HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Saliba, D. Articles by Rubenstein, L. Z. Search for Related Content PubMed PubMed Citation Articles by Saliba, D. Articles by Rubenstein, L. Z.

Identifying a Short Functional Disability Screen for Older Persons

^a Geriatric Research Education and Clinical Center, VA Greater Los Angeles Health Care System, California
^b Department of Medicine, University of California–Los Angeles Multicampus Program in Geriatrics
^c RAND Corporation, Santa Monica, California
^d University of California–Los Angeles Division of General Internal Medicine & Health Services Research

Debra Saliba, RAND, 1700 Main Street, Santa Monica, CA E-mail: saliba{at}rand.org.

Decision Editor: John E. Morley, MB, BCh

	Abstract

Top Abstract Methods Results Discussion References

 
Background. Disability in instrumental activities of daily living (IADLs) or activities of daily living (ADLs) is an indicator of health risk. The inclusion of these items in population screens may be limited by variation in item performance across gender and age groups. Further, identification of shortened lists may encourage inclusion of these items in screens.

Methods. We applied item response theory (IRT) methods to assess the responses of 9865 community-dwelling elders in the 1993 Medicare Current Beneficiary Survey to 11 IADL/ADL items. Items were classified as "receive help/ not receive help" for the overall population and stratified by age and gender. We assessed the same IADL/ADL items using responses classified as "difficulty/no difficulty." After eliminating items that performed poorly, we performed all-subsets analyses to identify abbreviated sets of items that would select the highest proportion of persons with IADL/ADL disability.

Results. Responses classified in receive help format showed consistency by gender and age group. Changing the response classification to difficulty/no difficulty influenced the reported order and relationship of IADL/ADL items. Receipt of help for any one of five items—shopping, doing light housework, walking, bathing, or managing finances—identified 93% of individuals receiving help with any IADL/ADL. A slightly different set of five items—walking, shopping, transferring, doing light housework, or bathing—identified 91% of persons reporting difficulty with any IADL or ADL.

Conclusions. The relationship of IADL and ADL items to the underlying construct of disability was similar for men and women. The relationship was also similar for oldest-old and younger-old individuals. This study also identified abbreviated lists of disability items that can be used to efficiently screen community-dwelling elders for the presence of IADL/ADL disability.

DISABILITY in basic activities of daily living (ADLs) and instrumental activities of daily living (IADLs) ^{(1) (2) (3)} predicts future functional decline, institutionalization, and death ^{(4) (5) (6) (7) (8) (9) (10)}. Most IADL/ADL lists used in surveys contain 10 to 15 items, and many item responses require follow-up questions, for example, identifying task modification ⁽¹¹⁾ and identifying the need for human assistance. Thus, a full set of IADL/ADL items can be cumbersome to incorporate in surveys. Having an efficient list of IADL and ADL items could facilitate consideration of disabled and high-risk individuals in population-based rapid health screens and individual risk assessments ^{(6) (12) (13) (14) (15)}.

Ideally, a screening instrument should perform similarly in different subgroups of the targeted population. However, the self-reported difficulty of IADL/ADL items may vary by gender because of traditional differences in role function ^{(16) (17)}. For example, Lawton ⁽³⁾ reported that food preparation, laundry, and housekeeping failed to form an ordered Guttman scale for men. Even if the question clarifies whether nonperformance is related to health, the hierarchy of IADL and ADL performance may vary by gender. Men are more likely than women to attribute receiving assistance with household tasks and shopping to traditional role assignment rather than to health ⁽¹⁷⁾. This may lead to underreporting of health-related disability because a significant percentage of men who attribute nonperformance to traditional division of labor have morbidity levels that suggest physical inability to perform the task ⁽¹⁷⁾.

Several findings also raise concern about whether the same screening instrument can be used for oldest-old individuals as is used for younger-old individuals. The prevalence of self-reported difficulty and self-reported need for help varies by age ⁽¹⁸⁾. While the prevalence of disability generally increases with age, the extent of decline varies among activities, and the hierarchical order of IADL items has been shown to vary between age subgroups ⁽⁶⁾. In addition, a recent meta-analysis showed that variation in the wording of non-IADL/ADL global health assessment questions results in different response distributions when comparing old-old to young-old groups ⁽¹⁹⁾. Thus, any abbreviated IADL/ADL list would have to consider whether items would perform differently for different gender or age categories.

