Predictive modeling helps target adherence outreach programming

Adherence intervention using predictive modeling can deliver success large enough to affect Medicare Star ratings

This study paired Prime’s predictive modeling with its Medicare pharmacist call program to impact adherence to drugs for chronic conditions.

The study focused on drugs included in Star rating categories (oral diabetes) high cholesterol (statins) and high blood pressure (RAS). Star ratings are used by the Center for Medicare and Medicaid Services (CMS) to measure the quality of health plans.

The predictive model identified and ranked members likely to become nonadherent. Then Prime targeted those members with proactive pharmacist calls for adherence support. Members that received calls were followed and then assessed against a control group of members that did not receive calls.

The pharmacist calls improved members’ adherence in each category. This outreach to members, informed by Prime’s predictive model, was associated with significantly improved adherence and a Star Ratings increase of 0.9 – 1.1 percentage points. A single percentage point increase can bump a 3 to 4, or a 4 to 5 Star Rating.

A high Star Rating attracts people to a plan during open enrollment. In addition, health plans with a 4- or 5-Star Rating are eligible to receive significant performance bonuses. This can equate to hundreds of millions of dollars that health plans can use to improve benefits and/or reduce costs for their members.

AMCP recognized Prime’s research with a gold award. Prime is increasingly using predictive modeling to develop more effective and targeted clinical programming.