| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the University of Texas Health Science Center at San Antonio, the Department of Medicine (H.P.H, M.B.G., C.D.M., M.J.L.), Division of Clinical Epidemiology (H.P.H.), Division of Geriatrics and Gerontology (M.B.G., M.J.L.), and Division of General Medicine (C.D.M.), the Aging Research and Education Center (S.L.), San Antonio, Texas; the Geriatric Research, Education, and Clinical Center (GRECC) The South Texas Veteran Health Care System, Audie L. Murphy Division (M.B.G., C.D.M., M.J.L.) and the Department of Veterans Affairs, Medical Research Service (M.B.G., C.D.M.), San Antonio, Texas.
Address reprint requests to: Helen P. Hazuda, PhD, The University of Texas Health Science Center at San Antonio, Department of Medicine, Division of Clinical Epidemiology, MC 7873, 7703 Floyd Curl Drive 78229-3900. Email: hazuda{at}uthscsa.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONOBJECTIVE 1: QUALITATIVE STUDYOBJECTIVE 2: QUANTITATIVE STUDYDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
METHOD: Semistructured interviews were conducted with 24 community-dwelling Mexican American elders (
65 years old) to elicit information about adaptations in performance of daily living tasks (eg, walking, dressing, and shopping) that may indicate presence of subclinical disability. This information was used to construct a quantitative self-report measure of subclinical disability administered to 207 older Mexican Americans. Item and factor analyses were performed to reduce the number of items and establish their underlying structure. Construct and discriminant validity of the reduced instrument was determined.
RESULTS: A framework comprised of nine categories of daily living tasks, three functional levels, and five adaptation types was generated from the qualitative data. The initial 133-item measure (named the ADAPT) was reduced to a 44-item scale with three subscales (physical, household, and social). ADAPT scores correlated significantly in the expected direction with standard functional status measures, but the shared variance was modest, indicating that the ADAPT captured substantial, unique variance. Mean ADAPT scores differed significantly and were monotonically lower across subgroups classified as independent, subclinically disabled, and disabled, respectively.
CONCLUSIONS: The ADAPT seems to have construct and discriminant validity as a measure of subclinical disability. Additional research is required to determine sensitivity to change and clinically significant cut points for varying risk of frank disability.
Key Words: activities of daily living, • adaptation, • disability, • elderly, • Mexican Americans, • methods.
Abbreviations: ADL = activities of daily living;; GDS = Geriatric Depression Scale;; HHAC = Hispanic Healthy Aging Center;; IADL = instrumental activities of daily living;; MMSE = Mini-Mental State Examination;; MOS SF-36 = Medical Outcomes Study Short Form-36;; OARS = Older Americans Resources and Services;; SALSA = San Antonio Longitudinal Study of Aging;; SES = socioeconomic status.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONOBJECTIVE 1: QUALITATIVE STUDYOBJECTIVE 2: QUANTITATIVE STUDYDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Researchers concerned with the epidemiology of disability among the US elderly have developed constructs that are closely aligned with the life-course perspective. In elaborating the pathway from chronic disease to disability, Verbrugge and Jette (4) identified two major components in the disablement process: 1) the main disease-to-disability pathway and 2) modifying factors outside that pathway that can slow, speed, prevent, or even reverse progression toward disability. The main disease-to-disability pathway is comprised of four stages. The first stage is pathology, defined as physiological and biochemical abnormalities that are detected and labeled as disease. Stage two is impairment, defined as significant structural abnormalities and dysfunctions in specific body systems such as the cardiovascular or musculoskeletal system. Stage three is functional limitation, defined as restrictions in basic physical and mental actions used in many different situations in daily life by an individual’s age-sex group. These basic actions involve the whole person but are not tied to a specific context. Examples include difficulty standing up from a seated position, being oriented in time and space, and feeling positive emotion. Stage four is disability, defined as difficulty in performing one’s expected social role activities in regular daily life due to a health or physical problem. These activities involve the whole person operating within a sociocultural and physical environment. Examples include difficulty eating, dressing, using the telephone, and managing finances.
Disease drives progression toward functional limitation via the mediating stage of impairment. Functional limitation, in turn, drives disability (5). Following person-environment fit models, disability is conceptualized as a gap between individual capability and demand of the task environment. The gap can be closed either by reducing the demand of the task environment (eg, making activity accommodations such as changing what we do, or how, how often, and how long we do it) or by increasing individual capability (eg, taking medications or changing our lifestyle and behavior). Both strategies represent adaptations to maintain function in the presence of biological change. Thus, individual differences in frank disability may result in large part from individual differences in adaptation to functional limitations.
