Case Study

Identifying At-Risk Members
by administrator



A Prescription Drug Plan commissioned Deft Research to help identify and classify members who were the most at risk for dis-enrolling.


Deft developed an at-risk model using multivariate statistical analysis on the client’s membership data. The model established which variables were correlated with current and dis-enrolled members. Variables that were the most correlated with dis-enrolled members were considered leading indicators. The logic was that if leading indicators characterize the experience of dis-enrolled members, then the leading indicators that were present in a current member’s experience would predict if that member is at-risk for dis-enrolling in the future.

Deft assisted the client’s internal research team in developing segments of customers using the leading indicators experienced by at-risk members as the basis.


The client was able to use the insights to:

  • Prioritize outreach efforts and keep expenditures within budget.
  • Tailor messages and plan change recommendations to address the unmet needs of at-risk members.
  • Leverage the new value of internal data (i.e. using the at-risk scores with demographic and plan-related attributes to develop specific messages and target groups).
  • Help customer service, product development, and operational teams to understand common problems and establish action plans.
  • Predict which members would be most at-risk if specific drugs were removed from the formulary.

The model allowed clients to classify or segment members into different groups based on their drug coverage needs. The client plans to append the segment data to their membership databases as another variable. This will allow them to direct specific messages to members who are the most at-risk within segments.

Follow us: