Early experience and lessons following the implementation of a Hospital‐at‐Home program

Healthcare is at a crossroads. Hospital admissions are increasing as staffing challenges remain—exacerbated by workforce departures, increased labor costs, and facility closures.1-3 Capacity constraints have led to increasing hospitalist censuses burnout, and uneven patient outcomes.4 Sustainably meeting demand for hospital care will require alternative approaches to acute care delivery. In contrast to surgery, the hospital medicine care model has evolved minimally over the last decade. As surgery transitioned from open to laparoscopic procedures and adopted enhanced recovery pathways to shift the surgical care continuum increasingly into the home, the practice of hospital medicine remained anchored in traditional brick-and-mortar (BaM) facilities, with domestic efforts to adopt Hospital-at-Home (HaH) limited by regulatory, financial, and operational constraints as well as by patient and provider preferences.5-7 The passage of the Acute Hospital Care at Home Program Waiver by the Centers for Medicare and Medicaid coupled with growing financial and staffing challenges, has renewed interest in shifting acute care into the home. Although the Waiver helped ease reimbursement barriers to HaH, the operational demands associated with designing and executing this acute care model are significant, and hospitalists are often the key physician leaders in these implementation efforts. Leveraging a hospitalist-led approach and partnership with an external company, we experienced rapid growth with over 450 admissions into the home during the first 270 days of implementation. We observed early HaH patient experience scores (n = 34) that surpassed BaM hospital care across multiple domains, including ease of care, staff collaboration, doctor communication, likelihood to recommend, and overall rating of care. Over this same period, hospitalist and nursing applications for positions in the HaH virtual command center outpaced available job openings. Yet, we also encountered numerous foreseen and unforeseen obstacles. We share our key learnings below as we navigated our health system's tripartite goal of improving care for patients, provider experience, and operational outcomes in the hope our experience may serve as a guide for others considering implementation of HaH. Launched in 2023, HaH was the culmination of 2 years of planning with key clinical, operational, financial, and information technology (IT) stakeholders. Our successful implementation was defined by several key steps: (1) Finding alignment between organizational objectives and the value proposition of HaH in defining measures of success; (2) Evaluating internal capabilities and assessing the need for external partnership to implement HaH; (3) Establishing a virtual command center with a dedicated hospitalist-led multidisciplinary team to manage hospitalized patients at home; (4) Developing patient selection criteria as well as workflows and electronic medical record (EMR)-integrated tools to support patient acquisition; (5) Embedding implementation scientists in operations to foster measurement, learning, and care model innovation; and (6) Engaging patients and providers in continuous improvement of the care model and operational infrastructure (Figure 1). A learning health system's roadmap for a hospital medicine-led Hospital-at-Home implementation. We adapted the RE-AIM framework to our organizational context, with these principles each continuously revisited as part of our commitment to use data to drive sustainable and scalable implementation. Our prior experience with patient acquisition and provider engagement challenges implementing home-based postacute care models informed our HaH approach to IT priorities, workflow design, and communication.8 We chose to partner with a company experienced in navigating the complexity of tethering ill patients in their home to a virtual clinical team, channeling native EMR orders to a platform accessible to a heterogeneous group of external service providers, and coordinating these clinical services across a broad geography. In addition, we embedded a research team to facilitate the mission of becoming a rapidly Learning Health System—using our data to improve clinical care and operational planning and to reinforce our provider and patient engagement efforts.9, 10 We underappreciated the distrust this new care model would encounter among frontline hospitalists, consultants, and ancillary staff. Providers questioned the safety and effectiveness of changing the paradigm of care for hospital medicine. To build trust, initial patient eligibility criteria were defined around Diagnosis Related Groups for which research supported noninferiority of HaH outcomes relative to BaM care. We were transparent about this rationale and continuously shared quality data with our care providers. The absence of well-defined home-based acute care pathways equivalent to those for common diagnoses managed in BaM settings introduced subjectivity and skepticism among our clinical teams that we counteracted by partnering with specialty leadership to create parallel home-based protocols; this prioritized patient safety while also leveraging our embedded implementation scientists to study and meet the growing need for publication in the space.11-13 Delivering acute care in the home posed unanticipated challenges ranging from medication administration to supply and pharmaceutical procurement. Standard inpatient practices wherein a nurse directly observes medication administration proved inefficient virtually, and often unnecessary, for patients at home, where they were already accustomed to managing their medications. Moreover, routine inpatient disease management—such as frequent nursing blood sugar checks and sliding scale insulin coverage for patients with diabetes—were impractical in a HaH setting. Patients often had their own diabetic supplies, which they preferred using; meanwhile, common inpatient medication formulations like insulin vials proved difficult to transport and self-administer. Hospitalists, nurses, and pharmacists had to reimagine medication ordering, distribution, verification, and administration for HaH. In turn, this led to more flexibility in our HaH formulary and clinical protocols, such as the addition of insulin pens to replace hospital syringes to facilitate easier home administration and the introduction of a new home medication verification process to enable medication use without centralized pharmacy distribution. For HaH to succeed, clinician engagement was central, starting with a shared understanding of the goals of HaH and the value it could bring to both patients and providers. Familiarizing hospitalists with the care model and its technological and logistical infrastructure was critical, but we found that the