Fried et al. (6, 7) have focused on identifying a stage of "preclinical disability," with primary emphasis on mobility-related disability. This stage is characterized by decreased functional ability, need for compensatory strategies, and increased risk of progressive functional decline and onset of disability (8). Essentially, preclinical disability is a state of "compensated disability." Although persons in this state have experienced real functional decrements, they perceive no difficulty in performing activities of daily living or mobility-related activities because modifications to task performance fully compensate for these decrements.
Two approaches have been used to assess preclinical disability. One is an objective approach in which persons with no self-reported ADL or mobility-related disability are given performance tests of physical function (eg, lower extremity physical performance battery) to determine presence of functional decrements. Prospective data based on this approach indicate that persons with poorer scores on performance tests are at 4.2 to 4.9 times greater risk of disability during 4 years of follow-up compared with those with the best scores (9). A second approach is to use self-reported task modifications among those reporting no mobility-related or ADL disability as an indicator of preclinical disability. Fried et al. have demonstrated that task modifications can be measured reliably by self-report (6) and that self-reported task modification predicts incident disability (ie, difficulty walking one-half mile or climbing up 10 stairs) at 18- and 35-month follow-ups (7).
The functional state of preclinical disability is significant because it may identify a subgroup of individuals who are functioning at a relatively high level but who are nonetheless at high risk of incident disability. If at some future point task demands exceed the reserve capacity required to invoke compensatory strategies, these individuals will become frankly disabled. Targeting early interventions to address underlying functional declines may eliminate the need for compensatory strategies and slow or prevent progression to frank disability. A substantial proportion of the elderly population may be at this stage in the disablement process. Fried et al. (7) found that 20% of high-functioning women 70 to 80 years old had preclinical mobility disability. Among similarly aged women in the lowest third of the functional spectrum, 46% to 51% had preclinical mobility disability (10).
This previous work on activity accommodations and preclinical disability has been carried out in predominantly white, non-Hispanic populations. In addition, work by Williamson and Fried (10) and Fried et al. (11) has been conducted primarily with women. Finally, although some attention was given to identifying diverse types of activity accommodations or task modifications, developing a systematic typology for classifying such strategies was not the primary focus. The purpose of our study was to extend this previous research by achieving two objectives: 1) to develop a comprehensive list of compensatory (or adaptive) strategies for preventing disability and organize these into a useful typology for future research and 2) to use this information to devise a quantitative, self-report measure of what we prefer to call "subclinical disability" (ie, a state of sustained independence in the presence of latent or manifest functional limitations). This work was originally undertaken as a pilot project within the Hispanic Healthy Aging Center, an Exploratory Center on Health Promotion in Older Minority Populations, funded by the National Institute on Aging. Thus, our initial goal was to develop a quantitative measure of subclinical disability that would be relevant for Mexican American elders. A long-term goal was to examine whether that measure, with or without modifications, was also relevant for white elders and could be used in cross-cultural research.
