Trends in the use of intensive procedures at the end of life

Despite widely publicized changes in attitudes toward the care of the dying and rapid growth in hospice care and other services targeted toward the dying, a remarkably stable 30 percent of total Medicare expenditures is spent on the 5 percent of Medicare beneficiaries over the age of 65 who die each year (Lubitz and Riley 1993). Since spending on behalf of decedents has kept pace with that for survivors (at a pace steeply above the rate of inflation), it does not appear that services are being disproportionately withheld from, or applied to, the dying. It is unknown whether the increase in service intensity at the end of life-or indeed, an-Long survivors-is associated with meaningful clinical or social gains. According to numerous anecdotal reports, many Americans die in hospitals with aggressive medical intervention and insufficient pain control or psychological support (Nuland 1996). More comprehensive evidence comes from the longitudinal Study to Understand the Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT), which was conducted in five U.S. teaching hospitals. SUPPORT revealed that 38 percent of patients who died spent at least ten days in an intensive care unit and half of conscious patients experienced moderate to severe pain at least half the time (SUPPORT 1995). However, a comprehensive review of Medicare costs in the last year of life found that "only a small proportion of decedents had the kind of high expenses that would suggest aggressive, high technology medical services such as the use of ICUs or respirators" (Scitovsky 1994). Retrospective analyses of Medicare claims data from 1976 to 1995 show a steady decline in the proportion of Medicare costs in the last year of life spent for hospitalization and an increase in the proportion spent on home health care and hospice. These trends may be related to the introduction of reimbursement for those services or perhaps a more complex interplay of social, legislative, and economic forces. However, this shift has not slowed the rate of growth of total expenditures for end-of-life care (Garber, MaCurdy, and McClellan 1998). No published study to date has specifically evaluated trends in the use of aggressive technologies among dying Medicare beneficiaries. Drawing from a 20 percent sample of Medicare claims for decedents and a 5 percent sample of claims from survivors, we studied the use of forty-five intensive procedures among Medicare beneficiaries in 1985, 1990, and 1995. Each of the procedures was performed more frequently among decedents than survivors, and the relative growth in the use of all forty-five procedures combined outpaced the rate of their use among survivors. However, a subset of procedures dependent on technologies introduced between 1985 and 1995 diffused less rapidly among decedents than survivors. Although a decedent is more likely to undergo an intensive procedure in any year, most procedures are performed in Medicare survivors because they vastly outnumber decedents. Exceptions include some life-support technologies: two-thirds of all intubations and tracheostomies performed among Medicare beneficiaries in 1995 were done in decedents, as were over half of all feeding-tube placements and cardiac-balloon assist devices. Expenditures associated with hospitalizations in which these forty-five intensive procedures were performed grew faster than overall inpatient expenditures, contemporaneous with a shift of less intensive procedures from the inpatient to the outpatient setting. This trend in increasing intensity of treatment among both decedents and survivors contributes to expenditure growth that outpaces changes in demographics and disease incidence. Furthermore, the rapid growth in the use of intensive procedures among decedents may explain why increased use of less expensive services such as hospice and home health care for some decedents has done little to slow the growth rate of expenditures at the end of life.