Do PDGM Clinical Groups Disadvantage People With Dementia? A Pre-Implementation Baseline Analysis

TL;DR

This study examines how a new Medicare payment model might affect people with dementia by analyzing how they are grouped for reimbursement before the model was implemented.

Contribution

The study provides the first pre-implementation baseline analysis of how Alzheimer’s Disease and Related Dementias (ADRD) patients are distributed across clinical groups under the PDGM.

Findings

Before matching, ADRD patients were disproportionately assigned to Neuro Rehab and Behavioral Health clinical groups.

After matching, most clinical groups had a balanced mix of ADRD and non-ADRD patients, but Behavioral Health and Neuro Rehab remained skewed toward ADRD.

The findings suggest reimbursement incentives under PDGM may differentially impact beneficiaries with ADRD.

Abstract

In January 2020, Medicare implemented the Patient-Driven Groupings Model (PDGM), linking 30-day home health payments to clinical (case-mix) groupings. How this affects individuals with Alzheimer’s Disease and Related Dementias (ADRD) is still not well understood. To establish a pre-PDGM baseline, we retrospectively applied PDGM rules to 2019 Medicare fee-for-service claims to simulate clinical grouping assignment for beneficiaries with and without ADRD (N = 1,432,374; 90% non-ADRD, 10% ADRD). We repeated this in a 1:1 propensity score-matched cohort stratified by ADRD status (129,428 per group), matching on demographics, comorbidity burden, prior health care use, social support, and baseline function/cognition. Before matching, ADRD patients were disproportionately represented in Neuro Rehab, while non-ADRD patients were concentrated in Surgical Aftercare. Notably, Behavioral Health, though small (0.94%), was the only clinical group with a majority of ADRD patients (65% vs. 3.1–17.7% in other groups). The top clinical groups in the stratified pre-match cohort were Musculoskeletal Rehabilitation (non-ADRD 34.0%; ADRD 20.1%) and Cardiac & Circulatory (non-ADRD 16.6%; ADRD 21.1%). After matching, most clinical groups converged toward a 50/50 mix of non-ADRD/ADRD patients. However, Behavioral Health and Neuro Rehab skewed toward ADRD (92.1%, 59.3%), while Infectious Disease, Surgical Aftercare, and Respiratory skewed toward non-ADRD (64.0%, 64.1%, 59.3%). These findings provide the first pre-PDGM characterization of ADRD patient distribution across clinical groupings, highlighting how reimbursement incentives may differentially affect beneficiaries with ADRD. This timely evidence establishes a foundation for evaluating whether the PDGM inadvertently disadvantages patients with ADRD, informing policy, clinical practice, and future research on equitable home health delivery.