Project HoPe: Achieving Home Discharge for institutionally-bound Patients with PROMs, AI, and the EHR - Unnecessary discharges from a hospital to a skilled nursing facility (SNF) are costly and may accelerate patients’ functional losses and requirement for long-term institutionalization. Patients with Alzheimer's Disease and Alzheimer's Disease Related Dementias (AD/ADRD) and other types of cognitive impairment are uniquely disadvantaged by this status quo in that they are twice as likely to be hospitalized, four times more likely to be discharged to SNFs with less than 50% returning to their homes. This situation can be addressed as it is the product of a typically rushed discharge planning process with inadequate time to discover, much less address, a patient’s barriers to home discharge. Recent reports suggest that as many as a third of patients dismissed to SNFs, including those with AD/ADRD, could return directly home if their post-acute care (PAC) needs and barriers were anticipated and addressed. Several key deficits prevent broad realization of a patients’ potential to discharge directly home, or their Home PAC Potential (HoPe). These include a limited ability to: 1) quantify factors that determine PAC needs, 2) identify and address remediable barriers to home discharge, and 3) mobilize stakeholders for advancement of individualized discharge plans. Collectively, these deficits prevent the timely initiation of acute care services that can realize a patient’s potential for home discharge, with PAC as necessary. Rehabilitation-focused, hospital-Home Healthcare Agency (HHA) partnerships have established that interdisciplinary care plans enacted early in a hospital stay with patient and caregiver involvement increase the likelihood of a patient’s return home. Our team developed an Epic electronic health record (EHR)-based discharge planning system that triangulates EHR, patient reported outcomes (PROs), and social determinants of health data to identify HoPe barriers and direct needs-matched rehabilitation service delivery. A pilot of the system among 358 patients increased the home discharge rate by over 25% and revealed high user acceptability. However, the pilot also identified the need to improve addressing of cognitive impairments, targeting of high-yield HoPe barriers, and engagement of non-clinical stakeholders. We propose to address these limitations by pursuing three Specific Aims: 1) Develop a low-burden computerized adaptive test PRO to assess the domains of functional cognition relevant to a safe home discharge; 2) Develop a machine learning algorithm to prioritize actionable HoPe barriers and estimate the degree of change needed for home discharge; and 3) Apply user-centered design principles to refine the EHR discharge planning system for optimal usability and enhanced EHR portal patient, caregiver, and HHA staff access. Our goal is to both integrate and pilot these deliverables in a mature and optimally usable EHR discharge planning system, and to evaluate the feasibility and acceptability of its implementation. We anticipate that the system will be scalable, and amenable to inter-institution transfer for testing in a multi-site pragmatic trial.
