American Journal of Epidemiology
Copyright © 1998 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 147, No. 11
Printed in U.S.A.
Higher Instrumental Activities of Daily Living Disability in Hispanics
Compared with Non-Hispanic Whites in Rural Colorado
The San Luis Valley Health and Aging Study
S. M. Shetterly,1 J. Baxter,1 N. E. Morgenstern,2 J. Grigsby,3 and R. F. Hamman1
This study examined Hispanic versus non-Hispanic white patterns of needing assistance with instrumental
activities of daily living (IADL). The authors interviewed 798 Hispanic and 614 non-Hispanic white residents of
rural Colorado, who were aged 60 years and older between 1993 and 1995. Seventy-five participants were
nursing home residents at the time of the interview. Community-dwelling Hispanics were 1.6 times as likely as
non-Hispanic whites to need assistance with at least one IADL task (95% confidence interval 1.25-2.13). A
larger proportion of disabled non-Hispanic whites were in nursing homes but, after including nursing home
residents, Hispanics remained significantly more likely to need assistance on at least one IADL task (odds
ratio = 1.49, 95% confidence interval 1.16-1.93). Hispanics were also more likely to have difficulty on
observed performance tasks. The Hispanic excess was not removed by adjusting for chronic disease, reported
difficulty walking, or income. English language proficiency adjustment lowered the Hispanic excess, but
adjusting for years of education or Mini-Mental State Examination scores more completely removed the ethnic
differences. Higher education was protective for both Hispanic and non-Hispanic white elderly. Efforts to
further investigate what facets or correlates of education are operating may offer useful insights into limiting
IADL difficulties in future cohorts. Am J Epidemiol 1998; 147:1019-27.
activities of daily living; aged; Hispanic Americans; prevalence
The US Hispanic population grew eight times more
rapidly than did the non-Hispanic white population
between 1980 and 1990, and future estimates suggest
that Hispanics may be the largest minority by the year
2020 (1). Hispanics are an increasingly prominent
subgroup of the elderly, yet disability estimates for
this minority population are uncommon. Several previous studies considering disability in Hispanics have
not had a concurrently measured comparison group
(2-4) or have focused on selected subgroups of the
elderly, such as persons with dementia or diabetes (5,
6). Mexican-American participants in the 1978-1980
Health Interview Survey were more likely than nonHispanic whites to report limitations in their activity
Received for publication June 25, 1997, and accepted for publication December 2, 1997.
Abbreviations: ADL, activities of daily living; IADL, instrumental
activities of daily living; MMSE, Mini-Mental State Examination
(Folstein's).
1
Department of Preventive Medicine and Biometrics, University
of Colorado School of Medicine, Denver, CO.
2
Division of Geriatrics, Department of Medicine, University of
Colorado Health Sciences Center, Denver, CO.
3
Center for Health Services Research, University of Colorado
Health Sciences Center, Denver, CO.
Reprint requests to Susan M. Shetterly, Department of Preventive Medicine and Biometrics, University of Colorado School of
Medicine, Box C245, 4200 E. 9th Ave., Denver, CO 80262.
(46.5 percent vs. 37.6 percent, respectively, aged >65
years) (7). Data from the 1987 National Medical Expenditure Survey suggested the opposite pattern, with
Hispanics reporting less functional limitations than
non-Hispanic whites in analyses that included Hispanics of Cuban and Puerto Rican origin as well as Mexican Americans (8). Haan and Weldon (9) presented
data suggesting that Hispanic disability may be more
evident among persons with at least two of the chronic
illnesses of diabetes, stroke and hypertension. To date,
it remains unclear if Hispanics tend to be more disabled and what factors affect their patterns of disability. The disparities and limited data sources encouraged funding of several studies that focus on patterns
of Hispanic disability. We report findings from one of
these studies, the San Luis Valley Health and Aging
Study, which provides population-based contrasts
of disability patterns for elderly Hispanic and nonHispanic whites living in two rural counties of southern Colorado.
Physical functioning is often measured by the ability
of elders to perform basic self-care activities (activities
of daily living (ADL)) or household chores and errands that facilitate independent functioning (instrumental activities of daily living (IADL)) (10, 11).
1019
1020
Shetterly et al.
We previously found that Hispanic elderly in the San
Luis Valley were as likely as non-Hispanic whites to
report "any" difficulty with ADL, but Hispanics had a
small excess of needing assistance with ADL tasks
(Hamman et al., University of Colorado School of
Medicine, unpublished manuscript). We report here on
the prevalence of IADL dependence to provide contrasts of Hispanics and non-Hispanic whites at this
higher level of physical functioning. Ethnic contrasts
for observed performance items are also presented,
since reported IADL items may be distorted by proxy
or language-based response differences. We also selected a limited number of explanatory factors to further explore pattern differences between the ethnic
groups.
MATERIALS AND METHODS
Population and sampling
The San Luis Valley Health and Aging Study is a
community-based study of health and disability among
Hispanic and non-Hispanic white elderly residents of
Alamosa and Conejos counties in southern Colorado.
