How well do ambulatory care groups predict expenditures on mental health and substance abuse patients?

This paper evaluates the ability of Ambulatory Care Groups (ACGs) to prospectively predict mental health and substance abuse expenditures and total health care expenditures of persons enrolled in the New Hampshire Medicaid Program during fiscal years 1993 and 1994. A series of multi-part models is estimated separately for adults and children and a synthetic R-squared and the mean absolute predictive error are calculated. The results show that with the exception of predicting total expenditures for children, ACGs do not perform as well as simple models containing various demographic and prior mental health/substance abuse utilization measures.