Journal of Gerontology: PSYCHOLOGICAL SCIENCES Copyright 1996 by The Gerontological Society of America

1996, Vol. 5IB, No. 5, P254-P26O

The Assessment of ADL Among Frail Elderly

in an Interview Survey:
Self-report versus Performance-Based Tests
and Determinants of Discrepancies
Gertrudis IJ.M. Kempen,1 Nardi Steverink,1 Johan Ormel,2 and Dorly J.H. Deeg3

'Northern Centre for Health Care Research and department of Psychiatry

and Department of Health Sciences, University of Groningen, The Netherlands.
department of Psychiatry and Department of Social Gerontology, Free University of Amsterdam, The Netherlands.

The impact of three sociodemographic, two cognitive, two affective, and four personality measures on the discrepan-
cies between self-reported and performance-based ADL in a sample of 753 frail elderly is studied by means of multiple
regression analyses. Underestimation (i.e., lower self-reported levels of ADL compared to performance-based levels)
occurs, in particular, among subjects with low perceptions of physical competence and mastery or personal control,
and high levels of depressive symptomatology. In contrast, the role of cognitive functioning and sociodemographic
variables in the discrepancies is a minor one. Although self-report ADL measures are easier to administer and less
sensitive to nonresponse than performance-based ADL measures, the confounding effects of perceived physical
competence, mastery, and depressive symptomatology on self-reported ADL should be considered in any application of
self-report measures of ADL among frail elderly.

A LTHOUGH the interest in performance-based measure-
•**• ment of activities of daily living (ADLs) has increased
during the last decade, relatively little is known about the
relationship between self-report and performance-based mea-
sures of ADL. Even less is known about the factors responsi-
ble for the discrepancies between them. The purpose of this
study was to identify determinants of discrepancies between
self-report and performance-based measures of ADL.
Recently, several studies showed only moderate to weak
relationships between self-report and performance-based
measures of functioning (Cress et al., 1995; Deeg, 1994;
Jette & Branch, 1985; Kelly-Hayes, Jette, Wolf, D'Agos-
tino, & Odell, 1992; Myers, Holliday, Harvey, & Hutchin-
son, 1993; Reuben, Valle, Hays, & Siu, 1995; Sager et al.,
1992). From a conceptual or theoretical point of view, the
discrepancies between both types of measurement can be
explained by at least three factors. Firstly, compared to self-
report measures, performance-based measures seem to be
more sensitive to (a) fluctuations in time and (b) differences
in equipment or aids used in the performance of daily
activities. Subjective evaluation of functioning reflects find-
ings over a period of at least a couple of days, and the
judgments of functioning are generally based on the experi-
ences with different aids or equipment. As was suggested by
Myers et al. (1993) and Guralnik, Branch, Cummings, and
Curb (1989), performance-based measures generally do not
reflect adaptations made in a person's everyday living situa-
tion (e.g., proper footwear, residential features such as types
of stairs and surfaces, compensatory mechanisms, e.g.,
handrails, and supportive relationships) that may affect the
lack of concordance between the types of measurement.
Secondly, discrepancies can be due to the differences in
concepts measured during the development from pathology
to its final consequences (e.g., impairments vs functional
limitations vs disability in Nagi's scheme [Nagi, 1991]).
Stronger associations can be expected among concepts
which are situated closer together in this development. Most
performance-based measures assess basic functional limita-
tions and are not expected to directly assess ADLs. In
contrast to most self-report measures, performance-based
measures are, typically, much more tightly focused on a
specific aspect of functioning and provide more precise
information regarding that aspect of functioning (Evans,
1993). Finally, previous research suggests that cognitive
functioning, affective functioning, personality characteris-
tics, and sociodemographic variables may reduce the con-
cordance between self-report and performance-based mea-
sures of ADL.
With respect to cognitive functioning, the lack of concor-
dance may reflect a problem of memory or of judgment
(Branch & Myers, 1987). Guralnik et al. (1989) stated that
performance-based measures are less sensitive to poor cog-
nitive functioning, educational level, and culture. Kelly-
Hayes et al. (1992) and Sager et al. (1992) found signifi-
cantly lower rates of agreement between performance-based
and self-report measures for subjects with cognitive impair-
ments. However, several studies showed positive associa-
tions of cognitive functioning with both self-reported func-
tioning (Barberger-Gateau et al., 1992; Cress et al., 1995;
Griffiths et al., 1987; Laukkanen, Kappinen, Era, &
Heikkinen, 1993; Rozzini, Frisoni, Bianchetti, Zanetti, &
Trabucchi, 1993) and performance-based functioning (Cress