These two counties cover about 2,000 square miles
(3,200 square km) with a 1990 US Census population
of 21,070 (12). In 1992 and 1993, all occupied households were enumerated (97.2 percent response rate).
Eligibility requirements included being age 60 years or
older, a current county resident, and of Hispanic or
non-Hispanic white ethnicity. Hispanic ethnicity was
defined by the 1980 US Census question, "Are you of
Spanish or Hispanic origin or descent?" (13). Nursing
homes in the two counties and all nearby counties
were enumerated between December 1994 and February 1995, and persons whose prior home address was
in the two-county study area were considered county
residents. Differential sampling within age and ethnic
strata was required to ensure appropriate numbers of
subjects for the planned ethnic contrasts. For both
community and nursing home residents, all eligible
Hispanics aged 65 years or older were invited to
participate as well as 49 percent of persons aged
60-64 years. Among non-Hispanic whites, we sampled 100 percent of persons aged 80 years or older,
58.5 percent of persons aged 65-79 years, and 37
percent of those aged 60-64 years. Overall, 2,067
persons were initially invited to participate in the
study. Before completing a visit, 310 persons became
ineligible: 171 died before a visit attempt, 125 moved
out of the study area, and 14 were ineligible because of
inaccurate age or ethnicity. Of the remaining 1,757
persons, 1,433 completed a study visit for a response
rate of 81.6 percent. Seventy-five participants were
nursing home residents when study visits were com-
pleted. The response among nursing home residents
was high (91.5 percent) and did not differ by ethnicity
(Hispanics, 92.3 percent; non-Hispanic whites, 91.1
percent). Refusers in the community were more likely
to be non-Hispanic white (response rate, 78.2 percent;
Hispanic response rate, 84.3 percent). Overall, refusers
were also less likely to have any reported ADL difficulty and were slightly more likely to be ^ 8 0 years of
age. Refusers did not significantly differ from respondents on sex, education, self-rated health, or hospitalization in the past year. Twenty-one of the 1,358
community-dwelling participants were missing data
on reported IADL items and are not included in these
analyses.
All interviewers were bilingual, and Spanishtranslated forms were available, though only 5.5 percent of Hispanics had "Spanish only" interviews. Participants more often "mixed" English and Spanish
when selecting non-English options. Cognitive status
was measured by Folstein's Mini-Mental State Examination (MMSE) (14). Participants scoring 18 or
higher completed the full protocol. Persons with literacy or vision difficulties scoring between 11 and 17
were evaluated by a trained interviewer, and a limited
number judged cognitively capable also completed the
protocol. All other participants scoring below 18 completed the examination and selected performance tasks
while the primary care giver or closest relative with
knowledge of recent functioning and medical history
was sought for self-reported items. Proxy responses
were used for 14.9 percent of participants and more
often for Hispanics (Hispanics, 18.2 percent; nonHispanic whites, 10.9 percent).
Participants were queried on their medical histories,
and a "count of chronic diseases" variable was created
using the following disease items: cancer, diabetes,
stroke, transient ischemic attacks, arthritis, high blood
pressure, Parkinson's disease, lung disease (chronic
obstructive pulmonary disease, emphysema, chronic
bronchitis), cirrhosis, kidney failure, osteoporosis, seizures, current migraines, current depression, angina,
heart failure, heart attack, or cardiac blood vessel
surgery.
Reported IADL dependence
Our IADL question set used the syntax and items
chosen for the 1984 National Health Interview Supplement on Aging (15). Participants were asked, "Because of a health or physical problem, do you have any
difficulty" with IADL tasks (preparing meals, managing money, and so on). Persons reporting any difficulty further rated their difficulty level as some, a lot,
able with help, or unable to do. Possible responses to
the initial difficulty questions included "does not do
Am J Epidemiol
Vol. 147, No. 11, 1998
Higher IADL Disability in Hispanics
for other reasons," and such responses are typically
considered nondisabled (16, 17), though this method
will underestimate IADL disability slightly (17, 18).
We confirmed that varying the methods of treating the
"does not do" responses (counting these responses as
nondisabled, excluding these responses, counting
ADL disabled persons as IADL disabled) did not alter
the ethnic contrasts of interest before selecting the
common method of counting "does not do" responses
as nondisabled for these analyses. IADL disability was
analyzed for this report at the level of "needing assistance or unable to do" tasks because of the higher
reliability of responses at this difficulty level (11). In
selected analyses using "any" difficulty with IADL
items as the outcome, Hispanics had similar excess
IADL disability (data not shown).
Vol. 147, No. 11, 1998
o> o) o)
ID Tt O)
co co
O> O
m
TO
q
TO
cvi CM
c
as
LU
W
§
Q
v
8
d
v
q
cvi cvi
ID
o>
* - i d •*CM Cvi Cvl
<0
in
1
ID
TO
TO TO O
O
O.
dig.
C
o
o
r-; in oo
't in
o
S 5 S
C5 S
d
V
CO
II.
I
>4m J Epidemiol
o
q
d
v
>- l ^ ••cvi cvi cvi
Participants also completed observed functioning
items from a shortened version of the Structured Assessment of Independent Living Skills (SAILS) (19).
