However, the contracting CMHCs submitted encounter forms which were used in reconciling outpatient reimbursements under the capitation rates during the first 2 1/2 years of the UPMHP and in documenting inpatient rates.
The financial incentives for CMHCs in the UPMHP suggest that under-reporting of outpatient utilization is not an issue for the first 2 1/2 years of the program.
Processing of data from the years after implementation of the UPMHP was more complicated because these data were located in two different places: the standard Medicaid claims files for the non-capitated sites and "shadow claims" files for the capitated sites.
The first section presents descriptive data on trends in expenditure and utilization rates for 36 months prior to the UPMHP and the first 42 months of the UPMHP.
To test the sensitivity of the model to the specification of time trends, we also examined an alternative specification that added two time variables that allow the effect of the UPMHP to change as the program matures during the partial and full capitation periods.
1) Equation estimated using non-contractng sites; used to predict values for UPMHP sites for detrending.
For all models, we examine the detrended expenditures and utilization rates for the UPMHP sites.
Figure 2 reveals an apparent effect of the UPMHP on inpatient expenditures for mental health care.
In both capitated and non-capitated sites, expenditures on outpatient mental health care were increasing before implementation of the UPMHP This trend continued after implementation of the UPMHP, with the difference between capitated and non-capitated sites increasing by a small amount.
Table 2 summarizes the findings related to the effect of participation in the UPMHP on total mental health expenditures per beneficiary per month.
While these findings suggest no UPMHP effect on overall mental health expenditures, had the capitated CMHCs been paid on a fee-for-service basis, there do appear to have been significant UPMHP effects on acute inpatient expenditures.
This UPMHP effect was less in the full capitation period, and was not statistically significant at the .