Discharge timing model identifies risk for readmission, death after COVID-19

Patients hospitalized with COVID-19 who have potentially reversible risk factors for death or readmission may benefit from a delayed discharge until those factors resolve, according to results of a retrospective study.

To determine the 30-day risk of readmission and death in patients hospitalized with COVID-19 and create a tool to better time hospital discharge, researchers assessed 62,195 COVID-19 patients (mean age, 61.9 years; 51.9% male) admitted to any of 180 U.S. hospitals associated with the HCA Healthcare system between March 2020 and February 2021. Findings were published by the Journal of General Internal Medicine on June 27.

Of the included patients, 7.2% were readmitted and 1.6% died within 30 days of discharge. Independent risk factors for death or readmission included fever within 72 hours of discharge; tachypnea, tachycardia, or lack of improvement in oxygen requirement in the last 24 hours; and lymphopenia or thrombocytopenia at the time of discharge. Being discharged seven or fewer days after their first positive COVID-19 test, receiving corticosteroids, having multiple comorbidities, and having a HOSPITAL readmission risk score of at least 5 were also linked with an increased risk of readmission or death. Patients who were readmitted or died were more likely to be older, male, White, and former smokers, with a shorter length of stay and discharge to a destination other than home.

Inpatient treatment with remdesivir or treatment-dose or prophylactic anticoagulation were associated with lower odds of readmission or death, as were more days in the ICU and discharge directly from the ICU. Researchers developed a risk prediction model for the outcomes; in an internal validation set (33% of the HCA cohort), it had an area under the receiver-operating characteristic curve (AUC) of 0.73 (95% CI, 0.71 to 0.74). The sensitivity of the model was 65%, specificity was 68%, and accuracy was 67%. Similar data from six academic centers were used for external validation, which showed an AUC of 0.66 (95% CI, 0.64 to 0.67).

The rates of readmission or death in the study were similar to those in other investigations of COVID-19, the study authors wrote, but readmission rates were lower than in typical medical patients. This could be because COVID-19 patients tended to be less complex than typical medical patients, they hypothesized.

Overall, many of the factors identified "could influence clinical decision making at the time of discharge in patients hospitalized for COVID-19," the researchers concluded.