Illustrating anti-selection

This week we announced the availability of a new paper that looks at the Healthy Indiana Plan (HIP), a voluntary Medicaid expansion program in Indiana. Among other things, the analysis assigns risk scores to enrollees based on their relative health and expected healthcare costs. This figure illustrates the average risk score for enrollees in the first months of the program. It looks at both of the populations modelled in the analysis: caretakers (those with dependents on Medicaid) and non-caretakers.

Average risk scores by month of enrollment


What does this mean? The paper provides some perspective:


HIP started in January 2008. In the first month there were very few people enrolled, so January numbers are not significant. But, starting in February, very meaningful numbers of people were enrolled. We charted those who enrolled in February, tracking their eventual risk scores. We did the same for each subsequent month of enrollees, tracking each group of new enrollees for the first six months and measuring their Medicaid Rx risk scores. The figure above shows that among those who enrolled in the first three to four months of HIP, caretakers had an average risk score ranging from 1.2 to 1.4; non-caretakers had a score between 1.6 and 1.8. Starting in May, June, July, and August, new enrollees were not as high risk or expensive, with risk scores 20% lower; the difference in health status primarily drove the decreased risk and cost.

What does this tell us? The first to enroll in this voluntary program had higher morbidity, with relatively healthier people enrolling a bit later.


Leave a Reply

Your email address will not be published. Required fields are marked *