Hospital‐at‐Home: Multistakeholder Considerations for Program Dissemination and Scale

Abstract

Policy Points

• Hospital‐at‐Home (HaH) is a home‐based alternative for acute care that has expanded significantly under COVID‐19 regulatory flexibilities.

• The post‐pandemic policy agenda for HaH will require consideration of multistakeholder perspectives, including patient, caregiver, provider, clinical operations, technology, equity, legal, quality, and payer.

• Key policy challenges include reaching a consensus on program standards, clarifying caregivers’ issues, creating sustainable reimbursement mechanisms, and mitigating potential equity concerns.

• Key policy prescriptions include creating a national surveillance system for quality and safety, clarifying legal standards for care in the home, and deploying payment reforms through value‐based models.

In 2019, us hospitals accounted for 36 million admissions and $1.2 trillion in spending (31% of national health expenditures). ¹ Acutely ill older adults, specifically Medicare Part A enrollees with chronic diseases, accounted for 10.5 million discharges in 2019. ² , ³ Inpatient care, however, is characterized by inefficiencies, as articulated in the Institute of Medicine's report, Crossing the Quality Chasm. ⁴ Research has also identified “post‐hospitalization syndrome,” which encompasses physical decline (e.g., decreased mobility), mental complications (e.g., delirium), and an increased risk of 30‐day readmissions, all of which are attributable to the allostatic stress of facility‐based care. ⁵ , ⁶ These gaps in care access and quality are amplified when examined through an equity lens, with disparate outcomes in readmissions, morbidity, chronic disease management, and overall health among racial minorities. ⁷

Nevertheless, population aging is expected to intensify the utilization of hospital services. This and the gaps in hospital quality exposed by the COVID‐19 pandemic highlight the need to redesign acute care to better serve older adults and minorities. Consequently, the National Academy of Medicine (NAM) convened a multistakeholder meeting to discuss the post‐pandemic future of acute care, with a particular focus on Hospital‐at‐Home (HaH) as an alternative to facility‐based approaches. The meeting's objectives were to (1) define the current scope of HaH care models, (2) create a shared understanding across stakeholders of the priority areas and challenges for HaH, and (3) identify sector‐specific levers to maximize HaH's benefits and scalability.

Hospital‐at‐Home: Definitions and Model Components

HaH programs are defined by the presence of several operational features. ⁸ First, the model must substitute a portion or the entirety of inpatient acute care with hospital‐level, at‐home treatment. This acuity criterion distinguishes HaH from other home‐based care models. For example, some home‐based models are specific to nonacute services (e.g., hospice care) or are intended to be longitudinal (e.g., home‐based primary care), as opposed to episode based. In terms of staffing, multidisciplinary teams composed of the same types of providers as in traditional inpatient care (e.g., physicians and nurses) are typically deployed, along with other allied professionals as needed (e.g., physical therapists). HaH clinicians also frequently leverage a combination of digital technologies, such as remote patient monitoring (RPM), mobile apps, and enhanced video visits. A potential benefit of HaH is its flexibility to integrate wraparound services (e.g., meal deliveries) tailored to a patient's needs and social context.

Second, HaH models must be limited to a specified geographic catchment area because in the event of clinical deterioration, proximity to the hospital facilitates timely care escalation. In contrast, other home‐based care models have a lower risk of rapid escalation due to the stage of intervention (e.g., chronic disease management) or level of service intensity (e.g., community‐level services).

Narrative Overview of Hospital‐at‐Home's Global and Domestic Evidence Base

HaH's evidence base is long‐standing, with the initial programs implemented by international health systems like those in Australia, New Zealand, and the United Kingdom. ⁹ In the United States, HaH was first piloted in the 1990s and then later studied nationally as part of a multisite demonstration model involving Medicare Advantage plans and a Veterans Affairs medical center. ¹⁰ , ¹¹ Evaluations of these early domestic and international programs were highly favorable, as patients treated under HaH experienced similar or improved outcomes relative to those of patients treated under facility‐based care, such as readmission rates, length of stay, and complication rates. ¹² Subsequent systematic reviews and meta‐analyses have also corroborated these associations, although causal inference is limited by the wide variability in progam structure (i.e., ranging from a complete multidisciplinary team to partial teams, to a single nurse with physician oversight) and significant representation of trials from non‐US settings. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ This latter point is significant because the organization, financing, and delivery of acute care in the United States are demonstrably different from those in other high‐income countries; specifically, acute care in the United States is less affordable and associated with more administrative inefficiencies, worse inequities, greater access barriers, and higher treatment intensity. ¹⁸

