Abstract

In this session we present findings from a mixed-methods, multi-arm research study. While we initially focused on the implementation and impacts of the Patient Driven Payment Model (PDPM), a system of reimbursement for skilled nursing facilities that was implemented throughout the United States on October 1, 2019, we shifted our focus and leveraged our unique study design to additionally examine how COVID-19 impacted skilled nursing facility operations and outcomes.

This session includes four sets of findings from this work. First, we present our qualitative examination of the impact of COVID-19 vaccinations, including vaccine hesitancy, mandates, and boosters, on skilled nursing facility staff and staffing levels and describe approaches and strategies used to improve staff COVID-19 vaccination rates. Second, we quantitatively show the effect of state-imposed vaccine mandates on staff vaccination coverage and associated staff shortages to examine whether effects differed between areas with different political preferences. Third, through 156 qualitative repeated interviews with 40 skilled nursing facility administrators over two years, we describe the implementation and impact of PDPM in skilled nursing facilities, as well as explore the effect of COVID-19 on PDPM’s implementation and outcomes. Fourth, we present findings regarding the extent to which there was a spillover effect of PDPM on therapy utilization from traditional fee-for-service Medicare to Medicare Advantage.

These mixed-methods findings provide unique insight on the fate and well-being of post-acute and long-term care systems in the United States. Explicitly, our work on skilled nursing facility staff COVID-19 vaccination highlights strategies to promote vaccination, but also presents nuanced perspectives regarding the dilemmas posed by vaccine mandates. While our work demonstrates how vaccine mandates have been an effective tool to improve staff vaccination coverage, ongoing monitoring of vaccination rates and staffing levels is needed. We also showcase the innovative ways that skilled nursing facilities prepared for and implemented PDPM, reflecting a substantial shift in how facilities are reimbursed. Our work shows the multi-faceted effects of PDPM, how policies in Medicare spilled over to the Medicare Advantage population, and how COVID-19 reordered priorities for skilled nursing facilities and their staff. These presentations highlight the daunting challenges facing skilled nursing facilities today. The strength of the rigorous approaches used here vividly demonstrates the complex ways in which the crisis of the pandemic of recent years has put post-acute and long-term care systems at enormous risk.

Presentation 1: A qualitative examination of the relationship between COVID-19 vaccination and skilled nursing facility staffing

Emily Gadbois, Brown University

Amy Meehan, Brown University
Joan Brazier, Brown University
Momotazur Rahman, Brown University
David Grabowski, Harvard Medical School
Renee Shield, Brown University

Background: Skilled nursing facilities (SNFs) have borne a heavy share of COVID-19 burden, as they provide post-acute and long-term care to older adults and others who are at risk of negative outcomes resulting from COVID-19. COVID-19 vaccination is a critical component of infection prevention in this setting, but uptake among SNF staff varies widely and remains suboptimal. It is necessary to understand staff attitudes toward COVID-19 vaccination to clarify the relationships among vaccination, promotion strategies, and staffing levels, in order to identify best practices and offer practical recommendations.

Objectives: To understand the impact of COVID-19 vaccination, including vaccine hesitancy, mandates, and boosters, on SNF staff and staffing levels and to clarify effective approaches to improve staff COVID-19 vaccination rates.

Methods: We conducted four repeated interviews at 3-month intervals with administrators from 40 SNFs in eight diverse healthcare markets across the United States, totaling 156 interviews. These in-depth interviews provide unique perspectives on the impact of COVID-19 on SNFs over time, including relationships between vaccination attitudes and staffing. We used thematic analysis to examine evolving administrator perspectives, including their developing insights about the effect of vaccinations on staff and staffing levels.

Results: The development of COVID-19 vaccines was initially met with both enthusiasm and skepticism by SNF staff. SNF administrators described varied reasons for staff hesitancy, including fears related to fertility, side effects, and the perceived haste with which the vaccine was developed; other reasons included specific conspiracy theories and overall distrust of the government. Before the federal vaccination mandate was announced in the United States, administrators reported their strategies to reduce staff vaccine hesitancy, including the use of incentives, one-on-one education, and reduced testing and personal protective equipment requirements. As staffing in SNFs has become increasingly difficult due to shortages, administrators have had to balance the priorities of protecting residents from COVID-19 on the one hand and maintaining appropriate staffing levels on the other. The vaccination mandate allowed administrators relief that they were not viewed as the “bad guys;” however, their worries about staffing levels persisted. They noted that the timing of the mandate injunction influenced vaccine-hesitant staff to receive the vaccination, while some who were expected to resign did not need to do so, providing the SNF administrators a temporary reprieve. The impact of boosters on SNF staffing meanwhile receded in emphasis compared to the effect of the initial vaccination.