Item wording, including the choice of how to dichotomize responses, might also influence the selection of IADL/ADL items for a screening instrument. The response categories used to assess IADL/ADL limitations include difficulty versus no difficulty, able versus not able, receive help versus not receive help, and does versus does not do. The choice of item wording and response categories can lead to different prevalence and outcome estimates across studies and in the same population ^{(20) (21) (22) (23)}. In addition, individual IADL/ADL items differ in the extent to which they vary across response categories ^{(18) (24)}. These variations affect the interpretation of IADL/ADL items and make comparison of ordered lists difficult.

Analytic approaches to examining item performance include Guttman scaling and item response theory (IRT) methods. Guttman scaling tests the hierarchical relationship of items and has been used to evaluate IADL/ADL item performance ^{(6) (12) (15)}. Guttman scaling is a deterministic model and assumes all items are equally related to the measured construct ⁽²⁵⁾. The IRT methodology, on the other hand, is probabilistic and can allow for the possibility that items may be differentially related to the underlying construct. In the case of IADL and ADL items, IRT analysis estimates the probability of reporting inability to perform an activity as a function of the respondent's degree of disability. IRT methods have been applied to National Long Term Care Survey data to show that a count of IADL/ADL deficits can summarize individual disability ⁽²⁶⁾.

To determine which IADL/ADL items were most appropriate to use in a general population-screening instrument, this study applied IRT methods to a representative community sample. This study estimated the relationship of each IADL/ADL item to the underlying construct of disability, examined the stability of the item properties across gender and age subgroups, and then identified efficient lists for IADL/ADL screening.

	Methods

Top Abstract Methods Results Discussion References

 
We applied IRT methods to data in the 1993 Medicare Current Beneficiary Survey (MCBS) public use files to (i) examine item performance for receipt of help with 11 IADL/ADL items, (ii) examine the relationship of each item to overall IADL/ADL disability for men compared with women, and (iii) examine the properties of these items for persons aged 85 or older compared with persons aged 65 to 84 years. We repeated steps i through iii using a common alternate approach to classifying responses—difficulty versus no difficulty. After eliminating items that performed poorly, we performed a best-subsets analysis to identify abbreviated sets of receive help items that would select the highest proportion of persons receiving assistance. We repeated the best-subset analysis for the difficulty list.

MCBS
The MCBS surveys a representative sample of all Medicare enrollees. We restricted our analysis to the 9865 community-dwelling respondents who were aged 65 or older in 1993 (i.e., we excluded MCBS respondents who were institutionalized or aged less than 65 years). Because we wanted to identify all community-dwelling elders with disability, we included proxy respondents. The MCBS oversampled persons age 85 and older, providing an opportunity to assess responses for this oldest-old group.

The MCBS included five IADL and six ADL questions that asked "because of a health or physical problem (do you/does study participant) have any difficulty ... ?": "using the telephone," "doing light housework," "preparing own meals," "shopping for personal items (such as toilet items or medicines)," "managing money," "bathing or showering," "dressing," "eating," "getting in or out of bed or chairs," "walking across room," and "using the toilet." The questions did not include an explicit time frame but were phrased in the present tense. Thus, a report of difficulty could reflect temporary/acute or chronic disability.

The response options included "yes," "no," "doesn't do." Respondents answering "doesn't do" were asked if this was for a health-related reason. Respondents reporting difficulty were asked if help was received from another person. We classified standby assistance as receive help ⁽²⁾. We classified item responses in two different ways: difficulty versus no difficulty and receive help versus not receive help. We classified persons who reported difficulty (with or without receipt of help) or nonperformance for health-related reasons as having difficulty. We classified persons who reported difficulty resulting in receipt of help or nonperformance for health-related reasons as receiving help. For the remaining analyses, the variable "difficulty" refers to any reported difficulty or nonperformance for a health reason and "receive help" refers to any reported receipt of human help or nonperformance for a health reason.

IRT Analysis
We used IRT methodology to assess the relationship of each of the 11 IADL/ADL items to the underlying construct of disability (i.e., to overall IADL/ADL disability). The IRT two-parameter logistic (2PL) model yields an item characteristic curve (ICC) that is described by the location (b) and slope (a) parameters. The b parameter is the point along the ICC at which the probability of a positive response for a dichotomous item is 50%. The larger the location parameter, the more of the measured construct a respondent must have to endorse that item. The a parameter represents the slope of the ICC at the value of the location parameter and indicates the extent to which the item is related to the underlying construct. A steeper slope indicates a closer relationship to the construct. Differential slopes for items indicate that the items are not uniformly related to the measured construct.