Four items were selected for these analyses: counting
out money (three amounts; analyzed as all incorrect
vs. at least one correct), making change (three trials;
analyzed as all incorrect vs. at least one correct), able
to manage medications (reading and understanding
medication label and able to open at least one of two
types of medication containers), and using a standard
phone (correctly reading and dialing a phone number).
IADL question items were not asked of nursing
home residents since nursing home staff complete
these tasks. However, non-Hispanic whites were more
likely to use nursing home care (see table 1), and
excluding nursing home residents could bias results.
While it might be reasonable to assume that all nursing
home residents require assistance with at least one
IADL task, whether residents would require assistance
on each of the eight IADL items was less clear. We
developed discriminant functions predicting individual
IADL outcomes using available community data to
estimate the likely IADL responses for nursing home
residents. These discriminant functions used ADL
functioning, MMSE score, hearing status, and vision
status as predictors of needing assistance on IADL
tasks. ADL functioning was the strongest predictor
and cognitive functioning an additional significant
predictor for all eight reported IADL items. Vision
status additionally improved prediction of problems
with driving, shopping, medication use, and managing
finances. Hearing status predicted problems with
phone use. Misclassification rates were low in these
q TO
TO" co
•o
Observed IADL disability
IADL and observed performance estimation
in
TO
1021
C
5
¥8e
fg
o o o
oID_ ^r_
o u>
o
<o_ cvi
o o
o in
O> ^
h^ TO
05
ID TO CO
y-
iri d 3
co h^ oo
in
r^
o
q
d
v
1
o
q
d
v
•o
c
a
as
Is
£
C8 C
If
Tt in
s
'S a
1
8111
fill
its
ID
II
c
a
a.
'5
Q.S
II
2
u P r
•a
°- c c c "r" 03
I
.$1 aa.0 s= - «1 &
5
O
1
CL
1022
Shetterly et al.
models (<12 percent), except for phone use (22 percent misclassified). Not surprisingly, over 90 percent
of the 75 nursing home residents were estimated to be
disabled for each reported IADL item using these
models. These estimated responses for nursing home
residents were used only in selected contrasts of Hispanic versus non-Hispanic white disability in the full
cohort.
The observational examination of the Structured
Assessment of Independent Living Skills was also not
completed in nursing homes because of time constraints in the institutional environment; it was also
missing for 72 community respondents. In similarly
developed discriminant function models, MMSE
scores were the strongest predictors for all four observed items, and ADL counts were the second strongest predictor of the money items and using a standard
phone. Vision status improved prediction of the
money items and the medication item. Misclassification rates in these models were higher than for selfreported items, ranging from 16 to 25 percent.
General statistical methods
Prevalence estimates were adjusted for age using
direct standardization with the full sample age distribution as the standard. Odds ratios for ethnic contrasts
were calculated from logistic regression models with
IADL items as the dependent variable (20), using the
SAS analysis package (SAS Institute, Inc., Cary,
North Carolina). Sex by ethnicity interactions were
tested for all models.
As expected, the prevalence of needing assistance
on at least one IADL task increased with age for both
ethnic groups (figure 1). Hispanics had higher prevalence estimates in all age groups for both men and
women, but differences in prevalence were notably
higher for Hispanic men aged ^85 years (p = 0.045)
and Hispanic women aged 80-84 years (p = 0.009).
No other sex- and age-specific differences were significant, though the number of participants is relatively small in each age-specific cell. Combining men
and women, Hispanics less than 80 years of age were
1.35 times as likely as non-Hispanic whites to need
assistance on IADL tasks (95 percent confidence interval 0.98-1.86), while Hispanics 80 years or older
were 2.62 times as likely (95 percent confidence interval 1.61-4.36).
The age-adjusted prevalence for needing assistance
with IADL tasks is provided in table 2 for communitydwelling residents. Nearly 15 percent of Hispanic females in this sample were disabled on five or more
IADL items compared with only 2.6 percent of nonHispanic white males. Hispanic males and nonHispanic white females were intermediate with 9.9
percent and 6.1 percent, respectively. With adjustment
for age and sex, Hispanics were 1.63 times as likely as
were non-Hispanic whites to need assistance on at
least one IADL task (95 percent confidence interval
1.25-2.13), and this excess was higher when assistance was needed on 5-8 IADL tasks (odds ratio =
5.16). For all eight IADL tasks, Hispanics had higher
rates of disability than non-Hispanic whites. The Hispanic excess was highest for "taking medications,"
RESULTS
Participants ranged in age from 60 years to 99 years,
with a mean age of 74.7 years. Fifty-seven percent of
participants were Hispanic. Table 1 presents selected
descriptive data for this cohort. Hispanics in this sample were slightly younger but had similar sex proportions. Seventy-five percent of Hispanic participants
reported speaking and understanding English well or
very well. Only 11 of the 199 Hispanics with poor
English were unable to speak or understand English at
all, and four of these were cognitively impaired participants whose proxies rated them as unable to understand or speak Spanish either. Hispanics had lower
household incomes, less education, and lower MMSE
scores. Non-Hispanic whites had higher chronic disease counts, but the two ethnic groups did not differ in
needing assistance walking across a room. Nursing
home residents were less likely to be Hispanic, and
further analyses confirmed that this univariate pattern
was not explained by the slight age differences between the ethnic groups.
