Clinical severity of COVID-19 has declined with widespread vaccination, therapeutics, and population immunity, but it remains important to quantify how previously identified risk factors continue to associate with severe outcomes. This retrospective cohort study used electronic health record data from PCORnet to evaluate how age and underlying medical conditions (UCs) relate to the incidence of hospitalization and mortality after a documented COVID-19 diagnosis between April 2022 and March 2025.
The primary outcomes were hospitalization within 16 days and death within 30 days of COVID-19 diagnosis. The analysis aimed to estimate adjusted cumulative incidence (aCI) and adjusted cumulative incidence ratios (aCIR) comparing individuals with and without specific UCs and numbers of UCs across defined age groups.
The study cohort comprised 1,662,679 patients aged 20 years and older with a recorded COVID-19 diagnosis in PCORnet EHRs during April 2022–March 2025. Hospitalization was defined as an admission occurring within 16 days of diagnosis; mortality was defined as death within 30 days of diagnosis. Data access and sharing followed PCORnet privacy policy; the analytic dataset is not publicly available.
Investigators used Poisson regression with cluster-robust sandwich standard errors to estimate aCI and aCIR. Models adjusted for sociodemographic and health characteristics; details of all covariates included in the final models were summarized in the source but are not reproduced in full in this summary. Comparisons were made between those with and without each UC and by counts of UCs.
Analyses were stratified by four age groups: 20–49, 50–64, 65–79, and 80+ years. Underlying conditions were grouped into categories for comparison and included: cancer, cardiac, hepatic, immunosuppression, metabolic, neurologic, pulmonary, renal, psychiatric, and other conditions. The number of UCs per individual was also evaluated (none, 1, 2, ≥3) to assess dose-response associations.
Cumulative incidence of both outcomes increased with age. Reported group-specific cumulative incidence values were:
Hospitalization: 4.7% (20–49 years), 6.3% (50–64), 10.8% (65–79), 22.6% (80+).
Mortality: 0.11% (20–49 years), 0.42% (50–64), 1.0% (65–79), 3.6% (80+).
These results indicate that older age remained a strong predictor of both hospitalization and death after COVID-19 diagnosis during the study period.
Within every age group, the aCI for hospitalization and mortality rose with the increasing number of UCs. Individuals with three or more UCs had the highest adjusted cumulative incidence for both outcomes. The authors report that having two or more UCs was generally associated with higher incidence of hospitalization and mortality across age groups, as reflected in aCIR estimates presented in the study.
The study examined aCI and aCIR for individual UC categories (cancer, cardiac, hepatic, immunosuppression, metabolic, neurologic, pulmonary, renal, psychiatric, and other). Across age groups, the presence of UCs—particularly multiple conditions—was associated with increased incidence of hospitalization and death. The summary reports that UCs, especially two or more, were associated with incidence of both outcomes in all age groups. The source document contains the detailed category-specific estimates; those numeric estimates and confidence intervals are not reproduced here.
Using a large multisite EHR cohort from PCORnet covering April 2022–March 2025, the study found that older age continued to be strongly associated with higher incidence of hospitalization and mortality after COVID-19 diagnosis. Adults with multiple underlying medical conditions—especially two or more—generally experienced higher adjusted incidence of both hospitalization and death across age strata. These findings indicate that age and multimorbidity remained important determinants of severe COVID-19 outcomes in this period despite population-level changes in immunity and treatment availability.
The study was reviewed by applicable Centers for Disease Control and Prevention authorities and was deemed not research under specified federal regulations; investigators report following applicable ethical guidelines. Data are not publicly available per PCORnet privacy policy. One author disclosed consultancy roles with Eli Lilly, Takeda, and Roche Genentech and ownership of Johnson & Johnson stock; other authors declared no conflicts. The source indicates that additional methodological and numerical details are available in the full preprint but are not restated in this summary.