Long COVID Persistence and Surveillance Gaps Across 58 US Hospitals

Area of Science:

• Infectious Diseases
• Public Health Surveillance
• Artificial Intelligence in Healthcare

Background:

• Postacute sequelae of SARS-CoV-2 infection (PASC) surveillance relies on diagnostic codes that capture less than half of affected individuals.
• This underestimation renders millions of PASC cases invisible to health systems and policymakers, hindering effective response.

Purpose of the Study:

• To quantify the gap between true PASC burden and diagnostic code-based estimates.
• To determine the proportion of PASC cases representing chronic disease.
• To characterize organ system heterogeneity and temporal trends of PASC across diverse populations.

Main Methods:

• Retrospective cohort study utilizing electronic health record data from 58 hospitals across 4 US regions (2017-2025).
• Inclusion of adults (≥18 years) with laboratory-confirmed SARS-CoV-2 infection or COVID-19 diagnosis code.
• Implementation of a custom artificial intelligence algorithm, Precision Phenotyping for Research Cohorts (P2RC), using federated infrastructure.

Main Results:

• The P2RC algorithm identified 74,560 PASC cases (16.28%) among 457,950 COVID-19 cases, over double the rate from code-based surveillance (<7%).
• Approximately 89.31% of identified PASC cases involved chronic conditions, representing a substantial, underestimated chronic disease burden.
• Observed significant organ system heterogeneity and a persistently increasing PASC prevalence through mid-2024, indicating an accumulating health burden.

Conclusions:

• Approximately 1 in 6 COVID-19 patients develop PASC, with the vast majority experiencing chronic conditions.
• Current diagnostic coding systems significantly undercount PASC cases, obscuring a major chronic disease burden.
• The increasing prevalence highlights an accumulating healthcare burden, necessitating investment in enhanced surveillance infrastructure and integrated care pathways.