Because IRT examines the relationship of individual's responses to each IADL/ADL item, we applied IRT methods to the actual study sample and did not weight the data to reflect the larger population. We compared the fit of the 2PL model with that of a 1PL model that assumes all items are equally related to the underlying construct (i.e., have the same slope) using the chi-square difference test. We selected a model to fit separately for the identified gender and age subgroups and calibrated the items for each subgroup. Parallel analyses were performed for the two types of response coding, difficulty versus receive help. The results of the two coding and four subgroup analyses were used to assess which items were comparable across subgroups and were therefore appropriate for inclusion in a general population screen.

All-Subsets Analyses
After eliminating any items whose slope indicated poor performance for identifying overall IADL/ADL disability, we performed all-subsets analyses to select the best abbreviated lists. Specifically, we created a minimum prediction rule to identify the combination of disability items that identified the highest proportion of subjects with any IADL/ADL disability. We identified all possible combinations of the IADL/ADL items and the percent of persons reporting any IADL or ADL disability that each combination identified. We thus considered list lengths ranging from single-item length to all- item length. We used MCBS cross-sectional weights to estimate the percent of the disabled population identified.

	Results

Top Abstract Methods Results Discussion References

 
Respondents' mean age was 76 (range 65–104); 14.5% of the sample was aged 85 or older; 41% of the sample was male. Proxies completed 10.8% of surveys. In Table 1 , columns 2 through 4 show the proportion of people choosing each IADL/ADL response category. When we classified respondents according to difficulty versus no difficulty (column 5), 25.6% and 32.3% had difficulty with any IADL and any ADL, respectively. When we classified respondents according to receive help versus not receive help (column 6), 21.1% and 11.5% received help with any IADL and any ADL, respectively.

Table 2 , column 1 shows the item order, according to the location parameter, b (column 3), for the 11 receive help items in the overall population. When men (columns 4–6) were compared with women (column 7–9), the ranges of both the location and slope parameters were similar. The item order differed only slightly. Only walking and using the telephone moved in relative ranking. Notably, although walking precedes managing finances for women, the location parameters (b) are essentially the same for walking (b = 1.70) and managing finances (b = 1.72).

We performed all-subsets analyses separately for the receive help coding and the difficulty coding using the remaining 10 items. For five potential list lengths, Table 5 presents the list that identifies the most persons endorsing any of the 10 items and the percentage of persons identified. The percentage of disabled persons endorsing any of 11 items, that is, the 10-item list plus telephone, is also presented. (Telephone was not a candidate screening question.) For each potential list length presented, the receive help list and the difficulty list differ by at least one item. The most efficient list of items for identifying more than 90% of persons receiving help with IADL/ADL was shopping, doing light housework, walking, bathing, and managing finances. This list identified 93% of persons receiving help with any of 11 IADL/ADL items and 97% of persons receiving help with any of the 10 items most closely related to the construct of disability (i.e., a list that excludes using the telephone). The efficient list of items for identifying more than 90% of persons reporting difficulty was walking, shopping, transferring, doing light housework, and bathing. This list identified 91% of persons reporting difficulty with any of 11 IADL/ADL items and 96% of persons reporting difficulty with the IADL/ADL list that excludes using the telephone.

	Discussion

Top Abstract Methods Results Discussion References

We used IRT methods to clarify the relationship of each IADL/ADL item to overall IADL/ADL disability. Notably, the IRT model that allowed IADL/ADL items to display different relationships to the underlying construct of disability performed better than a model that assumed that each item was equally related to disability. The IRT-derived order for the receive help ADL items (e.g., bathing first and feeding last) is consistent with prior Guttman-based proposed orders for ADL ^{(2) (25)}. A commonly accepted IADL order is not available for comparison since IADL order has been less consistent across studies ^{(3) (6)}.

