| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
1 University of Iowa and Iowa City Veterans Administration Medical Center.
2 Indiana University and Regenstrief Institute, Inc., Indianapolis.
3 University of Florida and Gainesville Veterans Administration Medical Center.
4 Saint Louis University, Missouri.
5 Washington University in St. Louis, Missouri.
Address correspondence to Fredric D. Wolinsky, College of Public Health, University of Iowa, E-205 General Hospital, 200 Hawkins Dr., Iowa City, IA 52242. E-mail: fredric-wolinsky{at}uiowa.edu
 |
Abstract |
|---|
 TOP Abstract MethodResultsDiscussionReferences |
|---|
Methods. Nine hundred ninety-eight African American men and women 49–65 years old in St. Louis, MO, received comprehensive in-home evaluations at baseline and two annual telephone follow-ups. Outcome measures included walking a half-mile, climbing steps, stooping–crouching–kneeling, lifting or carrying 10 lbs., and doing heavy housework.
Results. The baseline prevalence of subclinical status was 26.4% for walking a half-mile, 26.8% for climbing steps, 39.0% for stooping–crouching–kneeling, 29.1% for lifting or carrying 10 lbs., and 22.7% for doing heavy housework. The adjusted odds ratios for the task-specific subclinical status measure at baseline on developing difficulty 1–2 years later were 1.68 (p <.05) for walking a half-mile, 4.46 (p <.001) for climbing steps, 2.48 (p <.001) for stooping–crouching–kneeling, 2.51 (p <.001) for lifting or carrying 10 lbs., and 2.22 (p <.001) for doing heavy housework. Performance tests (tandem stand, chair stands, and preferred gait speed) did not have consistent independent effects on the onset of functional limitation or disability.
Conclusions. The subclinical status measures were the main predictors of the onset of difficulty in all tasks and functions 1–2 years later. Interventions to reduce frailty should focus on self-reported subclinical status as an early warning system.
In 1991, the Institute of Medicine (IOM, 1991)
identified the prevention of age-related functional limitation and disability in older adults as top priorities. Their importance has since skyrocketed because of growing evidence that functional limitation and disability result in frailty (Fried, Ettinger, Lind, Newman, & Gardin, 1994; Fried & Guralnik, 1997; Fried & Walston, 1998; Fried, Young, Rubin, & Bandeen-Roche, 2001; Walston & Fried, 1999), a condition characterized by multiple chronic conditions and inadequate reserve capacity to endure health setbacks (Fried et al., 2001). Two-fifths of Americans now live with frailty for several years before dying (Lunney, Lynn, & Hogan, 2002), and the average frail person has self-care disability that lasts for 2 years (Manton, 1989). As a result, in 2003, the IOM identified frailty as a top priority area for national action in transforming health care quality (IOM, 2003).
A critical issue in the prevention of frailty is the early detection of functional limitation and disability. Functional limitation involves difficulty in musculoskeletal performance (e.g., walking a half-mile, stooping–crouching–kneeling), whereas disability involves difficulty in performing specific activities of daily living (ADLs; e.g., bathing or showering, dressing) or instrumental ADLs (IADLs; e.g., shopping, money management) (Nagi, 1976, 1991; Verbrugge & Jette, 1994; Wolinsky & Miller, 2005). Ascertaining either can easily be done in face-to-face or telephone survey interviews. Subjects are simply asked if they have any difficulty in stooping–crouching–kneeling (a functional limitation example) because of a health or physical problem. Those responding "yes" are considered to be functionally limited in that task.
As Fried and colleagues have noted, however, this approach captures only the tip of the iceberg (Fried et al., 1996). Although it does provide an accurate reflection of functional limitation or disability, it is well known that there are many subjects who have performance decrements but who do not report having difficulty with physical function or ADLs or IADLs (Fried et al., 1996; Fried, Herdman, Kuhn, Rubin, & Turano, 1991; Gregory & Fried, 2003). Moreover, by focusing on what might be called clinical or manifest levels of difficulty, the opportunities for successful intervention are minimized, because attention is targeted on recovery rather than prevention (IOM, 2003). This is especially true when the affected task involves mobility, because mobility difficulties usually signal the onset of a pathway of progressive disablement (Fried et al., 1994).