US‐based trials and observational studies have reported clinical and statistically significant improvements in the outcomes and reductions in spending for patients enrolled in HaH programs. ¹⁹ , ²⁰ , ²¹ , ²² Several countries have since established HaH programs as a standard acute‐care option. For example, HaH accounts for more than 5% of total inpatient bed capacity in Victoria, Australia, and countrywide evaluations reveal lower mortality and readmission rates, longer lengths of stay, and more complex conditions for HaH patients. ²³ Despite a reassuring evidence base and unanimous endorsement by the Physician‐Focused Payment Model Technical Advisory Committee (PTAC), US implementation of HaH remained limited before the onset of COVID‐19. ²⁴

COVID‐19 Regulatory Flexibilities and Hospital‐at‐Home

In response to projected shortfalls in bed capacity and the anticipated number of severely ill COVID‐19 patients, the Centers for Medicare & Medicaid Services (CMS) in March 2020 initiated the “Hospitals Without Walls” (HWW) program. ²⁵ , ²⁶ This waiver allowed health systems to transfer lower‐acuity inpatients to alternative care sites (e.g., ambulatory surgery centers). ²⁶ While HWW did not encompass HaH, the model's shared ethos—decentralization of acute care into community‐based settings—led to strong policy interest in HaH as a pandemic response strategy. For example, the Mount Sinai Health System reported tripling monthly patient volumes for its HaH program in response to the more than 2,000 COVID‐19‐related admissions in the spring of 2020. ²⁷

As caseloads rose, CMS in November 2020 announced an expansion of HWW called the “Acute Hospital Care at Home” (AHCaH) program, which provided HaH‐specific regulatory flexibilities (Table 1). ²⁸ Any hospital applying for an AHCaH waiver was required to first demonstrate that its proposed model would comply with all of Medicare's Conditions of Participation for inpatient care, with the exception of 24/7 nursing service requirements, life safety code, and physical environment, which were waived for HaH. Hospitals also had to satisfy HaH‐specific operational criteria for eligibility: allowing twice daily in‐person visits, addressing in‐home emergency requests within 30 minutes, and reporting the volume of patients, unanticipated mortality, and escalation rates. Hospitals could admit patients into their HaH model from the emergency department or inpatient floor, with no diagnosis‐specific limits on eligibility. Approved providers were to be reimbursed at the same severity‐adjusted diagnosis‐related group rates as those for facility‐based admissions.

Table 1.

Key Features of the CMS Acute Hospital Care at Home Waiver

Key Components of Waiver
Eligible population	Medicare beneficiaries who require an acute inpatient admission Admission only from emergency departments or inpatient beds
Core operational requirements	HaH program must be integrated with a brick‐and‐mortar hospital In‐person physician evaluation required before initiation of HaH Utilization of established patient‐leveling process Daily nurse evaluation (in‐person or virtual) and twice‐daily in‐person home visit Capability to respond to decompensating patient within 30 minutes Establishment of patient safety committee
Reporting measures	Patient volume Unanticipated mortality during an acute episode of care Escalation rate Safety committee review Patient list
Reimbursement	Same payment process as a traditional admission under the Inpatient Prospective Payment System (Medicare Severity‐Diagnosis Related Group)

As of March 2022, the AHCaH program had expanded to 205 hospitals in 34 states, representing 92 unique health systems. Most of these providers delivered care to 1,878 patients during the waiver's first year of activity (Figure 1). ²⁹ Although AHCaH's flexibilities are time‐limited for the duration of the public health emergency, several national health systems and insurers have committed to scaling HaH beyond COVID‐19. ³⁰ Furthermore, the pandemic drew attention to the feasibility of leveraging the home as an effective site‐of‐care for both low‐ and high‐acuity conditions, entrenching an already documented preference among older adults to age in place.

Figure 1.

Health Systems Approved to Participate in the Acute Hospital Care at Home Waiver Program. [Colour figure can be viewed at wileyonlinelibrary.com]

Listed programs reflect health systems with an approved waiver from CMS as of March 30, 2022; however, not all approved programs may be active at this time. Health systems with HaH waivers covering hospitals in multiple states (Adventist Health, Baptist Memorial Health Care, ChristianaCare, Cleveland Clinic, Mayo Clinic, and ProMedica) are listed in the state where the health system is headquartered.

