OBJECTIVES: The field of pharmacoeconomics has evolved rapidly over the past two decades and the methodological rigor and sophistication found in contemporary pharmacoeconomic analyses are remarkable. However, many useful analytic methods have yet to be adopted or fully considered, techniques that may be particularly well suited for our field. One promising method, long used in economics and physics, is dynamic optimization, which can identify optimal treatment pathways for patients suffering from diseases with progressive treatment options, both types and doses, where the value of aggressive treatment is of interest. Using hyperlipidemia as a case study, we propose multiple mathematical and statistical methods to identify the optimal (both cost-minimizing and net benefit-maximizing) hyperlipidemia treatment trajectories that consider the options of diet and exercise and multiple-dose statin therapy to bring a patient, or population of patients, to a pre-determined cholesterol goal (e.g. their NCEP goal). METHODS: Discrete-time mathematical programming models, both deterministic and stochastic, are developed to optimize over the combinatorial set of clinically feasible hyperlipidemia treatment strategies. Contemporary cardiovascular risk equations, costs, and side effect risk data are used to populate the models and programs. RESULTS: Initial results suggest that optimal (economic) hyperlipidemia treatment strategy varies based upon patient profile and risk for CVD, but in the majority of cases more aggressive statin therapy results in higher expected net benefits and lower expected costs than current treatment approaches; specifically, gradual titration regimens. Analyses considering longer term horizons (e.g., > 1 year) tend to increases net benefit differences more than cost differences. CONCLUSION: The preliminary evidence in our research suggests that considerable cost savings (and net benefits gains) are possible if hyperlipidemia patients are treated aggressively to reduce their cardiovascular risk as quickly as possible. These methods are generalizable to other disease areas and are useful in economic evaluation of treatment patterns. 14497 modified by 192.77.198.12 on 1-7-2007-->