
mortality as the summary measure. When we repeated the
analyses for chair stands and balance (results not shown),
our findings were considerably weaker and less consistent
than those for the summary score or the short walk (results
available from the authors).
Why are performance measures so much more predictive
than self-report measures in our analysis? Performance
measures are fairly objective indicators of actual ability and
probably are an extreme manifestation of the disablement
process (24) when individuals reach the point when they are
unable to do a task. ADL measures are more subtle, but
more complex, and probably reflect individuals’ adaptation
to physical decline. Why ADLs appear to perform less well
in this sample of Mexican American elderly persons than
previously reported with other groups is not clear. Perhaps
they are subject to more error in this population, although
they correlate at similar levels with performance measures
as in the general population (22).
An important contribution of this study was the continued
strong influence of performance measures on mortality even
after controlling for life-threatening medical conditions.
Moreover, their influence on mortality was stronger than
that of any disease, including cancer, at least for the extreme
end of the distribution, which included those unable to per-
form a task. Clearly, performance measures of function are
useful for identifying high-risk persons in the elderly popu-
lation, which is better than simply relying on disease pro-
files. Identifying high-risk persons using objective standard-
ized tests will assist in providing them with better medical
care and preventive interventions.
DISCUSSION
As found in previous research, a simple three-task battery
of lower body function was a significant predictor of short-
term mortality in Mexican American elderly persons. In addi-
tion, the short walk alone was as predictive as the summary
measure and more predictive than the balance and chair stand
tasks. The predictive ability of both the summary measure
and short walk were reduced only minimally by the addition
of a self-reported lower body ADL measure and by the addi-
tion of life-threatening medical conditions. Finally, the self-
reported ADL measure was no longer a significant predictor
of mortality when the summary performance measure and
life-threatening medical conditions were entered into the
equation, as well as when the short walk and life-threatening
medical conditions were entered into the equation.
Additional research is needed to help us better understand
why poor physical function is so highly related to death. Ex-
cept for fatal falls and accidents, poor physical function does
not kill people. Obviously, it is a marker for poor health and
the adverse consequences of chronic disease. Although the
association between chronic disease and physical function in
the elderly is an area of extensive research, our understand-
ing of this relationship remains limited, especially among
Mexican Americans and other minority populations (13).
Our findings are consistent with previous literature show-
ing that simple physical performance tests of lower body
function are strong predictors of mortality (2,23). It also
supports the finding that lower body function may be a bet-
ter predictor of mortality than self-reported ADL function
(6). Although previous research has found that both perfor-
mance and ADL measures are significant independent pre-
dictors of mortality (5,6,23), we found that the predictive
power of self-reported ADL lower body function becomes
insignificant when performance measures and medical con-
ditions are entered into the equation.
Acknowledgments
This study was supported by Grant AG10939 (Hispanic Established Popu-
lation for Epidemiologic Studies of the Elderly) and Grant 17231 (The Uni-
versity of Texas Medical Branch Claude Pepper Older Americans Indepen-

BODY FUNCTION AND MORTALITY
M247
dence Center) both funded by the National Institute on Aging. The field
work was conducted by Harris Interactive, New York. This research was
presented at the 1998 meeting of The Gerontological Society of America.
tioning as a predictor of subsequent disability among older Mexican
Americans. J Gerontol Med Sci. 1998;53A:M1–M5.
13. Clark DO, Mungai SM, Stump TE, Wolinsky FD. Prevalence and im-
pact of risk factors of lower body difficulty among Mexican Americans,
African Americans, and Whites. J Gerontol Med Sci. 1997;52A:M97–
M105.
14. Rudkin L, Markides KS, Espino DV. Functional disability in older
Mexican Americans. Top Geriatr Rehabil. 1997;12:38–46.
15. Markides KS, Rudkin L, Angel RJ, Espino DV. Health status of His-
panic elderly. In: Martin LJ, Soldo B, eds. Racial and Ethnic Differences
in the Health of Older Americans. Washington, DC: National Acad-
emy Press; 1997.
