Understanding factors influencing the length of hospital stay among non-severe COVID-19 patients: A retrospective cohort study in a Fangcang shelter hospital

As a novel concept of responding to disease epidemics, Fangcang shelter hospitals were deployed to expand the health system’s capacity and provide medical services for non-severe COVID-19 patients during the outbreak in Wuhan. To give insights on patient management within Fangcang hospitals, we conducted a retrospective analysis to: 1) describe the characteristics of the patients admitted to Fangcang hospitals and 2) explore risk factors for longer length of stay (LOS). We enrolled 136 confirmed COVID-19 patients, including asymptomatic patients and those with mild symptoms, who were hospitalized in the Wuti Fangcang Hospital. 58 patients completed the treatment and discharged before 1 March 2020. After describing patients’ demographic and clinical characteristics, exposure history, treatment received and time course of the disease, we conducted linear regression analysis to identify factors influencing LOS. We found that patients having fever before admission were hospitalized 3.5 days (95%CI 1.39 to 5.63, p = 0.002) longer than those without fever and that patients having bilateral pneumonia were hospitalized 3.4 days (95%CI 0.49 to 6.25, p = 0.023) longer than those with normal CT scan results. We also found weak evidence suggesting that patients with diabetes were hospitalized 3.2 days longer than those without diabetes (95%CI -0.2 to 6.56, p = 0.065). However, we observed no significant differences in LOS between symptomatic and asymptomatic patients and between patients who received treatment and those without treatment. Longer duration of hospitalization among non-severe COVID-19 patients is associated with having fever, bilateral pneumonia on CT scan and diabetes. However, being asymptomatic and using supportive medications at the early stage of infection do not have significant influences on LOS. Our study is a single-centered study with relatively small sample size. The findings provide evidence for predicting hospital bed demand in a novel response scenario and may help decision-makers in preparing for ramping up the health system capacity.