While the experience to date under CMS's waiver has been positive, AHCaH's principal purpose was to advance the development of new care models in response to an exigent public health need, not to support a permanent transformation of acute care delivery. In a recent analysis, the agency noted that health systems’ experience with HaH during COVID‐19 could help “determine the next steps for this program within CMS's statutory authority once the [public health emergency] concludes.” ²⁹ To inform those next steps, policymakers may benefit from a multistakeholder consultative process to guide resource allocation, codify new systems of care, and inform the development of a robust infrastructure for payment reform, program standards development, quality measurement, credentialing, and oversight. ³⁰ , ³¹ Distilling the lessons learned from implementing care delivery innovations during the pandemic would also help policymaking. Last, greater vigilance will be necessary to mitigate any unintended risks from care inequities, professional liability, data privacy breaches, and patient safety issues in the home environment.

Sector‐Specific Program and Policy Considerations

The NAM convened a virtual expert meeting in July 2021 pursuant to the preceding objectives and sought to ensure a diversity of perspectives that were representative of the health care ecosystem. To this end, we invited experts in HaH implementation and adjacent sectors (health technology, payment policy, bioethics, health law, equity, and system strategy) to participate. This expansive framing allowed us to integrate the various perspectives necessary to best inform the evolution of the HaH program under CMS's statutory authority beyond the pandemic. ²⁹ A full list of the meeting's participants is provided in the acknowledgments. Each attendee presented a sector‐specific (e.g., patient, caregiver, provider, legal, equity, payer) perspective on HaH and then engaged in a “premortem” exercise in which he or she was asked to name a specific challenge and policy recommendation for HaH. ³² The participants then ranked the proposed considerations and interventions in a post‐meeting survey. NAM staff next distributed the survey results to the attendees and invited interested participants to create this article synthesizing the key discussion points from the meeting. The sector‐specific perspectives for HaH are described next and also are summarized in Table 2. Note that several of the opportunities and challenges listed in Table 2 are not unique to HaH and are also relevant to traditional hospital facilities.

Table 2.

Sector‐Specific Policy Considerations for Hospital‐at‐Home

Stakeholder Perspectives	Opportunities Under HaH	Challenges Under HaH	Key Program and Policy Considerations
Patient	Reduce the risk of hazards of hospitalization and post‐hospital syndrome Align with preferences to age in place and minimize disruptions to daily life	Concerns about quality and safety of HaH Concerns about the amenability of the built environment of the home to patient care	Incorporate culturally competent communication to address patient‐ and population‐specific concerns Develop informed financial consent policies to reduce surprise billing for patients Collect and report demographic‐level data and incorporate evaluations of patient‐reported outcomes and experience measures
Caregiver	Reduce the onset of additional complications requiring the caregiver's attention Increase the convenience of integrated home‐based care delivery	Risk of task‐shifting from providers to caregivers Concerns about hidden costs and further disruptions to caregivers’ daily lives	Assess caregiver needs and provide education on HaH Offer resources for caregivers' respite, training, and financial assistance Incorporate caregivers' well‐being into quality metrics
Provider	Reduce stress and burnout among clinicians, especially hospitalists	Coordinate with different provider teams Have experience with virtual and home‐based care delivery	Collaborate with professional societies to establish a training curriculum for “home hospitalists” Create best‐practice recommendations and redesign clinical workflows, particularly for patient handoffs and care escalation Measure stress and burnout among HaH staff
Clinical Operations	Provide a streamlined acute care experience for patients and providers	Risk of overtreatment and inappropriate admission of low‐acuity patients Challenges with coordinating across hospitals, providers, and contractors	Develop clear, site‐specific eligibility criteria for HaH Invest in infrastructure and enabling technologies
Technology	Improve access to care through virtualization and decentralization Improve efficiency of care coordination and patient‐provider communication	Limitations of technology for addressing upstream drivers of health Risk of device failure or cybersecurity breaches	Accommodate gaps in resources and accessibility for installing and using technologies in the home Ensure home‐based deployments include redundancies against single points of failure Develop consensus‐based guidelines for virtual visit protocols for HaH
Equity	Increase access to acute care services Improve quality and outcomes of acute care	Risk of recreating existing inequities in facility‐based acute care Assurance that an enabling infrastructure is in place for rural, safety net, and critical‐access hospital participation	Use HaH as a vehicle for integrating nonmedical services to address the social determinants of health Develop a transitional payment model to support resource‐constrained providers
Legal	Offer an opportunity to update tort law and statutory definitions of standard of care	Challenges from new sources of liability and with the distribution of liability Creation of new risks for privacy and data integrity following integration of informational inputs across platforms	Define a legal standard for care at the home Clarify how existing statutes (e.g., HIPAA) and rules (e.g., Conditions of Participation) will apply to HaH
Quality	Reduce the risk of complications and further functional decline for hospitalized patients	Challenges to managing quality of services and equipment during home‐based care delivery Evidence gaps regarding use cases and best practices for HaH	Coordinate across CMS and the Joint Commission to develop quality metrics for HaH Develop a national surveillance system for adverse events Invest in rapid cycle evaluation methodologies to generate evidence for HaH
Payer	Advance a lower‐cost model of acute care Decrease follow‐on acute care spending by reducing readmission rates	Challenges to managing institutional culture and process changes in CMS's payment methodologies to facilitate HaH Outstanding questions regarding episode pricing and reimbursement construct (e.g., FFS versus APM)	Advance a new reimbursement structure for HaH rooted in value‐ and‐risk‐based models Create a national learning network for HaH to support public reporting of outcomes and dissemination of best practices Clarify how Medicare's payer functions for HaH will operate in relation to existing prospective payment systems

