The Medicare Payment Advisory Commission (MedPAC) is launching a study to determine whether the Patient-Driven Groupings Model (PDGM) accurately aligns home health payments with patient care costs.
The study, which will analyze data on home health and skilled nursing facility (SNF) payments, will examine patient selection and coding practices that could inflate payments without corresponding increases in patient complexity, according to a Friday MedPAC meeting.
“This work is exploratory, but there have been concerns about patient selection in home health,” MedPAC analyst Evan Christman said in the presentation.
MedPAC is a nonpartisan agency to the legislative branch that advises Congress on Medicare.
At the time of PDGM’s implementation, some industry insiders suggested that PDGM would lead to disruptive cash flow pressures, while others said the model would not be as disruptive as other changes to the home health payment landscape.
CMS’ proposed CY2027 Medicare home health payment rule included a 3% temporary adjustment, tied to the agency’s effort to recoup what it considers PDGM-related overpayments. The home health community broadly decried the temporary adjustment.
MedPAC analysts will consider 2020 revisions to thee health payment systems which eliminated therapy-based incentives, as MedPAC had recommended, with the new systems relying more on patient characteristics to determine payment levels, Christman said.
Now is an opportune time to evaluate the payment system, with several years of data accumulated and providers having had time to adjust, Christman said. The probability of someone using home health in the last cycle has not changed significantly in 2023 compared to 2019, he added.
Still, Christman said that some home health stakeholders believe that the system favors post-hospital patients over community-admitted patients.
The study will examine patient selection to determine whether high-profit and low-profit case mix groups are distributed unevenly across providers, MedPAC analyst Carol Carter said. Carter dded that uneven distribution could arise from market characteristics or provider capabilities.
Analysts will also examine whether coding changes reflect genuine shifts in patient complexity or if documentation practices raised payments without a true change in underlying complexity, Carter said. For cases with a prior hospital stay, researchers intend to compare SNF and home health coding patterns against how the same cases were coded during the preceding hospitalization.
The researchers will then compare how coding has changed for patient assessment items used to create payments with those not used for payments. The difference could indicate changes in coding practices, Carter said.
The use of AI in coding and revenue cycle management emerged as a related concern. Commissioner Kenny Kan warned that providers may already be leveraging AI tools to drive up coding intensity or report patient complexity beyond what patient conditions warrant.
“I worry that many health systems and providers are way ahead of health plans in using AI to augment inappropriate coding intensity or reported patient complexity to optimize revenue cycle management,” Kan said. “This is likely inflationary for overall cost of healthcare long term.”
MedPAC plans to begin presenting results in fall of 2027.
