Evaluating the Impact of the Patient-Driven Payment Model on Medicare Coding Intensity, and Access, Care Delivery, and Outcomes for Metastatic Cancer Patients in Skilled Nursing Facilities

Abstract

The Patient-Driven Payment Model (PDPM), introduced in 2019, changed how Medicare pays for skilled nursing facility (SNF) care. Instead of primarily compensating for the amount of rehabilitative therapy delivered, PDPM bases payment on a patient’s clinical characteristics (e.g., diagnoses, functional ability, and clinical needs). This shift raised concerns about the accuracy of facility documentation of diagnoses, which can affect risk adjustment and the equitable implementation of the policy across facilities. PDPM was also designed to improve access for patients with serious and complex conditions. Under the prior volume-based payment system, people with metastatic cancer often experienced suboptimal SNF care. Cancer patients were often viewed as costly, which in some cases limited their access to SNFs. Of the cancer patients admitted to SNFs, many experienced high hospital readmission rates, low use of cancer treatments, and more aggressive end-of-life care. Although PDPM aimed to better align payment with patients’ clinical needs broadly, it is not focused on a single disease like cancer and does not directly link payment to health outcomes. It is therefore important to evaluate the impact of PDPM on outcomes, particularly for metastatic cancer patients, who stood to benefit significantly from the spirit of the policy. This study uses quasi-experimental methods specifically, a difference-in-difference design with a non-equivalent dependent control (Aim 1) and interrupted time-series analysis (Aim 2-3), to examine Medicare claims and clinical assessment data for all Traditional Medicare patients hospitalized between 2018 and 2021. It evaluates PDPM’s impact on documentation of clinical complexity for all patients (Aim 1) and on admissions, cancer-specific care delivery (Aim 2), and outcomes (Aim 3) for patients with metastatic cancer. Findings from Aim 1 show significant increases in diagnosis-based complexity measures — namely the Elixhauser score and number of diagnoses on the SNF claim — particularly among for-profit SNFs, as well as increased documentation of five highly reimbursable conditions, suggesting potential financially motivated coding changes. Aim 2 showed increased admissions of metastatic cancer patients to SNFs with little improvement in care delivery: chemotherapy and radiation use remained low, therapy volumes declined, and there was no change in length-of-stay. Aim 3 found no improvement, and some explicit worsening of cancer- and SNF-relevant health outcomes; specifically, higher in-facility mortality, higher all-cause hospital readmission rates, and no change in functional improvement or successful discharge to the community rates. These findings can inform future Medicare policy refinement, both patient-centered reforms to improve outcomes for metastatic cancer patients and broader adjustments to address behavioral responses to payment policies operationalized via documentation.