Quantifying utilization and practice variation of electroconvulsive therapy.

The objective of this study was to quantify inpatient electroconvulsive therapy (ECT) utilization and its practice variation within the State of Louisiana using Medicare data for beneficiaries age 65 years and older. The Louisiana Medicare claims (MedPar) history and the Medicare beneficiary denominator files for fiscal years 1993 and 1994 were used for analysis. Statistical techniques used were: chi 2 to determine significance of the proposed null hypothesis, and the modified systemic component of variance (SCV) to determine the magnitude of variation between the individual parish utilization rates for ECT. The ECT utilization rate for the Louisiana Medicare population was found to be 2.38 per 10,000 person-years, falling well within the range of previous ECT utilization studies in the United States. The chi 2 value was 0.0003 when comparing parishes, indicating the presence of significant nonrandom variation. The SCVs of inpatient treatment for major depression and impatient ECT were 0.47 and 1.34, respectively. Inpatient ECT in this population demonstrates high geographic variability. Further research is required to determine and quantitate the factors responsible for the geographic variation in inpatient ECT utilization within the Louisiana Medicare population.