Data-driven modelling of IRCU patient flow during the COVID-19 pandemic

Intermediate Respiratory Care Units (IRCUs) function as vital intermediaries between general wards and Intensive Care Units (ICUs), particularly during crises such as the COVID-19 pandemic. A unit's effectiveness depends on its structure, protocols, and clinical expertise. In this study, we assessed the clinical outcomes and operational dynamics of a new IRCU that implemented a specialist staffing model during the pandemic in Spain.

We conducted a prospective cohort study at the UHVN IRCU (Granada, Spain) from April to August 2021, enrolling 249 adult patients with COVID-19-associated respiratory failure. We collected data on patient demographics, Non-Invasive Ventilation (NIV) use, length of stay (LOS), and outcomes, including ICU transfer, mortality, and recovery. We then analysed these outcomes stratified by NIV status. Furthermore, we developed and calibrated a compartmental Ordinary Differential Equation (ODE) model and an empirical LOS-based convolution model to simulate patient flow dynamics under scenarios of admission surges and varying care efficiency.

The cohort's median age was 51 years, and 31 % (n=77) required NIV. Patients requiring NIV were significantly older than those who did not (median 61 vs 42 years, $p<0.001$ ). Overall, 8 % of patients (n=20) were subsequently transferred to the ICU, and 3 % (n=7) died within the IRCU. Notably, no patients managed without NIV required ICU transfer or died. Among the 77 high-risk patients who received NIV, 68 % recovered within the IRCU without needing ICU escalation. Our ODE modelling accurately reproduced aggregate outcomes and demonstrated that simulated admission surges placed the system under significant strain, which enhanced recovery efficiency partially mitigated. The LOS-based modelling yielded consistent peak occupancy estimates.

This IRCU, characterised by specialist clinical staffing, demonstrated effective management of severe COVID-19 respiratory failure. We observed high recovery rates, particularly among NIV patients, which eased pressure on ICU resources. Our dynamic modelling confirmed the unit's vulnerability to admission surges but also quantified the positive impact of efficient care. These findings underscore the importance of well-structured and expertly staffed IRCUs in pandemic response and the broader provision of respiratory care.