Why a new study says long COVID may be undercounted

ByBrian “Weence” Bateman

A May 27, 2026 study suggests long COVID can be missed when health systems rely on diagnosis codes alone. Here is why estimates differ, what CDC says about diagnosis, and when to get checked.

If long COVID estimates seem inconsistent, a new U.S. study offers one likely reason: many cases may be missed when health systems rely mainly on diagnosis codes instead of looking more broadly at what happens after infection.

In a JAMA Network Open study published on May 27, 2026, researchers analyzed electronic health record data from 457,950 adults with COVID-19 across 58 hospitals and affiliated clinics in four U.S. regions. Using a validated computable phenotyping method rather than diagnosis codes alone, they identified post-COVID conditions in 16.28% of patients, or about 1 in 6. That was more than twice the share found through code-based surveillance alone.

What the new study found

This was a retrospective cohort study, not a clinical trial. The researchers examined records from New England, Southeast Texas, Southern California, and Western Pennsylvania. Their goal was not to count only people who had already been labeled with a long COVID diagnosis code. Instead, they used an algorithm designed to identify patterns of post-COVID illness appearing at least three months after infection.

That matters because long COVID is often not one symptom, one lab result, or one specialist visit. In the study, 89.31% of the patients identified as having post-COVID conditions had at least one chronic condition that required ongoing clinical management. The authors also found that cumulative prevalence rose through mid-2024 in the participating regions.

Still, this does not mean 1 in 6 people in the United States currently have active long COVID. The study involved adults only, drew from four health-system regions rather than a nationally representative sample, and measured people who met the study definition in the records over time. It was not designed to estimate current nationwide prevalence.

Why long COVID counts can vary so much

Part of the problem is that long COVID is not diagnosed with one definitive test. CDC says there is no approved laboratory test that can determine whether symptoms are due to long COVID, and a prior positive COVID-19 test is not required for diagnosis. CDC also notes that routine blood tests, chest X-rays, and electrocardiograms can be normal in someone who still has long COVID.

That helps explain why different tracking methods can produce very different numbers. A code-based approach depends on a clinician recognizing the pattern, documenting it clearly, and assigning the diagnosis code. A broader phenotyping approach looks across the medical record for symptom clusters and timing after infection. Those methods will not capture exactly the same patients.

Definitions also vary from one study or health system to another. Some researchers focus on specific symptom clusters. Others require symptoms to last a certain length of time or exclude people with certain preexisting conditions. Follow-up matters too: people with fragmented care, fewer return visits, or limited access to healthcare may be harder to identify in electronic records at all.

Why documentation can matter for care

CDC’s clinical guidance says documenting prior SARS-CoV-2 infection and long COVID is important for public-health surveillance and may help patients receive needed care. That can matter for care coordination when primary care, rehabilitation, cardiology, pulmonology, neurology, or school and workplace accommodations are involved.

But documentation is not a guarantee. A diagnosis code or detailed note does not automatically mean a referral will be approved, that insurance will cover a service, or that disability paperwork will be accepted. Coverage rules, network limits, plan requirements, and state policies can still vary.

The same caution applies to families. This study included adults only, but CDC and the American Academy of Pediatrics both note that children and adolescents can also develop long COVID. Pediatric symptoms can look different from adult symptoms and may affect school attendance, exercise, concentration, and daily function.

What readers can do now

If you have symptoms that began after COVID-19 or have lasted for months after an infection, it is reasonable to schedule a medical evaluation rather than wait for one single test to settle the question. CDC and MedlinePlus both describe long COVID care as symptom-based and focused on improving function and quality of life. MedlinePlus also notes that there is no specific treatment for long COVID yet.

  • Write down when symptoms began and whether they followed a known or suspected COVID-19 infection.
  • Keep a short symptom diary, including fatigue, shortness of breath, dizziness, sleep problems, exercise intolerance, pain, or memory and concentration problems.
  • Note whether physical or mental activity seems to make symptoms worse afterward.
  • Bring a medication list and records from previous evaluations if you have them.
  • Stay up to date on COVID-19 vaccination, which CDC describes as the best available tool to help prevent long COVID.

Seek prompt medical evaluation if symptoms are worsening, limiting daily activities, or interfering with work, school, or caregiving. Seek urgent care right away for emergency warning signs such as severe trouble breathing, chest pain, fainting, new confusion, or other severe symptoms.

What this study does not answer

The paper strengthens the case that code-based surveillance can miss long COVID cases, but it does not settle every question about how common long COVID is. The authors note important limitations: the algorithm depended on the quality of electronic records, people with limited or fragmented healthcare engagement may have been underrepresented, several regions did not have site-specific chart-review validation, and the study did not include a COVID-negative comparison group to measure excess illness above background rates.

So this study is best read as a careful warning about surveillance blind spots, not the final word on prevalence. Even so, it helps explain why official counts, hospital coding, and patients’ lived experience do not always line up neatly.

Sources

Editorial note: Weence articles are researched from cited public-health, medical, regulatory, journal, and reputable news sources and may be drafted with AI assistance. They are checked for source support, clarity, and safety guardrails before publication.

This article is for general informational purposes only and is not medical advice. Research findings can be early or incomplete, and health guidance can change. Always talk with a qualified healthcare professional about personal symptoms, diagnosis, medications, vaccines, screenings, or treatment decisions. If you think you may have a medical emergency, call emergency services right away.

Related

Brian “Weence” Bateman is a health and medical writer who covers public health updates, healthcare policy, and medical developments affecting communities across the United States. He works closely with dentists and healthcare providers and focuses on making complex medical information clear, accurate, and useful for everyday readers.