P254

Downloaded from https://academic.oup.com/psychsocgerontology/article-abstract/51B/5/P254/586195/The-Assessment-of-ADL-among-Frail-Elderly-in-an

by guest
on 10 October 2017
 SELF-REPORTED VERSUS PERFORMANCE-BASED ADL
et al., 1995; Rozzini et al., 1993). The role of cognitive
functioning in the discrepancies between the types of mea-
surement is still unclear.
With respect to affective functioning, it was found that
depressive symptoms independently predict lower levels of
self-reported functioning among community-dwelling older
subjects, although depressive symptoms were also negatively
associated with performance-based functioning (Cress et al.,
1995). In contrast, Rozzini et al. (1993) found no significant
relationship between performance-based functioning and de-
pressive symptomatology. Several studies showed negative
associations between depressive symptomatology and self-
reported functioning (Griffiths et al., 1987; Kempen & Suur-
meijer, 1991; Laukkanen et al., 1993; Rozzini et al., 1993;
Turner & Noh, 1988; VonKorff, Ormel, Katon, & Lin, 1992;
Wells et al., 1989). Sager et al. (1992) reported that subjects
with depressive symptomatology tended to have lower rates
of agreement between self-report and performance-based
measures, although these differences were not significant
(Sager et al., 1992). These results suggest that depressive
symptomatology is more strongly associated with self-
reported functioning than with performance-based function-
ing. Subjects with depressive symptomatology tend to under-
estimate their level of functioning.
In the area of personality characteristics, research out-
comes supported the hypothesis that personality traits such
as neuroticism are related to the perception of health rather
than to objective health status (Smith & Williams, 1992),
although the positive correlation between neuroticism and
self-reported ADL dysfunctioning was not strong (Jorm et
al., 1993). Tinetti, Mendes de Leon, Doucette, and Baker
(1994) reported significant positive associations between
efficacy (as assessed with the Falls Efficacy Scale) and
ADL-IADL functioning while controlling for performance-
based levels of functioning. The latter result suggests that
subjects with lower levels of efficacy tend to underestimate
their level of functioning.
With respect to sociodemographic variables, previous
research showed that age was related to both self-reported
and performance-based functioning: older subjects reported
lower levels of functioning (Jette & Branch, 1985). The
same researchers showed that age and gender were still sig-
nificantly associated with self-reported functioning, while
performance-based levels of functioning are controlled.
Whether or not gender is related to self-reported and/or
performance-based ADL functioning seems to be partly
dependent on the items selected (Deeg, 1993). Parker,
Thorslund, and Lundberg (1994) found significant associa-
tions between social class (as an indicator of socioeconomic
status) and both self-reported and performance-based func-
tioning; in this case former white-collar workers had higher
levels of functioning than blue-collar workers.
The above findings suggest that levels of cognitive and
affective functioning, personality characteristics, and so-
ciodemographic variables may play an important role in the
lack of concordance between performance-based and self-
report measures of ADL.
The objective of this study was to investigate the relation-
ship between performance-based and self-report measures of
ADL among frail elderly persons and to identify factors
Downloaded from https://academic.oup.com/psychsocgerontology/article-abstract/51B/5/P254/586195/The-Assessment-of-ADL-among-Frail-Elderly-in-an
by guest
on 10 October 2017
P255
responsible for discrepancies between them. The starting
point is a set of performance-based ADL measures that can
be administered during an in-home interview session. We
expect (1) only moderate associations between self-reported
and performance-based ADLs, and (2) larger discrepancies
between self-reported and performance-based ADLs for sub-
jects who show lower levels of cognitive and/or affective
functioning (e.g., depressive symptomatology), lower lev-
els of self-efficacy (e.g., physical competence, mastery),
higher levels of neuroticism, and lower levels of education.
METHODS
Sample
The data are obtained from a subsample (n = 753) of the
baseline participants (N = 5,279) in the Groningen Longitu-
dinal Aging Study (GLAS). This is a population-based pro-
spective follow-up study of the determinants of health-related
quality of life of older people, with special emphasis on
physical and social disability and well-being (Ormel et al.,
1992). The primary objective of GLAS is to identify the
psychosocial factors that influence the trajectory of the qual-
ity of life, either independently or in interplay with disease-
related factors. The source population of GLAS is defined as
being people aged 5=57 years who lived in the north of the
Netherlands either independently or in adapted housing for
elderly people in 1993. Subjects with severe cognitive im-
pairments at baseline (MMSE < 17; Folstein, Folstein, &
McHugh, 1975) were excluded (n = 78). The subsample in
this paper comprises 753 frail elderly (14.3%) with the lowest
scores on the 6-item SF-20 physical functioning scale (Ste-
wart, Hays, & Ware, 1988). The selected subjects reported
four (35.9%), five (45.7%), or six (18.5%) physical limita-
tions on this scale. Additional data were collected in a face-
to-face follow-up interview, approximately three weeks after
the baseline. The subsample consisted of 544 females (mean
age: 73.6; range: 57-91; SD: 1.6) and 209 males (mean age:
71.9; range: 57-93; SD: 8.8). Of all the subjects, 133
(17.7%) were younger than 65 years of age, and 311 subjects
(41.3%) were 75 years of age or older.
Measures
As suggested by Guralnik et al. (1989), the application of
performance-based tests has, compared to self-report mea-
sures, both advantages and disadvantages. Performance-
based tests are time-consuming; they require the special
training of examiners or interviewers, adequate space, and
special equipment (e.g., stopwatches); also, there is a risk of
injury. As a consequence, not all tests can easily be imple-
mented during in-home interview sessions. On the other
hand, it was suggested that performance-based tests have
better reproducibility, have more face validity for the task
being performed, have greater sensitivity to change, and are
less influenced by poor cognitive functioning, culture, lan-
guage, and education. However, empirical results do not
lend support to the notions that performance-based measures
are more acceptable to patients, clinically feasible, repro-
ducible, or sensitive to change (Myers et al., 1993).
