Long COVID Incidence in a Large US Ambulatory Electronic Health Record System

Abstract

Estimates of the prevalence of long-term symptoms of coronavirus disease 2019 (COVID-19), referred to as long COVID, vary widely. This retrospective cohort study describes the incidence of long COVID symptoms 12-20 weeks postdiagnosis in a US ambulatory care setting and identifies potential risk factors. We identified patients with and without a diagnosis of or positive test for COVID-19 between January 1, 2020, and March 13, 2022, in the Veradigm (Veradigm LLC, Chicago, Illinois) electronic health record database. We captured data on patient demographic characteristics, clinical characteristics, and COVID-19 comorbidity in the 12-month baseline period. We compared long COVID symptoms between matched cases and controls 12-20 weeks after the index date (COVID-19 diagnosis date (cases) or median visit date (controls)). Multivariable logistic regression was used to examine associations between baseline COVID-19 comorbid conditions and long COVID symptoms. Among 916,894 patients with COVID-19, 14.8% had at least 1 long COVID symptom in the 12-20 weeks postindex as compared with 2.9% of patients without documented COVID-19. Commonly reported symptoms were joint stiffness (4.5%), cough (3.0%), and fatigue (2.7%). Among patients with COVID-19, the adjusted odds of long COVID symptoms were significantly higher among patients with a baseline COVID-19 comorbid condition (odds ratio = 1.91, 95% confidence interval: 1.88, 1.95). In particular, prior diagnosis of cognitive disorder, transient ischemic attack, hypertension, or obesity was associated with higher odds of long COVID symptoms.