100
80
60
c
0)
o
40
20
60-04 05-00 70-74 75-79 00-04 05 +
00-04 05-09 70*74 75-79 00-04 65-t-
Female
Male
FIGURE 1. Prevalence of any instrumental activities of daily living
dependency among community-dwelling residents of rural, southern Colorado, by age and ethnicity, San Luis Valley Health and
Aging Study, 1993-1995. • , non-Hispanic white; ^ , Hispanic.
Am J Epidemiol
Vol. 147, No. 11, 1998
Higher IADL Disability in Hispanics
1023
TABLE 2. Age-adjusted prevalence* of needing assistance with lADLf tasks by sex and ethnicity and
ethnic odds ratios for community sample and estimated total sample including nursing home residents,
San Luis Valley Health and Aging Study, 1993-1995
Prevalence (%)
Males
IADL disability
None
1
2-4
5-8
>1
Transportation
Shopping
Heavy housework
Light housework
Preparing meals
Using standard phone
Taking medications
Managing money
Hispanic vs. NHWt (odds ratio)*
Females
Hispanic
(n = 340)
NHW
(n = 240)
Hispanic
(n = 434)
NHW
(n = 323)
67.4
10.8
11.9
9.9
32.6
13.5
14.3
21.0
10.9
6.9
13.9
14.6
15.8
76.9
14.4
6.0
2.6
23.1
6.3
6.3
13.6
2.0
0.9
8.9
3.7
4.4
60.6
11.0
13.7
14.7
39.5
18.8
23.7
31.2
13.6
15.5
9.6
12.9
17.1
70.0
14.5
9.3
6.1
30.0
10.0
14.6
23.9
7.4
9.5
4.2
3.1
7.1
Reported
(n = 1,337)
Estimated?
(n= 1,412)
1.0
0.91 (0.65-1.29)11
2.13(1.42-3.20)
5.16 (2.90-9.20)
1.63(1.25-2.13)
2.52 (1.70-3.73)
2.30(1.60-3.31)
1.55(1.17-2.07)
3.04 (1.89-4.88)
4.08 (2.47-6.77)
2.13(1.36-3.33)
5.62 (3.21-9.82)
4.16 (2.63-6.57)
1.0
<).92 (0.65-1.30)
!2.16(1.44-3.25)
!2.43(1.57-3.75)
.49(1.16-1.93)
.68(1.20-2.34)
.68(1.22-2.32)
.33(1.02-1.74)
.61 (1.12-2.31)
.94(1.34-2.82)
.31 (0.92-1.88)
!2.20(1.50-3.22)
I2.26(1.58-3.24)
* Prevalence for community-dwelling resident sample.
t IADL, instrumental activities of daily living; NHW, non-Hispanic white.
$ Adjusted for age and sex.
§ Additionally included estimated responses for 75 nursing home residents.
H Numbers in parentheses, 95% confidence interval.
"managing money," and "preparing meals" items by
point estimates, although confidence intervals for most
items are overlapping. Sex by ethnicity interactions
were nonsignificant, suggesting that these patterns
were consistent in men and women.
The ethnic contrasts in our community sample do
not account for the differential pattern of nursing home
use by ethnicity. The final column in table 2 provides
estimates of the Hispanic versus non-Hispanic white
patterns including nursing home residents. Although
Hispanics remain significantly more disabled on all
items except using a standard phone, the magnitude of
the odds ratios was markedly reduced and suggested
more moderate excess disability in the combined community and institutional sample.
The age-adjusted prevalence of difficulty with the
four observed IADL items selected from the protocol
of the Structured Assessment of Independent Living
Skills is presented in table 3. Hispanics were 3-5 times
as likely as non-Hispanic whites to fail to complete
these tasks correctly. With addition of the estimated
responses for nursing home residents, the Hispanic
versus non-Hispanic white odds ratios dropped to
1.75-2.79 but remained statistically significant. All
four observed IADL items in these analyses primarily
measured cognitive functioning and required little mobility. We had no "observed" items to investigate
IADL tasks, such as light housework, that require
greater mobility but perhaps less cognitive functioning. However, on an observed ADL item of completAm J Epidemiol
Vol. 147, No. 11, 1998
ing a 10-foot (307 cm) walk, Hispanics were not
significantly more disabled than non-Hispanic whites
(odds ratio = 1.27, 95 percent confidence interval
0.98-1.63; age and sex adjusted).
We explored selected explanatory variables for the
ethnic differences: two markers more related to physical functioning (number of chronic diseases and reported difficulty walking across a room), a variable
indicating the ability to speak and understand English
"well" or "very well," household income, years of
education, and MMSE scores. Table 4 displays models
for three reported IADL outcomes (shopping, managing money, and taking medications) and two observed
IADL outcomes (counting money and managing medications) adding adjustment for the selected explanatory variables. Adjusting for the number of chronic
diseases generally increased the Hispanic versus nonHispanic white odds ratios, while adding "needs assistance to walk across a room" resulted in little
change to the estimates. Models adjusting for income
similarly explained little of the Hispanic excess. Accounting for English proficiency lowered the Hispanic
excess, and adjusting for years of education reduced
the odds ratios even further. For the combined estimates of community and nursing homes, all educationadjusted confidence intervals included 1.0, and the
highest odds point estimate was 1.5 for the observed
task of managing medications. Ethnicity-education interactions were nonsignificant in all models, suggesting that more highly educated participants of both
1024
Shetteiiy et al.
