Our team has worked with a major health insurer over the course of this project to develop a method for identifying which factors, measured by the insurer, have the greatest impact on whether a claimant will exhaust his or her benefits. Beginning with the 31,488 claims in the data provided by the insurer, we looked only at the 26,114 claims that had closed (which included the 3,948 claims that were closed due to benefit exhaustion). The dataset included a variety of information including age, gender, marital status, benefit period, state, diagnosis, etc., but was stripped of any personal information which might identify individual policyholders. The method of research that we decided was most important to our project compares claims closed due to exhaustion versus total claims closed, with respect to the frequency of any given factor. Next, we created a graphic we could reproduce for each factor based on this method. To verify that our method was viable, we tested all factors, including ones that we believed to not be relevant to the study. This gave us an indication of how much noise there might be when looking at the factors that we did believe to be relevant. From our study, we were able to classify the factors that we tested into 3 main groups—Predictive and Known during Underwriting, Not Predictive and Known during Underwriting, and Unknown during Underwriting. Although some of the factors may have had a strong correlation to benefit exhaustion, they may not be useful to the insurer when writing new policies (primarily because they were not knowable at the time the policy was underwritten and issued). Overall, the factors that we found to be the most useful in determining benefit exhaustion are those that are both Predictive and Known during Underwriting.