Medicare Substance Abuse Data Redacted for 4 Years, Fueling Flawed Research

During a crucial time for American healthcare, CMS redacted Medicare claims data regarding substance abuse diagnoses and procedure codes, a move that paralyzed some research and resulted in drastic underestimates of spending and disease prevalence, according to a new study by health policy experts.

The blackout occurred from 2013 to mid-2017, the same time when programs enshrined by the Affordable Care Act were going live and the opioid epidemic was gaining momentum, noted the researchers, who work for the Dartmouth Institute for Health Policy and Clinical Practice and the University of Michigan. Consequently, the redactions contributed to sinking studies before they were conducted and blemishing other research. In one case, they yielded a $6.8 billion underestimate on Medicare spending, according to the study.

“You don’t know what you’re missing,” Ellen R. Meara, PhD, a healthcare economics and policy expert at Dartmouth and a study co-author, told Healthcare Analytics News™. “The biggest issue is that at a time when things have been unfolding very rapidly, there are a lot of studies that didn’t get done.”

>> READ: Rise of the Anti-Opioid Algorithm

The study, published this week in the journal Health Affairs, was designed to determine the effects of these CMS redactions on prevalence estimates of conditions like hepatitis C and depression and inpatient use and spending. Investigators found that the frequency of hepatitis C was underestimated by 11.7%, with a more noticeable disconnect in the under-65 population. In total, 7.1% of Medicare’s 6.3 million beneficiaries during that period saw at least 1 of their claims go redacted. Ultimately, the gaps fueled a nearly $7 billion underestimate of inpatient Medicare spending, which subsequently declined, alongside admissions, by about 5%.

There was little discussion when the redaction went into effect because it was a new interpretation of a longstanding rule, Meara said. The move was designed to better protect patient privacy and align with existing regulations.

In mid-2017, CMS did away with the redaction policy and began working to restore substance abuse Medicare claims data for the blacked-out time period, Meara said. Although researchers may revisit that information now, it remains a costly and uniquely challenging task, she said.

Still, the return to transparency will likely improve and encourage research into opioid policies, patients with disabilities who are younger than 65, and other vulnerable populations, whose claims data were most affected by the redaction.

“It was a huge relief,” Meara said regarding CMS’s decision to no longer redact the data. “I was surprised given all the chaos that’s gone on, and you hear about all of the reductions that go on in government, I figured there was no way they’re going to have the workforce to go back and open it up.”

Still, how many studies the redaction corrupted or prevented from ever taking place remains unknown. Redactions also continue to plague data sets compiled and released by payers, and researchers typically don’t know how that information has been censored.

How the redaction affected health systems and other providers is also murky. Meara couldn’t point to any specific examples, but it’s possible that the bad data harmed hospitals’ efforts to conceive of, build, and assess new payment models.

This study serves as a warning for researchers, Meara noted. Remember that gaps exist, and big statistical swings should inspire skepticism.

“It’s been an incredible missed opportunity,” she added.

Get the best insights in healthcare analytics directly to your inbox.

Related

Why Healthcare Must Beware of Bunk Data

65% of Docs Prescribed More Opioids Than They Thought

Bad Data Is Inflating the Prevalence of Depression