Lead/Presenter: April Mohanty, COIN - Salt Lake City
All Authors: Mohanty AF (Salt Lake City IDEAS Center, University of Utah), Zhang Y (Salt Lake City IDEAS Center, University of Utah) Li H (Salt Lake City IDEAS Center, University of Utah) Jones B (Salt Lake City IDEAS Center, University of Utah) Haroldsen C (Salt Lake City IDEAS Center, University of Utah) Nevers M (Salt Lake City IDEAS Center, University of Utah) Adams B (Salt Lake City IDEAS Center, University of Utah) Zickmund S (Salt Lake City IDEAS Center, University of Utah) White PJ (Salt Lake City IDEAS Center, University of Utah)
Objectives:
Randomized clinical trials (RCTs) are the gold standard of scientific rigor but can lack generalizability due to failure to include underrepresented patient populations. Our objective is to (1) present and summarize patient characteristics, including social determinants of health (SDOH), of a VA cohort used for a comparative effectiveness study of remdesivir and to (2) compare our descriptive data with those of the remdesivir RCT by Beigel et al.
Methods:
Our study includes Veterans with VHA hospital admissions from March 2020 to December 2021 (N = 508,012 patients). We restricted our analysis to Veterans with a laboratory confirmed COVID-19 positive result two weeks before and up to 2 days after their hospitalization. We included the first VHA hospitalization due to COVID-19. We accessed Veteran sociodemographic, health, and healthcare data from the VHA COVID Shared Data Resource and the VA Corporate Data Warehouse. The United States Veterans Eligibility Trends and Statistics were used to ascertain SDOH data.
Results:
Compared to the remdesivir RCT, our cohort includes nearly twice as many patients aged ?65 years (64% vs 36% in the VA cohort vs RCT). The prevalence of some racial/ethnic minority groups is higher in the VA cohort (e.g., 28% vs 21% Black or African American and 2% vs 1% American Indian/Alaska Native and Native Hawaiian/Pacific Islander in the VA cohort vs RCT, respectively). At baseline, which for the VA cohort was the first 3 days after inpatient admission, the prevalence of severe hospitalization (i.e., requiring noninvasive ventilation/high-flow oxygen devices or receiving mechanical ventilation or extra-corporeal membrane oxygenation) was 19% vs 45% in the VA cohort vs RCT. The percent of VA inpatient admissions due to COVID-19 peaked in December 2020 during the study period, with the highest prevalence of 28% observed among Native Hawaiian/Other Pacific Islanders, followed by Asians (24%), American Indian/Alaska Native Veterans (22%) with the lowest prevalence observed among White Veterans 17% in the same month. Among VA participants, those treated with (N = 14,149) vs those treated without (N = 23,973) remdesivir had a lower prevalence of Black or African American race (22% vs 31%) or homelessness (6% vs 13%) and had a higher prevalence of White race (70% vs 61%) or rural residence (15% vs 13%).
Implications:
The VA cohort compared to the remdesivir RCT by Beigel et al., has a greater representation of older patients hospitalized due to COVID-19 and a higher prevalence of some racial/ethnic minority groups. Unadjusted rates of remdesivir use were lower for Black or African American and homeless Veterans.
Impacts:
Our study leverages clinical data from the electronic health record to explore the generalizability of findings from a RCT to underrepresented patients. In future work, we will evaluate efficacy among Veteran subgroups based on their SDOH and to focus on the Centers for Disease Control guideline recommended COVID-19+ population for remdesivir, i.e., hospitalized patients requiring no more than supplemental oxygen, and can provide real world evidence to address gaps in our understanding of the COVID-19 treatment effectiveness.