Several types of performance-based tests can be distin-
guished (for reviews see Guralnik et al., 1989; Myers et al.,
 P256 KEMPENETAL.
1993). They may differ considerably in the degree to which
they can be used during in-home interview sessions. In our
study we used three simple, performance-based ADL tests in
the follow-up. The tests were obtained from the Longitudi-
nal Aging Study Amsterdam (Deeg, 1994; Deeg, Knips-
cheer, & Van Tilburg, 1993). These consist of (a) putting on
and taking off a jacket, (b) walking 6 meters (including a
180° turn after 3 meters), and (c) standing up from a kitchen
chair and sitting down 5 times, without using the arms. The
number of seconds required for the performance of each test
is recorded as the score. Although putting on a jacket and
taking it off again were performed separately, the scores of
both tests were added up. Higher scores reflect poorer
performance. Subjects for whom no timed score was avail-
able (i.e., they tried but were unable to do the test, were told
to stop for safety reasons, or they refused) were assigned a
score equal to the worst actual score (putting on/taking off a
jacket: n = 22 of whom n = 16 refused; walking 6 meters: n
= 48 of whom 8 refused; standing up/sitting down from a
kitchen chair: n = 180 of whom n = 99 refused). (The
procedure conformed to Seeman et al., 1994.) Nonresponse
analyses showed significantly poorer self-reported ADL
functioning for nonresponders compared to responders for
each of the three performance-based tests (Student's Mests,
p< .001).
Self-reported ADL has been assessed at the baseline with
the 11 -item ADL subscale of the Groningen Activity Re-
striction Scale (GARS; Kempen, Miedema, Ormel, & Mole-
naar, in press; Kempen & Suurmeijer, 1990; Suurmeijer et
al., 1994). The items and response categories are presented
in Table 1. The theoretical range varies from 11 (no ADL
limitations) to 44 (maximum of ADL limitations). The
internal reliability estimate (Cronbach's alpha) was .84. The
results of previous studies showed that the GARS meets the
stochastic cumulative scalability criteria of the Mokken
model (Kempen et al., in press; Kempen & Suurmeijer,
1990; Suurmeijer et al., 1994). Three GARS items, which
will also be analyzed separately, correspond with the
performance-based ADL items: "Can you, fully indepen-
dently, dress yourself?" (item 1), "Can you, fully indepen-
dently, get around in the house [if necessary with a cane]?"
(item 8), and "Can you, fully independently, stand up from
sitting in a chair?" (item 3).
Educational level has been assessed according to the
International Standard Classification of Education (ISCED)
(UNESCO, 1976). The six levels of education vary from
"basic education, not finished" to "university level,
finished." The highest level of education reflects formal
education as well as occupational courses followed during
the course of life.
Two cognitive performance tests were administered in the
follow-up. One test, which is an adapted version of the
Auditory Verbal Learning Test (AVLT; Lezak, 1983), refers
to the acquisition and retention of new information. The
original test involves five presentations of 15 words with free
recall after each presentation, followed by delayed recall of
this list after 20 minutes. We have used a three-trial version
with free recall after every presentation. The score in this
study is the sum score of the correctly recalled words in the
three trials. The second cognitive performance test is part of
Downloaded from https://academic.oup.com/psychsocgerontology/article-abstract/51B/5/P254/586195/The-Assessment-of-ADL-among-Frail-Elderly-in-an
by guest
on 10 October 2017
Table 1. Eleven Self-report ADL Items and Four Response
Categories of Groningen Activity Restriction Scale
ADL items:
1. Can you, fully independently, dress yourself?
2. Can you, fully independently, get in and out of bed?
3. Can you, fully independently, stand up from sitting in a chair?
4. Can you, fully independently, wash your face and hands?
5. Can you, fully independently, wash and dry your whole body?
6. Can you, fully independently, get on and off the toilet?
7. Can you, fully independently, feed yourself?
8. Can you, fully independently, get around in the house (if necessary
with a cane)?
9. Can you, fully independently, go up and down the stairs?
10. Can you, fully independently, walk outdoors (if necessary with a
cane)?
11. Can you, fully independently, take care of your feet and toenails?
Response categories for each item:
1. Yes, I can do it fully independently without any difficulty.
2. Yes, I can do it fully independently but with some difficulty.
3. Yes, I can do it fully independently but with great difficulty.
4. No, I cannot do it fully independently, I can only do it with someone's
help.
the Groningen Intelligence Test (GIT; Luteijn & Van der
Ploeg, 1983); the test measures verbal comprehension (or
verbal intelligence). It is a 20-item multiple choice test
measuring synonyms of the 20 words; the test is not age-
dependent. The theoretical score ranges from 0 (no syno-
nyms correct) to 20 (all synonyms correct).
Affective functioning (e.g., depressive symptoms and
feelings of anxiety) was assessed using the two subscales of
the Hospital Anxiety and Depression Scale (HADS; Alyard,
Gooding, McKenna, & Snaith, 1987). Both were assessed at
the baseline. The theoretical range of these 7-item scales
varies from 0 to 21; higher scores indicate more symptoms.
The internal reliability estimates were .71 (depressive symp-
toms) and .84 (feelings of anxiety).
Four aspects of personality were measured: neuroticism
and extraversion (both using subscales of the revised version
of the Eysenck Personality Questionnaire (EPQ-R; Eysenck,
Eysenck, & Barrett, 1985), perceived physical competence
(Bosscher, Laurijssen, & De Boer, 1992; Ryckman, Rob-
bins, Thornton, & Cantrell, 1982), and mastery or personal
control (Pearlin & Schooler, 1978). Neuroticism, extraver-
sion, and mastery were assessed at the baseline, perceived
physical competence in the follow-up. Higher scores on the
personality measures indicate higher levels of neuroticism,
extraversion, perceived physical competence, and mastery,
respectively. The internal reliability estimates were .83, .80,
.87, and .73, respectively.
Statistical Analysis
Multiple regression analyses (SPSS/PC 5.0.1) were con-
ducted to examine (a) the extent of variance explained in
self-reported ADL by performance-based ADL only, and (b)
whether explained variance is added by each of the four
groups of determinants of discrepancies: sociodemographic,
cognitive, affective, and personality measures. For all cog-
nitive, affective, and personality measures, separate scores
 SELF-REPORTED VERSUS PERFORMANCE-BASED ADL P257

