Faster Information for Effective Long-Term Discharge: A Field Study in Adult Foster Care
Abstract
As the US population ages, a growing challenge is placing hospital patients who require long-term post-acute care into adult foster care facilities: small long-term nursing facilities that care for those unable to age in place because their care requirements exceed what can be delivered at home. A key challenge in patient placement is the dynamic matching process between hospital discharge coordinators looking to place patients and facilities looking for residents. We designed, built, deployed, and maintain a system to support decision making among a team of 6 discharge coordinators assisting in the discharge of 127 patients across 1,047 facilities in Hawai'i. Our system collects vacancy and capability data from facilities via conversational SMS and processes it to recommend facilities that discharge coordinators might contact. Findings from a 14 month deployment provide evidence for how timely, accurate information positively impacts matching efficacy. We close with lessons learned for information collection systems and provisioning platforms in similar contexts.
Introduction
As the US population ages, a growing body of CSCW and HCI research explores how technology might support the provision of care for older adults. Much of this research focuses on enabling older adults to remain at home or "age in place" [invalid citation][invalid citation][invalid citation], often with the help of formal [invalid citation][invalid citation] or informal caregivers [invalid citation][invalid citation]. However, 19% of adults over 50 believe their home is insufficient to facilitate aging in place [invalid citation], and over 1.5 million US adults over 65 currently live in nursing, residential, or adult foster care homes [invalid citation][invalid citation][invalid citation].
One prevalent avenue to no longer being able to remain at home is after a hospital stay: older adult patients often cannot be discharged back home even if they no longer need acute hospital care, and must instead be placed in a long-term care home. Such placement is a major challenge: in November 2022, 35 medical organizations, including the American Medical Association, wrote to the Biden administration detailing "gridlock" in Emergency Departments, and a "public health emergency" in part due to the inability to discharge patients [invalid citation]. Such gridlock leads to increased length of patient stays in hospital and substantial extra expense: an occupied bed can cost thousands of dollars per day, which is often paid by the hospital [invalid citation].
This paper contributes findings from a 14 month deployment of an intervention to improve the process of discharging patients in need of long-term care, from hospitals to adult foster care (AFC). We focus on patients whose hospitalization has lasted longer than 21 days; prior work has shown that such long-term stays are important; while they are only 2% of hospitalizations, they account for roughly 15% of overall hospital bed-days [invalid citation].
At hospitals, discharge is managed by Discharge Coordinators who work to find an appropriate home to discharge patients to. Coordinators often have limited control over a patient's care, including insurance coverage and care home availability; thus, a large part of their work is information discovery and communication. They may need to call hundreds of homes before finding one that can accept a given patient. Coordinators also seek to identify placements that are less likely to lead to readmission, which may increase costs and worsen patient outcomes—motivating an inclusion of social and cultural factors in the placement process. These factors combine to make coordinators' job important and challenging that, although fulfilling when a successful placement happens, often leads to high levels of stress and burnout.
On the other side are Adult Foster Care Homes, who must find suitable residents. In contrast to better-known nursing homes, AFC homes are smaller scale (often five or fewer residents), less institutionalized homes that are often run by, for example, a retired nurse who is also an older adult [invalid citation][invalid citation]. AFC homes are often more local to where their patients used to live—and thus more integrated within the community. However, finding suitable residents can be challenging. Homes have specific capabilities and preferences, for example with respect to the medical severity of a patient. Empty beds or ill-matched patients may have disproportionate impacts on small homes and their patients, compared to larger nursing homes.
In our setting, coordinators historically rely on State Department of Health data to decide which homes to contact about placement. This data principally serves to convey that a home is licensed to operate, to prevent placements into unlicensed homes, but also includes whether a home has a vacancy. As we'll show, this data is insufficient for effective matching: we find it is updated on average every 105 days, but actual vacancies change much faster. We show how the task of calling homes about patient placement is challenging, which can in-part be alleviated by systems designed to align with existing workflows, that enable faster information update intervals.
We designed and deployed a conversational SMS system in Hawai'i, in partnership with a local hospital, to exchange patient availability and home vacancy data. As a working system in a high-stakes domain, our system needs to continuously engage with stakeholders, balance their constraints and priorities, and fit into existing workflows. We worked closely with coordinators and care homes for over two years to understand the daily workflows and challenges associated with placement, and over multiple iterations of system design, to create tools that support their work. This paper describes a portion of our work which sends personalized surveys periodically to care homes via SMS. We use human-in-the-loop labeling to compile these messages into status updates for homes, which informs ranked call recommendations provided to a team of 6 coordinators, who use this list to decide which homes to call.
Findings from a 14 month study in Hawai'i provide evidence of the impact that timely, accurate information has on increasing matches between patients and homes. By reducing informational and coordination frictions, our system improves patient placement workflows. We show substantial engagement with the system throughout the population, both for individual surveys and over the duration of our study, indicating the potential for such systems to effectively collect and disseminate information. Homes frequently updated their information in our system, demonstrating the need for such systems and the desire from homes to participate.
We conclude by discussing challenges and opportunities associated with the long term viability of such a system. We highlight the importance of centering improvement in everyday workflows, rather than measuring success via specific end metrics. We also discuss how our system constitutes a coordination mechanism that aids coordination between a team of hospital discharge coordinators and a large, distributed network of independently-owned and operated care homes. Finally, we discuss questions of sustainability and potential scaling across contexts.
Author Positionality
We acknowledge that researchers' backgrounds and experiences shape their perspectives and that our research may include our personal biases [invalid citation]. Our team consists of interdisciplinary collaborators with expertise in HCI, Operations Research, and healthcare, and years of experience in long-term care placement in Hawai'i. Our work embraces a reflective approach necessitated by the sensitive nature of foster care placement of vulnerable patients.