mortality files and NMFBS). There is, however, substantial geographic variation in the site of death (Figure 3-3) (Pritchard et al., 1998; Wennberg, 1998). For example, Oregon has experienced a dramatic increase in the proportion of people dying at home (probably due to a number of factors, including closing of hospital beds and a vigorous public debate about physician aid in dying) (Tolle et al., 1999).
Based on the 1993 NMFBS, cancer patients are less likely to be functionally impaired in the last year of life and experience a more precipitous functional decline in the last five months of life than those dying from other causes (Figure 3-4), as measured by difficulty with activities of daily living (ADL: bathing, dressing, eating, transferring from a bed or chair, and using the toilet). In that year, nearly half of these deaths occurred in an acute care hospital, and 36 percent, at home. Only 19.7 percent of those who died from cancer in 1993 used hospice care. The functional trajectory measured as the number of ADL impairments in the last five months of life was associated with dying at home and with hospice involvement (Teno et al., 2001).