OBJECTIVES: The objective of this study is to empirically investigate how the medical loss ratio (MLR) changes over time and what factors contribute to that change for patients enrolled to Medicare Advantage Special Need Plans. METHODS: We used a data supplied by a Medicare Advantage Special Need Plan, which include enrollees' monthly premiums and expenses. The data also provide enrollees’ demographic information and CMS risk scores and information regarding facilities and providers participating in this plan. We defined MLR as the total amount of revenues from health insurance premiums that is spent to pay for the medical services covered by the plan. Next, we created a panel data structure and deployed multilevel econometric models in our analysis. Finally, we used Hausman test to evaluate the significance of fixed-effect estimators versus the alternative random-effect estimators. RESULTS: We found the following outcomes after controlling for individuals’ observed and unobserved characteristics together with their CMS risk scores: 1) Adjusted MLR spikes up at the same time as the share of enrollees who die also spikes upward; (2) Adjusted expenses paid by the plan on behalf of enrollees who died while enrolled are much higher than the CMS payments to the plan for the same enrollees in the month of death and in the 4 to 6 months prior to death; and 3) We found that death rates in each calendar month explain about 30 to 40% of adjusted MLR. CONCLUSIONS: The results show that Medicare Part A utilization and expenditures have a limited role in the increasing trend of adjusted MLR for beneficiaries who survive through the enrollment period. However, we find a positive, stronger and statistically significant correlation between inpatient visits and the expenses for enrollees who died while enrolled. The correlation remains significant up to 24 months prior to the death of an enrollee.