Discrepancies between self-reported and observed physical function in the elderly: the influence of response shift and other factors
Introduction
The World Health Organization defines disability as any restriction or lack of ability to perform an activity in the manner or within the range considered normal for a human being (Wood, 1980). It is recognized that a given level of ability, impairment or function may result in different degrees of disability depending on the expectations and social roles of the individual. Accurate measurement of disability has become an important social and medical issue. Dozens of scales have been developed to measure disability in performance of instrumental activities of daily living (McDowell and Newell, 1996). Despite criticisms, a number of these scales have demonstrated acceptable reliability and validity and have been widely adopted for use in population studies, randomized trials of treatments and follow-up of individual patients (McDowell and Newell, 1996).
The Health Assessment Questionnaire (HAQ), first developed for assessing instrumental activities of daily living in arthritis patients (Fries et al., 1980), is widely used in research settings, patient care, and general population surveys, including NHANES Harris et al., 1989, McDowell and Newell, 1996. A global Functional Disability Index score is created by averaging eight subscales (dressing and grooming, arising, eating, walking, hygiene, reaching, gripping and activities), derived from self-reported limitations in performing 20 activities. Although easy to administer, reliable and of proven validity in patient and general populations, the HAQ shares with similar scales a major drawback in that one's perception of disability is relative to his or her experience and expectations and is subject to response shift in reporting.
Response shift refers to a change in the meaning of one's self-evaluation of a target construct as a result of: a change in the respondent's internal standards of measurement (i.e. scale recalibration); a redefinition of the target construct (i.e. concept redefinition) or a change in the respondent's values (i.e. the importance of component domains constituting the target construct) (Schwartz and Sprangers, 1999). This may be in part derived from social considerations. For instance, given the same level of function, an older person might be less likely to report difficulty performing a given activity than a younger person, because gradual loss of function and reduced social expectations have led to reduced self-expectations. This would be characterized as a Beta (recalibration) response shift Armenakis, 1988, Breetvelt and Van Dam, 1991, Norman and Parker, 1996, Sprangers, 1996, in that the subject's internal calibration for difficulty of performance has changed over time. Other biases in reporting may be related to differing internal standards for `difficulty in performing' various activities according to gender, education, or psychosocial function. This study will focus on these biases and shifts in internal standards, which may compromise comparisons of disability between populations and tracking over the life-span.
Studies of rheumatic conditions and population surveys indicate that considerable discrepancies exist between self-reported limitations in function in independent activities of daily living and actual physical impairment (variably measured as observed function or disease severity) Ford et al., 1988, Rubenstein et al., 1984, Kelly-Hayes et al., 1992, Ramey et al., 1992, Daltroy et al., 1995, McDowell and Newell, 1996. A significant portion of variance in self-reported disability can be explained by demographic, cultural, social and psychological variables such as gender, familiarity with scale activities, depression and helplessness Blalock et al., 1988, Lorish et al., 1991, Hidding et al., 1994, McDowell and Newell, 1996. It is important to discern the determinants of self-reported disability, as disability may be addressed more effectively if it is understood to what extent it reflects true functional loss or impairment, versus psychosocial processes, including response shift. Both may need to be addressed. For instance, a true loss in some valued ability may loom out of proportion to a person with arthritis, who becomes depressed and inactive as a result. A caregiver may address the functional loss directly, with medicines and physical therapy, but the patient may also need to reassess the magnitude of the loss, perhaps with the help of a support group.
To explore the influence of social, psychological and medical factors on self-report of function, we compared a measure of function based on observed performance, the Physical Capacity Evaluation (PCE, Daltroy et al., 1995) with a self-reported measure of difficulty in performing instrumental activities of daily living, the Functional Disability Index score of the Health Assessment Questionnaire (HAQ, Fries et al., 1980).
