Journal of Gerontology: MEDICAL SCIENCES
2002, Vol. 57A, No. 4, M217–M222
Copyright 2002 by The Gerontological Society of America
Late Life Function and Disability Instrument:
II. Development and Evaluation of the
Function Component
Stephen M. Haley,1 Alan M. Jette,1 Wendy J. Coster,1 Jill T. Kooyoomjian,1 Suzette Levenson,2
Tim Heeren,2 and Jacqueline Ashba2
1Roybal
Center for Enhancement of Late-Life Function, Sargent College of Health and Rehabilitation Sciences, Boston
University, Massachusetts.
2School of Public Health, Boston University, Massachusetts.
Background. Self-reported capability in physical functioning has long been considered an important focus of research for older persons. Current measures have been criticized, however, for conceptual confusion, lack of sensitivity
to change, poor reproducibility, and inability to capture a wide range of upper and lower extremity functioning.
Methods. Using Nagi’s disablement model, we wrote physical functioning questionnaire items that assessed difficulty in 48 common daily tasks. We constructed the instrument using factor analysis and Rasch analytic techniques and
evaluated its validity and test-retest reliability with 150 ethnically and racially diverse adults aged 60 years and older
who had a range of functional limitations.
Results. Our analyses resulted in a 32-item function component with three dimensions—upper extremity, basic
lower extremity, and advanced lower extremity functions. Expected differences in summary scores of known-functional
limitation groups support its validity. Test-retest stability over a 1- to 3-week period was extremely high (intraclass correlation coefficients .91 to .98).
Conclusions. The Late-Life Function and Disability Instrument has potential to assess activity concepts related to
upper and lower extremity functioning across a wide variety of daily physical tasks and individual levels of physical
functioning.
P
HYSICAL functioning refers to performance of a variety of activities that are expected for community-dwelling persons, including walking, climbing stairs, and handling objects (1). Physical functioning is important because
it provides a foundation for many tasks required for living
independently (2,3). In this article, we use the term function
to refer to the person’s ability to perform specific activities
that require gross or fine motor actions.
Self-report measures have been found to be a reliable and
accurate methodology for obtaining information on functional status, describing the stages and severity of chronic
diseases (4,5), estimating demand for long-term care services (6,7), and examining the impact of intervention programs (8–10). A wide variety of measures with function
items have been published and are reviewed elsewhere
(6,11–16). Yet, lack of sensitivity to change, reproducibility, and inability to capture the full spectrum of functioning
have been common concerns regarding existing self-report
measures of physical functioning (17,18).
We have attempted to address two major concerns of current instruments. First, the structure and dimensionality of
function items have been poorly defined, which has led to
conceptual confusion. In accordance with Nagi and others
(19–22), our measure of physical function is focused specif-
ically on discrete activities. Second, many of the current activity of daily living (ADL) measures focus too narrowly on
basic physical skills, and thus are ineffective in capturing
variations in function in the general population of community-dwelling elderly (23,24).
The most widely used self-report measure of function is
the 10-item Physical Functioning Scale (PF-10), a component of the Short-Form 36 Health Survey (SF-36) (25,26).
The scale was designed to be brief in order to minimize response burden yet comprehensive so that it would sample a
core set of questions applicable for both general populations
and those with acute or chronic conditions. However, it was
not intended to be a general measure of change (27–29), in
response to interventions.
In response to the continuing need for a comprehensive
measure to examine functioning in older adults, we undertook the development of a new measure. This article
reports on the conceptual basis and construction of the
function component of the Late-Life Function and Disability Instrument (Late-Life FDI) and reports the results
of an initial field test of the measure’s reliability and validity. The development and evaluation of the disability
component of the Late-Life FDI are reported elsewhere
(30).
M217
M218
HALEY ET AL.