were used as predictors. To analyze whether self-reported
ADL is affected by the selected determinants of discrepancy
in a consistent manner (e.g., consistently higher or lower
levels of self-reported compared to performance-based lev-
els), partial correlation coefficients were computed between
each of the determinants and self-reported ADL, while
controlling for performance-based ADLs. To determine
whether the selected determinants are associated with the
magnitude of discrepancies (either positively or negatively,
indicating inaccuracy) between both types of measurement,
correlation coefficients were computed between each of the
selected sociodemographic, affective, cognitive, and per-
sonality measures and the absolute residual scores (obtained
after regression analysis with performance-based ADL as
the predictor and self-reported ADL as the outcome). Re-
gression analyses were conducted for the ADL sum scores
(11 items) as the outcome measure with all three test scores
as predictors, as well as for the three separate self-reported
ADL items with each of the three corresponding test scores
as the predictor. The latter were: (a) "putting on/taking off a
jacket'' as the predictor and the self-reported item ' 'can you,
fully independently, dress yourself?" as the outcome; (b)
"walking 6 meters" as the predictor and the self-reported
item "can you, fully independently, get around in the house
[with a cane if necessary]?'' as the outcome; and (c)' 'getting
up 5 times from a kitchen chair" as the predictor and the
self-reported item "can you, fully independently, stand up
from sitting in a chair?" as the outcome. Because of their
skewed distributions, all three self-reported items were
transformed logarithmically.
RESULTS
Table 2 shows the variance explained in self-reported
ADL by the performance-based ADL measures only, as well
as the additional variance explained by each group of deter-
minants. The three performance-based tests explain most
variance in the self-report GARS ADL sum score. The
contribution of personality and affective functioning, how-
ever, is also substantial. This suggests that objective capaci-
ties account for self-reported ADL but that other factors also
play a significant role.
Table 3 presents the partial correlation coefficients be-
tween the self-report ADL measures and the individual
sociodemographic, cognitive, affective and personality
measures while controlling for performance-based ADL.
Subjects with lower levels of perceived physical compe-
tence, and to some lesser extent mastery or personal control,
generally report lower levels of ADL functioning while
controlling for differences due to performance-based ADL
("underestimation"). The same holds for depressive symp-
toms, although for the ADL dressing item, the association is
not significant.
Univariate analyses (not in the table) showed that depres-
sive symptomatology was significantly associated with self-
reported ADL (correlation with GARS sum score, r =
.18), but not with the performance-based ADL measures.
In addition, the associations between self-reported ADL and
perceived physical competence and mastery were much
stronger (r = -.42 and r = -.24, respectively) than the
associations between both personality measures and the
three performance-based ADL measures (ranging from
- . 14 to -.35 for physical competence and from -.08 to - . 10
for mastery, respectively).
Table 4 shows Pearson correlation coefficients between
the absolute residuals of performance-based ADL on self-
reported ADL with the determinants of discrepancies. These
coefficients reflect the impact of the individual sociodemo-
graphic, cognitive, affective, and personality measures on
the accuracy ("over-" or "underestimation") of self-
reported ADL relative to performance-based ADL. The
results show that physical competence in particular and, to a
lesser extent, mastery, influence the accuracy of the self-
reported ADL. This means that among subjects with lower
levels of perceived physical competence and mastery, rela-
tively high discrepancies (either positive or negative) are
observed. The accuracy of the self-reported ADL "getting
around in the house'' item is also influenced by affective and
cognitive functioning and by age.
DISCUSSION
The results show that the association between self-report
and performance-based measures of ADL is not strong. This

