Physicians were more likely to be excluded from public insurance reimbursement programs due to fraud between 2007 and 2017 if they were older, male, or had osteopathic training.
A new study published in JAMA Network Open correlates physician characteristics with exclusion from Medicare and state public insurance programs for fraud, unlawful prescribing of controlled substances, or other healthcare-related crimes.
Exactly what these trends say about fraudulent billing and how they might be used to fight fraud remain open questions, at least for now.
Given the size of the issue, it’s a bit surprising that nobody has previously made an effort to dig into the characteristics of doctors who bill fraudulently, the study’s lead author, Alice Chen, Ph.D., M.B.A., M.Sc., an assistant professor at the USC Price School of Public Policy, told FierceHealthcare.
The study cites 2009 figures from the Institute of Medicine estimating roughly 3% of total healthcare spending in the United States went to fraud, or roughly $75 million. By contrast, the U.S. Department of Health and Human Services recovered $2.6 billion of fraudulent healthcare billings in 2017 despite rising numbers of exclusions over the previous decade.
“We spend a lot of money trying to detect fraud, and as a result we’ve seen increases in exclusion rates over time, and these exclusions really happen after physicians are convicted of a misdemeanor or felony crime,” Chen said. In that light, the 0.3% of physicians targeted for exclusion seems small. “I think that there’s a lot of fraudulent billing behavior going on that’s just not caught.”
While the study identified characteristics associated with doctors punished for fraud, its design precludes making inferences about motivation, or about why those traits might make physicians more likely either to commit fraud or to be caught for it.
The study also found geographic patterns in the data, but the influence of Medicare fraud “strike force” operations in these areas makes it difficult to determine whether the higher number of exclusions stems from a higher incidence of fraud in those areas which led to them being targeted, or whether the targeting itself generated a higher number of exclusions.
Chen points out that the Centers for Medicare & Medicaid Services uses a sophisticated and proprietary set of algorithms to search for fraud, so profiling and targeting physicians based on the characteristics identified in the study wouldn’t be of much use. However, she does believe CMS could use this information to fine-tune its algorithms. “There are errors in prediction, so they’ll find some docs that look like they’re doing something fraudulent, but it turns out they’re not, so improving the prediction model can help identify fraud,” suggests Chen.
In the long run, the data also represents a first step toward identifying the breadth of fraud and doing something about it. “We’ve done much more over time, which I think is good, and I think the more we learn about this problem the more we’ll be able to figure out the right solutions to fix it,” Chen says.