Some limitations of this study should be noted in considering our results. First, our analysis was restricted to those not in managed care, and understanding how transition rates differ for those in managed care may be informative and policy relevant. Second, our results are informative of factors that predict MME status and nursing home entry, but do not necessarily cause these transitions. Third, while we analyze general service utilization, we do not look at specific diagnoses or procedures that may be informative of more targeted policy prescriptions.
We also did not address state characteristics in this paper. When estimating models where we included variables for specific state characteristics instead of state fixed effects, we found mixed and contradictory results. One possibility is that the indicators we included may reflect other characteristics that are not captured in the models, but with which they are correlated. Without variation over time, it is difficult to disentangle these effects. Thus, future work is needed to fully understand how characteristics of Medicaid programs and state LTC systems influence transitions to dual eligibility and nursing home care.