TABLE 3. Age-adjusted prevalence* of difficulty with observed instrumental activities of daily living
(IADL) by sex and ethnicity and ethnic odds ratios for community sample and estimated for total sample
with nursing home residents, San Luis Valley Health and Aging Study, 1993-1995
Prevalence (%)
Males
Hispanic
(n = 314)
Difficulty dialing
number on standard
phone
Misreading medication
label or unable to
open bottle
Difficulty counting out
money (on all three
trials)
Difficulty giving
accurate change
(on all three trials)
Hispanic vs. NHWt (odds ratio)t
Females
Observed
(n = 1,264)
Estimated§
(n= 1,412)
2.0
3.20 (1.84-5.56)11
1.75(1.22-2.53)
17.6
8.2
3.33(2.16-5.14)
2.74(1.93-3.89)
1.5
7.1
3.0
3.10(1.56-6.17)
2.02(1.34-3.07)
1.9
12.9
2.9
5.22 (2.79-9.75)
2.79(1.89-4.12)
NHW
(n = 228)
Hispanic
(n = 409)
NHW
(n = 313)
10.0
7.4
10.1
16.5
5.7
5.5
8.8
* Prevalence for community-dwelling resident sample; 53 of the 1,337 community participants in these IADL
analyses were missing these observed items.
t NHW, non-Hispanic white.
i Odds ratio adjusted for age and sex.
§ Additionally includes estimated responses for nursing home residents and community subjects missing
observed items.
H Numbers in parentheses, 95% confidence interval.
ethnic groups were less likely to need assistance with
IADL tasks. Figure 2 illustrates the similar ethnic
patterns of disability within education strata. MMSE
adjustment also markedly lowered the odds ratio estimates and in some cases reversed the direction of the
estimates, though all remained nonsignificant at p =
0.05. Although "shopping" was the only more physical
IADL item presented in table 4, the patterns of odds
ratio changes were similar for the IADL items of
housekeeping, meals, and transportation.
DISCUSSION
We found an excess of IADL dependency among
Hispanic residents in rural, southern Colorado. Among
community-dwelling residents, Hispanics were 2-5
times as likely as non-Hispanic whites to need assistance with IADL tasks. However, a larger proportion
of disabled non-Hispanic whites were in nursing
homes, and estimates that included nursing home residents suggested a more modest Hispanic excess that
was generally less than twofold. Lower nursing home
use by Hispanics has been reported in other studies
(21, 22), which have suggested that cultural and attitude factors are more strongly linked than functionbased factors to lower Hispanic nursing home use. The
impact such differential patterns may have on ethnic
contrasts of disability is often not considered. Only 5.3
percent of our sample resided in nursing homes, yet
this relatively small percentage notably changed the
magnitudes of our estimates. If studies are meant to
consider the "health care needs for communitydwelling residents," then the larger excess of Hispanic
dependency in the community sample is a real need
that should be addressed. However, for the larger,
descriptive question, "Are Hispanics more disabled?,"
ignoring these nursing home use patterns will bias
estimates.
Non-Hispanic whites in our study had prevalence
estimates comparable with those of the 1984 Supplement on Aging for needing assistance with IADL
activities: 22.8 percent in the San Luis Valley versus
22.2 percent in the Supplement on Aging (aged S65
years and using six IADL items) (16). We found that
5.3 percent of elderly residents were in nursing homes
in the two-county study area, which is similar to the
National Long Term Care Surveys' estimates of 5.55.7 percent (23). Relatively few studies have estimates
for Hispanics that can be compared with our results. In
analyses reported here, San Luis Valley Hispanics had
only slightly lower estimates of "any" IADL difficulty
compared with Wallace and Lew-Ting's (2) nationally
based Mexican-American sample (47 percent vs. 54.5
percent, respectively; aged 3:65 years and using five
IADL items). In the latter study, Cuban Hispanics
reported fewer IADL difficulties than Mexican Americans or Puerto Ricans, but there was no non-Hispanic
white comparison group. The Health Interview Survey
in 1978-1980 reported slightly higher "major activity
limitations" for Mexican Americans compared with
non-Hispanic whites (46.5 percent vs. 37.6 percent),
Am J Epidemiol
Vol. 147, No. 11, 1998
OI
OJ 1—
o>
0
OJ
co 01
OJ
CO
n
T-
01 in
a
oi oi
O) p OJ
0
0
1-; in ^;
ri co'
2.0 0.9-4.3
0
CO
CO
n
I
•D
CO
0
0
Education (y ear 9 s 1 ,
FIGURE 2. Prevalence of number of instrumental activities of daily
living (IADL) dependencies by education and ethnicity, among
community-dwelling residents of rural, southern Colorado, San Luis
Valley Health and Aging Study, 1993-1995. NHW, non-Hispanic
white.
oi r~ oi
in cq OJ
odd
OJ CD CO
T^
d d
O
* -
• *
UZ
0
0
0
r*-
0
co co cq cq
T^ r-' d T^
oi oi
CD
oi oi c\i
8-2.0
CO
2.2
Q.