HaH: Hospital‐at‐home; HIPAA: Health Insurance Portability and Accountability Act; CMS: Center for Medicare and Medicaid Services; FFS: Fee for service; APM: Alternative payment model

Patient Perspective

Meaningful policy reforms must always be designed around patients’ needs. Although patients have generally expressed high satisfaction rates with HaH, pre‐pandemic studies indicate that patients’ participation rates may vary because of several factors. For example, some patients who opted against participating expressed concerns about the sufficiency, safety, or convenience of alternatives to facility‐based care. Others cited social factors like concern for the caregiver's burden and the amenability of their home's built environment to care delivery (e.g., space constraints) as reasons for declining to enroll in HaH. ³³ , ³⁴ Although demographic factors have generally not affected enrollment in randomized controlled trials for HaH, one retrospective analysis of HaH cohorts in Australia did identify differential participation rates according to socioeconomic status and non‐English speakers. ³⁵ To maximize access to and benefits of HaH for all populations, these programs will need to develop culturally and linguistically appropriate communication practices regarding the experience of HaH care, potential benefits, and escalation protocols to address safety concerns. HaH's leadership will also need to ensure that models are adapted to different operating environments (e.g., urban versus rural, specialty hospitals) and are given the resources to minimize disruptions to patients’ daily lives. Examples of policy interventions could include the development of a HaH‐specific version of the Consumer Assessment of Healthcare Providers and Systems survey.

Caregiver Perspective

Following an admission to a “brick‐and‐mortar” hospital, the processes related to hospitalization, discharge, and support of older adults with acute illness have been shown to place a substantial burden on caregivers. The average time commitment (24 hours per week) and out‐of‐pocket costs ($7,000 per year) of caregiving on top of professional and family responsibilities have led more than 30% of caregivers to report insufficient sleep and nearly 20% to report 14 or more physically or mentally unhealthy days during the past month. ³⁶ , ³⁷ Consequently, a potential benefit of HaH programs could be alleviating caregiver burden that is attributable to facility‐based acute care. ³³ Indeed, in the few HaH trials that have measured caregiver burden, there has generally been either no change or a reduction in caregiver stress. ¹⁵ , ³³ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ However, the exact responsibilities of caregivers under HaH remain inconsistently defined across programs, and some patients report perceiving additional caregiving burdens and costs (e.g., for utilities) with HaH. ³⁴ , ⁴² , ⁴³ Therefore, efforts to convert the home into a site‐of‐care should be accompanied by policies that minimize the risk of task‐shifting, additional expenses, and disruptions to daily life. Examples of program‐level interventions are routinely using validated assessment tools to determine caregivers’ burden, providing financial assistance to caregivers, implementing temporary relief from day‐to‐day responsibilities (i.e., respite for caregivers), and offering HaH‐related training. In addition, policymakers could better address concerns by explicitly identifying direct and indirect costs to caregivers (e.g., from food delivery, transportation, utilities) and codifying expectations for program coverage. Finally, policymakers could explore including caregiver well‐being in HaH quality metrics and could mandate alignment across all HaH programs with the core tenets and spirit of the CARE (Caregiver Advise, Record, Enable) Act. ⁴⁴ This state‐based legislation promotes the inclusion of caregivers and requires that health systems provide them with the necessary training in the medical tasks and care activities to be rendered in the home. ⁴⁴

