Background: Hospital length of stay (LOS) is a metric commonly used to estimate hospital efficiency and throughput. Various strategies to decrease hospital LOS have been described in the literature; however, less is known about the outlier patients with the longest LOS and how to manage their complex needs. At our tertiary care, academic medical center, we have created the Long Length of Stay (LLOS) multidisciplinary team comprising a group of physicians, case managers, hospital administrators, therapy services, and legal experts that collaborate weekly to facilitate safe, effective discharges from the hospital for LLOS patients. This study aims to describe LLOS patients and estimate impacts on hospital length of stay correlating with the work of the LLOS Team.
Methods: All adult (>18 years) inpatient encounters from a single site, 307 bed academic medical center in the Southeast US, from January 1, 2022, through December 31, 2023, were collected from the electronic health record. Acute organ or bone marrow transplant procedures or admissions for chemotherapy were excluded. The primary dependent variable was LLOS, defined as ≥ 10 days in hospital (approximately 90th percentile in the cohort).
Results: Patient and encounter characteristics were compared descriptively between LLOS vs non-LLOS groups and by year, comparing 2022 to 2023. Multivariate logistic regression compared LLOS vs non-LLOS patients. Overall, 3,588 patients (6.9% of the final cohort) experienced a LLOS. LLOS patients were more likely to be older (64.4 vs. 63.3 years), male (53.2% vs 49.2%), and identify as Black (16.0% vs 10.3%). LLOS patients were more likely to be admitted as a transfer from an outside hospital (12.5% vs. 2.5%) or emergently (54.4% vs. 32.7%); had more previous admissions in the prior year (1.15 vs. 0.72), more Elixhauser comorbidities (9.2 vs. 6.0), and higher rates of admission to ICU (39.9% vs. 7.5%) and PCU (27.6% vs. 7.2%). LLOS patients had higher rates of discharge to skilled nursing facility (8.4% vs. 3.3%), rehabilitation facility (19.4 vs 3.4%)) and home health service referrals (25.9% vs 12.9%).From 2022 to 2023, there was a significant decrease in the cohort LOS overall (4.0 to 3.6) and a small but significant reduction in odds of LLOS [OR (95% CI) 0.98 (0.98, 0.99)]. In multivariate logistic regression, patients admitted as a transfer from a different hospital (1.13 (1.12-1.15)) or being admitted to the ICU (1.17 (1.16-1.18) or PCU (1.10 (1.09-1.11) had higher adjusted odds of being LLOS. Based on care management notes, patients requiring dialysis (1.09 (1.07-1.10), IV antibiotics (1.12 (1.11-1.13), code status change during the encounter (1.24 (1.15-1.34), and having barriers to discharge including no medical clearance (OR (1.12 (1.11-1.13), or other barriers (1.28 (1.16-1.41) had higher adjusted odds of being LLOS.
Conclusions: LLOS patients are unique, with greater disease severity and higher resource utilization. Implementing a multidisciplinary LLOS team with hospitalist and PM&R physicians, case managers, and administrators correlated with reduced odds of hospitalized medical patients experiencing LLOS. Investing hospital resources to focus on LLOS, including staff and time, can facilitate safe and effective discharge for the most complex patients.