|
OBJECTIVE 1: QUALITATIVE STUDY |
|---|
|
TOPABSTRACTINTRODUCTIONOBJECTIVE 1: QUALITATIVE STUDYOBJECTIVE 2: QUANTITATIVE STUDYDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
65 years old). Subjects were purposively selected to include a broad spectrum of function and sociocultural status. Functional status was characterized using SALSA baseline measures collected within 6 months of the qualitative study. Physical functional limitations were measured with the Nagi index (12), which assesses degree of difficulty (none, a little, some, a lot, or just unable to do) in performing nine basic physical actions (pushing or pulling large objects; stooping, crouching, bending, or kneeling; lifting or carrying weights over 10 pounds; reaching or extending arms over shoulder level; writing, handling, or fingering small objects; standing in one place for long periods; getting up from a chair; standing on one foot; and picking up an object from the floor with one hand). Scores range from 9 to 45, with higher scores indicating more severe functional limitations. Disability (having difficulty) and dependence (needing help) in basic ADLs were measured with the modified Katz ADL scale (13), which assesses performance of seven activities required for personal maintenance (walking across a small room, bathing, personal grooming, dressing, eating, getting from a bed to a chair, and using the toilet). Dependence in IADLs was measured with the OARS IADL Scale (14), which assesses performance of seven activities required for household maintenance and integration with the community (using the telephone, driving or traveling alone, shopping, preparing meals, doing housework, taking medicine, and handling money). Individual items for the ADL and IADL scales were scored as 0 for no disability or dependence and 1 for any disability (a little, some, or a lot) or dependence ("yes" to needing help). Total scale scores ranged from 0 to 7. Health-related functioning was also characterized using three subscales from the MOS SF-36 (15). The Role Physical subscale assesses whether a person’s physical health caused them to cut down on the amount of time spent on work or other activities, accomplish less than desired, or limited the kind of work or other activities. Similarly, the Role Emotional subscale assesses whether any emotional problems led a person to cut down on the amount of time spent on work or other activities, accomplish less than desired, or do work or other activities less carefully than usual. The Social Functioning subscale assesses the extent (not at all, slightly, moderately, quite a bit, or extremely) to which either physical health or emotional problems interfered with one’s normal social activities with family, friends, neighbors, or groups. All three subscales are tied to behavior during the past 4 weeks and are standardized to a score of 0 to 100. Higher scores indicate better function.
We used neighborhood type (barrio, transitional, and suburb) as an indicator of level of SES (lower, middle, and upper) and assimilation to the broader American society (low, moderate, and high); we selected an equal number of participants from each neighborhood. Because we depended on the subjects’ ability to verbalize their experiences and worldviews, we also tried to select subjects who were reasonably articulate.
To construct the qualitative interview schedule, we expanded tasks encompassed by standardly measured ADL and IADL activities (13, 14) to include three levels of function. The essential level includes tasks required for personal maintenance, the intermediate level includes tasks required to maintain integration into the larger social community, and the discretionary level includes tasks that enrich the quality or scope of everyday living but may be discontinued without endangering independent functioning (16). The expanded range of daily living tasks was then reorganized into nine broad categories (Table 1).
|
Results
Average age of participants was 69 years. Only half (52%) had completed high school, and almost 40% did the interview totally or partially in Spanish. On average, participants were dependent in less than one ADL and IADL and had mild to moderate physical functional limitation. Average MOS scores were between the 25th and 50th percentile based on published population norms for persons 65 to 74 years old (15).
Content analysis of the transcripts identified five broad types of adaptations: physical (changes in frequency, completion time, or effort), cognitive (changes in planning), affective (restructuring of affect), social (changes in delegating and sharing), and environmental (altering environment and using devices). Table 2 provides examples of each type by category of daily living tasks.
|
Rank sum scores for the various types of adaptations were generally similar across neighborhoods. The single exception occurred for cognitive adaptations, which were reported almost twice as frequently by suburban participants than by those residing in the transitional neighborhood or barrio (Wilcoxon rank sum scores: 17.6, 10.8, and 8.3 for suburban, transitional, and barrio neighborhoods, respectively; adjusted p value = .04).
|
OBJECTIVE 2: QUANTITATIVE STUDY |
|---|
|
TOPABSTRACTINTRODUCTIONOBJECTIVE 1: QUALITATIVE STUDYOBJECTIVE 2: QUANTITATIVE STUDYDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Item and function analyses were performed on the pilot data to reduce the number of items and establish their underlying structure. Before factor analyses, items were eliminated based on mean score <2 or >4, SD < 0.95, coefficient of variation < 20%, "not applicable" response 25%, or skewed distribution for men or women (50% in the two lowest or highest response categories).
Principal components exploratory factor analyses with oblique rotation and oblique multiple groups component confirmatory factor analyses were performed to account for correlated factors. Items with communality <0.2 or loadings <0.40 in factor analyses were eliminated, as were those with item-total correlation <0.20 for the overall scale. Internal consistency of scale items was examined using Cronbach coefficient 
(17). Model goodness-of-fit for confirmatory factor analyses was evaluated by the root-mean-square residual and percent of variance of the model (18–20).
Construct validity of the measure was examined by testing two hypotheses. First, the measure will be significantly correlated, in the expected direction, with standard self-report measures of functional status. Second, shared variance with other functional status measures will be in the low to moderate range (<40%), indicating that the new measure captures a substantial proportion of unique variance.