Table 2. Multiple Regression Analyses: Amount of Variance Explained in Self-Reported ADL Sum Score
and (logarithmically transformed) ADL Items by Performance-Based Measures of ADL and Variance Added
by Sociodemographic, Cognitive, Affective, and Personality Measures Each to Performance-Based Measures
Getting Around Standing Up From
Self-report: ADL Sum Score Dressing in the House Sitting in Chair
R2:
Performance-based" 38.8%* 12.5%* 15.3%* 13.5%*
2
R -change after performance-based measures are included in the regression equation:
Sociodemographic 0.7% 0.3% 0.1% 1.2%
Cognitive functioning 0.6% 0.5% 0.6% 0.3%
Affective functioning 1.6%* 0.6% 1.7%* 1.6%*
Personality 6.5%* 8.2%* 5.6%* 2.7%*

"Percentages reflect amount of variance explained by performance-based ADL measures in self-report ADL measures. For ADL sum scores all three
performance-based measures have been included in the regression equation as independent variables. For self-report ADL items only, the equivalent
performance-based ADL tests have been included as independent variables.
*p< .01.

Downloaded from https://academic.oup.com/psychsocgerontology/article-abstract/51B/5/P254/586195/The-Assessment-of-ADL-among-Frail-Elderly-in-an

by guest
on 10 October 2017
 P258 KEMPENETAL.

Table 3. Partial Correlation Coefficients of Self-Reported ADL Sum Score and (logarithmically transformed) ADL Items
with Sociodemographic, Cognitive, Affective, and Personality Measures While Controlling for Performance-Based Measures of ADL

Getting Around Standing Up From

Self-report:" ADL Sum Score Dressing in the House Sitting in Chair
Sociodemographic:
Age .09 .04 .03 -.03
Gender" .03 -.04 -.01 .11*
Educational levelc -.07 .00 -.01 -.03
Cognitive functioning:
AVLT total score of 3 trials'1 -.10* -.06 -.05 -.06
GIT verbal comprehension11 -.03 -.01 .05 -.03
Affective functioning:
Depressive symptoms' .16* .08 .13* .13*
Feelings of anxiety0 .06 .04 .12* .09
Personality:
Neuroticism" -.01 -.03 .03 .04
Extraversionf -.07 -.01 -.03 -.04
Physical competencef -.28* -.27* -.23* -.17*
Mastery' -.20* -.16* -.16* -.08

"Higher self-report scores indicate poorer functioning. Coefficients are correlation coefficients between self-report ADL and the 11 selected predictors of
discrepancies while controlling for all three performance-based ADL measures in case of self-report ADL sum scores and while controlling for each
equivalent performance-based ADL measure in the case of self-report ADL items.
b
l = male, 2 = female.
c
Higher scores indicate higher educational levels.
d
Higher scores indicate better functioning.
'Higher scores indicate poorer functioning.
'Higher scores indicate higher levels of extra version, physical competence, or mastery/personal control.
*p< .01.

Table 4. Correlation Coefficients of Absolute Residuals of Self-Reported ADL Sum Score and (logarithmically transformed) ADL Items
and Performance-Based Measures of ADL with Sociodemographic, Cognitive, Affective, and Personality Measures

Getting Around Standing Up From

Self-report:" ADL Sum Score Dressing in the House Sitting in Chair
Sociodemographic:
Age -.04 .08 .10* .00
Gender* -.05 -.06 -.01 .06
Educational levelc -.05 .00 -.07 -.04
Cognitive functioning:
AVLT total score of 3 trials'1 -.04 -.10* -.10* -.06
GIT verbal comprehension"1 -.06 -.02 -.10* .00
Affective functioning:
Depressive symptoms' .12* .08 .16* .07
Feelings of anxietye .05 .01 .10* .05
Personality:
Neuroticism' .01 -.06 .01 .02
Extraversionf -.04 -.01 -.04 -.02
Physical competence' -.18* -.26* -.31* -.14*
Mastery' -.18* -.16* -.18* -.09

•Higher scores indicate poorer functioning. Coefficients are Pearson correlation coefficients between the 11 selected predictors of discrepancies and the
absolute residuals obtained after four separate multiple regression analyses with performance-based and self-report ADL measures: one analysis with all three
performance-based ADL measures as independent variables and the self-report ADL sum score as dependent variable, and three analyses each with one
performance-based ADL measure as independent variable and one equivalent self-report ADL item as dependent variable.
b
l = male, 2 = female.
°Higher scores indicate higher educational levels.
d
Higher scores indicate better functioning.
e
Higher scores indicate poorer functioning.
'Higher scores indicate higher levels of extra version, physical competence, or mastery/personal control,
*p< .01.