Thus, Fried and colleagues have offered an alternative conceptualization of the ascertainment model (Fried et al., 1991, 1994, 1996, 2001; Fried, Bandeen-Roche, Chaves, & Johnson, 2000; Fried & Guralnik, 1997; Fried & Walston, 1998; Gregory & Fried, 2003). They introduced the notion of "subclinical (or preclinical)" status as an intermediary stage along a continuum that has difficulty and high functioning as its polar anchors. To ascertain whether the subject has subclinical status, two additional questions are asked (Fried et al., 1996). One focuses on modifying the method of task performance (e.g., "Because of health or physical problems, have you changed the method you use to bathe or shower?"), and the other focuses on modifying the frequency of task performance (e.g., "Because of health or physical problems, do you bathe or shower less often now?"). With these questions, one can identify subjects who report no difficulty with task performance but who have compensated for their ability decrement by modifying the method or frequency of task performance.
In a cross-sectional study of 231 men and women 60 years old or older, Fried and colleagues (1996) have shown the feasibility of ascertaining subclinical status and provided evidence of its construct validity. They found that the number of chronic conditions increased monotonically across the three levels of self-reported performance (i.e., high function, task modification, and difficulty). Similar results were observed for performance data on walking and stair-climbing speeds. In a longitudinal study of 436 women 70–80 years old from the Women's Health and Aging Study, subclinical status in two tasks—walking a half-mile or more and climbing up 10 steps—were significant predictors of incident difficulty in those tasks 18 months later (Fried et al., 2000). Moreover, even after adjustment for physical performance testing (i.e., walking and step-climbing speeds), the significant (p <.001) adjusted odds ratios associated with subclinical status were 3.67 and 3.84 for the onset of difficulty in walking and step climbing, respectively.
The results of Fried and colleagues (1996, 2000), however, have some limitations. The cross-sectional study involved a convenience sample recruited primarily from senior citizen centers, and the longitudinal study had a low (49.5%) participation rate. Few African Americans or other minorities were included in either study. Additionally, subclinical status was assessed for only two tasks (walking a half-mile and climbing steps) in the longitudinal study. Thus, further research is needed to generalize these promising results to (a) larger representative samples of community-dwelling men as well as women, (b) disadvantaged minority populations like African Americans (Department of Health and Human Services, 2000), (c) functional tasks other than walking a half-mile and climbing steps, and (d) disability items like ADLs and IADLs.
The purpose of this article is to do just that. Using data from the 998 subjects in the African American Health (AAH) Project, we first determine the prevalence of subclinical status at baseline for 12 tasks (5 physical functions, 3 ADLs, and 4 IADLs). Then, we use subclinical status at baseline to predict the onset of difficulty either 1 or 2 years later in five of these tasks (i.e., walking a half-mile, climbing steps, stooping–crouching–kneeling, lifting or carrying 10 lbs., and heavy housework). These five tasks include the two reported by Fried and colleagues (1996, 2000), and all five tasks (a) are especially sensitive to mobility decline, (b) signal the beginning of the progressive disablement process (Fried et al., 1994; Fried & Guralnik, 1997), (c) are most prevalent at baseline in terms of subclinical status, and (d) have good test–retest reliability (kappa statistics >.60; unpublished data).
|
METHOD |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
; Wolinsky, Miller, Andresen, Malmstrom, & Miller, 2004). AAH includes 998 African Americans who were born in 1936 through 1950 and who lived in either a poor, inner-city area previously studied by the investigative team (Miller et al., 1996) or the near northwest suburbs. To recruit equal numbers of subjects from both strata, unequal sampling proportions were set; therefore, in all analyses reported here, weighted data are used. Beyond age, inclusion criteria involved the ability and willingness to sign informed consent, self-reported Black or African American race, and Mini-Mental State Examination scores of 
16 (Molloy et al., 1996). Baseline evaluations averaged 2.5 hours and occurred in the subject's home between September 2000 and July 2001. The response rate was 76% (77% in the inner-city area and 75% in the suburban area).