The similar patterns of receive help items for men and women in our analysis supports the creation of a single screening instrument for both groups. The male and female subgroups demonstrated similar item order when that order was determined by each item's relationship to overall IADL/ADL disability. For both men and women, IADL and ADL integrate physical and cognitive abilities ^{(12) (24)} and represent self-maintenance tasks important for living independently ⁽²⁷⁾. Qualifying IADL nonperformance as health-related decreases the misattribution of preferred or traditional nonperformance to disability. Also, environmental factors (e.g., marital status) and/or comorbidity may be more important determinants of dependency than gender alone ⁽¹⁷⁾.

The overall order and performance of receive help items also were similar for persons aged 65 to 84 years compared with persons aged 85 and older. Importantly, differences in the order of several items were determined by very small differences in location parameters. Failure to consider the actual parameter values might have led to an overexaggeration of the change in item order. Such slight differences may explain age-based inconsistencies in other studies ⁽⁶⁾.

Item performance and overall order differed for items coded as receive help compared with items coded as difficulty. This difference supports contentions that difficulty responses complement help responses ^{(23) (24)}, that response classification influences item order ^{(26) (28) (29)}, and that difficulty is not a simple midpoint value on a scale extending from no difficulty to receive help ⁽²⁰⁾. Differentiating difficulty from receive help provides a different and more complete characterization of disability ⁽¹¹⁾. Our study shows that this difference stems from more than differences in point prevalence ⁽²⁵⁾, and that difficulty and receive help vary in their relationship to the underlying construct of disability for different population subgroups.

After reviewing item performance, we eliminated using the telephone from consideration for the abbreviated screening instrument. Using the telephone has been problematic in other studies. Fillenbaum ⁽⁶⁾ found that using the telephone failed to load on the same factor as five other IADL items. When Longitudinal Study on Aging respondents identified the underlying condition causing difficulty with specific tasks, 53% selected hearing/vision problems as the cause for difficulty using the telephone, a markedly different explanation from that given for other IADLs ⁽³⁰⁾. In addition, respondents may be considering different embedded tasks, for example, answering the phone, reaching the phone, placing calls, recalling numbers, using a written directory, and using directory assistance ⁽²⁴⁾.

After excluding using the telephone as a screening item, we identified efficient lists for disability screening. The lists in Table 5 serve as a resource for selecting an abbreviated list to screen for IADL and ADL disability. The lists range in length from a single item—shopping—that identified 69% of persons who received help with any of 11 IADL/ADL items, to five items—shopping, doing light housework, walking, bathing, and managing finances—that identified 93%. Thus, the table illustrates the potential effect that adding or subtracting items has on the ability to identify all persons with IADL/ADL disability. As an example, a query about receiving help with shopping might be supplemented by a query about doing light housework. The addition of this one item would increase detection of IADL/ADL disability by 9%.

The MCBS format and content may have influenced our results. The MCBS asked if the respondent received help, not if the respondent needed help. Unmet need could explain some of the difference between having difficulty and receiving help. However, unmet need could also lessen the difference because assistive devices can avert the need for assistance, and unmet need for these devices may hasten dependency. Despite these potential resource issues, the receive help response format is a common metric for ADL scales, including the original classification ⁽²⁾.

The disability items do not include ability to arrange transportation or to manage medications. The ability to manage medications ⁽⁶⁾ and transportation ^{(6) (15)} has been found to have a questionable relationship to other IADL items. Women are more likely than men to attribute nonperformance of transportation to traditional role performance ⁽¹⁷⁾.

An elder may fail to recognize or admit when health prevents the performance of a task. Such underreporting, if influenced by gender-based preferences ⁽¹⁷⁾, would tend to increase the differences in item performance between men and women. Because our analyses found no important gender differences, possible underreporting does not alter our conclusion that men and women can be assessed with the same IADL/ADL screen.

We investigated variation only by gender and age. Cultural differences in IADL may be significant ^{(6) (31)}. Unfortunately, the MCBS data did not allow us to ascertain whether cultural differences influence IADL/ADL performance. Future studies should examine item performance across ethnic groups and cultures.

Our analyses included proxy responses. Proxy and respondent estimation of overall disability may differ ⁽³²⁾. However, it is not clear that these differences would alter the relationship between the items. On a practical level, we wanted to select items for a disability survey. In the absence of direct observation, proxy responses are needed to avoid excluding many subjects with severe disability ⁽³³⁾.

