Trending Now: Learning from Medicaid Spending

Feb 27, 2017 | Christopher F. Koller, President

From 1997 to 2005, I ran Neighborhood Health Plan of Rhode Island (NHPRI), the state’s largest Medicaid managed care organization. We worked hard to help coordinate high-value health care and we thought we did pretty well. Recently compiled Medicaid spending data, however, points out that success for states may not mean success for the federal government, and might explain why the state-federal “Medicaid partnership” is under such scrutiny.

During the time I was at the helm of NHPRI, the state expanded Medicaid eligibility several times in RIte Care, its Medicaid managed care program, to include higher-income moms and eventually families. RIte Care was a success.  Patient quality measures were markedly better when compared to fee-for-service, and they improved over time. Members gave NHPRI strong reviews. The program was hailed nationally and was a point of state pride.

A trove of analyses by Manatt Health recently released by the Robert Wood Johnson Foundation’s State Health Reform Assistance Network (SHRAN) indicates that success did not come cheap, however.

The analyses look at 11 years of state and federal Medicaid spending data, organized by state and by the four Medicaid eligibility categories—children, adults, disabled enrollees, and the elderly. This categorization is important—any health policy nerd will tell you these eligibility groups have very different health needs, benefit packages, and resulting costs. The strategies that work for improving the value of care for one will not work for another.

The bulk of Medicaid spending increases in recent years has been due to increases in the number of people enrolled—a policy choice of states and the Affordable Care Act. But what about per person costs? In terms of annual rate of inflation on a per enrollee basis, RIte Care may not look so right, based on these analyses. Between 2000 and 2011, our 9.4% and 10.7% rates of increase for expenses for child and adult Medicaid enrollees were almost twice the national averages of 5.3% and 5.6%, respectively.

The SHRAN data yield many similar findings worth investigating. Figure 1 shows Manatt’s calculation of annual rates of increase for each eligibility category and the states with the five highest and lowest trend rates.

Figure 1
State Ranking – Growth in Medicaid Spending Per Full Benefit Enrollee by Eligibility Group, FY 2000-2011

The data also point out the significant variations in trend rates—twentyfold or more—that exist among states. Low- and high-trend states are not consistently so in all categories and defy characterization by management strategy (fee-for-service or Medicaid managed care), region, or political persuasion.

It is too simplistic to assume this analysis identifies well-managed Medicaid programs. No cash-strapped legislatures were about to sign off on high per enrollee trend rates—and we at NHPRI were certainly not seeing 9.4% increases in our premiums for children.  Reviewing the data with former Medicaid officials helped identify policy choices and incentives for state Medicaid programs that explain some factors in high-trend states during this time.

Changes in enrollment mix can explain some of the trend increases. During this time in Rhode Island, all foster children and disabled children on Medicaid were enrolled in managed care, improving access and raising our average costs. Medicaid eligibility expansions in some states meant more older adults were being enrolled, raising costs in that category.

During this time, aggressive state Medicaid officials were constantly identifying state-funded services eligible for matching federal funds under Medicaid. Where federal money replaced state money, Medicaid program expenses went up.  New or increased provider taxes to generate new federal revenues were also being implemented in some states. Such initiatives are good fiscal policy for the states—and a cause of consternation for some federal officials.

Low-trend states may or may not be more successful Medicaid administrators.  Perhaps they have systematically limited access to needed benefits, elected not to expand coverage to sicker populations or shortchanged providers. High-trend states may have made targeted investments for specific populations or specific classes of providers.

The variation in state per enrollee Medicaid trend rates point to the complex interplay of state policy decisions and program management. In the short run, data like these can inform states as they model the impacts of congressional proposals to limit federal budget exposure for these decisions and cap federal expenses to states for Medicaid. The number of categories used to calculate federal payments and the starting base and rates of increase matter in such proposals—and will make winners and losers.

A successful partnership must work for both parties. Congress and the states are currently trying to figure out what constitutes success in Medicaid, and financial sustainability will be part of the definition. Regardless of the outcome of the current per capita discussion, anyone concerned with population health should pay attention to these cross-state comparisons. No one analysis can be considered authoritative, and qualitative investigations have to follow.  Cost studies must be paired with assessments of quality.

A five percent trend, however, only looks good compared to the rest of health care. Coupled with increasing poverty rates that make more people eligible for Medicaid, health care price inflation forces disinvestment in other services like housing, employment, training, and education that can have a far greater impact on population health.  In the future, states will likely have an even greater need to identify and learn from the places that are systematically improving the value of Medicaid spending. Analyses such as this ensure greater accountability for state Medicaid agencies and their managed care subcontractors.