Downloaded from https://academic.oup.com/psychsocgerontology/article-abstract/51B/5/P254/586195/The-Assessment-of-ADL-among-Frail-Elderly-in-an

by guest
on 10 October 2017
 SELF-REPORTED VERSUS PERFORMANCE-BASED ADL
is consistent with recent reports (Cress et al., 1995; Reuben
et al., 1995). We found that depressive symptoms, in addi-
tion to perceived physical competence and mastery, enlarge
discrepancies between self-reported and performance-based
ADL. Overestimation, i.e., higher self-reported levels of
ADL compared to performance-based levels, occurs particu-
larly among subjects with high perceptions of physical
competence and mastery, and, to a lesser extent, with low
levels of depressive symptomatology. This indicates that
subjects with lower levels of perceived physical competence
and mastery (or personal control), and higher levels of
depressive symptomatology, tend to report lower levels of
ADL compared to their performance-based ADL level. Fur-
thermore, our results show that the perception of physical
competence in particular and, to a lesser extent, mastery,
affect the accuracy of the self-perceived level of ADL
functioning. This means that among subjects with lower
levels of perceived physical competence and mastery, rela-
tively high discrepancies (either positive or negative) are
observed. In contrast to what was previously suggested
(Guralnik et al., 1989), we found no effects of education on
the discrepancy between performance-based and self-report
measures.
Our results suggest that discrepancies between
performance-based and self-reported ADL are not strongly
associated with cognitive functioning. However, 78 subjects
with severe cognitive impairments (MMSE < 17) were
excluded at the baseline. Although the MMSE scores in our
subsample ranged from 17 to 30 (mean: 27.2; SE: 2.2) and
the number of subjects excluded at baseline was rather low,
the association of the size of the discrepancies with cognitive
functioning is probably reduced by this baseline exclusion
criterion.
Subjects for whom no timed score was available were
assigned a performance-based test score equal to the worst
actual score. In contrast to performance-based ADL, no
missing scores were observed for self-reported ADL. Nonre-
sponse analyses showed significantly poorer self-reported
ADL functioning for nonresponders compared to respond-
ers, for each of the three performance-based tests. This
supports the hypothesis that, in particular, elderly people
with high "frailty" levels were not able or refused to
participate in the performance-based tests. In contrast,
Tinetti and Ginter (1988) and Myers et al. (1993) found no
difference in frailty between responders and nonresponders,
but nonresponse rates in their studies were considerably
higher than in our study. However, the level of frailty was
probably higher in our sample.
It should be noted that only three very basic ADL items
were involved in our study. The main reason for this was that
we intended to focus on self-report and performance-based
measures which reflect identical aspects of functioning
(e.g., disability). However, it is of interest whether the
selected determinants of discrepancies also affect discrepan-
cies between self-report and performance-based measures of
more complex activities (i.e., instrumental activities of daily
living; IADLs). Additional research is necessary to study
this issue.
The question arises as to which type of measurement is
most appropriate in which circumstance. Recently, it was
Downloaded from https://academic.oup.com/psychsocgerontology/article-abstract/51B/5/P254/586195/The-Assessment-of-ADL-among-Frail-Elderly-in-an
by guest
on 10 October 2017
P259
shown that performance-based measures should not be
viewed as being "superior" (i.e., more acceptable to pa-
tients, clinically feasible, reproducible, sensitive to change)
compared to self-report measures (Myers et al., 1993).
Because performance-based tests are less easy to carry out in
most data-collection circumstances (e.g., special equipment
and special training is needed), self-report measures will
continue to be used. The results showed that self-reported
levels of ADL were independently influenced by depressive
symptomatology, mastery, and perceived physical compe-
tence. In contrast to the results of Cress et al. (1995), but in
line with the results of Rozzini et al. (1993), we found no
significant associations between depressive symptomatol-
ogy and performance-based measures of ADL. Self-reported
ADL and depressive symptoms were significantly associ-
ated. In addition, we found some evidence that cognitive
functioning plays a role in the concordance between self-
report and performance-based measures of ADL, but the
effects were relatively small. The effect of age, gender, and
educational level on the discrepancies is even smaller or
absent. Stronger associations of perceived physical compe-
tence and mastery with self-reported ADL were found than
with performance-based ADL. These results suggest that
self-report measures of ADL are more strongly "con-
founded" by personality and affective factors than