1-
T-
0
l—OO
i n co ^- •*
011•^ oj
3.1 1.5-6.2
I
T-
CO
do
p
i- r^ r^ O
•*. en cq TT
co o i T^ o i
O) CO
LO IT)
O
•D
C
f^
n
8|
•D
I |
oS
r-^
^ ^ p
en r*»; i n i n
CNJ
C- C-C- 2-2-2-2f»OJ
inoi
OJ r^
^ r^ o> p
i^ ^ O ^
jtn E
p"
cn'ST
w'p'co'in 1
f
co,
o j , ^ cc-2-2-
p
in
cq O)
^- c\i
1
cq*
^*c»r
oq'in'^'or
•s
Oi
Oi OJ
»- r1 r^ r^
I
Q
s)
o
3
o
11
I
(A
C
O
t a an
4;
co p
i?
7£ « to ;
If
CD cq co OJ
c\i ^ r^ r^
en r^ in to
*~
d-C-
2-2-22-
p
OJ
cq in
r^ r1
WONO
i- r1 O r^
p"
•S'cq'
to* 0*^*0*
&>
J ) OJ
d-
C^zL C-2-2-2-
r-m
oi
OJ CO, CD. CD
co oq
oi i-
N co o>»y- r d i-
?
§
CO
O
|88
2|2
1 Ej
oo
2
Am J Epidemiol
1025
CO CO
1.2
5-3.4
,3.
CO
i1
6-4.0
in in
OJ
I
t- 0
cn U>
1^-
8-2
3
O O T
3-2
2
CO
0-2
8-2
7-2
•<t co
2-5.2
CM"
1-4
7-4
Higher IADL Disability in Hispanics
Vol. 147, No. 11, 1998
but only 39.8 percent of "other" Hispanics were limited (7). In contrast, Hispanics were less likely than
non-Hispanic whites to report difficulty with IADL in
the 1978-1980 National Medical Expenditure Survey
(12.5 percent vs. 17.5 percent) (8), though lack of age
adjustment in their tables could have affected the
ethnic patterns (24). Varying regional or Hispanic
subgroup patterns may also be due to the different
socioeconomic, immigration, and education histories
of the populations studied, making generalizations difficult. In our rural population, most Hispanics have
resided in the area for generations, and only 35 percent
self-identify themselves in the subcategory of Mexican
Americans, most preferring the category of Spanish/
other Hispanic. This pattern of identification is typical
of many parts of the Southwest where Hispanics are
not recent immigrants. Markides et al. (25) presented
estimates of needing assistance with ADL for Hispanics from five southwestern states based on interviews
between 1993 and 1994. Our ADL estimates are remarkably similar (Hamman et al., University of Colorado School of Medicine, unpublished manuscript)
and suggest that Hispanic disability patterns may be
generally consistent in this region.
We investigated both reported and observed IADL
outcomes in this study. One reason that we included
observed IADL outcomes was to limit the effect of
differential proxy response. In our study, Hispanics
were more likely to have proxy respondents for the
IADL questions than were non-Hispanic whites.
Nearly 88 percent of persons with proxy respondents
had low cognitive function (Folstein's MMSE scores
1026
Shetterly et al.
<18), although frailty, illness, or time constraints
were also occasional reasons for using proxy respondents. By contrast, observed performance tasks were
completed by participants at all levels of cognitive
status. Previous studies have reported high agreement
between proxy and subject disability reports, but they
have also shown that proxy respondents typically report slightly higher levels of disability (26, 27). The
observed tasks allowed us to investigate Hispanic versus non-Hispanic white patterns among participants
who span the range of cognitive abilities, many of
whom had proxy responses for question items. Bilingual interviewers administered the observed tasks, further ensuring that language constraints would minimally affect performance. The congruent results
between the observed tasks and reported IADL in our
study offer strength to the finding of Hispanic excess
IADL disability in this population.
We investigated several explanatory variables for
the IADL ethnic differences: number of chronic diseases, reported needing assistance walking across a
room, income, English proficiency, years of education,
and MMSE score. Adding adjustment for number of
chronic diseases or needing assistance walking across
a room generally increased the Hispanic excess
slightly. Our finding that chronic disease counts failed
to explain the Hispanic excess is congruent with that
of Haan and Weldon, who reported that greater Hispanic IADL limitation was "neither confounded nor
modified by prevalent stroke, diabetes, or hypertension" (9, p. 395). Similarly, income failed to explain
the Hispanic excess in our models. This differs from
the finding of Kington and Smith (28), who reported
that socioeconomic status markers explained the
poorer functional status of Hispanics in their analyses
of the 1992 Health and Retirement Survey. While this
discrepancy may be due to their stronger "wealth"
variable, their functional status score also didn't include IADL items but rather summarized ADL and
other items focused primarily on physical abilities. In
our study, adjusting for English proficiency lowered
the Hispanic excess, and adjusting for either years of
education or MMSE score generally removed the Hispanic excess of needing IADL assistance. Point estimates for the Hispanic excess were highest for the
items of money management and medication management that are conceptually the least "physical" components of the IADL (29, 30). The importance of
the cognitive elements in IADL has been noted by
Barberger-Gateau et al. (31), who predicted incident
dementia using four IADL items (phone use, transportation, money, and medication management). While
there may be general agreement that cognitive aspects
of IADL are strong, interpreting the meaning of the
"explanatory" effect of education or MMSE is less
clear. Education does not solely mark early learning,
but it also correlates with language, income, and lifestyle choices that impact health. Increased education
was associated with lower IADL disability in both
ethnic groups, suggesting that education was not
solely marking English proficiency in this study.