METHODS
Instrument Development
We constructed questionnaire items that encompass a
wide variety of physical activities encountered in daily routines. After pilot-testing an initial draft of 54 items with 50
community-dwelling older adults, subsequent review by
project staff, and consideration of suggestions from community focus groups, 48 physical functioning items were retained for the prototype instrument. The function items on
the Late-Life FDI assess self-reported difficulty in performing a discrete physical activity. Questions are phrased, “how
much difficulty do you have doing a particular activity
without the help of someone else and without the use of assistive devices?” with response options of “none,” “a little,”
“some,” “quite a lot,” and “cannot do.” Factors that may influence difficulty in task performance (e.g., pain, fatigue,
fear, weakness, soreness, ailments, health conditions, and
disabilities) were identified for participants before the measure was administered.
Sampling Procedures
To develop a function scale designed to be sensitive
across a wide range of ability levels, we recruited a convenience sample of adults who were 60 years of age and older,
from diverse ethnic and racial backgrounds, with a range of
functional limitations (30). A random subset of participants
(n 15) completed a second interview within 1 to 3 weeks
after the initial interview to assess test-retest reliability of
the function component.
Instrument Construction and Analyses
A series of factor analyses was used to identify the number and nature of latent factors that could be responsible for
the covariation in the functioning data. A one-factor and a
three-factor model were tested. A series of one-parameter
Rasch rating scale analyses was performed to estimate item
locations (calibrations) along a common scale (31). By our
convention, the items on which most people express difficulty are associated with scores of increasing magnitude.
Thus, the continuum of 0–100 reflects increasing function,
Table 1. Estimates of Factor Loading and Cronbach Alphas for Three- and One-Factor Models for Difficulty in Function
Three-Factor
Items
Hike a few miles including hills
Carry while climb stairs
Walk a brisk mile
Go up and down 1 flight, no rails
Walk 1 mile with rests
Run to catch bus
Walk on slippery surface
Go up and down 3 flights inside
Walk several blocks
Run one-half mile
Get up from floor
Walk around one floor of home
Pick up a kitchen chair
Get into and out of car
Reach overhead while standing
Wash dishes while standing
Up and down from a curb
Put on and take off coat
Open heavy outside door
On and off bus
Make bed
Bend over from standing position
Go up and down a flight of stairs
On and off a step stool
Stand up from a low soft couch
Remove wrapping with hands only
Unscrew lid without assistive device
Pour from a large pitcher
Hold full glass of water in 1 hand
Put on and take off pants
Use common utensils
Reach behind back
ML test: (Ho: 1 factor model is sufficient)
% variance explained by model
Cronbach Alphas
Advanced Lower
Extremity Function
.86
.85
.84
.83
.78
.77
.77
.75
.73
.72
.65
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
.96
Basic Lower
Extremity Function
Upper Extremity
Function
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
.88
—
.81
—
.78
—
.73
—
.70
—
.68
—
.66
—
.66
—
.66
—
.65
—
.64
—
.62
—
.61
—
.52
—
—
.75
—
.71
—
.70
—
.67
—
.54
—
.50
—
.45
ML 2 1579.1 (df 496, p .0001)
69.1%
.96
.86
One-Factor
.70
.79
.76
.82
.83
.78
.76
.77
.83
.56
.85
.74
.78
.79
.79
.61
.86
.71
.84
.85
.75
.75
.83
.83
.79
.58
.53
.62
.50
.66
.65
.64
55.2%
.97
LATE-LIFE FUNCTION INSTRUMENT
with scores approaching 0 indicating poor ability and scores
approaching 100 characterizing good function. The shortterm stability of the Late-Life FDI was assessed by calculating intraclass correlation coefficients (32) between scores at
the initial assessment and follow-up scores. We used the
method of known-groups validity to test the ability of the
overall measure and each identified dimension to discriminate between groups known to differ in functional status as
measured by the physical functioning items (PF-10) of the
SF-36 (33,34). These methods are described in detail in a
companion article (30).
RESULTS
Sample Characteristics
The final sample was composed of 150 communitydwelling older volunteers. See companion article (30) for a
full description of the sample.
