A newly released study casts doubt on whether the shift to Medicare Patient Driven Payment Model (PDPM) succeeded in the goal of curbing so-called “upcoding” practices among nursing homes.

Conducted by researchers from the University of Rochester and the University of Texas at Austin, the analysis spans seven years and found patterns of exploitation that affect resident care. The study, which has not been peer reviewed, drew criticism from the American Health Care Association/National Center for Assisted Living (AHCA/NCAL).

The Centers for Medicare and Medicaid Services (CMS) introduced the PDPM case-mix classification in fall of 2019. PDPM bases reimbursement on patient characteristics and comorbidities rather than therapy levels, as the previous RUG-IV model did.

The change from RUG-IV to PDPM was designed to improve outcomes for care while reducing costs, but some “opportunistic” nursing home systems appear to have found ways to continue gaming the system, Alex Priest, assistant professor of finance at the University of Rochester, told Skilled Nursing News.

“Changes to the reimbursement scheme alone, such as the change from RUG-IV to PDPM, seem unlikely to be successful in eliminating fraud,” said Priest.

Since the PDPM rollout, these skilled nursing facilities (SNFs) are linked to an additional 35,000 hospitalizations and 30,000 deaths since PDPM was enacted, while overbilling Medicare by $4.3 billion, according to the study’s findings. These facilities are expanding at more than 2.5 times the rate of those with accurate billing practices, the study also finds.

Advocacy groups push back on study

Advocacy groups for nursing homes question the accuracy of the study’s data and the validity of its findings.

“This paper does not provide a reliable analysis of nursing home quality, nor does it paint the full picture in the context of recent global events and the multitude of detailed reports and policies that govern reimbursement,” AHCA/NCAL shared in a statement to SNN. “The fact is that the vast majority of nursing homes provide high quality care to their residents and operate in compliance with programs that make this possible. They employ passionate caregivers who love what they do and who treat their residents as family.”

Indeed, the report is not peer-reviewed, a process by which research is evaluated by a group of experts in the appropriate field.

However, Priest explained that in academic papers in finance or economics, since the peer review process often takes several years, it is standard practice to release working papers as a “preprint,” using platforms such as the Social Science Research Network (SSRN).

“We believe our findings are too important for the public not to know,” Priest said, adding that his co-author’s well-established publishing record speaks to his credibility.

The main data used by Priest and co-author John Griffin is patient-level Limited Data Sets from CMS. “These are proprietary data sets in that they are not public or free,” Priest said.

The study drew from a dataset, mainly claims and revenue files, available for purchase from CMS, he said. That dataset included more than 7 million patients and 14 million nursing home stays. 

PDPM’s intended impact versus reality

Despite a change to the system from RUG-IV to PDPM, the data suggests that the same providers who were manipulating the RUG system learned how to exploit PDPM, said Priest. They learned to exploit the new system by “upcoding” patients to increase reimbursements, he said. This can often result in hundreds of dollars in additional funding per patient.

“We find that the same skilled nursing systems which had gamed RUG-IV by targeting just enough therapy minutes to qualify for the highest level quickly began upcoding patients under PDPM to increase patient revenue,” Priest said. “We [completed] a large number of tests to validate these findings. Essentially, despite a large change in incentives presented by PDPM, we find similar problems even after the change in payment scheme.”

The researchers found a “strong correlation” between what they characterized as “system level excess rehab” between January 2016 and September 2019 under RUG-IV and coding intensity from October 2019 to December 2022 under PDPM.

Under the RUG-IV system, these “opportunistic systems” were 5.28 times more likely to provide weekly therapy for a 10-minute period above the highest threshold for reimbursement versus the entire 210-minute range below that threshold.

Following the transition to PDPM, many of these same opportunistic systems increased billing in ways that do not seem tied to actual patient comorbidities, the researchers found. For instance, billing for the highest-compensating comorbidities increased rapidly “precisely” when PDPM was enacted, and there was no increase in these same comorbidities observed on the referring hospital side.

Furthermore, “use of the highest-compensating patient codes increases significantly following acquisitions by opportunistic SNF systems despite no changes in patient characteristics,” the researchers stated.

CMS could improve its policies by increasing enforcement, Priest said.

“Our findings suggest many areas for potential reform and further investigation. First, the fact that SNF systems have such widespread and persistent differences in fraud and health outcomes indicates that better care at lower prices is feasible by many systems. Enforcement and policy should focus more on system-wide practices due to greater precision and more powerful tests,” the researchers note.

The researchers also suggest that attention needs to be given on measuring and quantifying patient health outcomes.

“Academic research can assist by providing more transparency and monitoring, which would be more efficient if CMS made additional granular data available,” Priest said.

The authors also recommend increasing the fines and adding criminal charges to those manipulating the system where management is complicit. Priest said that the industry itself should push for greater information disclosure and transparency.

“While not all of these reforms may be necessary, our results point to significant changes being needed to enact change,” he said.