Association of Non-medically Long Stay Patients with Hospital Flow and Quality: A Retrospective Observational Study

Background: Patient flow in inpatient medical-surgical units (MDU) is shaped not only by overall occupancy but also by the composition of patients occupying beds. Non-medically Long-stay patients (NLSP) are often viewed as bottlenecks that impede flow; however, their broader spillover effects on regular patients and downstream outcomes remain poorly understood. We examined how NLSP presence and overall MDU occupancy affect Length of Stay (LOS) and readmission rates among regular patients.

Methods: We conducted a retrospective observational, system-level study of MDUs across two large U.S. academic hospitals between January 1, 2018 and October 31, 2019. We included adult patients (18 years or older) admitted to MDUs, excluding visits to non-MDU units, patients who died during their stay, admissions to women and children's service lines, and encounters with missing data. Using administrative and clinical data, we quantified unit-level occupancy and the proportion of NLSPs, examining their associations with LOS and 30-day readmission rates among regular patients. Multivariable regression models were employed to disentangle the independent effects of overall occupancy from NLSP occupancy, adjusting for temporal trends and unit characteristics.

Results: Among 7,123 patient encounters, multivariable analysis showed that higher overall MDU occupancy was significantly associated with prolonged LOS among regular patients, with Quartiles 3 and 4 showing 44.2% (β=0.366, 95% CI: 0.201–0.532) and 57.8% (β=0.456, 95% CI: 0.291–0.622) longer LOS relative to the lowest quartile (both p<0.001) in the fully adjusted model. Conversely, higher NLSP proportion was associated with substantial LOS reductions of 74–76% across all quartiles (all p<0.001), suggesting a behavioral operations response whereby clinical teams accelerate discharge decisions under perceived capacity pressure. Importantly, the presence of NLSP did not significantly affect 30-day readmission rates among regular patients, indicating that faster discharges did not translate into worse short-term outcomes for this group.

Conclusions: This study provides novel evidence that patient mix, beyond simple occupancy metrics, plays a critical role in shaping inpatient flow and outcomes. NLSPs generate contrasting spillover effects—constraining capacity while simultaneously accelerating flow for other patients. These findings have important implications for hospital administrators, highlighting opportunities to optimize case-mix management, bed allocation strategies, and discharge planning without compromising quality of care.