Depression and substance abuse treatment hospitalizations declined for Medicare beneficiaries over a decade
Research Activities, April 2009
The 1990s brought many changes in psychiatric care, including tighter admission criteria for hospital stays and a wealth of drug therapies available for conditions such as depression. To examine empirically how hospitalization rates might differ in two time periods, researchers from Rutgers' Institute for Health, Health Care Policy and Aging Research, compared Medicare data from 1992 and 2002 for patients over 65 who had psychiatric conditions and were insured by fee-for-service plans.
They found that, although epidemiological work by other investigators did not report a significant decrease in mental disorders over the decade in question, overall rates of psychiatric admissions were 28 percent lower. A reduction in stays for depression accounted for nearly half of that decrease, and stays for substance use disorders declined as well. However, inpatient stays for bipolar disorder and schizophrenia did not change during the study period, most likely because these patients exhibit severe symptoms, such as psychosis or lack of behavior control. The authors observe that this pattern of use fits with an "intensive care" model (i.e., a model that describes the treatment philosophy of psychiatrics) in which hospital use may be reserved for patients who are severely dysfunctional or may harm themselves.
Average hospital stay dropped from nearly 18 days in 1992 to just 12 in 2002, and costs per stay went from about $6,500 to $6,000. Though many policymakers worried that shorter stays would result in repeat hospitalizations, the authors did not find evidence that shortened stays resulted in readmissions later.
This study was funded in part by the Agency for Healthcare Research and Quality (HS16097).
See "Hospitalizations for psychiatric illness among community-dwelling elderly persons in 1992 and 2002," by Ayse Akincigil, Ph.D., Donald R. Hoover, Ph.D., James T. Walkup, Ph.D., and others in the September 2008 Psychiatric Services 59(9), 1046-1048.