To prepare the U.S. healthcare workforce to meet forthcoming challenges of population aging, it is essential to understand linkages between ADL limitations and other health concepts. In this nationally representative sample of community-dwelling people aged 70 and older, global perceived health status, types of physical and mental conditions, and perceptions of home environmental barriers were all associated with ADL stage in complex ways, consistent with the HEI13
framework. Concordance with other studies and clinical plausibility of the relationships support the internal validity of the stages.20,24,30–34
An enhanced understanding of associations between ADL stage and other health-related concepts has the potential to improve the care of populations and persons with disabilities in several ways. First, knowledge of patterns of association across explicitly defined ADL stages can help direct research toward identifying the contributing factors behind the associations. Second, policy-makers and clinicians may be better able to plan and project disability-related needs for persons with chronic health conditions as the population ages.
Consistent with a review of 24 studies demonstrating that the combination of chronic conditions and the overall severity of conditions were associated with poorer functioning or quality of life,24
the findings of the current study suggest that knowledge of global perceived health, along with the type(s) of mental and physical condition(s), is essential to understanding linkages between illness and disability. The association between perceptions of poorer health and stage was strongly ordered, suggesting that self-reported global health assessments are closely related to functional stage.34
Although less is known about associations with function, global perceived health status is known to predict mortality with greater precision than even physician ratings, suggesting that people may have important self-knowledge that physicians do not.35,36
Lack of a one-to-one correspondence between perceptions of health and ADL stage highlights that severe illness is not always associated with severe disability. A small proportion of persons (4.4%) whose health was rated as very good or excellent were at the most disabled stage. Conversely, 18.3% of those whose health was rated as fair or poor had no ADL difficulty, highlighting the importance of recognizing disability and illness as related but distinct.
All types of mental and physical health conditions studied were independently associated with stage, after removing the effects of demographic differences and other conditions, but the magnitude of effects and patterns of effects across the stages differed markedly and in expected ways. Stroke and need for proxy use because of dementia were the most strongly associated conditions. The likelihood of being at each more-limited stage increased from Stage 0 in a dramatically ordered fashion. The proportions of elderly people with stroke and dementia can be expected to increase sharply with stage, becoming highest at the complete ADL difficulty Stage IV. Diabetes mellitus showed a similar pattern but was less ordered. Osteoporosis, card-iopulmonary disorders, and arthritis were moderately associated with stage. People with these conditions have much higher probabilities of being at Stages I, II, or III than Stage 0 but not of being at Stage IV. People with major mental illness showed a similar pattern for Stages I and II compared with Stage 0. Hypertension and cancer were not strongly associated with stage. Although stroke and dementia are often catastrophically disabling, the findings highlighted that most chronic conditions can be expected to manifest in partial rather than complete functional loss (cluster at intermediate ADL stages).
Results indicate that perception of unmet needs for home accessibility features can be expected in a small percentage of elderly persons living in the community even before onset of ADL limitation at Stage 0, to increase progressively in Stage I and II, peak in Stage III, and then decline at Stage IV. Reasons for relative decline at Stage IV are uncertain, but it may be because people are so dependent they ignore home accessibility concerns. Alternatively, at the stage of total dependency, an individual's abilities to interact with the environment becomes so limited that further accessibility features may not seem helpful. Recent longitudinal evidence supports the assumption that environmental modifications are most effective when supplied early in the aging process.14
Applying propensity score methods, one study recently demonstrated that older people with accessibility features in their homes were less likely to decline in ADL limitation. It further noted that the benefit of having residential modifications may differ according to subgroup, suggesting that people may pass a functional threshold when these modifications are no longer beneficial.20
The prevalence of disability in community-dwelling people aged 70 and older has decreased over the past decades in the United States,37
although the decline occurred primarily in the more-complex instrumental ADLs than in basic ADLs.38
Rates of ADL limitation remained remarkably constant from 1983 to 2005. Assuming that these trends continue and distributions of persons according to stage remain constant, based on current census data,39
we estimate that the numbers of people living in the community at the heaviest care Stages III and IV nationwide will jump from 5.5 million in 2010 to approximately 7.1 million by 2020, representing an increasing burden on families and society.