performance-based measures. Although self-report ADL
measures are easier to administer and less sensitive to nonre-
sponse than performance-based ADL measures, the effects
of perceived physical competence, mastery, and depressive
symptomatology should be considered in any application of
self-report ADL measures among frail elderly.
ACKNOWLEDGMENTS
This research is part of the Groningen Longitudinal Aging Study
(GLAS), conducted by the Northern Centre for Health Care Research
(NCG) and various departments of the University of Groningen (RUG).
The primary departments involved are Health Sciences (Prof.drW.J.A. van
den Heuvel), Family Medicine (Prof.dr B. Meyboom-de Jong), Psychiatry
(Prof.dr J. Ormel), Sociology (ICS) (Prof.dr S.M. Lindenberg) and Human
Movement Sciences (Prof.dr P. Rispens). Director of GLAS is prof, dr J.
Ormel (Health Sciences and Psychiatry). GLAS and its substudies are
financially supported by the Dutch government (through NESTOR), the
University of Groningen, the School of Medicine, the Dutch Cancer
Foundation (NKB/KWF), and the Netherlands Organization for Scientific
Research (NWO).
Address correspondence to Dr G.I.J.M. Kempen, Northern Centre for
Health Care Research, School of Medicine, University of Groningen,
A. Deusinglaan 1, 9713 AV Groningen, The Netherlands. E-mail:
G.I.J.M.Kempen@med.rug.nl
REFERENCES
Alyard, P. R., Gooding, J. H., McKenna, P. J., & Snaith, R. P. (1987). A
validation study of three anxiety and depression self-assessment scales.
Journal of Psychosomatic Research, 31, 261-268.
Barberger-Gateau, P., Commengs, D., Gagnon, M., Letenneur, L.,
Sauvel, C , & Dartigues, J. F. (1992). Instrumental activities of daily
living as a screening tool for cognitive impairment and dementia in
elderly community dwellers. Journal of the American Geriatrics Soci-
ety, 40, 1129-1134.
Bosscher, R. J., Laurijssen, L., & DeBoer, E. (1992). Competentie op
latere leeftijd: Een exploratieve studie. Bewegen & Hulpverlening, 9,
255-265.
Branch, L. G., & Meyers, A. R. (1987). Assessment of physical function-
ing in the elderly. Clinics in Geriatric Medicine, 3, 29-51.
 P260 KEMPENETAL.
Cress, M. E., Schechtman, K. B., Mulrow, C. D., Fiatarone, M. A.,
Gerety, M. B., & Buchner, D. M. (1995). Relationship between
physical performance and self-perceived physical function. Journal of
the American Geriatrics Society, 43, 93-101.
Deeg, D. J. H. (1993). Sex differences in IADL in the Netherlands:
Functional and situational disability. In J. M. Robine, C. D. Mathers,
M. R. Bone, & I. Romieu (Eds.), Calculation of health expectancies:
Harmonization, consensus achieved and future perspectives (pp. 203-
213). Montrouge/London: Colloque INSERM/John Libbey Eurotext
Ltd., vol. 226.
Deeg, D. J. H. (1994). Performance tests of physical disability. In D. J. H.
Deeg & M. Westendorp-de Seriere (Eds.), Autonomy and well-being in
the aging population. Report from the Longitudinal Aging Study Am-
sterdam 1992-1993 (pp. 21-29). Amsterdam, The Netherlands: VU
University Press.
Deeg, D. J. H., Knipscheer, C. P. M., & Van Tilburg, W. (1993).
Autonomy and well-being in the aging population. Concepts and design
of the Longitudinal Aging Study Amsterdam. Bunnik, The Netherlands:
Netherlands Institute of Gerontology.
Evans, D. A. (1993). Methodological issues in population studies of physical
function. In D. J. H. Deeg, C. P. M. Knipscheer, & W. Van Tilburg
(Eds.), Autonomy and well-being in the aging population. Concepts and
design of the Longitudinal Aging Study Amsterdam (pp. 151-159).
Bunnik, The Netherlands: Netherlands Institute of Gerontology.
Eysenck, S. B. G., Eysenck, H. J., & Barrett, P. (1985). A revised version
of the psychoticism scale. Personality and Individual Differences, 6,
21-29.
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-Mental
State: A practical method for grading the cognitive state of patients for
the clinician. Journal of Psychiatric Research, 12, 189-198.
Griffiths, R. A., Good, W. R., Watson, N. P., O'Donnell, H. F., Fell, P.
J., & Shakespeare, J. M. (1987). Depression, dementia and disability in
the elderly. British Journal of Psychiatry, 150, 482-493.
Guralnik, J. M., Branch, L. G., Cummings, S. R., & Curb, J. D. (1989).
Physical performance measures in aging research. Journal of Gerontol-
ogy: Medical Sciences, 44, M141-M146.
Jette, A. M., & Branch, L. G. (1985). Impairment and disability in the
aged. Journal of Chronic Diseases, 38, 59-65.
Jorm, A. F., Christensen, H., Henderson, S., Korten, A. E., Mackinnon,
A. J., & Scott, R. (1993). Neuroticism and self-reported health in an
elderly community sample. Personality and Individual Differences, 15,
515-521.
Kelly-Hayes, M., Jette, A. M., Wolf, P. A., D'Agostino, R. B., & Odell,
P. M. (1992). Functional limitations and disability among elders in the
Framingham study. American Journal of Public Health, 82, 841-845.
Kempen, G. I. J. M., Miedema, I., Ormel, J., & Molenaar, W. (in press).
The assessment of disability with the Groningen Activity Restriction
Scale. Conceptual framework and psychometric properties. Social Sci-
ence and Medicine.
Kempen, G. I. J. M., & Suurmeijer, Th. P. B. M. (1990). The development
