WASHINGTON – Early analysis of the implementation of Medicaid work requirements suggests administrative burden on beneficiaries is leading to decreased coverage.
according to state reports.
A study of the early months after Arkansas put its work requirement into effect in June 2018 found those most affected were people already working but unaware of or unable to meet the administrative requirements to demonstrate they were working. These people lost coverage.
Researchers identified “lots of confusion,” Anna Goldman, MD, of Harvard University, Boston, told attendees at the annual meeting of the Society of General Internal Medicine 2019 conference. “People didn’t know if they should report, or if they thought they should, they didn’t know how.”
Additionally, there is no evidence that the implementation of the work rules caused an increase of employment within the Medicaid-eligible population in Arkansas within the first 6 months of the law going into effect.
Arkansas’s rules required individuals aged 30-49 years (applied during the first 6 months of the rollout) to work at least 80 hours a month in a job or doing something equivalent to a job, such as job skills training. Exemptions were available to those who were disabled, pregnant, or caregiving; had student status; or were receiving treatment for substance abuse. Individuals were required to report their work hours each month. Failure to do so for 3 months in a year would cause Medicaid benefits to be suspended for the balance of the year.
Dr. Goldman noted that phone surveys were done in four states (Arkansas, Kentucky, Louisiana, and Texas) of low-income adults earning up to 138% of the federal poverty limit. About 6,000 people were surveyed across the four states, with half coming from 2016 and the other from a survey in 2018. In both years, the surveys were conducted in November and December of survey years. She acknowledged a low response rate of 14%, but she said that weighting was used to minimize nonresponse bias.
With the introduction of the work requirements, researchers found a 13.2% decrease in Medicaid or marketplace insurance and a 7.1% increase in the uninsured rate. No change in employment rates were found.
She noted that the findings that employment didn’t change were not surprising “given that 95% of people were either exempt or working, according to our survey results.”
Dr. Goldman said the survey completed for this study was consistent with studies done in Arkansas that showed most people that would be covered by the work rule were either working or would be exempt from the work requirements laid forth by state regulation.
“Those studies had raised an initial concern that most of the coverage losses would be related to nonreporting and administrative issues as opposed to just removing people who were healthy and didn’t want to get a job,” she explained.
Dr. Goldman noted that this newer study found that of the people required to report their work activities to the state, half didn’t because they either had no Internet access or were confused and could not figure out how to use the reporting system, even though they met the work requirements necessary to keep their Medicaid coverage.
In general, about 52% of the Arkansas population surveyed had heard something about the work requirements, but of those in the age range targeted by the rules, only 22% thought they were subject to the work requirements. Of those informed by the state that they needed to report, only half said they were doing so, Dr. Goldman added.
The survey results were consistent with state reports on the number of people experiencing coverage loss after the rollout, she said.
“One important thing to take away is that information barriers and administrative hassles are key reasons why people are losing coverage. ... Therefore, anything that states can do to automatically exempt people using state data that they already have can possibly reduce unnecessary coverage losses,” Dr. Goldman concluded.
Arkansas’s Medicaid work requirements were blocked by a federal judge in March 2019. Appeals are ongoing.
The authors reported no disclosures.