Editor’s note: Part 1 of this post provided background on Medicaid managed care and outlined the pressing questions and challenges facing states that—increasingly—use it. Part 2 of the post, which appears below, examines in more depth the results Medicaid managed care has produced. Brief portions of introductory material from Part 1 are included below for context.
Enrollment of the nation’s 74 million Medicaid recipients in managed care plans continues to increase. By 2016, an estimated 71 percent of Medicaid recipients were receiving their care via private health plans, both investor-owned and nonprofit. The theory behind this shift is that managed care plans can do things that state Medicaid agencies cannot, such as use sophisticated network contracting, information technology, and utilization management systems to squeeze out low-value care and improve the health of beneficiaries.
Yet, as will be seen in Exhibit 1 below, there is tremendous variation in the financial performance of Medicaid programs in their managed care organization (MCO) contracting. Our analysis of Medicaid MCOs’ insurance regulatory filings reveals a five-fold variation in the amount of administrative overhead among states’ managed care plans, and a 37-percentage-point spread in the percentage of MCO revenues that are spent on actual care. All of this is coupled with a more than five-fold variation in state agencies’ own Medicaid administrative expenses.
The variation in Medicaid MCO contracting performance is significant enough to raise important and complex questions. What are the causes of this variation? Is this variation simply a reflection of differing operating models and environments, or is it an indication that taxpayer dollars in some states could be better invested? We don’t know the answers yet. What we do know is that there are a lot of important questions that need to be answered. And we believe that the answers will differ meaningfully from state to state.
What we also know is that governors, state Medicaid agencies, and their contracting plans will be on the hook to answer these questions. Medicaid spending claims almost 20 percent of state general funds (not counting federal matching funds), a percentage certain to rise in any economic downturn. With such a large share of state revenues going to health plans, both managed care plans and state agencies will need to be prepared to demonstrate that Medicaid managed care dollars are being spent appropriately.
Exhibit 1: Medicaid Managed Care Performance Measures
Sources: National Association of Insurance Commissioners Statutory Filings, 2016, and Medicaid and CHIP Payment and Access Commission’s MACStats: Medicaid and CHIP Data Book, December 2017. Notes: Averages are weighted by multiplying Medicaid revenue by the medical loss ratio (MLR) for each plan in a state and dividing by the total Medicaid revenue for all plans in the state. Analysis completed on companies with $10 million in Medicaid revenue, excluding behavioral health special plans and dual-eligible plans. The District of Columbia is included as a state. Eleven states do not have managed Medicaid programs: Alabama, Arkansas, Connecticut, Idaho, Maine, Montana, North Carolina, Oklahoma, South Dakota, Vermont, and Wyoming. Alaska, Arizona, California, Colorado, Delaware, New Hampshire, and Oregon are not included because of incomplete data. The state of domicile was adjusted for a small number of plans to reflect Medicaid enrollment location.
Are State Administrative Expenses Reduced By Contracting Out To MCOs?
One might expect that states no longer so heavily burdened with network contracting, claims processing, and provider utilization management, and so forth would have much lower Medicaid administrative overhead. The volume of state administrative activities such as utilization review and prior authorization, as well as claims payment, do, indeed, decrease with MCO contracting.
Indeed, states that contract out their programs do have lower average administrative overhead: 4.9 percent versus 5.8 percent for states that directly contract with providers. However, as can be seen in Exhibit 1, some states that contract out to MCOs have administrative overhead as high as 10.8 percent. (It should be noted that the highest overhead states, North Dakota and Rhode Island, are also among the smallest in population.)
One factor inhibiting administrative savings is that, with few exceptions, states that contract with MCOs continue to operate a residual Medicaid fee-for-service program in geographic areas where the plans do not operate. Thus, they must maintain information technology infrastructure and continue to manage provider networks and process claims while their MCOs enroll providers in their health plan-specific networks (an inherent duplicative administrative burden). States must also manage their health plan contracts.
Of course, MCOs do a lot more than merely contract with providers and process claims. They also attempt to reduce inappropriate service use, review provider credentials, and perform other quality assurance activities. They also provide member outreach and patient education, scheduling, and care coordination and disease management, all intended to improve the health of the Medicaid population. These are services many state agencies struggle to provide on their own. Nevertheless, governors and state legislators ought to expect their Medicaid agencies to be smaller and better focused if they contract out to MCOs.
How Much Risk Is Actually Transferred To The Plans?
