Background: Association between COVID-19 and tobacco use is not completely understood. The NCI Electronic Health Record (EHR) COVID-19 project collects data on COVID-19 positive patients from 21 health systems nationally to explore the relationship between the COVID-19 infection and tobacco use. The University of Maryland Greenebaum Comprehensive Cancer Center participants in this project, collecting data from the 13 hospitals at the University of Maryland Medical System (UMMS).
Application to Family Medicine: Research encompasses population seen by family practice providers and provides insight into the association of tobacco use on COVID-19.
Objective: Analyze demographic and clinical data for the UMMS patients tested positive for COVID-19 between February 1, 2020 – April 30, 2021 and explore the relationship between COVID-19 severity at first positive diagnostic encounter and tobacco use.
Methods: De-identified data extracted from Epic EHR across the entire UMMS using the data code shared by the NCI through the University of Wisconsin Center for Tobacco Research and Intervention. COVID-19 severity was measured in mutually exclusive categories (ambulatory, ED only visit, hospital admission) based on symptom expression and oxygen requirement. Ambulatory or ED patients were further categorized as no symptoms present, symptoms but no shortness of breath or pneumonia, symptoms and/or shortness of breath and pneumonia. Hospital admission patients were categorized by use of room air only, non-invasive positive pressure, required supplemental oxygen, invasive intubation, and death.
Results: Data for 40,918 COVID-19 positive patients were analyzed including 62% ambulatory, 20% ED only, and 18% inpatient encounters. Contingency and frequency tables were used to describe demographic and study parameters of interest. In the study population, mean age 47±19.3 years; 37.2% were Black Non-Hispanic, 46.2% White Non-Hispanic, and 7.8% Hispanic; 55.9% were female. In the bivariate analysis, COVID-19 severity was positively associated with tobacco use status and age in ambulatory patients (p<0.0001).
Outcomes to be reported: Association between COVID-19 severity and tobacco use will be assessed by ordinal multivariable logistic regression modeling. Co-variates will include demographics (age, sex, race/ethnicity), socio-economic (insurance, residency in zip codes with high level of poverty), and clinical (presence and number of co-morbidities) factors.