Treatment of Major Depression Before and After Implementation of a Behavioral Health Carve-Out Plan

OBJECTIVE: The study examined utilization, payments, and quality indicators for treatment of major depressive disorder before and after the 1993 implementation of a behavioral health care carve-out plan for Massachusetts state employees who received medical coverage through indemnity plans or preferred provider organizations. METHODS: The sample of 2,259 enrollees with claims for treatment of major depressive disorder was drawn from the group of 39,541 persons continuously enrolled in preferred provider organizations or indemnity plans for fiscal years 1992 to 1995. A subsample of 243 users of inpatient services accounted for 352 admissions. Bivariate tests were used to compare utilization and quality indicators before and after implementation of the carve-out plan. Simple comparisons of current-year dollars were used. RESULTS: The proportion of enrollees with claims for treatment of major depressive disorder increased significantly under the carve-out plan. Inpatient utilization decreased substantially, mostly due to a significantly lower average length of stay (16 days before implementation of the carve-out plan and nine days after). Net inpatient payments fell 71 percent overall, 65 percent per admission, and 40 percent per day. The unadjusted proportion of discharged patients treated for major depressive disorder who were readmitted within 15 and 30 days did not change significantly. The unadjusted proportion of cases receiving follow-up within those time frames increased significantly. CONCLUSIONS: Implementation of a behavioral health carve-out plan may be accompanied by substantial reductions in inpatient utilization and payments for treatment of major depressive disorder. Descriptive findings suggest that such reductions may not have a detrimental impact on readmission and follow-up treatment rates within 30 days. However, this analysis did not control for patient characteristics, used short follow-up periods, and did not include some relevant outcome measures.