Certainly, states gain budget predictability by locking in per-member-per-month (PMPM) payment capitation rates to MCOs for the term of their contracts. The main source of uncertainty in state program cost then becomes enrollment---the number of beneficiaries to which the rates apply. However, that predictability may come at the expense of actually exerting market discipline on program cost.
In the bidding process, many states actually state in their request for proposal the PMPM they are willing to pay, instead of receiving competitive bids on the actual capitation rates. In these states, contractors are chosen based on how much “value” they can create within the rates through activities such as care coordination and management, enhanced services, member and provider services, and a whole host of other operational functions.
Many legislators and even Medicaid program leaders are mistaken in believing that the Centers for Medicare and Medicaid Services (CMS) requirement that rates be “actuarially sound” provides some level of assurance that the rates are market-based or akin to the “Blue Book Value” of care. In fact, although the requirement has been in place for two decades, the Actuarial Standards Board defined the term only recently (March 2015), and it simply means “projected capitation rates and other revenue sources provide for all reasonable, appropriate, and attainable costs.”
It is possible for MCOs to lose money if the rates they accept in contracts are grossly inadequate to meet the care needs of their enrolled populations, as appears to have happened recently in Iowa. And plans often experience losses during the start-up period. But the actuarial soundness requirements do provide some degree of protection for the plans by establishing a floor under their capitation rates that, to an extent, limits their business risk.
Since Medicaid MCOs do not typically competitively bid on capitation rates, managed care savings are achieved only to the extent that the states and their actuaries are willing to exert downward pressure through managed care rate adjustments; that is, assumptions about changes in service use that MCOs can achieve relative to historical usage and built into the contracted rates. It is the states’ degree of aggression in pushing these adjustments that determines the amount of real risk transferred to the MCOs, by lowering the targeted cap rate.
States that make more aggressive assumptions about these rate adjustments will achieve greater managed care savings, to the potential detriment of health plan profitability. Moreover, when plans achieve PMPM savings, their capitation rates are eventually reduced to reflect those savings. Since plans’ shareholders and boards of directors expect their managements to maintain or grow their revenues year over year, or even quarter over quarter, the interests of the plans and the states are inherently at odds.
Some states have attempted to moderate these conflicting interests by calculating PMPM rates based on a rolling three-year average, thereby creating a shared savings arrangement of sorts. Or they have implemented incentive payment programs (with incentive pools capped at an amount equal to 5 percent of the total capitation premium payments) as a means of aligning plans’ incentives with their own.
One might wonder how actual service use and health outcomes might change if Medicaid health plans were subject to a fixed capitation rate for 10 years and permitted to retain as profit any surpluses, regardless of medical loss ratio (MLR), provided that quality of care improved. This would recognize the reality that it may take more than a few years to realize managed care’s long-term promise of better health for beneficiaries. The state agency-MCO relationship should be managed not merely on a transactional basis but as a long-term partnership.
Does Medicaid Managed Care Contracting Actually Save Money?
Advocates of contracting out Medicaid to managed care plans often cite instilling market discipline and bringing competitive market forces to bear on lowering program spending, familiar arguments for outsourcing any government function. Presumably, therefore, Medicaid MCO contracting ought to reduce program spending.
For such a high-profile policy question, the evidence for savings is surprisingly thin. There are no peer-reviewed studies that we can find documenting Medicaid MCO savings, but rather waves of studies by consulting firms most of which worked for the state Medicaid agencies. A frequently cited meta-analysis conducted by the Lewin Group in 2004 (updated in 2009) reviewed 24 studies, many using data from the 1980s or 1990s, that compared actual MCO expense trends with “projected” spending under fee-for-service either local or national. Many of these studies point to reductions in inpatient hospital use versus expectations. But significant overall reductions in inpatient use were taking place in the overall US population during the same time period.
In the Lewin analysis, there was one standout success story—Pennsylvania in the mid-1990s—based on a study that compared select urban counties’ experience under managed care versus the fee-for-service trend in the balance of the state during the same time period. But there was also a five-state analysis by Mathematica (in 2001) that found no savings versus national trends.
A 2011 analysis by Mark Duggan from the Wharton School and Tamar Hayford from the Congressional Budget Office of the financial results achieved by states’ switching from fee-for-service to managed care in Medicaid reported: “Our baseline estimates suggest that the average effect on Medicaid spending of shifting recipients from FFS to managed care is close to zero. This result holds for both HMO contracting and other types of MMC, and suggests that the policy-induced shift of millions of Medicaid recipients from FFS to managed care during our study period did little to reduce the strain on the typical state’s budget.”
