Applied cost-outcome analysis for mental health management.

The objective of this study is to demonstrate an evaluation procedure for use by program managers in the local public mental health system which combines the measure of resources consumed (cost) with the related outcome achieved (cost-outcome) that when combined with management’s values (cost-effectiveness) will routinely facilitate programs evaluation and management decision making. It was also postulated that cost-outcome data would specifically be useful to administrators as not only a measure of accountability but also a tool for assessing continued fiscal viability. Both cost and outcome data were readily available from the information system of Granite Community Mental Health Center in Salt Lake City, Utah. Annually, in concert with other mental health centers in Utah, Granite performed extensive cost-analysis to determine their cost per unit of service. The Center also had integrated ad Level of Functioning Scale (LOF) with their routine reporting of staff time. This scale, adapted from the Spitzer’s and Endicott’s Global Assessment Scale by a multi-disciplinary committee, used client functioning descriptors primarily reflecting the psychodynamic focus and training of the staff. Thus, their data system was designed so that each time a client was seen, the corresponding cost and an assessment of functioning was recorded. Cost and outcome data were gathered on 258 outpatient clients who were admitted and discharged during a six-month period and who had attended at least four therapy visits. These data were then complied into matrices reflecting the clientsâ€â€™ initial and ending LOF scores as well as the corresponding costs. The total study population was, then, subdivided into those receiving only individual therapy, those receiving only group or family therapy, and those receiving a combination of individual and group or family. This enabled a demonstration of the utility of the matrix technique by posing the null hypothesis that it is as cost-effective to treat clients individually as it is in groups. Since the data was client specific, it was also possible to beans of a unique identifier to further analyze the data referencing client demographics such as diagnosis and sex. The combined data from the matrices were statistically analyzed by means of Chi-Square and ANOVA but did not appear to support the hypothesis since while there was little difference demonstrated in outcome results there was definite difference in corresponding costs. Limitation and strengths of the study were discussed. Recommendations were made for those who supply and use these data. And, suggestions for further research were offered.