A 70-year-old woman with Medicare insurance and a history of mild dementia and chronic bronchiectasis was hospitalized for acute respiratory failure due to influenza. She was treated in the intensive care unit (ICU) for 2 days, received mechanical ventilation, and was subsequently extubated and weaned to high-flow nasal cannula (HFNC) at 8 liters of oxygen per minute and noninvasive ventilation at bedtime. She had otherwise stable cognition and required no other medical or nursing therapies. For recovery, she was referred to a skilled nursing facility (SNF) for respiratory support and rehabilitation but was declined due to HFNC use, noninvasive ventilation, and mild dementia. Instead, she was transferred to a long-term acute care hospital (LTACH) for respiratory support. In the context of major post-acute care (PAC) policy changes, where should—and could—this patient go to recover after hospitalization?
BACKGROUND AND HISTORY
In 2018, 44% of hospitalized patients with fee-for-service Medicare (herein referred to as Medicare) were discharged to PAC, accounting for nearly $60 billion in annual Medicare spending.1 PAC includes four levels of care—home health agencies (HHAs), SNFs, inpatient rehabilitation facilities (IRFs), and LTACHs—which vary in intensity and complexity of the medical, skilled nursing, and rehabilitative services they provide; use separate reimbursement systems; employ different quality metrics; and have different regulatory requirements (Table 1). Because hospitalists care for the majority of these patients and commonly serve in leadership roles for transitions of care and PAC use, PAC policy is important, as it has direct implications on discharge patterns and the quality and nature of patient care after discharge.
HHAs, the most commonly used PAC setting, provide skilled nursing or therapy to homebound beneficiaries.1 HHAs were historically reimbursed a standardized 60-day episode payment based on casemix, which was highly dependent on the number of therapy visits provided, with extremely little contribution from nontherapy services, such as skilled nursing and home health aide visits.2
SNFs, which comprise nearly half of PAC spending, provide short-term skilled nursing and rehabilitative services following hospitalization. SNFs are reimbursed on a per diem basis by Medicare, with reimbursement historically determined by the intensity of the dominant service furnished to the patient—either nursing, ancillary care (which includes medications, supplies/equipment, and diagnostic testing), or rehabilitation.3 Due to strong financial incentives, payment for more than 90% of SNF days was based solely on rehabilitation therapy furnished, with 33% of SNF patients receiving ultra-high rehabilitation (>720 minutes/week),3even if it was not considered beneficial or within the patient’s goals of care.4
IRFs provide intensive rehabilitation to patients who are able to participate in at least 3 hours of multidisciplinary therapy per day.1 IRF admissions are paid a bundled rate by Medicare based on the patient’s primary reason for rehabilitation, their age, and their level of functioning and cognition.
LTACHs, the most intensive and expensive PAC setting, care for patients with a range of complex hospital-level care needs, including intravenous (IV) infusions, complex wound care, and respiratory support. Since 2002, the only requirements for LTACHs have been to meet Medicare’s requirements for hospital accreditation and maintain an average length of stay of 25 days for their population.5 LTACH stays are paid a bundled rate by Medicare based on diagnosis.
POLICIES IN CLINICAL PRACTICE
Due to considerable variation in PAC use, with concerns that similar patients can be treated in different PAC settings,6,7 the Centers for Medicare & Medicaid Services (CMS) recently introduced several major policy changes for HHAs, SNFs, and LTACHs (Table 1).1 No major policy changes were made for IRFs.
For HHAs and SNFs, CMS implemented new payment models to better align payment with patients’ care needs rather than the provision of rehabilitation therapy.1 For SNFs, the Patient
Driven Payment Model (PDPM) was implemented October 1, 2019, and for HHAs, the Patient-Driven Groupings Model (PDGM) was implemented January 1, 2020. These policies increase payment for patients who have nursing or ancillary care needs, such as IV medications, wound care, and respiratory support. For example, the per diem payment to SNFs is projected to increase 10% to 30% for patients needing dialysis, IV medications, wound care, and respiratory support, such as tracheostomy care.8 These policies also increase payment for patients with greater severity and complexity, such as patients with severe cognitive impairment and multimorbidity. Importantly, these policies pay HHAs and SNFs based on patients’ clinical needs and not solely based on the amount of rehabilitation therapy delivered, which could increase both the number and complexity of patients that SNFs accept.