patient-centeredness and humanism at the center of HaH design held deeper appeal. Virtual interactions with patients in their homes proved surprisingly intimate and afforded hospitalists the opportunity to better understand patients' social needs, living situations, and family dynamics in ways that enriched their care and improved job satisfaction. Clinicians identified issues in home environments—from lights too dim to read medication labels to inadequate meals for caregivers—that they found creative ways to address, while also discovering the satisfaction of coaching patients in how to navigate in-home challenges and increase their self-efficacy with chronic disease management. Similarly, our nursing team reported that HaH enabled a renewed focus on patient needs, allowing their days to be rebalanced around spending time talking to patients and their families. Surprisingly, we observed the intimacy of in-home monitoring and frequent communication created more openness to caring for complex patient populations in the home. We anticipated that provider education efforts would be continuous and developing comfort within this new clinical paradigm would take time. To address this, we created standardized screening tools and automated demographic and geographic eligibility flags. Hospitalists and nurses were responsible for completing clinical appropriateness and social support screening questionnaires embedded within the EMR for this patient population to determine eligibility for admission into the home. We identified significant variation in how hospitalists, nurses, and emergency room providers applied these tools. Our experience with these variations during the first 270 days of implementation is described in Figure 2a. Interestingly, more than 70% of patients with an eligible ZIP code and payor failed to receive a documented clinical screening by a hospitalist and represented the largest leak in the patient acquisition funnel. Similar to the published experience of patients (n = 13) who declined HaH admission during the COVID-19 pandemic, we conducted an analysis of 84 patients who declined HaH admission and offer strategies to address these concerns in Figure 2b.14 We identified variation in hospitalist screening practices, HaH communication, and interpretation of clinical appropriateness as key provider-driven barriers to patient acquisition (Figure 2b). To address these, we further standardized screening workflows, encouraged clinicians to introduce the program early, and coached hospitalists on how to make a clear recommendation when HaH was appropriate. Hospitals are under financial pressure to optimize revenue and expenses, which can create resistance to shifting care away from BaM facilities into the home. Confronting this tension required frank discussions around organizational mission, financial health, and workforce needs. Beyond aligning with our health system's desire to be a healthcare innovator, it was critical to reframe how acute care delivery within the home could offer financial benefit for our organization via (1) expansion of bed capacity to meet increasing patient demand without the capital constraints inherent in acquiring additional physical capacity, (2) extension of nursing productivity facilitated by the lower staffing ratios needed for virtual care coupled with demand for these types of nursing positions, and (3) projected optimization of long-term margins by shifting targeted medical admissions to lower cost-of-care settings while opening beds in BaM facilities for higher margin admissions. Important to our implementation and change management strategy was continuously setting and re-enforcing financial expectations with organizational leadership, particularly around when the virtual hospital would be operating cash flow positive and when it would be expected to break even following the initial upfront investment. We are closely monitoring the validity of these financial assumptions underlying the long-term financial viability of HaH and, while still early on, are seeing better-than-expected results. Despite a desire to internally build an end-to-end HaH program, a multiyear assessment of our internal capabilities revealed areas where we lacked technical expertise, operational experience, and in-home service capacity. In balancing our organization's core business interests and the size and scope of the investment initially required to develop a model for in-home acute care delivery, we decided to partner with a company who had experience in successfully implementing HaH and could provide in-home patient monitoring technology, support for contracting and management of in-home service providers, and a census-based flexible contracting structure. We selected a partner to support in-home service delivery based on alignment with these goals. Strategically, this enabled us to more quickly implement and scale HaH, while allowing simultaneous development of internal operational expertise and time to assess whether these novel elements of HaH would become core to our patient care delivery model. Figure 3 displays the structure of our HaH program. To ensure accountability, we developed a service provider network scorecard to monitor performance. Given the logistical complexity of coordinating internal and external clinical services in the home, this tool proved critical to identifying opportunities for improvement—such as persistent late arrivals in the home—and to creating the transparency and accountability needed for collaborative problem-solving. The evolution of HaH presents a unique opportunity for hospital medicine leaders to shape the development of this model of care and spearhead the movement toward expanded competencies in the field. The historical role of a hospitalist has been grounded in the BaM setting; growth of HaH represents an opportunity for hospital medicine to be forward-looking and responsive to how patients increasingly want their care needs met in more convenient, and often home-based, settings when possible. This, in turn, will necessitate the development of more familiarity with home-based acute care, including attention to how patients are safely triaged and managed in the home and how technology can efficiently be leveraged to support this. Given the uncertainty surrounding the Acute Hospital Care at Home waiver program and Medicare telehealth waivers beyond 2024, hospitalists are also well positioned to advocate for and study this care delivery model to inform future policy surrounding HaH. The authors declare no conflict of interest. The manuscript conforms to the International Committee of Medical Journals Editors regarding authorship. It has not been published, nor is it currently under consideration for publication in any other journal. The manuscript has been seen and approved by all named authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.