study risk factors for the onset of disability in a population or subpopulation that was free of disability and then examine the development of incident cases of disability and the risk factors that predicted its onset. The condition that I will discuss is mobility disability, defined here as an inability to walk a quarter mile and an inability to climb a set of stairs.
When my colleagues and I started our study, 72 percent of the cohort of 10,000 individuals was free of mobility problems at the baseline (Guralnik et al., 1993). Over a period that included four annual follow-ups, 53 percent of this group maintained mobility, 35 percent lost mobility, and a small percentage died without any evidence of mobility loss. At the baseline, we collected data on a number of chronic conditions that we hypothesized might predict mobility loss. We found the odds ratio for the loss of mobility to be in the range of about 1.2 to 1.5 for people with baseline reports of heart attack, stroke, diabetes, dyspnea, or exertional leg pain compared with the risk for people free of these conditions.
A considerable amount of cross-sectional and longitudinal research, including some studies documented in these appendixes, has investigated a variety of potential risk factors for disability. A range of physical and behavioral risk factors have been shown to be associated with disability. These factors include low levels of physical activity, smoking, high and low body mass index, weight loss, heavy and no alcohol consumption, high levels of medication use, poor self-rated health, and reduced social contacts. Andreas Stuck did a very nice job of summarizing this body of research in a 1999 paper (Stuck, 1999).
Among the chronic conditions that have been shown in epidemiologic studies to be associated with disability are heart disease and stroke, osteoarthritis, hip fracture, diabetes, peripheral artery disease, chronic obstructive pulmonary disease, cancer, visual impairment, depression, and cognitive impairment. This list of conditions is in no particular order, and people frequently ask what conditions are the primary causes of disability. This is actually a much more complicated question than it initially appears to be.
Some of the issues involved in assessments of the overall impact of a chronic condition on disability include the strength of the association between the condition (risk factor) and a particular disability; the prevalence of the risk factor; and—putting these together—something called an “attributable risk,” which has been assessed for some conditions. Also, it is important to consider the disability outcome of interest, as Alan Jette has done (Jette, 1994). For example, are you assessing the difficulty of performing a certain activity or, more narrowly, the human assistance required to perform the task?
Another important issue is population characteristics, such as age and gender. The main causes of disability may be different in men and women. For example, Ettinger and colleagues found arthritis-musculoskeletal dis-