Impending morbidity of the aging baby boomer generation could overwhelm the healthcare system unless high-quality, cost-effective chronic care strategies are established.40
Reductions in the proportion of people with disabilities living in congregate care facilities reflect proposed Healthy People 2020 objectives.41
The optimal matching of resources to personal and environmental needs will become increasingly difficult to achieve with the declining proportion of younger persons in the population that are potentially available to provide care for older adults.4
Stages capture the severity and type(s) of difficulties people experience with self-care, thus indicating the nature of support that people need to remain living in their homes. The sequence of increasing difficulties from less- to more-severe stages of limitation follow the well-established ADL hierarchy, defining thresholds of retained functional ability that are clinically significant.7,8,29
People at Stage I, for example, are by definition able to eat and toilet without difficulty but can be expected to have some problems with the more-complex ADLs such as dressing. Although people at Stage I are still able to perform all of the ADLs but with difficulty, their status is in sharp contrast to those at Stage IV, at which people are unable to perform any of the ADLs and require total care. ADL stages as aggregated measures of functioning might, in future efforts, be shown to be appropriate for population surveillance or screening in clinical settings where periodic self-reports (or proxy reports) of functioning are obtained. A reduction in ADL status to a lower stage could trigger system responses and a more-detailed needs assessment. It is hoped that, by simultaneously expressing the types and severity of ADL limitations, ADL stages will help clinicians, healthcare managers, and policymakers address functional needs with greater specificity than typical score- or count-based indicators that obscure the nature of disability.
This study has several limitations. First, self-reported status and functioning could be biased. The major National Institutes of Health Patient Reported Outcomes Measurement Information System initiative which is developing standard self-report batteries of functional measures for trials, however, supports a general movement toward the reporting of self-reported functioning.42
Self- or close proxy–reported functioning is particularly valuable when the objective is to understand the challenges that people living in the community face. Second, approximately one-third of responses in the LSOA II were by proxy. Proxy informants may not always reflect self-perceptions and could generate systematic bias in population prevalence estimates,43
although correlations between self-administered and proxy-reported ADL questions have been shown to be high and were demonstrated to be from moderate to good even in people with strokes.44
When self-reported and caregiver ratings were compared with physical therapist ratings, self-reported function more closely correlated with therapist ratings than that caregiver-reported function, although all three (self, caregiver, and therapist) ratings of function were highly correlated.45
In contrast, self- and caregiver-reported psychosocial status were not highly correlated.46
Consequently, ratings of psychological well-being were not included. Previous analyses of NHIS-D data found large differences in prevalence of ADL limitations in people who report for themselves and those for whom proxies respond. Consequently, it was believed that inclusion of proxy reports would result in less-biased prevalence estimations than excluding them, so proxy information was included to avoid eliminating the most-vulnerable segments of the population.47
Third, the age of the LSOA II data is troubling, but the study would have been impossible without these data because no other source as rich in disability-related information could be found. Even though data were used from the mid-1990s, the findings are still valid because ADL limitation prevalence in the elderly population has remained remarkably stable for more than 2 decades.38
Although environmental details have not been re-collected in a linkable format, the ongoing MCBS includes the same ADL questions.48
The prevalence estimates from the LSOA II were concordant with later estimates from the MCBS. Although it was estimated that 28.4% of the U.S. population aged 70 and older had ADL limitations in 1995, the MCBS estimate was 26% of those aged 65 and older a decade later.1
Fourth, a cross-sectional design of association cannot be considered causal. It is impossible to know which conditions most explained the older people's distribution across the various stages. Future efforts will need to address longitudinal associations between stage and a variety of outcomes. Finally, prevalence estimates according to stage can be expected to generalize only to community-dwelling older persons. People in nursing homes were not included in the LSOA II.
This study presents three important findings. First, individuals’ intrinsic mental and physical conditions, along with external environmental factors, recognized in the HEI theoretical model as co-determinants of disability, were correlated with SPs’ stages of ADL limitation. Second, when addressing linkages between chronic illness and disability, it appears essential to measure global illness severity, as well as the specific type(s) of mental and physical condition(s) the individual is experiencing. Third, there are striking population-level associations between ADL stage and perceptions of unmet need for home accessibility features, with vulnerability appearing greatest at intermediate stages of severity. It is hoped that future work will establish ways of using staging to help identify integrative healthcare strategies that incorporate the most effective and achievable means of improving health status and reducing the incidence or progression of preventable disabilities. Staging might prove helpful in projecting the future needs and challenges associated with disabilities, encouraging much-needed shifts in the national health agenda toward life enrichment.