Follow-Up Sample
Telephone follow-up interviews were conducted at 12 and 24 months. Of the 998 original subjects, 932 were successfully reinterviewed at 12 months, and 888 were successfully reinterviewed at 24 months. Because 18 subjects died between baseline and the first follow-up and an additional 13 died between the first and second follow-up, the response rates for surviving subjects were 95% and 92%, respectively. Follow-ups averaged 12 minutes and focused on monitoring ADLs, IADLs, lower body function, chronic conditions, self-rated health, and physical activity.
Subclinical Status
Subjects were asked a comprehensive series of standard functional performance, ADL, and IADL questions taken from the Second Longitudinal Study on Aging that was part of the 1994 National Health Interview Survey (for complete wording of the items, see Andresen, Malmstrom, Miller, Miller, & Wolinsky [2005]). If the subject expressed any difficulty (or was unable) to perform the function or task, s/he was considered to be limited in that functional performance, ADL, or IADL item. For the purposes of this study, subjects reporting no difficulty at baseline for a specific functional performance, ADL, or IADL item, but who subsequently reported difficulty at either the 12- or the 24-month follow-up interview, were considered to have developed incident difficulty. Because this study focuses on the cumulative incidence of difficulty, subjects who recovered from their 12-month incident difficulty by their 24-month interview were treated analytically as having incident difficulty.
The two follow-on subclinical status questions (method and frequency; Fried et al., 1991) were asked only if the subject reported having no difficulty in performing the targeted task or function. Subclinical status was assessed at baseline for 12 functional performance tasks, ADLs, or IADLs. The five functional performance tasks were walking a half-mile, climbing steps, stooping–crouching––kneeling, using the fingers to grasp or handle, and lifting or carrying something as heavy as 10 lbs. The three ADL tasks were bathing and showering, dressing, and getting in and out of bed and chairs. The four IADL tasks were meal preparation, light housework, heavy housework, and medication management. For each of these tasks or functions, the time anchor was "compared to when you were 40 years old." Thus, the minimum reference period approached 10 years (for those about to reach their 50th birthday), and the maximum reference period was 25 years (for those who had just reached their 65th birthday). Subjects reporting no difficulty with a task or function but who either changed the method of or decreased the frequency of performing it were defined as having subclinical status.
Covariates
Numerous risk factors have been identified for the onset of difficulty in functional tasks, ADLs, and IADLs (Stuck et al., 1999; Wolinsky, Stump, Callahan, & Johnson, 1996). Accordingly, we included four sets of covariates (i.e., demographic characteristics, socioeconomic status, comorbidity, and physical performance tests) to determine the independent effect of having subclinical status at baseline on developing difficulty either 1 or 2 years later. Demographic characteristics included age (coded in years), a binary marker for gender (1 = men, 0 = women), and a set of three dummy variables for marital status (divorced or separated, widowed, or single versus the reference category of married). Socioeconomic status indicators included education (coded in years), a set of two dummy variables for subjective income (having a comfortable income or not enough to get by versus the reference category of having just enough to get by), and a binary marker for sampling stratum (1 = inner city, 0 = suburb). Comorbidity was measured by a count of the number of nine chronic conditions self-reported by the subject (hypertension, diabetes mellitus, cancer other than a minor skin cancer, chronic airway obstruction, coronary artery disease, congestive heart failure, arthritis, stroke, and chronic kidney disease) and by a binary marker for whether the subject reported a clinically relevant level of depressive symptoms using the 11-item Center for Epidemiological Studies Depression Scale (Kohout, Berkman, Evans, & Cornoni-Huntley, 1993; Miller et al., 2004).
