A national analysis of trends in COVID-19 infection and clinical management in Veterans Health Administration medical facilities

Abstract

OBJECTIVEWe explored longitudinal trends in sociodemographic characteristics, reported symptoms, laboratory findings, pharmacological and non-pharmacological treatment, comorbidities, and 30-day in-hospital mortality among hospitalized patients with coronavirus disease 2019 (COVID-19). METHODSThis retrospective cohort study included 43,267 patients diagnosed with COVID-19 in the Veterans Health Administration between 03/01/20 and 08/31/20 and followed until 09/30/20. We focused our analysis on patients that were subsequently hospitalized, and categorized them into groups based on the month of hospitalization. We summarized our findings through descriptive statistics. We used a nonparametric rank-sum test for trend to examine any differences in the distribution of our study variables across the six months. RESULTSDuring our study period, 8,240 patients were hospitalized, and 1,081 (13.1%) died within 30 days of admission. Hospitalizations increased over time, but the proportion of patients that died consistently declined from March (N=221/890, 24.8%) to August (N=111/1,396, 8.0%). Patients hospitalized in March compared to August were younger on average, mostly black, and symptomatic. They also had a higher frequency of baseline comorbidities, including hypertension and diabetes, and were more likely to present with abnormal laboratory findings including low lymphocyte counts and elevated creatinine. Lastly, receipt of mechanical ventilation and Hydroxychloroquine declined from March to August, while treatment with Dexamethasone and Remdesivir increased. CONCLUSIONWe found evidence of declining COVID-19 severity and fatality over time within a national health care system.