of a hierarchical polychotomous ADL-IADL scale for noninstitutiona-
lized elders. The Gerontologist, 30, 497-502.
Kempen, G. I. J. M., & Suurmeijer, Th. P. B. M. (1991). Professional
home care for the elderly: An application of the Andersen-Newman
model in the Netherlands. Social Science and Medicine, 33, 1081 —
1089.
Laukkanen, P., Kappinen, M., Era, P., & Heikkinen, E. (1993). Factors
related to coping with physical and instrumental activities of daily living
among people born in 1904-1923. International Journal of Geriatric
Psychiatry, 8, 287-296.
Lezak, M. D. (1983). Neuropsychological assessment. New York: Oxford
University Press.
Luteijn, F., & Van der Ploeg, F. A. E. (1983). GIT: GroningerIntelligence
Test; Handleiding. Lisse, The Netherlands: Swets & Zeitlinger.
Myers, A. M., Holliday, P. J., Harvey, K. A., & Hutchinson, K. S.
(1993). Functional performance measures: Are they superior to self-
assessments? Journal of Gerontology: Medical Sciences, 48,
M196-M206.
Nagi, S. Z. (1991). Disability concepts revisited: Implications for preven-
Downloaded from https://academic.oup.com/psychsocgerontology/article-abstract/51B/5/P254/586195/The-Assessment-of-ADL-among-Frail-Elderly-in-an
by guest
on 10 October 2017
tion. In A. M. Pope & A. R. Tarlov (Eds.), Disability in America:
Toward a national agenda for prevention (pp. 309-327). Washington,
DC: National Academy Press.
Ormel, J., Kempen, G. I. J. M., Steverink, B. J. M., Van Eijk, L. M.,
Brilman, E. I., Wolffensperger, E. W., & Meyboom-de Jong, B.
(1992). The Groningen Longitudinal Aging Study on functional status
and need for care; outline of a NESTOR research program. Groningen,
The Netherlands: Northern Center for Health Care Research, University
of Groningen.
Parker, M. G., Thorslund, M., & Lundberg, O. (1994). Physical function
and social class among Swedish oldest old. Journal of Gerontology:
Social Sciences, 49, S196-S201.
Pearlin, L. I., & Schooler, C. (1978). The structure of coping. Journal of
Health and Social Behavior, 19, 2-21.
Reuben, D. B., Valle, L. A., Hays, R. D., & Siu, A. L. (1995). Measuring
physical function in community dwelling older persons: A comparison
of self-administered, interviewer-administered, and performance based
measures. Journal of the American Geriatrics Society, 43, 17-23.
Rozzini, R., Frisoni, G. B., Bianchetti, A., Zanetti, O., & Trabucchi, M.
(1993). Physical performance test and activities of daily living scales in
the assessment of health status in elderly people. Journal of the Ameri-
can Geriatrics Society, 41, 1109-1113.
Ryckman, R. M., Robbins, M. A., Thornton, B., & Cantrell, P. (1982).
Development and validation of a physical self-efficacy scale. Journal of
Personality and Social Psychology, 42, 891-900.
Sager, M. A., Dunham, N. C , Schwantes, A., Mecum, L., Halverson, K.,
& Harlowe, D. (1992). Measurement of activities of daily living in
hospitalized elderly: A comparison of self-report and performance
based methods. Journal of the American Geriatrics Society, 40,
447-462.
Seeman, T. E., Charpentier, P. A., Berkman, L. F., Tinetti, M. E.,
Guralnik, J. M., Albert, M., Blazer, D., & Rowe, J. W. (1994).
Predicting changes in physical performance in a high-functioning el-
derly cohort: MacArthur studies of successful aging. Journal of Geron-
tology: Medical Sciences, 49, M97-M108.
Smith, T. W., & Williams, P. G. (1992). Personality and health: Advan-
tages and limitations of the five-factor model. Journal of Personality,
60, 395-423.
Stewart, A. L., Hays, R. D., & Ware, J. E. (1988). The MOS Short-form
General Health Survey: Reliability and validity in a patient population.
Medical Care, 26, 724-735.
Suurmeijer, Th. P. B. M., Doeglas, D. M., Mourn, T., Brian§on, S., Krol,
B., Sanderman, R.,Guillemin, F., Bjelle, A.,& vandenHeuvel, W. J.
A. (1994). The Groningen Activity Restriction Scale for measuring
disability: Its utility in international comparisons. American Journal of
Public Health, 84, 1270-1273.
Tinetti, M. E., & Ginter, S. F. (1988). Identifying mobility dysfunctions in
elderly patients. Standard neuromuscular examination or direct assess-
ment? Journal of the American Medical Association, 259, 1190-1193.
Tinetti, M. E., Mendes de Leon, C. F., Doucette, J. T., & Baker, D. I.
(1994). Fear of falling and fall-related efficacy in relationship to
functioning among community-living elders. Journal of Gerontology:
Medical Sciences, 49, M140-M147.
Turner, R. J., & Noh, S. (1988). Physical disability and depression: A
longitudinal analysis. Journal of Health and Social Behavior, 29,
23-37.
UNESCO. (1976). International standard classification of education
(ISCED). Paris, France.
VonKorff, M., Ormel, J., Katon, W., & Lin, E. H. B. (1992). Disability
and depression among high utilizers of health care: A longitudinal
analysis. Archives of General Psychiatry, 49, 91-100.
Wells, K. B., Stewart, A., Hays, R. D., Burnham, A., Rogers, W.,
Daniels, M., Greenfield, S., & Ware, J. (1989). The functioning and
well-being of depressed patients. Journal of the American Medical
Association, 262, 914-919.
Received June 28, 1995
Accepted March 5, 1996