Provider Perspective

A major development in American medicine has been the advent of the hospitalist, a clinician workforce trained specifically to manage the care of hospitalized patients. ⁴⁵ Successfully moving acute care into the home may require a new model of “home hospitalists,” that is, providers possessing additional competencies with home visits, technology‐enabled services, and nonmedical service coordination. ⁴⁶ Likewise, the prevailing model of graduate medical education must be modernized to include formal training in clinical informatics, health care operations, implementation science, and telemedicine services. ⁴⁷ To prepare for health care's future, orientation and exposure to in‐person and remote house calls across a wide range of clinical encounters, from synchronous one‐on‐one visits to complex, multidisciplinary team care, are a must. ⁴⁸

Scaling HaH will also require program administrators to invest substantially in optimizing clinical workflows (e.g., managing care escalations) and in training multidisciplinary teams (e.g., navigating patients’ handoffs from “admission” to “discharge”). But a shift to HaH may also portend positive spillover effects for traditional facility‐based providers. For example, evidence from pilot programs suggests that implementing HaH is associated with low rates of burnout and high rates of job satisfaction among clinical staff. ⁴⁹ Plausible mechanisms include low patient census, a flat hierarchical structure, and effective communication practices. ⁴⁹ This is in contrast to the substantial research documenting greater burnout among hospital‐based clinicians, a trend that has been exacerbated by COVID‐19. ^{50 . 51} While further evidence is needed, this development highlights the potential for advances in clinical practice and new paradigms of collaborative, team‐based care to improve clinician well‐being. Professional societies and certifying boards should work to standardize the training and accreditation process for HaH clinicians (e.g., physicians, skilled nurses), to address issues related to state‐based licensing, and to collaborate with quality stewards to develop metrics related to provider burnout. Last, professional societies and participating HaH providers should work together to inform future iterations of the care model by creating consensus‐based recommendations that standardize clinical workflows, outline preferred frequencies of virtual visits, define professional conduct, and help create key performance indicators specific to a virtualized care encounter. ⁵²

Clinical Operations Perspective

HaH care necessitates standardized protocols, adherence to well‐defined quality and safety standards, and complex reimbursement procedures. As the US experience expands in scope, program leadership must contend with the challenges of contextualizing these processes for different patient populations (e.g., cancer, pediatric, psychiatric, and post‐surgical patients). ⁵³ Adaptations include condition‐specific workflows, care escalation triggers, exclusion criteria, and clinical protocols. Procedures must also be developed for “admitting” patients from multiple starting points beyond the emergency room (e.g., home, infusion centers, and ambulatory clinics). During the pandemic, CMS also required any hospital participating in the AHCaH waiver to use defined eligibility criteria, with hospitals either designing their own protocols or drawing from existed, validated methods. ⁵⁴ , ⁵⁵

Accelerator programs like the partnership between CaroNova and Ariadne labs will be an important way to disseminate HaH care models via expert coaching and a collaborative ecosystem of hospitals that are also implementing HaH. ⁵⁶ A key area of focus will be the use of exclusion criteria to appropriately segment or identify patients who may be too sick for HaH, either as a complete hospital substitution or a length‐of‐stay reduction strategy (i.e., transferring from inpatient care to finishing hospital‐level care at home). By optimizing site‐of‐care decision making, we will ensure that the right patient is receiving the right care, at the right time, and in the right setting; be it in HaH or a traditional hospital facility.

Second, hospitals must invest in the necessary information technology (IT) infrastructure to support HaH. This includes the development of an interoperable electronic medical record (EMR) that can be accessed remotely by partner organizations and perhaps also by patients and caregivers (e.g., documenting medication administration). Likewise, hospitals need to ensure that enabling technologies (e.g., Wi‐Fi for RPM) and medical equipment (e.g., for oxygen delivery) are made available in the patient's home. Depending on the HaH institution's operating environment, these requirements could pose a challenge. For instance, safety‐net and critical access hospitals may face resource constraints related to acquiring the infrastructure (e.g., IT, digital technologies) needed to set up a HaH program.

Third, many hospitals are likely to contract with outside entities (e.g., home health agencies) as part of their model and consequently must be prepared to ensure that relevant partnerships are grounded in effective communication practices and adhere to applicable regulations. For example, health systems may encounter state‐specific challenges regarding licensure for the provision of home‐based care that may impose an added regulatory hurdle for HaH programs.