We performed a population-based analysis to identify short lists useful for rapid screening. Abbreviated lists may fail to identify all IADL/ADL disabilities and may miss smaller areas of dysfunction within an item. Not all persons follow the same path to disability, and the underlying cause of disability (e.g., dementia vs stroke) may differentially affect individual performance of specific tasks ⁽²⁵⁾. Likewise, our findings related to gender performance do not obviate the need for individualized resource planning that considers traditional role function ⁽¹⁷⁾.

Conclusions
IADL/ADL disability for men, women, and oldest-old individuals can be correctly classified using the same survey. Five IADL/ADL items identified 93% of persons that receive help with any IADL or ADL, while a slightly different five items identified 91% of persons that have difficulty with any IADL or ADL. The abbreviated lists may facilitate the inclusion of IADL/ADL items in more settings. These abbreviated lists can be used with confidence that the items relate to overall IADL/ADL disability in community-dwelling elders.

	Acknowledgments

 
Supported by a contract from Pfizer Pharmaceuticals, Inc. Global Outcomes Research, Dr. Saliba is a recipient of a UCLA Claude D. Pepper Older Americans Independence Center Career Development Award AG10415. We thank David Reuben for his review of the manuscript. We also thank Lisa Wong and Yingying Ma for programming assistance and Patty Smith for her assistance in manuscript preparation. The ACOVE research team that encouraged the development of this work includes David Solomon, Robert Brook, Roy Young, Paul Shekelle, John Schnelle, Catherine MacLean, Elizabeth Sloss, Caren Kamberg, and Carol Roth.

Received January 14, 2000

Accepted February 17, 2000

	References

Top Abstract Methods Results Discussion References

 