Discriminant validity was established by examining whether scores on the measure differed significantly among subjects classified as independent, subclinically disabled, and disabled based on self-report measures. Independence was defined as the absence of functional limitation (physical, affective, and cognitive), disability, and dependence. Subclinical disability was defined as the presence of any functional limitation (physical, affective, or cognitive) in the absence of disability or dependence. Disability was defined as the presence of any disability or dependence. Physical functional limitation, ADL disability and dependence, and IADL dependence were assessed using the measures and cut points described previously. Cognitive functional limitation was measured with the Folstein MMSE (scale range = 0–30; cut point for limitation <24) (21). Affective functional limitation was assessed with the GDS (score range = 0–30, cut point for presence of limitation 11) (22).
Results
The mean age of participants was 68 years; 57% of the sample was female. Slightly fewer than half (46%) had completed high school, and about one-fourth (24%) did the interview in Spanish. Half of the participants (52%) lived in the barrio; the other half was equally distributed between the transitional neighborhood (23%) and suburbs (25%). Participants represented a broad range of function (Table 3). However, most participants fell in the upper quartile of health-related functioning on MOS subscales (15) and in the low to moderate range of functional limitation and disability. The distribution of participants across Nagi categories of physical functional limitation were none, 22.3% (45 of 202); low to moderate, 68.3% (138 of 202); and severe, 9.4% (19 of 202). This closely approximated the distribution among Mexican Americans in the total SALSA sample.
|
Item analysis resulted in a 44% reduction in the number of items (from 133 to 75); most were eliminated on the basis of multiple criteria. A five-factor model corresponding to the types of adaptive strategies was tested initially using confirmatory factor analysis, but the variance explained by the model was no better than the minimum variance explained by a random model. Exploratory factor analyses suggested that a three-factor model (physical, household, and social) was more consistent with the data and that additional items should be eliminated. A confirmatory factor analysis was performed to evaluate the three-factor model in the remaining pool of 44 items. The items encompassed all task categories, levels of function, and adaptive strategies. As shown in Table 4, 25 items had their highest correlations (factor loadings) with the physical factor, nine with the household factor, and 10 with the social factor. Goodness-of-fit statistics indicated that the model provided a good fit to the data. The model explained 45% of the variance, closely approximating the maximum possible, and the root-mean-square residual was 0.055.
|
|
|
As shown in Table 5, coefficients reflected high internal consistency and met the standard for individual-level predictions ( 0.8). Deletion of any subscale lowered the coefficient for the total scale. Mean and median scores were similar and near the midpoint of the scale range. Floor and ceiling effects are the percentage of respondents achieving the lowest and highest possible scale scores, respectively (15). Floor effects were 0% for the ADAPT and subscales. Ceiling effects were 0% for the total scale and household subscale and 0.5% and 1%, respectively, for the physical and social subscales.
|
|
.0001). The prevalence of functional limitations among subclinically disabled participants was 93% for physical limitations, 75% for cognitive limitations, and 23% for affective limitations. Correlations between the ADAPT and severity of functional limitations were all in the expected direction (ie, lower ADAPT scores were associated with greater levels of functional limitation). Further analysis indicated that these associations all had significant linear trends (data not shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONOBJECTIVE 1: QUALITATIVE STUDYOBJECTIVE 2: QUANTITATIVE STUDYDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
The work presented here is innovative in that it expands the construct of subclinical disability in several ways. First, we posited that adaptations suggestive of subclinical disability occur at all levels of function (essential, or ADL; intermediate, or IADL; and discretionary, or AADL) (16) and that adaptations at the discretionary level may be the earliest indicators of subclinical disability. Consistent with these hypotheses, the final reduced ADAPT incorporated tasks at all three functional levels. Furthermore, only 18% of the items measured task adaptations at the essential level, whereas 43% measured intermediate level tasks and 39% measured discretionary level tasks. Tasks measured by the social subscale were primarily at the discretionary level (7 of 10 items). To our knowledge, the ADAPT represents the first attempt to examine essential, intermediate, and discretionary levels of task performance within a single instrument.