While education might also be postulated to affect
health care access and use, we have found similar
health care use by ethnicity in this population (32).
These results suggest that correlates or components of
education apart from language and health care access
appear to be operating in this population. MMSE is a
marker of current cognitive function, but it is well
known that MMSE scores differ by educational status
(33, 34), and the similar results for these two variables
are not surprising. Further investigations are needed to
adequately understand how these general markers are
acting as "explanatory factors." Nevertheless, these
current analyses support the body of literature that
continues to suggest that higher educational attainment
early in life may be associated with beneficial effects
on health, cognition, and mortality throughout the later
decades of life (33, 35, 36).
The strengths of our study include a concurrently
measured non-Hispanic white comparison group, high
response rates from a population-based sample, and
the availability of both observed performance and
self-reported IADL items. Refusers were less disabled
and more often non-Hispanic white, suggesting that
the reported Hispanic excess was conservative. Like
many studies of the elderly, we were limited by the
need for proxy respondents, although congruent observed performance patterns suggested that proxy response differences did not account for the general
ethnic patterns of this analysis. Interpretations of our
findings are restricted to the selected explanatory variables explored in this work, and our analyses could not
distinctly separate the effects of correlated factors,
such as education and MMSE. Because of the crosssectional design, temporal patterns between postulated
risk factors and IADL could not be discerned, and
survivorship bias is a concern in this elderly cohort.
Our results should be generalizable to similar Hispanic
populations in the Southwest, though generalizing to
Hispanics throughout the US is less tenable.
Our results indicated a greater need for assistance
with IADL tasks among southwestern Hispanic elderly, and they suggest that communities should ensure that home care and elderly support systems are
alerted and sensitive to the needs of this population. Of
all our findings, the strength of education in explaining
IADL differences may be the most interesting. Educational differences between the ethnic groups are
Am J Epidemiol
Vol. 147, No. 11, 1998
Higher IADL Disability in Hispanics
smaller among the young elderly in our study, and
such a continuing trend could eliminate ethnic differences in future cohorts. Examination of incident disability patterns and further exploration of such correlates of education as physical activity, diet, adult
reading habits, or other life-style choices may improve
our understanding of the factors that influence differing ethnic patterns of disability.
ACKNOWLEDGMENTS
This research was supported by the National Institute of
Aging, grant AG10940.
The authors thank the skilled staff of the San Luis Valley
Health and Aging Study: Carolyn Swenson, Olivia
Martinez, Rose Martinez, Anna Trujillo, Elaine Trujillo,
Roberta Vialpando, and Rodney Sandoval.
16.
17.
18.
19.
20.
21.
22.
REFERENCES
1. Vega WA, Amaro H. Latino outlook: good health, uncertain
prognosis. Annu Rev Public Health 1994;15:39-67.
2. Wallace SP, Lew-Ting CY. Cross-cultural medicine a decade
later. Getting by at home, community-based long-term care of
Latino elders. West J Med 1992; 157:337-44.
3. Markides KS, Lee DJ. Predictors of health status in middleaged and older Mexican Americans. J Gerontol 1991;46:S243-9.
4. Ailinger RL. Self-assessed health of Hispanic elderly persons.
J Community Health Nurs 1989;6:113-18.
5. Loewenstein DA, Ardila A, Rosselli M, et al. A comparative
analysis of functional status among Spanish- and Englishspeaking patients with dementia. J Gerontol 1992;47:P389-94.
6. Mitchell BD, Stern MP, Haffner SM, et al. Functional impairment in Mexican Americans and non-Hispanic whites with
diabetes. J Clin Epidemiol 1990;4:319-27.
7. Trevino FM, Moss AJ. Health indicators for Hispanic, black,
and white Americans. Hyattsville, MD: National Center for
Health Statistics, 1984:1-88. (Vital and health statistics, Series 10, no. 148) (DHHS publication no. (PHS) 84-1576).
8. Leon J, Lair T. Functional status of the noninstitutionalized
elderly: estimates of ADL and IADL difficulties. National
Medicare Care Expenditure Survey research findings 4. Rockville, MD: US Department of Health and Human Services,
1990. (DHHS publication no. (PHS) 90-3462).
