Responding to my recent Twitter solicitation for things to blog about, David Ramsey wanted to know about health outcomes of traditional Medicaid vs. privatized versions.
Though there is some literature on this subject, I stopped searching for it when I came across two reviews that seem fairly comprehensive. One is by Katie Rosingana, Paul Saucier, and Karen Pearson (Word doc) of the Muskie School of Public Service, University of Southern Maine (April 2010). The other is Michael Sparer’s Robert Wood Johnson Foundation Synthesis Project report (September 2012). Because it is more recent and more thorough, what follows is based on the latter.
First of all, let’s move through the issue of cost (or spending) quickly. David didn’t ask about it, and I covered it, though briefly, in a prior post. Sparer’s conclusion is no different than in that post: Medicaid managed care is unlikely to lead to savings, in general.
Turning to access and quality, prepare to be disappointed. As Sparer notes, “the academic literature  is rather thin,” and the evidence is inconclusive. There are reasons for this. Medicaid programs vary tremendously, challenging generalization. Most studies “focus either on a single state (or even a single county or region within a state) or on a small subset of states,” Sparer writes. Moreover, studies typically examine a single subpopulation of Medicaid beneficiaries, like high-risk pregnant women or the elderly.
Based on the evidence available, on access, Sparer concludes that Medicaid managed care is likely to have a “mixed impact” because
Managed care is explicitly designed to have a mixed impact on access, increasing access to a usual source of primary care while reducing inpatient and emergency room utilization.
Health plans employ a variety of provider reimbursement methodologies, from capitation (which in theory can create incentives to underserve) to fee-for-service (which in theory can create incentives for unnecessary utilization). Different managed care models are thus likely to have different impacts on access.
Managed care plans often have limited provider networks, which can make it difficult for vulnerable and high-risk populations to maintain access to needed specialists and specialized services. At the same time, managed care plans can at times provide more access to specialists than is available in traditional fee-for-service programs.
States were using prior authorization and utilization review in their traditional fee-for-service Medicaid programs, lessening the odds that health plans could use such tools to change utilization patterns.
Health plans cannot impose significant co-payments on Medicaid beneficiaries, thereby complicating efforts to encourage beneficiaries to change care-seeking habits.
The health delivery system for the poor is entrenched and decentralized, and health plans generally lack the leverage to ensure systemwide changes.
On quality, conclusions are hard to reach because
Health outcomes are produced by a complex combination of factors, including various social determinants (such as education, housing and culture), making it hard to identify the impact of any particular intervention (such as Medicaid managed care).
Much of the data on quality available to researchers are process measures that are less helpful in identifying and evaluating changes in health outcomes.
The complex combination of factors impacting quality makes it more feasible to conduct case studies on particular quality initiatives than to conduct large-scale (and presumably more generalizable) studies.
The literature on quality improvements in Medicaid Advantage and the commercial markets is also thin, suggesting the problem of evaluating quality in managed care goes well beyond Medicaid.
I told you to expect to be disappointed. The flip side is that there are opportunities for future work. Sparer points to some ways in which researchers have not exploited available resources.
Notwithstanding the weak conclusions, Sparer’s review is worth a look if this subject area interests you. There are a lot of important details that can’t make it into a ~600-word post, and he organizes the landscape and literature nicely.