Second, rather than focusing on adaptations attributed exclusively to health-related causes, we took a broad approach to task adaptation without reference to specific causal factors. Our experience in the qualitative study suggested that any reference to age- or health-related changes was counterproductive and that identifying "the reason" for specific adaptive behaviors was difficult for our participants. This suggests that older persons may make many adaptations without tying them consciously to health- or age-related changes. Although this is unlikely to occur when adaptations are made in response to catastrophic events, such as injuries or stroke, many other health- or age-related adaptations may occur so gradually that their causes may be imperceptible to the persons making them. An advantage of not grounding our approach to changes in health is that it may lead to broader insights into the range and utility of adaptive strategies used to maintain independence in later life. An apparent disadvantage is that we cannot distinguish adaptations made consciously or unconsciously in response to functional limitations from adaptations made for other reasons. If adaptation is driven primarily by age-related declines, as posited by life-course theories of aging, this distinction may be unimportant.
Third, the ADAPT successfully distinguished persons defined as subclinically disabled, on the basis of objective external criteria, from those defined as either independent or frankly disabled. A closer examination of the relationship between ADAPT scores and measures of physical, cognitive, and affective functional limitations among persons with subclinical disability suggested that individuals increase adaptive strategies that preserve independence in response to increases in functional limitation. However, mean ADAPT scores among persons classified as independent fell below 100, indicating that although they reported no functional limitations, they had also made adaptations in task performance. Thus, it seems that the construct of subclinical disability may comprise a broad spectrum of adaptations related to functional limitation, including adaptations to prevent latent functional limitations from becoming manifest and adaptations in the presence of manifest functional limitations to prevent disability. From this perspective, subclinical disability represents a continuum of adaptations in the presence of underlying physiologic declines with some critical threshold marking passage to frank disability. This conceptualization of subclinical disability in relation to stages of the disablement process model is illustrated in Figure 2.
|
Fourth, the instrument was developed using qualitative techniques to assure its relevance to older Mexican Americans from diverse sociocultural backgrounds. This qualitative approach allowed us to construct a typology of adaptive strategies that included two types of adaptations (ie, cognitive and affective) previously unrecognized in the literature. We also found suggestive evidence that some types of adaptive strategies may vary systematically by sociocultural status. More specifically, suburban Mexican Americans were significantly more likely than those in transitional and barrio neighborhoods to use cognitive strategies. This neighborhood difference may reflect varying preferences or intra-individual resources acquired over the life course and/or differences in extra-individual resources across sociocultural contexts.
It should be noted that task adaptations identified by older Mexican Americans in our sample seem equally relevant for older white subjects and perhaps other ethnic groups as well. One might argue that this occurred because we purposefully selected subjects who were reasonably able to articulate their experiences. Given the representation of subjects across neighborhoods, their range of function, and the number of Spanish-speaking and bilingual subjects in the sample, however, we believe that this is unlikely. Furthermore, there is no reason to feel that an articulate subject’s worldview is substantially different from that of inarticulate subject within the same cultural subgroup. Given the increasingly multiethnic composition of American society and the need for functional status instruments with broad applicability, it will be important to determine whether the ADAPT, in its current or a modified form, may be suitable for use with other ethnic groups as well.
Finally, in addition to providing a marker for subclinical disability, adaptive strategies may themselves be useful interventions to help individuals with uncompensated functional limitations or disability regain independent functioning. Further examination of adaptive strategies within the context of the disablement process may help identify those that are most useful for particular daily living tasks. It would also be useful to link adaptive strategies to specific changes in environmental demands and individual capacity.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONOBJECTIVE 1: QUALITATIVE STUDYOBJECTIVE 2: QUANTITATIVE STUDYDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONOBJECTIVE 1: QUALITATIVE STUDYOBJECTIVE 2: QUANTITATIVE STUDYDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication September 26, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONOBJECTIVE 1: QUALITATIVE STUDYOBJECTIVE 2: QUANTITATIVE STUDYDISCUSSIONAPPENDIXACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  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]
|
|
|
|
 |
|
 D. K. Miller, F. D. Wolinsky, T. K. Malmstrom, E. M. Andresen, and J. P. Miller Inner City, Middle-Aged African Americans Have Excess Frank and Subclinical Disability J. Gerontol. A Biol. Sci. Med. Sci., February 1, 2005; 60(2): 207 - 212. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. A. McHorney Ten Recommendations for Advancing Patient-Centered Outcomes Measurement for Older Persons Ann Intern Med, September 2, 2003; 139(5_Part_2): 403 - 409. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. G. Ory and M. Chesney Aging and the Life-Course: Advancing Psychosomatic Medicine Research Psychosom Med, May 1, 2002; 64(3): 367 - 369. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
