This study tested incorporating psychological prediction models of treatment response within an economic model of a diabetes education program (DAFNE). Regression models were used to predict patient response to DAFNE based on psychological characteristics. These prediction models were then integrated into a simulation model to evaluate policies of only providing DAFNE to predicted responders or offering follow-up support to predicted non-responders. The results found that providing DAFNE only to predicted responders increased costs and decreased health benefits compared to current practice. Providing follow-up support to predicted non-responders dominated current practice, but was sensitive to assumptions about follow-up benefit. This methodology demonstrates how psychological factors can be better incorporated into economic models of behavioral interventions.