Background: Increased hospital capacity causes significant strain on medical institutions. Patients who are clinically ready for discharge but “stuck” awaiting post-acute resources are thought to contribute to this capacity strain. Here, we aim to provide a clinically relevant measurement of the prevalence (proportion of patients) and weight (proportion of days) for patients who have spent prolonged time awaiting discharge on general Medicine services at hospitals across the U.S. using information gathered from the attending physicians of record.
Methods: This study used a cross-sectional “point-in-time” approach. At each of 16 U.S. hospitals, on a single weekday in fall 2022, a site lead administered a structured data collection tool with any attending physician of record responsible for a general Medicine service. Respondents were asked to run through the patients under their care to identify whether they would consider each patient “clinically stable for lower level of care,” estimate for how long they had been so, and describe what support or discharge resources were needed. Patients who had spent a week or longer clinically ready for lower level of care were classified as “prolonged discharge delay” patients. The number of patients who qualified for this designation is summarized by count and percentage of total patients sampled. Because ranges were provided for respondents to estimate the time spent awaiting discharge, two measures of patient-days in this status are provided: lowest-possible (the smallest number of days in the range selected) and mid-point (the middle number of days in the range selected). Data is presented for the total population sampled and site-based ranges (lowest to highest site percentages). In addition, a binomial logistic regression model was constructed using hospital characteristics to predict the outcome of prolonged discharge delay.
Results: Of all patients sampled, 9.8% (189/1933; site range: 0%-28.5%) experienced prolonged discharge delay, meaning they had been clinically ready for a lower level of care for at least one week. Using the lowest possible time estimate, this population represented 20.7% (5427/26234; site range: 0%-37.7%) of total patient-days and 35.3% (9264/26234; site range: 0%-65.1%) using a mid-point estimate. Patients with prolonged discharge delay often required skilled therapy, ADLs, and/or iADLs at discharge (159/189; 84.1%) and were awaiting a discharge resource considered external to hospital control (155/189; 82.0%). In a binary logistic regression model, admission to a safety-net hospital was a strong positive predictor of prolonged discharge delay (OR 3.21; 95%CI 2.07-4.99, p < 0.001).
Conclusions: Nearly all hospitals sampled from across the U.S. reported that patients on their acute care general Medicine services experienced prolonged discharge delays, though the prevalence of this occurrence (measured by number and percentage of total patients) as well as the weight of this occurrence (measured by number and percentage of total patient-days) varied dramatically by site. At some sites, over half of all patient-days may come from patients in this situation. Furthermore, this appears to represent a particularly vulnerable, marginalized group as the majority of patients with prolonged discharge delay had support care needs (such as mobility, cognitive supervision, etc.) and were admitted to safety-net hospitals. Future investigations should focus on better understanding barriers to access and facilitating equitable outcomes.