Causes of increasing mortality in a nursing home population

OBJECTIVE: To examine the roles played by changes in case-mix, quality of care, and aggressiveness of care in explaining the 42% increase in mortality of the Medicaid nursing home population of Hennepin County, Minnesota, between 1984 and 1988. DESIGN: Retrospective chart review. SETTING: All nursing homes in Hennepin County, MN, that care for Medicaid patients. PATIENTS: A random sample of 1605 Medicaid nursing home residents from 1984 and 1988 stratified by year and by whether the resident died in that year. Sampling was disproportionate to allow approximately 400 individuals per stratum. A total of 1405 charts (87%) were reviewed; the remainder were either lost or destroyed. MAIN OUTCOME MEASURES: Measures included case-mix (Charlson index, functional status, implicit reviewer assigned severity [range 1-4]), aggressiveness of care (orders limiting care), quality of care (process of care for tracer conditions [range 1-5], falls), and resident death. RESULTS: Implicitly rated severity of illness worsened between 1984 and 1988 (2.77 vs 2.91; P = .009), but other measures of case-mix were unchanged. A greater percentage of residents had a DNR order in 1988 (12% in 1984 vs 37% in 1988; P < .001), and more received less aggressive care (31% vs 40%; P = .006). Overall process of care improved between 1984 and 1988 (2.88 vs 3.01; P < .05). With adjustment of the mortality rates and with logistic regression controlling for age and gender, changes in quality of care alone accounted for less than 5% of the mortality rate change between 1984 and 1988, case-mix alone accounted for 49%, and aggressiveness of care alone accounted for nearly 100%. CONCLUSIONS: The nursing home population became sicker between 1984 and 1988, but process of care improved. These changes had a modest effect on the mortality rate. The increase in less aggressive care between 1984 and 1988 accounts for nearly all of the increase in mortality.