1. Staff of The Benjamin Rose Hospital1959. Multidisciplinary studies of illness in aged persons: II. A new classification of functional status in activities of daily living. J Chron Dis. 9:55-62. [Medline]
2. Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW, 1963. Studies of illness in the aged the index of ADL: a standardized measure of biological and psychosocial function. JAMA. 185:914-919.
3. Lawton MP, Brody EM, 1969. Assessment of older people: self-maintaining and instrumental activities of daily living. The Gerontologist 9:179-186. [Medline]
4. Katz S, Ford AB, Chinn AB, Newill VA, 1966. Prognosis after strokes. Part II. Long-term course of 159 patients. Medicine. 45:236-246.
5. Katz S, Downs TD, Cash HR, Grotz RC, 1970. Progress in development of the index of ADL. The Gerontologist. 10:20-30. [Medline]
6. Fillenbaum GG, 1985. Screening the elderly: a brief instrumental activities of daily living measure. J Am Geriatr Soc 33:698-706. [Medline]
7. Mor V, Wilcox V, Rakowski W, Hiris J, 1994. Functional transitions among the elderly: patterns, predictors, and related hospital use. Am J Pub Health 84:1274-1280. [Abstract/Free Full Text]
8. Reuben DB, Rubenstein LV, Hirsch SH, Hays RD, 1992. Value of functional status as a predictor of mortality: results of a prospective study. Am J Med 93:663-669. [Medline]
9. Fried LP, Kronmal RA, Newman AB, et al. 1998. Risk factors for 5-year mortality in older adults: the cardiovascular health study. JAMA 279:585-592. [Abstract/Free Full Text]
10. Brody KK, Johnson RE, Ried LD, 1997. Evaluation of a self-report screening instrument to predict frailty outcomes in aging populations. The Gerontologist 37:182-191. [Abstract]
11. Fried LP, Bandeen-Roche K, Williamson JD, et al. 1996. Functional decline in older adults: expanding methods of ascertainment. J Gerontol Med Sci. 51A:M206-M214. [Abstract]
12. Siu AL, Reuben DB, 1990. Hierarchical measures of physical function in ambulatory geriatrics. J Am Geriatr Soc 38:1113-1119. [Medline]
13. Lachs MS, Feinstein AR, Cooney LM, et al. 1990. A simple procedure for general screening for functional disability in elderly patients. Ann Intern Med 112:699-706.
14. Lawton MP, Moss M, Fulcomer M, Kleban MH, 1982. A research and service oriented multilevel assessment instrument. J Gerontol 37:91-99. [Medline]
15. Spector WD, Katz S, Murphy JB, Fulton JP, 1987. The hierarchical relationship between activities of daily living and instrumental activities of daily living. J Chron Dis 4:481-489.
16. Kempen GIJM, Miedema I, Ormel J, Molenaar W, 1996. The assessment of disability with the Groningen Activity Restriction scale: conceptual framework and psychometric properties. Soc Sci Med 43:1601-1610.
17. Allen SM, Mor V, Raveis V, Houts P, 1993. Measurement of need for assistance with daily activities: quantifying the influence of gender roles. J Gerontol Soc Sci 48:S204-S211.
18. Jette AM, 1994. How measurement techniques influence estimates of disability in older populations. Soc Sci Med 38:937-942.
19. Roberts G, 1999. Age effects and health appraisal: a meta-analysis. J. Gerontol Biol Sci 54B:S24-S30.
20. Wiener JM, Hanley RJ, Clark R, Van Nostrand JF, 1990. Measuring the activities of daily living: comparisons across national surveys. J Gerontol Soc Sci 45:S229-S237.
21. Langlois JA, Maggi S, Harris T, et al. 1996. Self-report of difficulty in performing functional activities identifies a broad range of disability in old age. J Am Geriatr Soc 44:1421-1428. [Medline]
22. Fried LP, Herdman SJ, Kuhn KE, Rubin G, Turano K, 1991. Preclinical disability: hypotheses about the bottom of the iceberg. J Aging Health 3:285-300. [Abstract/Free Full Text]
23. Gill TM, Robison JT, Tinetti ME, 1998. Difficulty and dependence: two components of the disability continuum among community-living older persons. Ann Intern Med. 128:96-101. [Abstract/Free Full Text]
24. Myers AM, 1992. The clinical swiss army knife: empirical evidence on the validity of IADL functional status measures. Medical Care 30: (suppl) MS96-MS111. [Medline]
25. Lazaridis EN, Rudberg MA, Furner SE, Cassel CK, 1994. Do activities of daily living have a hierarchical structure? An analysis using the longitudinal study of aging. J Gerontol Med Sci 49:M47-M51.
26. Spector WD, Fleishman JA, 1998. Combining activities of daily living with instrumental activities of daily living to measure functional disability. J Gerontol Biol Sci 53B:S46-S57.
27. LaPlante MP, 1991. The demographics of disability. Milbank Q 69: (suppl 1–2) 55-77.
28. Kempen GIJM, Suurmeijer TPBM, 1990. The development of a hierarchical polychotomous ADL-IADL scale for noninstitutionalized elders. The Gerontologist 30:497-501. [Abstract]
29. Kempen GIJM, Myers AM, Powell LE, 1995. Hierarchical structure in ADL and IADL: analytical assumptions and applications for clinicians and researchers. J Clin Epidemiol 48:1299-1305. [Medline]
30. Crimmins EM, Saito Y, 1993. Getting better and getting worse: transitions in functional status among older Americans. J Aging Health 5:3-36. [Abstract/Free Full Text]
31. Alonso J, Black C, Norregaard JC, et al. 1998. Cross-cultural differences in the reporting of global functional capacity: an example in cataract patients. Med Care 36:868-878. [Medline]
32. Rothman ML, Hedrick SC, Bulcroft KA, Hickam DH, Rubenstein LZ, 1991. The validity of proxy-generated scores as measures of patient health status. Med Care 29:115-124. [Medline]
33. McBean AM, Turner CF, Fitterman LK, et al. 1999. Monitoring the health status and impact of treatment on Americans: the medicare beneficiary health status registry. Med Care 37:189-203. [Medline]

This article has been cited by other articles:

				M. Extermann and A. Hurria Comprehensive Geriatric Assessment for Older Patients With Cancer J. Clin. Oncol., May 10, 2007; 25(14): 1824 - 1831. [Abstract] [Full Text] [PDF]

				F. Lally and P. Crome Understanding frailty Postgrad. Med. J., January 1, 2007; 83(975): 16 - 20. [Abstract] [Full Text] [PDF]

				J. E. Morley Editorial. Mobility Performance: A High-Tech Test for Geriatricians J. Gerontol. A Biol. Sci. Med. Sci., August 1, 2003; 58(8): M712 - 714. [Full Text] [PDF]

				A. J. Sinclair Commentary J. Gerontol. A Biol. Sci. Med. Sci., December 1, 2002; 57(12): M811 - 812. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Saliba, D. Articles by Rubenstein, L. Z. Search for Related Content PubMed PubMed Citation Articles by Saliba, D. Articles by Rubenstein, L. Z.