GAO: States Spend $100B on Medicaid Demonstrations, with Limited Results

Many of those programs found to have “significant methodological weaknesses and gaps in results."

A new report from the Government Accountability Office (GAO) found that roughly a third of Medicaid dollars spent in 2015 went into a demonstration program. In total, that’s about $100 billion worth.

Under the Social Security Act, states can request approval for such programs from the Secretary of HHS. They can range from programs to allow use of managed care systems for long-term in-home care to incentive initiatives for population health or value-based care, and they can be important tools for states looking to reform their Medicaid care delivery.

In a majority of states, GAO found that there were either no ongoing demonstration expenditures (11 states) or that such expenditures were less than a quarter of Medicaid spending (22 states). In 6 states, however, between 50% and 75% of Medicaid money was being spent on demonstration programs, and in 10 states—including Texas, Tennessee, and New Jersey, among others—more than 75% of Medicaid spending was going into demonstration programs.

The amount of money that went into demonstration programs can be substantial. In New York in 2014 alone, $19 billion worth of Medicaid money was spent on demonstrations—and that was only the 3rd highest total in the country that year.

The problems identified in today’s report lie in a lack of reporting standards, which can severely negate their ability to prove efficacy and inform policy, or a lack of efficacy altogether. The watchdog office looked specifically at efforts in 8 states while compiling the report, and it identified “significant methodological weaknesses and gaps in results,” in state-led demonstrations.

An initiative in Arkansas, for example, used Medicaid funds to purchase private coverage for 200,000 individuals. Its evaluation, however, failed to address the central question of whether doing so would improve continuity of coverage for the beneficiaries—which was the central hypothesis of the program.

In programs in Arkansas, California, Indiana, and Maryland, a contractor hired by CMS to evaluate certain programs found that demonstrations had insufficient control groups, if they had any at all. Other states reported state-level resource gaps that made serious evaluations difficult, or other changes in the state healthcare system that clouded regulators’ ability to isolate the effects of a demonstration.

Some demonstrations included multiple states, but inconsistency in data reporting practices again created situations that compromised usability. At times, CMS was unable to even get access to the necessary data from the states.

Because so much of the funding behind the programs is federal, GAO made recommendations to improve programs that currently do nothing to inform federal policy.

First, GAO recommends that CMS institute written procedures that require submission of a final evaluation following each demonstration cycle. It also suggested the agency issue written criteria for when state can and cannot limit their evaluations. Finally, the report recommends CMS determine a format and timeline for the public release of federal evaluation findings to encourage transparency.

“By not making the results of the federal evaluations public in a timely manner, CMS is missing an opportunity to inform important policy discussions happening at the state and federal levels,” the report concludes. It acknowledges, however, that CMS has said it is currently developing procedures that address all 3 of the report’s recommendations.

Related Coverage:

GAO Details Failed VA EHR Initiatives as Agency Requests New Interoperability Rule

To Fight Opioid Abuse, Report Says CMS Will Need More Data

CMS Cancels Another Innovation Center Model