Fourth, although HaH programs under the CMS waiver are reimbursed under the inpatient prospective payment system, hospitals and health systems still need to collaborate with commercial payers to determine the most appropriate contractual vehicle for reimbursing home hospitalizations (e.g., bundled payment, total capitation) to both cover the start‐up costs associated with launching HaH and deliver on the promise of savings. These reimbursement methodologies should ideally be aligned with a patient's existing benefits structure to avoid exposure to surprise billing and “financial toxicity.” Although surprise billing is not unique to HaH‐enabled care, the underlying mechanism is distinct, as it remains a delivery model that is not recognized or sufficiently explained in many commercial payers’ existing rules and regulations. While our hope is that this policy vacuum is short‐lived, it still poses the real risk of catastrophic costs being passed on by a third‐party provider in the absence of a prudent screening and preapproval process. The development of processes for “financial informed consent” that ascertain at the point of HaH admission whether potential enrollees are liable for surprise bills, will go a long way to engendering trust in a purportedly lower‐cost and equivalent‐outcome acute care model. ⁵⁷

Finally, several operational questions may also require action through rule making and guidance. For example, Medicare Advantage plans may benefit from policymakers’ clarification of benefit structures, and state Medicaid programs and professional clinical societies may need to collaborate to address issues around provider licensing.

Technology Perspective

HaH also represents a paradigm shift away from wholly in‐person care to hybrid models that virtualize some components of the care journey. Key technologies (e.g., RPM and an interoperable EMR) can improve patients’ access and care efficiency by enabling real‐time communication and feedback among patients, caregivers, and their provider team. However, the adoption of digital technology in health care—from EMRs to telemedicine—has historically been difficult, requiring the realignment of financial incentives, capital investments, and a commitment to transparency and regulatory safeguards. In the case of HaH, providers will need to ensure that technologies are able to accommodate potential resource gaps (e.g., patients lacking a robust internet connection), languages other than English, and shared caregiving responsibilities. An important concern will be system resiliency, ensuring that the home‐based technology infrastructure include redundancy against “single points of failure” (e.g., backup routers) and robust cybersecurity protection against data breaches of protected health information. Policymakers may seek to issue expectations about product security, the performance evaluation of digital tools in the home, and best practices for complying with data protection.

Equity Perspective

HaH may provide a platform for addressing structural inequities in health care, with evaluations of prominent US‐based programs reporting that HaH participants are more likely to be enrolled in publicly financed insurance programs and that underserved populations may well have similar and sometimes superior outcomes. ¹⁹ , ²² , ⁵⁸ As programs scale up, policymakers and HaH administrators should proactively monitor and mitigate equity risks, such as those reported in Australia, where patients living in the most disadvantaged areas were 50% less likely to participate in HaH. ³⁵ Policymakers should mandate rigorous data collection and public reporting, in a standardized manner, of HaH program outcomes stratified by race/ethnicity, age, gender, primary language, and social risk factors.

HaH programs could consider how this model might address existing shortcomings in facility‐based care, such as the ability to identify and tackle the social determinants of health (i.e., access to nutritious foods and clean water). In addition, by bringing care into the home, HaH inverts the traditional power dynamics of a care encounter (i.e., treating someone in the context of his or her life) and, in the process, engenders more trust among patients, especially those from marginalized communities. ³¹

At the facility level, providers’ finances may limit the implementation of HaH due to the cost of necessary upfront investments in infrastructure (e.g., at‐home supply chain), technology platforms, and personnel training. Such investments may be infeasible for safety‐net, rural, and critical‐access hospitals, given that they often lack financial flexibility, and further exacerbate the gap in acute care access for economically deprived communities. To ensure that HaH programs cover all patients and providers, policymakers could consider financing reforms. For example, reimbursement for HaH programs sponsored by rural hospitals could be treated as a part of the existing total cost of care models (e.g., the Pennsylvania Rural Health Model, Maryland All‐payer Total Cost of Care Model), and urban safety‐net HaH programs could receive upfront resources and a prepayment model (e.g., analogous to the Advanced Payment ACO model). ⁵⁸ , ⁵⁹ While some of these demonstrations were ultimately discontinued by CMS, they demonstrated the feasibility of advanced payments to support the transformation of delivery models in historically underserved settings and may be applicable to HaH following the provision of technical support for program development and the appropriate refinement of timelines for participation and expected savings. ⁵⁹ Finally, policymakers should also consider how payment reforms for HaH could (1) better integrate the delivery of social and behavioral health services and (2) best support the development of necessary partnership models.

