EHR criteria alter the composition of Long COVID research cohorts

A study of U.S. National COVID Cohort Collaborative electronic health records examined how additional eligibility criteria changed cohorts identified by a U09.9 Long COVID diagnosis. The findings show that requirements related to COVID-19 documentation, hospitalization, and medical-history length were associated with differences in demographic representation and healthcare utilization.

EHR eligibility criteria changed cohort demographics and utilization estimates.
A two-year history requirement reduced representation of non-White groups.
Hospitalization criteria were associated with higher utilization and fewer women.
Findings reflect N3C sites and data processes.

Researchers analyzed harmonized electronic health record data from the National COVID Cohort Collaborative for nondeceased individuals with at least one U09.9 Long COVID diagnosis between October 2021 and October 2023. They assessed how three additional criteria—a documented COVID-19 diagnosis, likely COVID-19-related hospitalization, and at least two years of prior medical history—affected cohort composition and estimates of healthcare utilization. Models accounted for contributing data partners and examined sex, race and ethnicity, area-level poverty, insurance coverage, and utilization; sensitivity analyses assessed regional variation and compared results with a COVID-19-only cohort.

Cohort size, demographic composition, and utilization estimates varied substantially across criteria. Requiring two years of history increased representation of women and people with higher outpatient utilization, but reduced representation of people from areas with lower insurance coverage. Hospitalization was associated with more inpatient and non-inpatient use, including emergency visits, and greater representation of non-Hispanic Black individuals, but fewer women. Requiring a documented COVID-19 diagnosis enriched cohorts for people from higher-poverty and lower-insurance-coverage areas and for greater utilization. Medical-history requirements were negatively associated with inclusion of all non-White racial and ethnic groups.

The authors conclude that EHR inclusion criteria are not analytically neutral and may affect equity assessments and the generalizability of Long COVID findings. The analysis used de-identified N3C data, so results reflect that dataset’s participating sites and data processes and may not apply uniformly to other EHR systems.

In practice

For clinicians and researchers interpreting EHR-based Long COVID studies, cohort eligibility criteria can materially shape who is represented and how healthcare utilization is estimated. In this N3C analysis, requiring two years of prior records reduced inclusion of non-White groups, while documented COVID-19 and hospitalization criteria shifted demographic and utilization patterns. These findings support caution when comparing estimates across cohorts or applying them to different populations; they do not establish clinical effects or justify changes to patient care, and may not generalize beyond N3C.

AI-generated clinical interpretation · Not editor-reviewed

AI summary · Not yet editor-reviewed

Is this summary clinically accurate?

Help fellow clinicians: your rating sends inaccurate summaries straight to our editors.

Source

BMC Medicine: Assessing the impact of additional inclusion criteria on the representativeness of EHR-based long COVID cohorts identified using U09.9 ↗

This is an automated AI-condensed summary that has not yet been reviewed by an editor. Always consult the full item at the original source.