The NLTCS is a nationally-representative longitudinal survey of the health and disability profile population aged 65 and over. The first NLTCS survey wave was conducted in 1982 and subsequent surveys were administered in 1984, 1989, 1994 and 1999. Each survey wave began with a screener that collected information on whether the respondent reported difficulty conducting the 6 ADLs and 8 IADLs without help (i.e. help from another person or special equipment), and whether any difficulty had lasted or was expected to last at least 3 months. The screener also collected demographic information on marital status, race and age. alta white teeth whitening
Respondents who reported difficulty with any ADLs or IADLs for at least 3 months on the screener were asked to complete a detailed survey. Disability status was determined by responses to questions about use of help and inability to conduct the ADLs and IADLs on the detailed survey. Sampling and weighting issues are described fully elsewhere (Manton, Corder et al. 1993; Manton, Corder et al. 1997; Manton, Stallard et al. 1997; Singer and Manton 1998; Manton and Gu 2001).
We obtained Medicare-linked data for all NLTCS participants, including data on date of death from the denominator files. We used inpatient claims for all analyses, because Medicare claims files for part B and other non-hospital services were incomplete prior to 1991. We also obtained data on zip code of residence at the most recent interview for all NLTCS survey respondents.
Basic data on disability among the elderly population is shown in Figure 1. For reasons that will become clear briefly, we report disability for the population that is aged 70 and older in each of three years: 1989, 1994, and 1999. The share of the elderly population that is disabled declined markedly in the decade of the 1990s. The cumulative decline was 6.3 percentage points, or 2.1 percent per year.
To examine the role of cardiovascular disease in explaining this reduction in disability, we form a population sample likely affected by the condition. We start by looking at the population aged 70 and older in the 1989, 1994, and 1999 surveys. For each of these cohorts, we group all hospitalizations over the preceding 5 years into one of 32 categories (the five year look back is the reason for the restriction to people over 70). These categories were designed to pick up relatively homogenous clinical conditions that would be predictive of disability. The set of 32 categories is shown in Table 1, along with the rate of disability for people hospitalized with each condition from the 1989 survey cohort. The relevant categories for cardiovascular disease are stroke, hypertension, ischemic heart disease, heart failure and arrhythmia, peripheral vascular disease, and other circulatory diseases.