Legal and Regulatory Perspective

Establishing a HaH program will require thoughtful deliberation about how existing legal issues for hospital care are applicable (e.g., provider credentialing, facility accreditation, standards for patient privacy). While to our knowledge there has been no evidence of increased malpractice issues for HaH, it is worth unpacking the additional legal complexity that hospitals may need to consider as HaH programs scale up. For example, do the legal standards for acute care differ in the community‐based setting, and is an accreditation process needed for the home environment? ⁶⁰ In the case of data integrity, some HaH models leverage interoperable EMRs capable of integrating information from multiple entities, which can increase the efficiency of care delivery but also introduce new risks related to data breaches. These issues are salient because the integration of wearables and RPM devices increases the vulnerability of health data to cyber criminals. ⁶¹ , ⁶² Unfortunately, the existing privacy protections and regulatory oversight of home‐monitoring technologies (e.g., minimum cybersecurity standards, commercialization of health data collected from patients by technology vendors) in the United States are not as comprehensive as the European Union's General Data Protection Regulation. ⁶² Policymakers will need to examine how the application of existing statues (e.g., Health Insurance Portability and Accountability Act) and rules (e.g., Conditions of Participation [COP] for 24‐hour nursing availability) may differ for HaH, and what new frameworks may be needed. ⁶⁰

Updating existing hospital policies and procedures to account for HaH's virtual beds may obviate the need for HaH‐specific COP related to certain basic hospital functions, particularly the requirements for physical environment and performance improvement programs. ⁶³ Finally, state‐level regulatory definitions for hospital‐level care, licensed hospital beds, Clinical Laboratory Improvement Amendments (CLIA) exclusions, and oversight of controlled medications, among others, must be harmonized. ⁶³ Clarifying these issues would be enormously helpful for health system leadership as they weigh strategic decisions regarding the scope of programs and interstate expansion.

Quality Perspective

Evidence from the extensive domestic and international studies conducted to date shows HaH to be a safe and highly effective intervention with associated mortality rates that are less than facility‐based care. Nevertheless, policymakers will likely require more evidence at the population level to fully vet program safety and to validate quality measures, with CMS highlighting the limitations of drawing conclusions from AHCaH for nonpandemic care. ²⁹ Participating programs in the AHCaH waiver must report two quality metrics: the care escalation rate (defined as the rate of patients transferring to a traditional hospital facility because of their preference or deterioration in medical condition) and unexpected mortality. ²⁹ According to a recent commentary article published by CMS, the overall escalation rate of 7.14% and unexpected mortality rate of 0.43%, for the AHCaH experience to date, represent improvements from published reports ²⁹ of established HaH programs and facility‐based care, respectively.

The rate at which hospitals and payers adopted HaH during the pandemic and independent of CMS's actions suggests that the traditional health services research model of randomized controlled trials with multiyear time frames will be unsuitable for generating the kind of timely data needed to inform regulatory decision making. Instead, policymakers could use HaH as an opportunity to channel the ethos of a learning health system by conducting natural, rapid cycle experiments. ⁶⁴ , ⁶⁵ As the evidence accumulates, quality stewards like CMS will need to develop structural requirements and process‐of‐care measures tailored to the nuances of the home environment. The HaH Users Group, a collaborative of North American HaH programs and industry and academic partners that helped inform CMS's development of AHCaH, has produced a series of guidance documents and standards for HaH programs that will be useful for setting future quality standards and regulations. ²⁹ , ⁶⁶ Last, policymakers could use quality measurement as a lever for integrating an equity lens (e.g., separating outcomes by zip code and race) and the perspectives of different stakeholders (e.g., requiring the use of patient‐ and caregiver‐reported outcomes).

Payer Perspective

The durable expansion of HaH beyond the pandemic will require the creation of a long‐term payment structure. First, payers need to determine the appropriate pricing of a HaH episode of care and whether reimbursement for those costs will continue to occur under a fee‐for‐service (FFS) framework or transition to alternative payment models. Although subjecting HaH to reimbursement under FFS Medicare was necessary in order to jumpstart program development during the pandemic, paying for HaH at rates equivalent to those for facility‐based care according to patient volume would leave HaH programs vulnerable to demand volatility, negate the purported savings that could be achieved from using less costly care sites, and significantly heighten concerns about the imminent risk of insolvency of Medicare's hospital trust fund. ⁶⁷

In contrast, the use of alternative payment models—either a HaH‐specific, episode‐based payment model that is adjusted for quality as suggested in the PTAC‐endorsed proposal or the nesting of HaH within existing constructs like Accountable Care Organizations (ACOs) or global budgets, could provide more sustainable revenue streams. Data gathered via the AHCaH waiver will offer insight into the typical resource burden of a HaH admission, which in turn can inform the development of reimbursement policies.

Second, CMS will need to consider how payment arrangements can account for the potential savings from home hospitalizations while recognizing that providers will face start‐up costs when establishing HaH programs. Policymakers could explore the possibility of a shared savings model for HaH that phases in risk over time.