Questionnaire Content
Based on an initial item review consisting of a preliminary factor analysis, feedback from the participants, Rasch
analysis to examine content redundancy and item fit, 16 of
the original 48 items were deleted from further analyses. Six
items were eliminated because they were not applicable to a
significant number of participants, six were removed due to
redundant content, one item was deleted due to respondent
difficulty interpreting item content, and three items were removed because of ambiguous loadings on one or more factors.
Function Domains
Table 1 presents the factor loading estimates and Cronbach alpha values for the one- and three-factor analysis
models for the function component. A maximum likelihood
chi-square test showed that one factor was not sufficient to
adequately explain the data (2 1579.1, p .0001). Three
factors were identified by Eigenvalues greater than 1.0. A
three-factor solution explained 69.1% of the variance. We
labeled the first factor advanced lower extremity functioning because it reflected physical activities that involve a
high level of physical ability and endurance. A second factor was labeled basic lower extremity functioning because it
is composed of activities that primarily involve standing,
stooping, and fundamental walking activities. A third factor
M219
reflected activities of the hands and arms and was labeled
upper extremity functioning. The correlation among factors
for advanced versus basic lower extremity was r .87 ( p .001), advanced lower extremity versus upper extremity
was r .64 ( p .001), and basic lower extremity versus
upper extremity was r .69 ( p .001); all indicating some
overlap.
Function Instrument Scaling
The hierarchical order of items on the function component of the Late-Life FDI is graphically depicted in Figures
1 and 2 for the one- and three-factor solutions, respectively.
Item separation statistics for the overall function scale
(item separation 10.1) and its subscales [advanced
lower extremity functioning (item separation 8.6), basic
lower extremity functioning (item separation 5.2), and
upper extremity functioning (item separation 6.2)] all exceed a value of 3.0, which is a good indicator of spread of
items along a scale.
Five of the 32 items in the overall function component
have item fit scores equal to or greater than 2.0; however,
only one is quite large (“unscrew lid off jar”; infit 6.2).
Across the three-factor solution, only four items exceed the
misfit cutoff score of equal to or greater than 2.0, and only
one (“get up from the floor”) approximated a value of 3.0.
However, because the content was found to be central to the
instrument, all were retained.
Instrument Reliability and Validity
Mean total function scores were significantly different on
all comparisons between functional limitation categories, as
indicated in Table 2. In most cases, the separate scores were
also significantly different between groups. The exceptions
were the scores for basic lower extremity functioning and
upper extremity functioning on which the slight functional
limitation group and the nonfunctional limitation group did
not differ significantly. Test-retest intraclass correlations
from 15 individuals on each summary score (Table 3) were
very high (Shrout-Fleiss random set ICC .91–.98). There
was an average of 12 days (SD 8.3) between test administrations.
DISCUSSION
Physical functioning can best be described as having one
dimension (difficulty) and three separate domains (ad-
Figure 1. Rasch model for one solution of functioning. *Identifies items with fit scores 2.0. See text for details.
M220
HALEY ET AL.
Figure 2. Rasch models for three solutions of functioning. *Identifies items with fit scores 2.0. See text for details.
vanced lower extremity functioning, basic lower extremity
functioning, and upper extremity functioning). However,
both the one- and three-factor solutions are reasonable hierarchical scales, although fit problems are more evident in
the one-factor solution. The choice of which model to use
will be determined by the particular application intended.
Domain scores may be most valuable in circumstances
where it is expected that there will be significantly different
degrees of change across the three domains.
Items on the function scales include daily activities that
sample key indicators of functioning. We found that the
items were not structured along basic ADL (BADL) or instrumental ADL (IADL) categories, but rather along dimensions that reflect different physical capabilities. For exam-
ple, the advanced lower extremity dimension had a mixture
of mobility items (e.g., “walk a brisk mile”) and physical
activities usually classified as IADLs such as shopping and
carrying groceries (e.g., “carry objects while walking upstairs”).