16. Perkowski LC, Stroup-Benham C, Markides KS, et al. Lower-extrem-
ity function in older Mexican Americans and its association with medical
problems. J Am Geriatr Soc. 1998;46:411–418.
17. Markides KS, Stroup-Benham CA, Goodwin JS, Perkowski LC, Lich-
tenstein M, Ray LA. The effects of medical conditions on the functional
limitations of Mexican American elderly. Ann Epidemiol. 1996;6:386–
391.
18. Markides KS, Stroup-Benham CA, Black SA, Satish S, Perkowski LC,
Ostir G. The health of Mexican American elderly: Selected findings
from the Hispanic EPESE. In: Wykle ML, Ford AB, eds. Serving Mi-
nority Elders in the 21st Century. New York: Springer; 1999.
19. Jette AM, Branch LG. Impairment and disability in the aged. J Chron
Dis. 1985;38:59–65.
Address correspondence to Kyriakos S. Markides, PhD, Department of
Preventive Medicine & Community Health, University of Texas Medical
Branch, Galveston, TX 77555-1153. E-mail: Kmarkide@utmb.edu
References
1. Guralnik JM, Ferrucci L, Simonsick EM, Salive ME, Wallace RB.
Lower-extremity function in persons over the age of 70 years as a pre-
dictor of subsequent disability. N Engl J Med. 1995;332:556–561.
2. Guralnik JM, Ferrucci F, Pieper CF, et al. Lower extremity function
and subsequent disability: consistency across studies, predictive mod-
els, and value of gait speed alone compared to the short physical
performance battery. J Gerontol Med Sci. 2000;55A:M221–M231.
3. Johnson RJ, Wolinski FD. The structure of health status among older
adults: disease, disability, functional limitations, and perceived health.
J Health Soc Behav. 1993;34:105–121.
4. Stump TE. The structure of health status among Hispanic, African
American, and White older adults. J Gerontol Psych Soc Sci. 1997;
52B(special issue):49–60.
5. Guralnik JM, Simonsick EM, Ferrucci L, et al. A short physical per-
formance battery assessing lower extremity function: association with
self-reported disability and prediction of mortality and nursing home
admission. J Gerontol Med Sci. 1994;49:M85–M94.
20. Tinetti ME. Performance oriented assessment of mobility problems in
elderly patients. J Am Geriatr Soc. 1986;34:119–126.
6. Brock DB, Lemke JH, Branch LG, Evans DA, Berkman LF. Mortality
and physical functioning in epidemiologic studies of three older popu-
lations. J Aging Soc Policy. 1994;6:21–37.
7. Manton KG. A longitudinal study of functional change and mortality
in the United States. J Gerontol Soc Sci. 1988;43:S153–S161.
8. Rosow I, Breslau N. A Guttman health scale for the aged. J Gerontol.
1966;21:556–559.
9. Katz S, Downs TD, Cash HR, et al. Progress in development of the in-
dex of ADL. Gerontologist. 1970;10:20–30.
10. Branch LG, Katz S, Kniepmann K, Papsidero JA. A prospective study
of functional status among community elders. Am J Public Health.
1984;74:266–268.
21. Sonn U. Longitudinal studies of dependence in daily life activities
among elderly persons. Scand J Rehab Med. 1996;28(suppl 34):2–32.
22. Reuben DB, Siu AL. An objective measure of physical function of el-
derly outpatients: the physical performance test. J Am Geriatr Soc. 1990;
38:1105–1112.
23. Reuben DB, Siu AL, Kimpau S. The predictive validity of self-report
and performance-based measures of function and health. J Gerontol
Med Sci. 1992;47:M106–M110.
24. Verbrugge LM, Jette AM. The disablement process. Soc Sci Med.
1994;38:1–14.
11. Nagi SZ. An epidemiology of disability among adults in the United
States. Milbank Memorial Fund Quarterly. 1976;54:439–468.
12. Ostir GV, Markides KS, Black SA, Goodwin JS. Lower body func-
Received September 9, 1999
Accepted March 29, 2000
Decision Editor: William B. Ershler, MD