Section snippets
Study design and population
The basic methods of this cross-sectional study have been described by Daltroy et al. (1995). In brief, a convenience sample of community-dwelling elderly persons 65 years of age and over was recruited from low-to-moderate-income housing sites for elderly people, rest homes, retirement communities and private homes. Subjects were balanced on age (65–74, 75–84, 85+), gender and geographic region (urban, suburban, rural). Subjects were assessed in their homes by trained interviewers (CP, HE)
Population characteristics
We recruited 289 subjects aged 65 to 97 years of age; 31% were 65–74 years, 36% 75–84 years and 33% 85–97 years. The sample profile is reported in Table 2.
Measures of functional ability
The mean HAQ score was 0.63 (S.D.=0.72, range 0 to 2.875). Psychometric assessment of the HAQ indicated that it was highly reliable (Cronbach's alpha=0.90, test–retest r=0.96) and that it represented a single domain (there was one dominant factor) (Daltroy et al., 1995). 35% of subjects reported no limitations in any activity (score of 0). We
Discussion
Consistent with other research cited earlier (e.g. McDowell and Newell, 1996), a standardized test of observed function was able to explain only about half of self-reported disability in an elderly population. In a multiple regression model, persons characterized by current joint pain or stiffness, use of prescription medications, urban dwelling, depression, female gender, lack of memory problems, arthritis and lack of exercise were more likely to report higher levels of disability, controlling
Acknowledgements
This research was supported in part by NIH Grant No. AR36308 and AG07459 and the Massachusetts Health Research Institute's Public Health Research Fellowship of the Medical Foundation.
References (36)
- et al.
Underreporting by cancer patients: the case of response shift
Social Science & Medicine
(1991) Social comparison as a mediator of response shift
Social Science & Medicine
(1999)- et al.
Response shift phenomenon: a challenge to longitudinal research in health and behavior
Social Science & Medicine
(1999) Response-shift bias: a challenge to the assessment of patients' quality of life in cancer clinical trials
Cancer Treatment Reviews
(1996)A review of research on the change typology
Research in Organizational Change and Development
(1988)- et al.
Self-evaluation processes and adjustment to rheumatoid arthritis
Arthritis and Rheumatism
(1988) - et al.
Psychological well-being among people with recently diagnosed rheumatoid arthritis: do self-perceptions of abilities make a difference?
Arthritis and Rheumatism
(1992) - et al.
Objectively measuring physical ability in elderly persons: the physical capacity evaluation
American Journal of Public Health
(1995) - et al.
Knee osteoarthritis and physical functioning: evidence from the NHANES I Epidemiologic Follow-up study
Journal of Rheumatology
(1991) - et al.
Health and function in the old and very old
Journal of the American Geriatric Society
(1988)
Measurement of patient outcome in arthritis
Arthritis and Rheumatism
Longitudinal study of physical ability in the oldest old
American Journal of Public Health
Adaptation Level Theory
Comparison between self-report measures and clinical observations of functional disability in ankylosing spondylitis, rheumatoid arthritis, and fibromyalgia
The Journal of Rheumatology
Gender, race and health: the structure of health status among older adults
The Gerontologist
Functional limitations and disability among elders in the Framingham Study
American Journal of Public Health
Disease and psychosocial factors related to physical functioning in rheumatoid arthritis
The Journal of Rheumatology
Cited by (161)
The impact of widowhood on wellbeing, health, and care use: A longitudinal analysis across Europe
2021, Economics and Human BiologyCitation Excerpt :Thus, it seems that this adaptation effect (Mctaggart‐Cowan et al., 2011) may vary depending on the type of health shock and chronic condition. Moreover, such adaptive effect on the perception of an individual's health does not have to keep a direct relationship with objective conditions of health improvements (Riis et al., 2005; Daltroy et al., 1999; Buick and Petrie, 2002; Brockmann and Klein, 2004). In general, our results are in line with the hedonic treadmill models for wellbeing, mental health, and deaths.
The Income-Health Gradient: Evidence From Self-Reported Health and Biomarkers in Understanding Society
2019, Panel Data Econometrics: Empirical ApplicationsUsing structural equation modeling to detect response shift in quality of life in patients with Alzheimer's disease
2019, International Psychogeriatrics