9. Haan MN, Weldon M. The influence of diabetes, hypertension, and stroke on ethnic differences in physical and cognitive functioning in an ethnically diverse older population. Ann
Epidemiol 1996;6:392-8.
10. Branch LG, Meyers AR. Assessing physical function in the
elderly. Clin Geriatr Med 1987;3:29-51.
11. Fillenbaum GG. Screening the elderly. A brief instrumental
activities of daily living measure. J Am Geriatr Soc 1985;33:
698-706.
12. US Department of Commerce Bureau of the Census. 1990
census of population and housing. Summary social, economic,
and housing characteristics. Colorado. Washington, DC: US
Government Printing Office, 1992.
13. US Department of Commerce Bureau of the Census. 1980
census of population. Vol I. General social and economic
characteristics of the population, table 180. Washington, DC:
US Government Printing Office, 1983.
14. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state."
A practical method for grading the cognitive state of patients
for the clinician. J Psychiatr Res 1975;12:189-98.
15. Fitti JE, Kovar MG. The Supplement on Aging to the 1984
Am J Epidemiol
Vol. 147, No. 11, 1998
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
1027
National Health Interview Survey. Hyattsville, MD: National
Center for Health Statistics, 1987:1-115. (Vital and health
statistics, Series 1, no. 21) (DHHS publication no. (PHS)
87-1323).
Dawson D, Hendershot G, Fulton J. Aging in the eighties.
Functional limitations of individuals age 65 years and over.
Hyattsville, MD: National Center for Health Statistics, 1987:
1-11. (Advance data from vital and health statistics, Series 133).
Mermelstein R, Miller B, Prohaska T, et al. Health data on
older Americans: United States, 1992. Measures of health.
Hyattsville, MD: National Center for Health Statistics, 1993:
9-21. (Vital and health statistics, Series 3, no. 27) (DHHS
publication no. (PHS) 93-1441).
Allen SM, Mor V, Raveis V, et al. Measurement of need for
assistance with daily activities: quantifying the influence of
gender roles. J Gerontol 1993;48:S204-ll.
Mahurin RK, DeBettignies BH, Pirozzolo FJ. Structured assessment of independent living skills: preliminary report of a
performance measure of functional abilities in dementia. J
Gerontol 1991;46:P58-66.
Hosmer DW, Lemeshow S. Applied logistic regression. New
York: John Wiley & Sons, 1989.
Mui AC, Burnette D. Long-term care service use by frail
elders: is ethnicity a factor? Gerontologist 1994;34:190-8.
Eribes RA, Bradley-Rawls M. The underutilization of nursing
home facilities by Mexican-American elderly in the Southwest. Gerontologist 1978;18:363-71.
Manton KG, Corder LS, Stallard E. Estimates of change in
chronic disability and institutional incidence and prevalence
rates in the U.S. elderly population from the 1982, 1984, and
1989 National Long Term Care Survey. J Gerontol 1993;48:
S153-66.
Markides KS, Mindel CH. Aging and ethnicity. Newbury
Park: Sage Publications, 1987.
Markides KS, Stroup-Benham CA, Goodwin JS, et al. The
effect of medical conditions on the functional limitations of
Mexican-American elderly. Ann Epidemiol 1996;6:386-91.
Magaziner J, Bassett SS, Hebel JR, et al. Use of proxies to
measure health and functional status in epidemiologic studies
of community-dwelling women aged 65 years and older. Am J
Epidemiol 1996; 143:283-92.
Elam JT, Graney MJ, Beaver T, et al. Comparison of subjective ratings of function with observed functional ability of frail
older persons. Am J Public Health 1991;81:1127-30.
Kington RS, Smith JP. Socioeconomic status and racial and
ethnic differences in functional status associated with chronic
diseases. Am J Public Health 1997;87:805-10.
Wolinsky FD, Johnson RJ. The use of health services by older
adults. J Gerontol 1991;46:S345-57.
Judge JO, Schechtman K, Cress E, et al. The relationship
between physical performance measures and independence in
instrumental activities of daily living. The FICSIT Group.
Frailty and injury: cooperative studies of intervention trials.
J Am Geriatr Soc 1996;44:1332-41.
Barberger-Gateau P, Dartigues J, Letenneur L. Four instrumental activities of daily living score as a predictor of oneyear incident dementia. Age Ageing 1993;22:457-63.
Baxter J, Scarbro SL, Cortese CF, et al. Ethnic patterns of
health care use—are rural Hispanic and NHW elderly different? The San Luis Valley Health and Aging Study (SLVHAS).
(Abstract). Gerontologist 1995;35:177.
Farmer ME, Kittner SJ, Rae DS, et al. Education and change
in cognitive function. The Epidemiologic Catchment Area
Study. Ann Epidemiol 1995;5:l-7.
Crum RM, Anthony JC, Bassett SS, et al. Population-based
norms for the Mini-Mental State Examination by age and
education level. JAMA 1993;269:2386-91.
Ross CE, Wu C. Education, age, and the cumulative advantage
in health. J Health Soc Behav 1996;37:104-20.
Feldman JJ, Makuc DM, Kleinman JC, et al. National trends
in educational differentials in mortality. Am J Epidemiol
1989;129:919-33.