These results are in contrast to a previous questionnaire
that yielded three functionally related dimensions: (i) basic
ADLs (walk indoors, feed oneself), (ii) mobility (stairs,
carry a heavy object), and (iii) instrumental ADLs (cooking,
housework) (35). Analyses from instruments that measure
dependency in goal-directed activities are often consistent
with ADL and IADL factor structures (9). We surmise that
these differences in structure between our data and the factor structure of ADL scales may be related to the discrete
Table 2. Average Function Summary Scores and Mean Differences Between Adjacent Functional Limitation Groups
Means (SD)
Functioning total
Advanced lower extremity
Basic lower extremity
Upper extremity
Mean Differences (SE)
No. of
Items
Severe Functional
Limitations
(n 27)
Moderate Functional
Limitations
(n 45)
Slight Functional
Limitations
(n 57)
No Functional
Limitations
(n 21)
Severe to
Moderate
Moderate to
Slight
Slight
to None
32
11
14
7
41.7 (7.0)
9.7 (13.6)
45.6 (11.9)
64.4 (16.9)
53.2 (6.2)
35.1 (13.2)
65.6 (12.1)
73.2 (12.5)
65.6 (6.9)
57.7 (9.7)
81.5 (12.9)
84.6 (10.7)
75.6 (11.0)
72.1 (14.3)
89.3 (11.7)
88.4 (9.6)
11.5* (1.8)
25.5* (3.0)
20.0* (3.0)
8.7** (3.0)
12.4* (1.5)
22.6* (2.4)
15.9* (2.5)
11.5* (2.5)
10.0* (1.9)
14.4* (3.1)
7.8 (3.1)
3.8 (3.2)
Notes: The difference between the means was tested using the least squared differences t test, which controls the Type 1 comparisonwise error rate. The alpha level
for statistical significance was adjusted for the three pairwise comparisons of interest.
*p .0167 (alpha 0.05/3 comparisons).
**p .05 (alpha 0.01/3 comparisons).
LATE-LIFE FUNCTION INSTRUMENT
M221
Table 3. Test-Retest Reliability
Mean Scores
Functioning total
Advanced lower extremity
Basic lower extremity
Upper extremity
Reliability
No. of Items
Test 1 (SD)
Test 2 (SD)
Difference (1–2)
ICCs*
p Values
32
11
14
7
62.9 (13.0)
51.0 (22.3)
76.55 (19.8)
82.3 (15.5)
61.1 (12.3)
48.4 (21.7)
74.3 (20.0)
81.1 (15.3)
1.9
2.6
2.2
1.3
.960
.966
.976
.912
.001
.001
.001
.001
*Intraclass correlations (ICCs) are calculated using Shrout-Fleiss formulae assuming the two time points are from a random set of all possible time points.
nature of the activities that are part of the function component of the Late-Life FDI. In addition, our items asked respondents to rate their experienced degree of difficulty performing these activities without help, whereas many other
instruments have used dependency or assistance levels as
the measure of function. Adults may report that they perform a particular functional activity without help, yet experience quite different degrees of difficulty. Our results are
similar to the three major dimensions of function proposed
by Avlund (upper limb function, lower limb function, and
mobility) (36) from data analyses of an instrument that used
a tiredness scale as the evaluative criterion.
Results of the known-groups analyses confirm the content spread of each scale and suggest that the scales will be
able to discriminate among groups of older persons with
different levels of functioning. These results are consistent
with previous findings that self-reported difficulty in functional tasks identified distinct levels of functional ability
(37). The results of the present analyses also suggest that the
new measure has successfully avoided the difficulty with
floor or ceiling effects identified on many self-report measures (21). Only one individual in our sample (less than 1%)
reached a maximum score on the overall functioning scale,
and no persons received the lowest score possible (summary
score mean 59.0, SD 13.2). In contrast, on the PF-10, 21
(14%) of our sample had a maximum score (ceiling effect),
and one individual had the lowest possible score (floor effect).
Test-retest reliability correlations for self-report ADLs
and IADLs generally have been reported to be high (r .90
or above) (38). The total function summary scores and separate dimension scores on the Late-Life FDI showed equally
high consistency, including the upper extremity functioning
score (ICC .91), which is derived from a small number of
items (n 7).