Physical performance was measured using three standard tests: tandem stand with eyes open (30-second maximum), five repeated chair stands, and preferred gait speed over a 3- or 4-meter course (Guralnik et al., 1994; Wolinsky, Miller, Andresen, Malmstrom, & Miller, 2005). Each test was demonstrated by the interviewer before the subject attempted to perform it. Electronic stopwatches accurate to 0.01 second were used to record timings. If the subject or the interviewer had concerns about the subject's ability to safely perform the test, then the test was not attempted. For the tandem ("tight rope") stand with eyes open, a set of two dummy variables are used (stand held for the full 30 seconds or stand not attempted versus the reference category of attempted but not held for the full 30 seconds). For the repeated chair stands, a set of two dummy variables are used (test not attempted or attempted and completed in the faster half of timings versus the reference category of attempted but completed in the slower half of timings). For preferred gait speed, a set of two dummy variables are used (test attempted but completed in the slower half of timings or attempted and completed in the faster half of timings versus the reference category of not attempted [most often owing to limited unobstructed space in the home]).
Analytic Model
Six multiple logistic regression models are used, with a separate model estimated for each of the five difficulty questions that include the corresponding subclinical status marker at baseline and the potential confounders. Each of these models is estimated only for those subjects who did not report difficulty at baseline for that specific task or function. The sixth model assesses the effect of having subclinical status on any of these five tasks or functions at baseline on developing difficulty in any of the five tasks or functions within 1–2 years and is estimated only for subjects who had no difficulties in any of the five tasks or functions at baseline. Model fit is assessed using the area under the receiver-operating characteristics curve (C statistic; Hanley & McNeil, 1982).
|
RESULTS |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
Prevalence of Difficulty and Subclinical Status
Table 1 shows the number and percentage of subjects at baseline who reported difficulty performing each of the 12 physical functions, ADLs, or IADLs. Also shown are the number and percentage of those subjects who did not report difficulty at baseline but who did report subclinical status for each of the functions or tasks. As shown, the prevalence of difficulty ranges from a low of 2.3% for medication management to a high of 39.8% for stooping, crouching, or kneeling. Similarly, the prevalence of subclinical status among those without difficulty ranges from a low of 10.5% for medication management to a high of 39.0% for stooping, crouching, or kneeling.
|
|
0.70 indicative of a good-fit) (Hanley & McNeil, 1982), these results show substantial predictability. Among the covariates, only three variables have independent effects in three or more of the six models. Being widowed (compared with being married) was associated with adjusted odds ratios for developing difficulty in walking a half-mile of 2.21 (p <.05), in climbing steps of 2.09 (p <.05), and in stooping–crouching–kneeling of 2.08 (p <.05). The adjusted odds ratio for each additional chronic disease for developing difficulty in walking a half-mile was 1.27 (p <.001), for climbing steps was 1.34 (p <.001), for lifting or carrying 10 lbs. was 1.40 (p <.001), and for doing heavy housework was 1.29 (p <.001). The adjusted odds ratio for having clinically relevant levels of depressive symptoms for developing difficulty in climbing steps was 2.54 (p <.001), in lifting or carrying 10 lbs. was 1.78 (p <.05), and for doing heavy housework was 1.80 (p <.05). No consistent independent effects are identified for any of the physical performance tests.
|
DISCUSSION |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
of identifying whether subjects who report no difficulty with a particular task or function have nonetheless modified either the method or the frequency of performing it. This approach is entirely consistent with traditional conceptualizations of the disablement process (Nagi, 1976, 1991) and reflects what Verbrugge and Jette (1994) describe as "activity accommodations." Moreover, subclinical status ascertainment yields important clinical information. In this study of late-middle-aged African Americans, the prevalence of subclinical status among those without difficulty was substantial, ranging from a low of 10.5% for medication management to a high of 39.0% for stooping, crouching, or kneeling. This distribution approximates well the prevalence of difficulty on these tasks and functions and is also quite consistent with the prevalence of subclinical status in older persons reported previously (Fried et al., 1996).
