JAMA Netw Open

Long COVID prevalence estimates vary widely depending on definition used

August 14, 2025

Study details: A prospective, multicenter cohort study analyzed 4,575 adults (mean age, 40 years; 68% female) from the INSPIRE registry across 8 U.S. sites, enrolled with symptomatic illness at the time of SARS-CoV-2 testing. The study systematically applied the five most widely used published definitions of long COVID to estimate prevalence at 3 and 6 months post-infection and compared these to self-reported long COVID as a criterion standard.

Results: Prevalence of long COVID among COVID-19–positive participants varied substantially by definition: 31% to 42% at 3 months and 14% to 22% at 6 months. Comparator studies in the literature reported even wider ranges (2.6% to 47.4% at 1–5 months; 10% to 62% at ≥6 months). Sensitivity of published definitions vs. self-report was low to moderate (up to 66% at 3 months, 46% at 6 months), while specificity was high (up to 81% at 3 months, 94% at 6 months).

Clinical impact: The wide variability in long COVID prevalence across definitions highlights the urgent need for a standardized, validated diagnostic framework. Without consensus, clinicians risk under- or over-diagnosing patients, complicating care and hindering research comparability. A unified definition would improve diagnostic accuracy, treatment planning, and public health surveillance.

Source:

Wisk LE, et al; INSPIRE Group. (2025, August 12). JAMA Netw Open. Variability in Long COVID Definitions and Validation of Published Prevalence Rates. https://pubmed.ncbi.nlm.nih.gov/40794409/