The true cost of long-term COVID-19 infections could be twice as much as current estimates, and is hidden from current surveillance systems that rely on capturing diagnostic codes, according to a new study led by Army Gen. Brigham. Using a new AI algorithm, researchers reviewed the medical records of nearly 460,000 COVID-19 patients across 58 U.S. hospitals and found that about 1 in 6 people, or about 16%, developed long-term COVID-19 infection. These incidence rates represent more than 18 million Americans, twice as high as current estimates, and reflect the increased cumulative burden of chronic disease post-COVID-19. Results are posted below JAMA network open.
More than 10 million people with long-term COVID-19 infection will never be detected by the diagnostic codes that health systems and policymakers rely on to track the burden of the disease. The numbers we reveal are almost certainly an underestimate. ”
Dr. Hossein Estiri, lead study author, faculty member at Brigham General Hospital, Massachusetts
Current diagnostic coding, including ICD code U09.9 designated for post-COVID-19 symptoms, captures less than 7% of patients with long-term COVID-19 infection.
Researchers at Massachusetts General Brigham have introduced a new “precision phenotyping” algorithm specifically designed to identify long-standing coronaviruses in long-term electronic medical records by analyzing a time series of clinical events from hundreds of thousands of patients with COVID-19. The algorithm has previously been validated to identify long-term cases of COVID-19 as a diagnosis of exclusion, identifying conditions that emerge after infection with COVID-19 and cannot be explained by diseases already present in the patient’s medical history.
Researchers analyzed electronic health records of 457,950 patients who had previously tested positive for COVID-19 in four regions of the United States: New England, southeast Texas, Southern California, and western Pennsylvania. They identified long-term COVID-19 infection in 16.3% of patients overall, with rates ranging from 13.6% to 22.7% across the region. Across the study cohort, 14.5% (66,587 people) of COVID-19 patients developed chronic conditions that required ongoing clinical care. The study also revealed regional differences in the long-term clinical manifestations of COVID-19, including dramatically different rates of prediabetes, a new aftereffect of long-term COVID-19, across the United States.
Contrary to the assumption that long-term COVID-19 infections are a holdover from earlier waves of the pandemic, the researchers also found that cumulative prevalence continued to increase in all regions studied. This indicates that the virus continues to act as a catalyst for new long-term chronic health conditions that affect various systems within the body. Statistical modeling shows significant quarterly increases in New England, Southern California, and Western Pennsylvania, and trends indicate continued growth over the next decade if current patterns persist.
“This study demonstrates how longitudinal clinical data in health systems can be structured and analyzed to support more consistent identification of complex post-viral infection conditions,” said study co-author Sean Murphy, MD, chief research and information officer at the University of Washington. “Clinical AI has great potential if it is designed for public health and integrated across real-world medical settings.”
Researchers note that the findings do not include illicit infections, which have become the majority since large-scale testing ended, and exclude patients without long-term medical records. These restrictions suggest that the overall number of people infected with the long-lasting coronavirus could be even higher.
“These patients are not missing clinical care; they are missing the diagnosis codes that identify them as long-term COVID-19 patients,” said study lead author Jiazi Tian, Ph.D., a data scientist in the Clinical Augmented Intelligence Group at Massachusetts General Brigham University. “Cardiologists have discovered new autonomic imbalances, endocrinologists have discovered new metabolic disorders, and neurologists have discovered unexplained cognitive disorders, including long COVID-19 infections that arrived without a label linking them to COVID-19.”
“This study shows how hospitals can leverage AI to help fill surveillance gaps that public health agencies are no longer tracking. What I’m most excited about is what happens next with this new surveillance data,” Estilli said. “Being able to distinguish between the different clinical and organ-specific symptoms of long-lasting coronavirus will allow us to initiate new clinical trials and test targeted treatments in the right patients.”
sauce:
Reference magazines:
Tian, J. others. (2026). Long-term persistence of the novel coronavirus and surveillance gaps in 58 U.S. hospitals. JAMA network open. DOI: 10.1001/jamanetworkopen.2026.14909. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2849452