These findings indicate that the function component of
the Late-Life FDI has potential to assess functioning across
a wide variety of daily activities with a reasonable set of
items and with stability over short periods of time. In subsequent field tests, we will evaluate the ability of the function
component of the Late-Life FDI to detect meaningful
changes in functional status across dimensions in a larger
sample of community-dwelling older persons and compare
its psychometric characteristics to those of other functional
instruments commonly used in gerontological intervention
research.
Acknowledgments
This project was funded by the National Institutes of Health, National
Institute on Aging, Grant AG11669. We thank all of the volunteers who
participated in our study and New England Research Institute for recruiting
50 subjects for the pilot test of the initial instrument draft.
To obtain a copy of the Late-Life FDI, please access www.bu.edu/
roybal/products on the World Wide Web.
Address correspondence to Stephen M. Haley, Sargent College of
Health and Rehabilitation Sciences, Boston University, Boston, MA
02215. E-mail: [email protected]
References
1. Stewart A, Ware JE, Brook RH. The meaning of health: understanding
functional limitations. Med Care. 1977;XV(11):939–952.
2. Kaplan GA, Strawbridge WJ, Camacho T, Cohen RD. Factors associated with change in physical functioning in the elderly. J Aging
Health. 1993;5(1):140–153.
3. Gill T, Williams C, Tinetti M. Assessing risk for the onset of functional dependence among older adults: the role of physical performance.
J Am Geriatr Soc. 1995;43:603–609.
4. Siu A, Reuben D, Hays R. Hierarchical measures of physical function
in ambulatory geriatrics. J Am Geriatr Soc. 1990;38(10):1113–1119.
5. Spector WD, Katz S, Murphy JB, Fulton JP. The hierarchical relationship between activities of daily living and instrumental activities of
daily living. J Chronic Dis. 1987;40(6):481–489.
6. Rodgers W, Miller B. A comparative analysis of ADL questions in
surveys of older people. J Gerontol Psychol Sci Soc Sci. 1997;
52B(special issue):21–36.
7. Winograd C, Lindenberger E, Chavez C, Mauricio M, Shi H, Bloch D.
Identifying hospitalized older patients at varying risks for physical performance decline: a new approach. J Am Geriatr Soc. 1997;45:604–609.
8. Pols MA, Peeters PH, Ocke MC, Slimani N, Bueno-de-Mesquita HB,
Collette HJ. Estimation of reproducibility and relative validity of the
questions included in the EPIC Physical Activity Questionnaire. Int J
Epidemiology. 1997;26[1(suppl 1)]:S181–S189.
9. Avlund K, Kreiner S, Schultz-Larsen K. Functional ability scales for
the elderly: a validation study. Eur J Public Health. 1996;6:35–42.
10. Avlund K, Davidsen M, Schroll M, Gause-Nilsson I, Laukkanen P,
Heikkinen E. Self-reported functional ability in three Nordic localities:
a comparative study of activities of daily living among 75-year-old
men and women in Glostrup, Gothenburg and Jyvaskyla. Int J Health
Sci. 1994;5(3):119–128.
11. Kempen GI, Suurmeijer TP. The development of a hierarchical
polychotomous ADL-IADL scale for noninstitutionalized elders. Gerontologist. 1990;30:497–502.
12. Ren X, Kazis L, Lee A, et al. Comparing generic and disease-specific
measures of physical and role functioning. Results from the Veterans
Health Study. Med Care. 1998;36(2):155–166.
13. Thomas VS, Rockwood K, McDowell I. Multidimensionality in instrumental and basic activities of daily living. J Clin Epidemiol. 1998;
51(4):315–321.
14. Dencker K, Gottfries C. Activities of daily living ratings of elderly
people using Katz’s ADL Index and the GBS-M Scale. Scand J Caring Sci. 1995;9(1):35–40.
M222
HALEY ET AL.
15. Kelly S, Jessop EG. A comparison of measures of disability and health
status in people with physical disabilities undergoing vocational rehabilitation. J Public Health Med. 1996;18(2):169–174.