Second, these data confirm the clinically relevant association between subclinical status reported at baseline and the onset of difficulty 1–2 years later. Fried and colleagues (2000) reported statistically significant (p <.001) adjusted odds ratios for incident difficulty after 18 months in walking a half-mile of 3.67 for those with subclinical status in that task and for incident difficulty in step climbing of 3.84 for those with subclinical status in that task. The adjusted odds ratios obtained in these AAH subjects after 1–2 years were 1.68 (p <.05) for walking a half-mile and 4.46 (p <.001) for step climbing. Moreover, the adjusted odds ratios for incident difficulty obtained in these AAH subjects were 2.48 (p <.001) for stooping–crouching–kneeling, 2.51 (p <.001) for lifting or carrying 10 lbs., and 2.22 (p <.001) for heavy housework. Furthermore, the effects of the subclinical status markers in the AAH were the only risk factors to yield significant and clinically relevant effects on all of the incident difficulty measures. Thus, the ability of subclinical status to predict the onset of difficulty generalizes to men as well as women, to African Americans, to functional tasks other than those previously studied, and to at least one disability item (i.e., heavy housework).
In combination, these two findings have implications for national health policy. Simply put, they suggest that interventions to reduce frailty should focus on self-reported subclinical status as an early warning system. After all, the 2003 IOM report distinctly identifies the reduction of frailty associated with older age as a top national priority. If interventions continue to focus on difficulty rather than on subclinical status, the opportunities for successful intervention will be minimized, because attention will remain targeted on recovery rather than on prevention. This is especially true when the affected tasks involve mobility, because mobility difficulties usually signal the onset of a pathway of progressive disablement (Fried et al., 1994). Thus, given the relative ease of identifying subclinical status using self-reports and the ability of those self-reports to predict the onset of subsequent difficulty, national health policy should capitalize on the opportunity to move the focus on interventions further upstream in the disablement process to more effectively address prevention rather than recovery.
Not all of our findings, however, are consistent with those reported by Fried and colleagues (1996, 2000). In our study, the objective physical performance tests at baseline generally did not have significant associations with the onset of difficulty 1–2 years later. At first glance, this might appear somewhat surprising, inasmuch as very similar subclinical status, difficulty, and physical performance measures were used. The analyses of Fried and colleagues, however, were based on a response rate of <50% from a specially selected, high-functioning population of 70- to 80-year-old women, most of whom were White. In contrast, the AAH enjoyed a very high (76%) response rate from a randomly selected sample of community-dwelling African American men and women aged 49–65 years old. Alternatively, the absence of consistent effects of physical performance on the onset of difficulty in our study may, in part, reflect the fact that the subclinical status and difficulty measures use the same self-report methodology (with no objective assessment of their veracity), whereas the physical performance measures are objective assessments. In any event, further research is needed to sort out this important difference in the findings from these two studies. The need for that research, along with further confirmatory studies involving other age ranges, ethnic groups, and locations notwithstanding, the evidence supporting the importance of subclinical status as a concept for functional limitations and disability, and the self-reported ascertainment approach used here, is substantial.
|
Acknowledgments |
|---|
The opinions expressed here are those of the authors and do not necessarily reflect those of the National Institutes of Health or any of the academic or governmental institutions involved.
|
Footnotes |
|---|
Received for publication October 21, 2004. Accepted for publication December 16, 2004.
|
References |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. M. Clough-Gorr, T. Erpen, G. Gillmann, W. von Renteln-Kruse, S. Iliffe, J. C. Beck, and A. E. Stuck Multidimensional Geriatric Assessment: Back to the Future Preclinical Disability as a Risk Factor for Falls in Community-Dwelling Older Adults J. Gerontol. A Biol. Sci. Med. Sci., March 1, 2008; 63(3): 314 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. Porter Scales and Tales: Older Women's Difficulty With Daily Tasks J. Gerontol. B. Psychol. Sci. Soc. Sci., May 1, 2007; 62(3): S153 - S159. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kuh, R. Hardy, S. Butterworth, L. Okell, M. Richards, M. Wadsworth, C. Cooper, and A. A. Sayer Developmental Origins of Midlife Physical Performance: Evidence from a British Birth Cohort Am. J. Epidemiol., July 15, 2006; 164(2): 110 - 121. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. K. Miller, E. M. Andresen, T. K. Malmstrom, J. P. Miller, and F. D. Wolinsky Test-Retest Reliability of Subclinical Status for Functional Limitation and Disability J. Gerontol. B. Psychol. Sci. Soc. Sci., January 1, 2006; 61(1): S52 - S56. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
