GAO: State Evaluations of Medicaid Waivers Lack Key Data, Findings

The Trump Administration’s decision to let states make unprecedented and radical changes to Medicaid through demonstration projects, such as requiring that beneficiaries work or engage in work-related activities, reinforces the need for well-designed, timely evaluations of state demonstrations. A recent Government Accountability Office (GAO) report bolsters concerns that there will be little federal or state oversight of how the Administration’s changed criteria for demonstrations and willingness to approve work requirements, coverage lockouts, and other changes affect beneficiaries, despite evidence that indicates these changes will do harm.

Section 1115 of the Social Security Act gives the Secretary of Health and Human Services (HHS) authority to waive certain Medicaid requirements when necessary to carry out demonstration projects (“waivers”) that promote Medicaid’s objectives. State demonstrations are supposed to test innovative approaches to providing coverage and delivering care, so it’s important to know whether they’re helping or hurting beneficiaries before HHS extends them or other states replicate them.

GAO found big gaps in state-led evaluations of state demonstration projects in eight states — Arizona, Arkansas, California, Indiana, Kansas, Maryland, Massachusetts, and New York:

• Methodological limitations. In four states, evaluations lacked comparison groups needed to measure the demonstration’s impact. Survey sample sizes and response rates were insufficient in Indiana, and analyses of the demonstration’s impact on costs raised concerns in Arkansas, California, and Maryland.
• Failure to evaluate key policy areas. Massachusetts failed to evaluate whether hospitals that received large payments to improve care achieved their intended outcomes. Arkansas’ Medicaid expansion, which provided premium assistance to hundreds of thousands of beneficiaries to enroll in marketplace plans, was intended to improve continuity of coverage across Medicaid and the marketplace, but the state didn’t provide data on continuity of care. Arizona didn’t provide data on hospital readmission rates and quality of life for adults and children with disabilities in its managed long-term services and supports program.
• Lack of timely results. HHS didn’t require final evaluation reports before extending demonstrations and, when it extended them, it pushed the due date for final reports beyond the extension’s end date.

GAO also reviewed progress on a federal multi-state evaluation of state demonstration projects and found that the lack of accurate, complete, and timely Medicaid data limited its scope and timeliness. GAO also noted that the Centers for Medicare & Medicaid Services (CMS) — the HHS agency that performs the work on demonstrations — hasn’t released 15 reports on various aspects of the demonstrations such as their strategies to engage beneficiaries and three interim evaluation reports from the multi-state evaluation.

GAO also looked at federal evaluations of Indiana and Montana demonstrations but found no federal progress on the Indiana evaluation because the state hasn’t provided necessary data to the federal evaluators. The interim Montana report is expected this September but HHS won’t make it public until after a CMS review.

Even with these weaknesses, waiver evaluations have provided important insights, for example finding that both Medicaid enrollees and providers understood little about the requirements of Iowa’s demonstration (the Healthy Behaviors Program), which put enrollees at great risk of losing coverage for not paying premiums.

But there’s no evidence that HHS is considering these evaluations in deciding whether to extend demonstrations. HHS recently extended Indiana’s demonstration (the Healthy Indiana Plan, or HIP) without changes in its premium structure despite evaluations showing that its design confuses beneficiaries and has likely prompted fewer people to enroll in and maintain coverage than if Indiana expanded Medicaid without a waiver. Kentucky’s waiver replicates Indiana’s design in many respects, and HHS has approved both Kentucky’s and Indiana’s plans to implement work requirements and coverage lock-outs for even minor violations of waiver rules that will likely harm large numbers of beneficiaries.

With HHS approving new, unprecedented policies as part of Medicaid demonstrations, rigorous evaluations are even more important. But instead of improving on past evaluation practices, HHS approved these new demonstrations without even requiring states to submit their evaluation plans. Instead, the agency is giving states up to six months after demonstration approval to submit their draft evaluation designs to CMS, and another 60 days to submit a revised draft responding to CMS comments. Kentucky’s work requirement and other lockout provisions take effect this July, however, making it likely that people will start losing coverage well before its evaluation plan is in place.