Evaluation of Clinical and Financial Outcomes in a Medication Therapy Management Program in a Publicly Insured Adult Population

The current literature on MTM suggests that many patients receiving MTM counseling see improved health outcomes that include:  1) better medication adherence, 2) reduced exposure to potential drug-drug or drug-disease interactions, 3) reduced instances of over or under medication, and 3) better control of their conditions as reflected by fewer inpatient hospitalizations and visits to the ED.  Payers have reportedly observed lower medical and prescription drug reimbursements for populations that receive an MTM intervention.  The majority of the published literature evaluating MTM programs was conducted on populations of working age adults covered by private insurance.

 The purpose of this project was to evaluate the effectiveness of an MTM intervention in minimizing expenditures and reducing inpatient hospital admissions and outpatient emergency department visits in a disabled working-age (21-64 years) adult population eligible for Medicaid via a Section 1115 Waiver from the Centers for Medicare & Medicaid Services.  A MTM program was chosen as the required demonstration project for this waiver.

 We used two fundamental approaches to construct comparison groups and to adjust statistically for remaining differences between the intervention and comparison group.   The pool for CG members included anyone from the MEG3 population that had not received the MTM intervention. 

Method 1 relied on careful editing of the pool of MTM-P and potential CG members to remove any observed enrolled days which should have been excluded.  The MTM-P and CG constructed with Method 1 was then analyzed using a difference-in-difference (DiD) models to control for baseline differences in the four outcomes of interest.  Although the sign on the DiD coefficient was often negative, indicating improved outcomes in the MTM-P intervention group, across of all of the models, none of the coefficients was statistically significant. 

Method 2 used propensity score matching to explicitly choose a distinct CG for each outcome measure.  The CG was selected based on age, sex, length of enrollment, and ACG morbidity score and an average treatment effect was computed.  The results were consistent with Method 1, there was no statistically significant difference between the MTM-P and CG across any of the four outcomes.

Positive results for MTM interventions have been widely published MTM evaluative literature is still in its infancy.  More research is needed to determine its effectiveness in a Medicaid population.