Third, irrespective of the specific payment arrangement, CMS will need to develop infrastructure for evaluating costs and savings and to monitor potential adverse consequences for health equity based on the experience of providers caring for underserved populations with other value‐based models. ⁶⁸ Last, addressing each of these technical considerations for HaH reimbursement will require agency officials to consider how foundational CMS processes (e.g., rate setting, claims adjudication) may need to evolve to support a care paradigm operating beyond the scope envisioned by the prospective payment system.

Post‐pandemic Policy Agenda

Before the AHCaH waivers expire, CMS will need to offer guidance for a permanent regulatory and payment framework as well as quality standards for HaH. It would be helpful to review data generated from HaH models established during the pandemic and to provide a public report of key trends regarding participating patient populations (e.g., stratified by race, diagnoses used at admissions), service utilization (e.g., most common diagnosis‐related groups), and clinical outcomes (e.g., length of stay, care escalation rates). As part of this evaluation process, CMS should also convene HaH stakeholders in order to learn from their experiences designing, implementing, and operating this model. This will help identify outstanding research and operational questions and could evolve into a formal community for promoting continuous improvement in HaH, in a way analogous to the Health Care Payment Learning & Action Network. ⁶⁹

The next step in transitioning HaH programs into a national demonstration model would be initiating rapid‐cycle, pseudo‐randomized mandatory evaluations, an approach informed by the agency's experience in which hospitals in selected metropolitan statistical areas were randomly assigned to participate in the orthopedics bundled payment model. ⁷⁰ This approach will generate adequate causal inference and generalizability regarding the redesign of acute care. Ideally, the implementation should be multipayer (i.e., commercial and Medicaid) and be inclusive of both the legacy and the pandemic‐borne HaH programs.

Evidence from these evaluations can guide the development of national quality standards that build on conceptual work by the HaH Users Group and are integrated into prospective payment reforms. CMS can use HaH to accelerate the transition to value‐based payment, using either existing models of episode‐based or capitated payment as references, and drawing from Mount Sinai's PTAC‐endorsed proposal as an example. ²⁴ While the specifics of the payment model—pricing of the episode, application of risk adjustment, distribution of shared savings—ill require iteration over time, a core focus of the agency should be establishing a permanent mechanism outside FFS that encourages provider participation, offers accommodations to the different types of health systems, and can serve as a model for multipayer reform. In turn, the payment should drive the broad adoption of quality standards. CMS can consider using HaH as a “burning platform” for transforming the quality measurements of both inpatient care (e.g., greater attention to the hazards of hospitalization) and community‐based care (e.g., defining relevant metrics unique to the home environment).

Conclusions

America's aging population and the increasing fraction of national health expenditures attributed to hospital care have generated an impetus for transforming the care of acute illness. The COVID‐19 pandemic has called attention to HaH as a promising model for reimagining acute care. However, realizing the potential benefits of HaH requires consideration of the risks associated with diffusing a new care model, from the impact on specific stakeholders to the adaptation of regulations for payment, quality, technology use, and legal considerations. By reviewing the key opportunities and challenges from the perspective of different sectors, policymakers can be better positioned to guide the regulation and implementation of HaH after the pandemic.

Acknowledgments: The authors thank each of the participants at the NAM's 2021 virtual expert meeting on Hospital‐at‐Home for their insight. The participants’ names and affiliations at the time of the meeting are lied in alphabetical oer: Doug Clarke, MD (Medical Officer, CMMI); I. Glenn Cohen, JD (Deputy Dean, Harvard Law School); Patrick Conway, MD, MSc (CEO, Care Solutions at Optum); Lee Fleisher, MD (CMO, CMS); Harlan Krumholz, MD, SM (Professor, Yale School of Medicine); Omar Lateef, DO (CEO, Rush University Medical Center); Lewis Levy, MD (CMO, Teladoc); Michelle Mello, JD, PhD (Professor, Stanford Law School); Travis Messina, MBA (CEO, Contessa Health); Margaret O'Kane, MHA (President, NCQA); Nirav Shah, MD, MPH (CMO, Sharecare); Albert Siu, MD (Professor, Mount Sinai Icahn School of Medicine); James Weinstein, DO (SVP, Microsoft Healthcare); Consuelo Wilkins, MD, MSCI (SVP, Vanderbilt University Medical Center).

Conflict of Interest Disclosures: This perspective has not been endorsed by the National Academy of Medicine and is not intended to be a factual reproduction or to imply the consensus of the planning meeting's participants.

Funding/Support: The authors of this perspective had no sources of support.