A new CMS-funded study suggests surgeons are overpaid for certain bundled procedures and proposes a solution it says could save the agency billions of dollars annually.
The New England Journal of Medicine report found that just a fraction of post-operative visits the CMS pays for as part of procedure bundles actually take place. The report says reducing the payments accordingly would have saved Medicare $2.6 billion in 2018 by decreasing payments for 10- and 90-day global procedures by 28%.
The findings have "huge" implications for physician revenue, but also for Medicare patients, who face a 20% co-pay under Medicare Part B, which includes post-operative visits, said Andrew Mulcahy, lead author of the study and senior health policy researcher with RAND Corp.
"Of the $2.6 billion, 20% of that on paper would be the patient's responsibility," he said, adding that many beneficiaries have Medigap or supplement plans to offset that cost.
Post-operative visits account for roughly 25% of Medicare payments to physicians for procedures with bundled post-operative care, which totaled $9.9 billion in 2017, the study found.
The findings are based on the volume of post-operative visits reported by physicians and other practitioners under a 2017 CMS requirement designed to strengthen the agency's purview into the number of visits actually delivered after surgeries.
Medicare's bundled payment rates are based on physicians' survey responses that estimate the number and level of post-operative visits a typical patient would require. But recognizing that there's no way to verify whether physicians actually deliver that many visits, the CMS now requires certain physicians and other practitioners in nine states to report each post-operative visit using a "no pay" code.
RAND Corp. published initial findings on the data in a trio of CMS-funded studies last year.
"Up until last year, Medicare didn't know how many visits were happening," Mulcahy said. "Now they do. They're overpaying."
The data showed that post-operative visits took place in just 4% of 10-day global periods for minor procedures, such as a dermatologist removing a skin tag. For more complex procedures with 90-day global periods, 39% of the visits that were assumed to have taken place under Medicare's payment valuation actually took place, according to the RAND study.
Dr. Ateev Mehrotra, an author of the study and associate professor of medicine at Harvard Medical School, said he doesn't think it's because patients aren't getting the post-operative care they need.
"Rather, I think clinical patterns and so forth have changed over time potentially, and therefore that post-operative care is not necessary," he said.
Still, Mulcahy said he wouldn't be surprised if some doctors tend to round up in their survey estimates of how many post-op visits they provide. And when surgeons provide visits after the 10- or 90-day follow-up window, they're allowed to bill for those separately, he added.
The new report also notes that post-operative care is increasingly being shifted to hospitalists and intensivists, who bill separately from the bundled payment for the initial procedure.
"In that case, the surgeon is getting paid and then the other practitioner that did that extra visit is also getting paid," Mulcahy said, "so Medicare double pays."
The study notes that because of Medicare's budget-neutral payment policy, if the CMS were to lower payments for surgical procedures, it would result in across-the-board pay increases for all other physician services, such as evaluation and management. Mehrotra said that would shift a significant amount of payments from surgeons to primary-care physicians.
Unsurprisingly, surgeon specialty groups had strong reactions to RAND's first set of studies, and the overall idea that Medicare was overpaying for procedure bundles.
The American College of Surgeons, for example, objected to counting the number of "no pay" codes submitted to tally the number of post-operative visits provided. Vinita Mujumdar, the ACS' manager of regulatory affairs, said it's possible doctors could forget to submit the code or have trouble reporting it due to barriers in the hospital or inadequate software.
"There are many steps along the way where the code could have been prevented from getting to CMS and being counted," said Mujumdar, who had not seen the most recent RAND study because it was under embargo.
RAND's analyses use only "clean" procedures, or those in which an individual patient did not have an overlapping 10- or 90-day global period. The ACS estimates that excludes about 40% of procedures, possibly with a bias toward less-complicated procedures.
Mulcahy said he expects to hear similar criticism as he did following the earlier studies, mostly that the data are imperfect. Still, he said the new data provides an understanding of post-operative care that's "leaps and bounds" ahead of what was understood previously.
"Even if we were missing some chunk of visits, you'd have to be missing an enormous chunk to change the punchline that Medicare is overpaying," he said.