Background: To facilitate evidence-based clinical and diagnostic guidelines that respond to the evolving presentation of COVID-19, it is important to identify temporal changes in patient characteristics at diagnosis.

Objectives: Describe temporal changes in patient baseline patient demographic factors, comorbidities, and medications at COVID-19 diagnosis.

Methods: HealthVerity claims data were used to identify COVID-19 patients (diagnosis or positive SARS-CoV-2 lab result) between April - Sept 2020, using the first COVID-19 confirmation to identify index date and stratify patients by month and diagnosis setting (inpatient, IP; outpatient, OP). Mean age at diagnosis (standard deviation) is reported. Baseline variables were assessed 90 days prior (index-inclusive for chronic conditions; index-exclusive for medication use), reported as percentages. Heterogeneity in the distribution of patient characteristics over time were evaluated using Cochran-Armitage trend test.

Results: The cohort included 1,218,866 COVID-19 patients (80% OP; from 64% in April to 85% in Sept). Mean age at diagnosis (overall: 50 [22], IP: 65 [20], OP: 46 [21]) decreased from April to Sept (overall: 57 [20] to 48 [23]; IP: 68 [17] to 66 [20]; OP: 52 [19] to 45 [22]). A Charlson-Quan Comorbidity Score ≥ 2 was observed for 21% overall, 56% IP, and 13% OP, which decreased from April to Sept (overall: 34% to 17%, p<0.01; IP: 62% to 51%, p<0.01; OP: 18% to 11%, p<0.01). The most common baseline comorbidity was cardiovascular disease (36%), which decreased from April to Sept (overall: 51% to 31%, p<0.01; IP: 82% to 72%, p<0.01; OP: 35% to 23%, p<0.01). Similarly, hypertension (30%) and diabetes mellitus (18%) decreased (HTN overall: 43% to 25%, p<0.01; IP: 70% to 60%, p<0.01; OP: 29% to 19%, p<0.01 and DM overall: 26% to 15%, p<0.01; IP: 43% to 35%, p<0.01; OP: 17% to 11%, p<0.01). The same trends were observed in use of concordant medications (antihypertensives, antidiabetics, statins; p<0.01). The most common baseline COVID-associated medication class was antibiotics (excluding azithromycin; 19%), which decreased from April to Sept (overall: 23% to 18%, p<0.01; IP: 22% to 20%, p<0.01; OP: 23% to 18%, p<0.01). Similarly, azithromycin (8%) also decreased (overall: 13% to 6%, p<0.01; IP: 11% to 7%, p<0.01; OP: 14% to 6%, p<0.01).

Conclusions: In this population, the comorbidity burden at COVID-19 diagnosis decreased over time in IP and OP settings, which may reflect the increasing availability of testing in the US during the study period. Nonetheless, to inform future clinical and diagnostic guidelines, this trend should be monitored as the pandemic further evolves due to infection spikes, new virus variants, and vaccine availability.