16. Asberg K, Sonn U. The cumulative structure of personal and instrumental ADL: a study of elderly people in a health service district.
Scand J Rehabil Med. 1988;21:171–177.
17. Reuben DB. What’s wrong with ADLs? J Am Geriatr Soc. 1995;
43(8):936–937.
18. Linn BS, Linn MW. Objective and self-assessed health in the old and
very old. Soc Sci Med. 1980;14A:311–315.
19. Nagi SZ. Some conceptual issues in disability and rehabilitation. In:
Sussman MB, ed. Sociology and Rehabilitation. Washington, DC:
U.S. Department of Health, Education, and Welfare; 1965:100–113.
20. Nagi SZ. Disability concepts revisited: implications for prevention. In:
Pope AM, Tarlov AR, eds. Disability in America: Toward a National
Agenda for Prevention. Washington, DC: National Academy Press;
1991:309–327.
21. Sherman S, Reuben D. Measures of functional status in communitydwelling elders. J Gen Intern Med. 1998;13:817–823.
22. Stewart A, Ware J. Measuring Functioning and Well-Being: The Medical Outcomes Approach. Durham, NC: Duke University Press; 1992.
23. Jette A, Branch L. The Framingham Disability Study: II. Physical disability among the aging. Am J Public Health. 1981;71(11):1211–1216.
24. Frederiks C, te Wierik M, Visser A, Sturmans F. A scale for functional
status of the elderly living at home. J Adv Nurs. 1991;16:287–292.
25. Ware J, Sherbourne C. The MOS 36-item Short Form Health Survey
(SF-36). Med Care. 1992;30(6):473–483.
26. Lyons RA, Perry HM, Littlepage BNC. Evidence for the validity of
the Short-Form 36 questionnaire (SF-36) in an elderly population. Age
Ageing. 1994;23:182–184.
27. Prieto L, Alonso J, Ferrer M, Anto J. Are results of the SF-36 Health
Survey and the Nottingham Health Profile similar? A comparison in
COPD patients. J Clin Epidemiol. 1997;50(4):463–473.
28. O’Mahony P, Rodgers H, Thomson R, Dobson R, James O. Is the SF-
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
36 suitable for assessing health status of older stroke patients? Age
Ageing. 1998;27(1):19–22.
Brazier J, Harper R, Jones N, et al. Validating the SF-36 Health Survey Questionnaire: new outcome measure for primary care. BMJ.
1992;305:106–164.
Jette A, Haley S, Coster W, et al. Late Life Function and Disability Instrument: I. Development and evaluation of the disability component.
J Gerontol Med Sci. 2002;57A:M209–M216.
Wright B, Masters G. Rating Scale Analysis. Chicago: MESA Press; 1982.
Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420–428.
Kerlinger F. Foundations for Behavioral Research. New York: Holt,
Rinehart, & Winston; 1973.
McHorney CA. Methodological and psychometric issues in health status assessment across populations and applications. In: Albrecht GL,
Fitzpatrick R, eds. Advances in Medical Sociology, Vol. 5: Quality of
Life in Health Care. Greenwich, CT: JAI Press, LTD; 1994.
Hoeymans N, Feskens EJ, van den Bos GA, Kromhout D. Measuring
functional status: cross-sectional and longitudinal associations between performance and self-report (Zutphen Elderly Study, 1990–
1993). J Clin Epidemiol. 1996;49(10):1103–1110.
Avlund K, Kreiner S, Schultz-Larsen K. Construct validation and the
Rasch model: functional ability of health elderly people. Scand J Soc
Med. 1993;21(4):233–246.
Langlois J, Maggi S, Harris T, et al. Self-report of difficulty in performing functional activities identifies a broad range of disability in
old age. J Am Geriatr Soc. 1996;44:1421–1428.
Ottenbacher KJ, Mann WC, Granger CV, Tomita M, Hurren D, Charvat
B. Inter-rater agreement and stability of functional assessment in the community-based elderly. Arch Phys Med Rehabil. 1994;75(12):1297–1301.
Received September 11, 2001
Accepted October 25, 2001