Conclusions: SNF staff perspectives regarding vaccinations have evolved over the course of the pandemic, and legislation around vaccinations has impacted SNF staffing in several unpredictable ways. This study provides a unique longitudinal perspective describing the evolving stressors SNF administrators have experienced during the pandemic, including shifting SNF staff perspectives on vaccinations, legislation of vaccinations, and the consequent impact on staffing levels. Qualitative examination of these shifting SNF dynamics reveal important implications of administrator perspectives to help identify strategies and maintain high levels of vaccinated SNF staff.

Presentation 2: Nursing home COVID-19 vaccine mandates, staff vaccination coverage and staff shortages in the United States

Brian McGarry, University of Rochester

Ashwin Gandhi, University of California Los Angeles
Maggie Syme, Hebrew Senior Life
Sarah Berry, Hebrew Senior Life
Elizabeth White, Brown University
David Grabowski, Harvard Medical School

Background: High staff COVID-19 vaccination coverage is critical to reduce nursing home COVID-19 outbreaks and death, yet many staff in United States nursing homes remained unvaccinated months after vaccines became available. In response, several states implemented vaccine mandates for nursing home employees.

Objectives: To estimate the effect of state-imposed vaccine mandates on staff vaccination coverage and reported staff shortages and examine whether effects differed between areas with different political preferences.

Methods: We compiled data on state vaccine mandates, including announcement and implementation dates, and whether the state allowed staff to submit to frequent COVID-19 testing in lieu of vaccination (“test-out option”). We measured weekly facility-level outcomes from the National Healthcare Safety Network data between June 6, 2021, and November 14, 2021. Key outcomes were the percent of staff with at least one COVID-19 vaccine dose and whether the facility reported a staff shortage. We also measured the political leanings of the nursing home’s county by the vote share for the Republican candidate in the 2020 presidential election.

Results: Among 37 states with available data, 26 had no mandate, 4 had a mandate with a test-out option, and 7 had a mandate with no test-out option. Ten weeks or more following mandate announcement, nursing homes in states with a mandate and no test-out option experienced a 6.6 percentage point (pp) increase in staff vaccination coverage (P=0.078) while nursing homes in mandate states with a test-out option experienced a 3.1 pp increase (P =0.022) relative to facilities in states with no mandate. We detected no significant increases in the frequency of reported staff shortages following mandate announcement.

Mandate effects on vaccination coverages were larger in Republican-leaning counties without evidence of increased reported staff shortages. Among nursing homes in Republican leaning counties, those in states with a mandate and no test-out option experienced a 14.1 pp increase in staff vaccination coverage ten or more weeks following announcement relative to Republican-leaning counties in non-mandate states (P<0.001).

Conclusions: State-level vaccine mandates increased staff vaccination coverage without increases in reported staff shortages in the United States. Increases were largest when mandates had no test-out option, indicating that strict mandates are an effective policy tool to improve staff vaccination coverages. Mandate effects were also larger in Republican-leaning counties, which had lower average vaccination coverage at baseline, suggesting that mandates are effective in areas where there may be political opposition to such policies. Ongoing monitoring of vaccination coverage and staffing levels will be essential as all United States nursing homes become subject to the federal vaccine mandate recently upheld by the United States Supreme Court.

Presentation 3: Implementation of the patient driven payment model and the impact of COVID-19

Amy Meehan, Brown University

Emily Gadbois, Brown University
Joan Brazier, Brown University
Momotazur Rahman1, Brown University
David Grabowski, Harvard Medical School
Renee Shield, Brown University

Background: On October 1, 2019, the Patient Driven Payment Model (PDPM), a new case-mix classification model of reimbursement for skilled nursing facilities (SNFs), was implemented across the United States in an attempt to shift payment incentives to appropriately support care provision for medically complex needs. PDPM replaced the prior Resource Utilization Groups (RUGs) system, which incentivized SNFs to prioritize therapy. Substantial shifts in SNF staffing, patient acuity, and strategic planning were anticipated as a result of the implementation of PDPM, but COVID-19 devastated the SNF setting six months into implementation.

Objectives: To qualitatively examine the implementation and impact of PDPM in SNFs, as well as explore the effect of COVID-19 on implementation and outcomes.

Methods: We conducted four repeated interviews at 3-month intervals with administrators from 40 SNFs in eight diverse healthcare markets across the United States, totaling 156 interviews. Interviews were designed to understand implementation of PDPM, including how SNFs prepared, what administrators expected prior to PDPM’s start compared to their evolving views, initial implementation, and the impact of COVID-19 on how PDPM was carried out. We used thematic analysis to examine administrator insights.