To discourage LTACH use by patients who are unlikely to benefit from this level of care, CMS fully implemented the site-neutral payment policy on October 1, 2020 (although it is paused during the coronavirus disease 2019 [COVID-19] pandemic), which substantially decreased payment to LTACHs for patients who either did not have an ICU stay of 3 or more days preceding the transfer or did not receive prolonged mechanical ventilation in the LTACH for 96 or more hours.
COMMENTARY AND RECOMMENDATIONS
Historically, PAC payment policy has not properly incentivized the appropriate amount of care to be delivered in the appropriate setting.9 The recent HHA, SNF, and LTACH policy changes not only shift the discharge of patients across PAC settings, but also change the amount and type of care that occurs at each PAC site (Table 2). The potential benefit of these new policies is that they will help to align the right level of PAC with patients’ needs by discouraging inappropriate use and unnecessary services. Under the new HHA and SNF payment models, initial media reports suggest a decline in therapy services has occurred, which could be beneficial if the therapy was excessive and not indicated.4,10,11 Similarly, LTACHs are experiencing a large decline in admissions as fewer patients meet the new payment criteria.1 As with all policy changes, the potential exists for unintended consequences. Because HHAs and SNFs are no longer incentivized to provide therapy, they might withhold the provision of needed rehabilitation therapy.10 Furthermore, because payments are based on patient coding by HHA and SNF providers under the new payment models, coding practices may change in order to optimize their payments. Indeed, the PDGM policy for HHAs includes a “behavioral adjustment” to account for anticipated changes in improved documentation by HHAs. Because LTACHs will be less likely to admit patients without prolonged mechanical ventilation or a qualifying ICU stay of 3 or more days, these patients might remain in the hospital for longer periods of time if they are too sick or their care needs are too complex for other PAC settings. Given these possible unintended consequences, the implications for hospital discharge patterns, PAC access, and quality of care will need to be closely monitored, as it is unclear how these PAC policy changes will impact patient care.
In terms of broader payment reform, the four PAC settings are still fragmented, with little effort to unify payment, regulation, and quality across the PAC continuum. As required by the Improving Medicare Post-Acute Care Transformation (IMPACT) Act of 2014, we would encourage the adoption of a unified PAC payment system that spans the four settings, with payments based on patient characteristics and needs rather than site of service.12 This type of reform would also harmonize regulation and quality measurement and reward payments across settings. Currently, CMS is standardizing patient assessment data and quality metrics across the four PAC settings. Given the COVID-19 pandemic, the transition to a unified PAC payment system is likely several years away.
WHAT SHOULD I TELL MY PATIENT?
For our patient who was transferred to an LTACH after referrals to SNFs were denied, PAC options now differ following these major PAC policy reforms, and SNF transfer would be an option. This is because SNFs will receive higher payment for providing respiratory support under the PDPM, and LTACHs will receive considerably lower reimbursement because the patient did not have a qualifying ICU stay or require prolonged mechanical ventilation. Furthermore, hospitals participating in accountable care organizations would achieve greater savings, given that LTACHs cost at least three times as much as SNFs for comparable diagnoses.
Instead of referring this patient to a LTACH, the care team (hospitalist, discharge navigator, and case manager) should inform and educate the patient about discharge options to SNFs for weaning from respiratory support. To help patients and caregivers choose a facility, the discharge planning team should provide data about the quality of SNFs (eg, CMS Star Ratings scores) instead of simply providing a list of names and locations.13,14 Discharge planning should start as soon as possible to permit caregivers an opportunity to visit facilities and for the providers to coordinate the transfer as seamlessly as possible.
Recent major PAC policy changes will change where hospitals discharge medically complex patients and the services they will receive at these PAC settings. Historically, reduction in PAC use has been a key source for savings in alternative payment models that encourage value over volume, such as accountable care organizations and episode-based (“bundled”) payment models.15 We anticipate these PAC policy changes are a step in the right direction to further enable hospitals to achieve value by more closely aligning PAC incentives with patients’ needs.
Dr Makam received travel expenses from the National Association of Long Term Hospitals (NALTH) for an invited presentation in February 2020. Dr Makam has also received a research grant from NALTH. Dr Grabowski received research grants from the National Institute on Aging, the Agency for Healthcare Research and Quality, The Donaghue Foundation, the Warren Alpert Foundation, and the Arnold Foundation. He received personal fees from naviHealth, the Medicare Payment Advisory Commission, Compass Lexecon, Analysis Group, Abt Associates, and the Research Triangle Institute. The authors have no other conflicts of interest to disclose, financial or otherwise.
Dr Makam was supported by the National Institute on Aging (K23AG052603). The study sponsors had no role in the preparation, review, or approval of this manuscript.