OBJECTIVES: The objective of this study was to evaluate the relationship between a physician-based intervention and the reduction in poly-pharmacy for a segment of a national managed care Medicare Part D cohort. METHODS: Providers of Medicare Part D members in 2009 were sent a letter if one or more of their patients met the criteria for both medication therapy management (MTM) and poly-pharmacy. The providers were asked to review the member's prescription profile and make appropriate changes based solely on the provider's clinical judgement. Two follow-up quarters were defined for each member in 2009, and two baseline quarters were defined in 2008. From prescription claims, fields were collected allowing the creation of change variables.  Changes in claim count, drug (GPI) count, and medication cost were evaluated using feasible generalized least squares and a linear probability model (change in claim and drug count).  The intervention cohort was compared to a control cohort, consisting of Medicare Part D members who met poly-pharmacy criteria but not MTM program criteria. RESULTS: The analysis revealed a significantly greater reduction of 3.72 claims (second baseline quarter 2008 versus second follow-up quarter 2009. p <0.001) for the intervention. There was a significantly greater reduction of 1.67 drugs (GPI8) for the intervention (p <0.001). Medication costs for the intervention were decreased $204 (ingredient cost, $149 in amount paid for medications, p <0.001 for both medication costs). The linear probability model showed the intervention was associated with a 4.4%, 6.5%, and 12.9% significantly greater probability of a three to five, six to eight, and a greater than eight, respectively, claim count reduction (p <0.001). The intervention was also associated with a significantly greater probability of drug count reduction (p <0.001). CONCLUSIONS: The intervention was associated with a statistically significant marginal reduction in claim count, drug count, and medication costs compared to a control Medicare Part D cohort.