Why do new long COVID estimates look so different?

One recent long COVID headline may sound reassuring, while another sounds alarming. That does not necessarily mean one study is right and the other is wrong. In this case, the studies are measuring different things in different populations with different methods.

That is the most useful starting point for readers. There is not one single settled number for long COVID that works in every setting. Estimates can change a lot depending on who is counted, how long COVID is defined, and whether researchers ask people directly about symptoms or try to find cases through medical records.

Study 1 asked U.S. adults directly about symptoms and recovery

The first study, published March 2, 2026, in JAMA Network Open, used the National Health Interview Survey, a nationally representative survey of U.S. adults. It looked at adults who reported a prior COVID-19 infection and asked whether they had symptoms lasting 3 months or longer that were not present before COVID-19.

Among adults who reported prior infection, the weighted share reporting long COVID since the start of the pandemic fell from 19.7% in 2022 to 13.7% in 2024. Reported recovery rose from 51.2% to 59.7%. In 2024, the authors estimated that 8.3% of U.S. adults, or about 21.3 million people, had ever had long COVID.

This study also found higher odds of long COVID among women, adults ages 35 to 64, and people with lower household income. Recovery was less likely among adults age 35 and older.

Still, this was a repeated cross-sectional survey based on self-report, not a clinical examination of the same people over time. The authors noted important limits: symptoms and recovery could be misclassified, people with stop-and-start symptoms might appear recovered, and the design cannot track one person’s recovery course from year to year.

Study 2 used hospital-linked records and a broader phenotyping method

The second study, published May 27, 2026, in JAMA Network Open, analyzed 457,950 adult COVID-19 cases across 58 U.S. hospitals and affiliated clinics. Instead of asking patients whether they had long COVID, researchers used a broader electronic health record phenotyping approach to identify post-acute sequelae of SARS-CoV-2 infection, often shortened to PASC.

Using that method, the researchers identified PASC in 16.28% of documented adult COVID-19 cases, or about 1 in 6. The paper also found that this was more than twice the level captured by code-based surveillance alone, suggesting that routine coding may miss many cases.

That estimate sounds more severe, but it answers a different question. This was a retrospective health-system study of adults who appeared in linked medical records, not a household survey of the U.S. public. Its results depend on documentation quality, can underrepresent people with limited or fragmented healthcare use, and do not include a COVID-negative comparison group to show how much illness exceeded background rates.

Why the numbers differ so much

The two estimates look far apart because the studies are built differently from the ground up.

Different populations: one study sampled U.S. adults living in the community; the other analyzed adults connected to hospital-linked records.
Different denominators: one can be framed around all adults or adults with prior infection, while the other is based on documented COVID-19 cases in participating health systems.
Different definitions: one used self-reported symptoms lasting 3 months or longer; the other used a broader EHR-based phenotyping framework for post-acute sequelae.
Different case-finding methods: asking people directly and searching records for clinical patterns will not capture the same patients.

So the safe conclusion is not that one study disproves the other. It is that they are answering different questions.

Why the surveillance gap matters

The May 2026 paper carries an important public-health warning. If health systems, researchers, or policymakers rely mostly on billing and diagnosis codes, they may miss many people whose symptoms fit long COVID or related post-acute illness but were never cleanly coded that way.

That matters because undercounting can make the problem look smaller than it is. It can also affect planning for clinics, rehabilitation, disability support, workplace accommodations, and follow-up care.

What long COVID diagnosis and care look like now

CDC says there is currently no laboratory test that can definitively diagnose or rule out long COVID. The agency also says objective laboratory or imaging findings should not be used as the only measure of how a patient is doing.

In plain language, that means a person can have persistent symptoms and impaired daily function even if standard testing looks normal. Diagnosis is still clinical, based on history, symptoms, physical examination, and directed testing when needed.

CDC’s current guidance describes care as patient-centered, symptom-based, and focused on function and quality of life. That often means treating the symptoms that matter most to the patient first, building a rehabilitation plan, and managing related conditions over time.

AAPM&R has also described long-COVID care as a multidisciplinary effort. That reflects the reality that symptoms can affect more than one organ system and may involve primary care, rehabilitation, pulmonology, cardiology, neurology, mental health, or other specialties depending on the patient’s needs.

What readers can do with conflicting headlines

Read the study type first: A survey, a claims analysis, and an EHR study may all produce different-looking estimates without actually contradicting each other.
Focus on the denominator: Ask whether the number refers to all adults, previously infected adults, hospital-linked patients, or some other group.
Do not overread recovery claims: Better survey recovery estimates do not mean long COVID is fully understood or easy to treat.
Be cautious about cure headlines: Randomized treatment research is still ongoing, so simple cure claims can get ahead of the evidence.
Know the prevention message: CDC says COVID-19 vaccination remains the best available means of preventing long COVID.

When to seek care

If symptoms started or worsened after COVID-19 and have lasted for weeks or months, it is reasonable to bring a clear timeline to a clinician. Write down when symptoms began, what makes them worse, and how they affect work, school, sleep, exercise, or daily activities.

Seek urgent medical care for emergency warning signs such as chest pain, trouble breathing, fainting, signs of stroke, or new severe confusion. Those symptoms should not be dismissed as routine post-viral problems.

Bottom line

The new long COVID estimates look different because the studies are measuring different populations with different definitions and different tools. One offers a national self-report snapshot among U.S. adults. The other offers a health-system burden estimate using a broader record-based phenotyping approach.

Both studies are useful. Neither supports sweeping claims that long COVID is either mostly over or fully understood. For readers, the safest takeaway is to be skeptical of overly simple headlines, especially ones that promise certainty about risk, recovery, or treatment when the evidence is still evolving.

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.