As Duggan points out, utilization reductions achieved by managed care may be washed out by the dilutive effect of having multiple MCOs replace the single-payer state agency as the contracting entity, resulting in higher unit payments than the state would otherwise have made. Some states may have too many contracting MCOs, diluting the potential savings. The result in traditionally low payment states is that overall program costs actually increase in the shift to managed care.
There was a much more recent review by the Government Accountability Office (GAO) of the effectiveness of managed care programs under Medicaid 1115 waivers. The GAO was highly critical both of CMS and state evaluations of the managed care demonstrations. The report concluded: “Selected states’ evaluations of these demonstrations often had significant limitations that affected their usefulness in informing policy decisions.” The GAO found only one clear cut case (Oregon) of documented effective managed care implementation under the 1115 waiver program.
Another key question is whether the health status of managed care beneficiaries is better. The main argument supporting managed care in general is that managing the care---for example, determining that care provided is appropriate and necessary---results in better care and improved outcomes. Intuitively, this should be the case.
However, again, we can find no peer-reviewed evidence that this is so. A report prepared by the Robert Wood Johnson Foundation concluded: “Findings on Medicaid managed care quality outcomes are scarce and have mixed results. Based on a review of the literature prior to 2012, there is also limited evidence that managed care improves quality of care relative to Medicaid FFS.”
There is significant evidence of service gains---for example, enrollment in primary care case management services, access to prenatal care, family planning services, care coordination, disease management services, and so forth---all of which are easier to provide in a managed care environment. However, the linkage of these services to improved health of managed care beneficiaries remains unproven.
Finally, What Effect Will Significant Cuts In Medicaid Budgets Have On The Willingness Of MCOs To Contract With State Agencies?
We cannot assume that current economic growth continues indefinitely. It is reasonable to assume that sometime in the future, we will have a recession. When the next recession strikes, it is entirely possible given current eligibility standards that as many as 80--82 million people may enroll in Medicaid. Recessions also damage state revenues at the very time that Medicaid enrollment grows. In the aftermath of the 2008 recession, Congress temporarily raised the federal Medicaid match rate, sparing states from having to make major Medicaid funding cuts. It is extremely unlikely that this will happen again.
How will states facing declining revenues and rising Medicaid enrollments respond to the resulting fiscal crisis when the majority of Medicaid beneficiaries are covered under MCO contracts? With MCO contracts in place, states will be locked into the rates they agreed on until contract renewal, reducing their fiscal flexibility and tilting policy toward eligibility reductions.
Medicaid risk contracts could become less attractive to investor-owned plans than they are now, at the top of the economic cycle. And some states with significant financial problems may find that when they rebid their MCO contracts that the pool of contractors either shrinks or disappears altogether.
If this happens, could states re-establish provider contracting and payment? And can they maintain managed care-type discipline if they decide to do this? Perhaps, as Alabama recently considered, some states may try to stand up statewide networks of provider-sponsored accountable care organizations (ACOs). Massachusetts’ MassHealth recently switched from MCO contracting to direct contracting with provider-based ACO’s for roughly two-thirds of its Medicaid beneficiaries. If states renew their existing MCO contracts at lower rates under financial pressure, and compel their MCOS to push back hard on their networks, will they compromise network adequacy in the process?
Moving Forward
Conceptually, managed care has the potential to both reduce health costs and improve the quality of care delivered to Medicaid beneficiaries. This is when the actual return on the states’ investment should come. Most states seem not to have pressed MCOs for the information needed to determine if this return on investment exists in practice. States are going to need to up their game, both in regulation and contract management. As prudent purchasers, states are going to need to be much more aggressive in using their market leverage and in understanding health plan operations, not just “managing the contracts.”
As “laboratories of democracy,” state governments should benefit from copying the best practices of the states with the longest and best experiences in managed care contracting. If contracting out Medicaid services to private health plans is to be sustainable policy, legislators, provider groups, and patient advocates should demand that their states not only gather but also publicize information on MCO performance in their states, and policy makers should use that information to make better decisions.
Authors’ Notes
The authors would like to thank Sara Rosenbaum, Brad Gilbert, Michael Koetting, and Gary Taylor for reading and commenting on drafts of this posting. Lauren Linn provided stellar research support.