Results: Our interviews with administrators revealed several themes. First, administrators usually described significant efforts to prepare for PDPM, often beginning over a year before implementation. These efforts primarily included staff training, especially for Minimum Data Set (MDS) coordinators, responsible for implementing the new coding systems. Less frequently, these efforts to prepare included reconfiguring staffing and assessing capability to admit more medically acute patients with concomitant higher reimbursement. As a result of this preparation, implementation of PDPM was largely reported as primarily smooth and without undue surprise. In most facilities, the shift to PDPM was associated with a mild increase in reimbursement. In some facilities, especially those that focused more efforts on rehabilitative therapy, some reductions were reported, while in others, especially those that had consistently cared for more acute populations with high nursing needs, PDPM was associated with increased reimbursement. However, COVID-19 caused a significant shift in priorities such that large-scale strategic planning regarding PDPM was placed on hold. For example, facilities that had planned staffing and education changes to take on more medically complex, highly reimbursed patients were not able to do so. COVID-19 did, however, result in a bump in reimbursement for SNFs that provided care to patients with COVID-19, which meant that substantial decreases in reimbursement and changes to staffing models were not realized. As the pandemic waned, administrators began to shift their focus back to PDPM, but reported a need to reconsider their approaches.

Conclusions: Administrators described how they planned for and implemented PDPM, and also how they had to pivot to respond to COVID-19. Findings demonstrate the flexible ways that SNFs implemented a substantial national-scale policy change, as well as their strategies to continue to function during unprecedented challenges.

Presentation 4: Spillover effect of the patient driven payment model on therapy delivery among Medicare advantage enrollees in skilled nursing facilities

Brian McGarry, University of Rochester

Momotazur Rahman, Brown University
David Meyers, Brown University
Elizabeth White, Brown University
David Grabowski, Harvard Medical School

Background: Medicare is the public health insurance program for older adults and younger persons with disability in the United States. Currently, about 45% of Medicare beneficiaries receive Medicare benefits through managed care plans known as Medicare Advantage (MA) as opposed to traditional Medicare (TM). The impact that the two types of insurance have on one another remains unclear. In October 2019, traditional Medicare adopted the Patient Driven Payment Model (PDPM) for skilled nursing facility (SNF) care that changed the financial incentives regarding the volume of therapy fee-for-service enrollees. Adoption of PDPM resulted in a substantial reduction in the volume of therapy for TM enrollees. However, the spillover effects of PDPM on MA enrollees remain unknown.

Objectives: To assess how the PDPM policy change intended for TM beneficiaries affected therapy utilization of MA enrollees.

Methods: We used a regression discontinuity (RD) approach linking the Medicare enrollment file and Minimum Data Set (MDS). Our cohort included 3.7 million Medicare enrollees aged 65 years or older who were admitted to a SNF between January 2018 and June 2020 and did not have a nursing home stay in the prior one year. Twenty-six percent of these individuals were admitted after PDPM was implemented. Twenty-nine percent of these individuals were enrolled in MA. Our outcomes included the number of individual and non-individual (concurrent and group) minutes for physical, occupational, and speech therapy reported in the first scheduled MDS assessment following admission. We fit separate regression discontinuity models for MA and TM enrollees, estimating outcomes as a function of continuous time and its fourth-degree polynomials and an indicator of admission post-PDPM. In this specification, the coefficient of the post-PDPM indicator captured any discrete jump in trend associated with PDPM. Other variables included age, gender, race, calendar month dummies, day of the week dummies, and SNF fixed effects.

Results: Before PDPM, MA enrollees received less individual therapy (69 vs. 86 minutes per day) and more non-individual therapy (2.8 vs. 0.4 minutes per day) compared to TM enrollees. We observed a discrete decrease in reported individual therapy minutes and an increase in non-individual therapy minutes for both TM and MA enrollees, but the change was much smaller for MA enrollees. The RD estimates for total therapy use in the first week following admission were 9 minutes per day for TM enrollees and 3 minutes for MA enrollees. The RD estimates for TM enrollees did not vary across SNFs with different MA penetration. However, RD estimates for MA beneficiaries were 4 minutes per day in SNFs in the lowest MA penetration quartile and 2 minutes per day in SNFs in the highest MA penetration quartile.

Conclusions: Our results suggest that PDPM had similar effects on MA enrollees as it had on TM enrollees, but the effect sizes were smaller for MA beneficiaries. This substantial spillover effect of TM policies on the MA population imply that MA plans follow traditional fee-for-service care practices for post-acute care in skilled nursing facilities.

Slides