Long Covid Underdiagnosis Poses Broader Care, Coverage Hurdles

A new study estimates that long Covid is underrepresented in medical records, illustrating documentation issues affecting medical research, insurance coverage, and patient care.

The study, published in the Journal of the American Medical Association, found that roughly 16% of people with Covid-19 developed long Covid, more than double the 7% identified by the condition’s diagnosis code.

The researchers developed an AI algorithm—which demonstrated an 80% accuracy rate in a previous study—to search medical records of nearly 458,000 Covid patients across 58 hospitals around the US, identifying long Covid symptoms not explained by pre-existing conditions.

The data signal how the chronic burden of Covid-19 plagues patients fighting for recognition and answers from doctors, insurers, and researchers, six years after the virus emerged. There is broad recognition that long Covid’s diagnosis code is underused, which translates to more hurdles for patients, scarcer data for researchers, and less evidence for insurers making coverage decisions.

Diagnosis codes are used to document patient conditions in medical records and bill insurance companies for treatment. Detailed medical records are critical for obtaining coverage for care and disability insurance, and an official diagnosis is often needed to gain admittance to specialty long Covid clinics. But there is no dedicated treatment for long Covid, and doctors say it can be difficult to get covered.

One reason the diagnosis is not well captured is because insurers are more likely to deny it, said Gary Kaplan, who treats long Covid patients as founder of the Kaplan Center for Integrative Medicine.

“Especially if we’ve got somebody who’s going to go on disability, they would be much more accepting of an ME-CFS [Myalgic Encephalomyelitis/Chronic Fatigue Syndrome] diagnosis,” he said.

Dozens of long Covid patients have sued insurers for denying disability claims, facing down a system courts have repeatedly criticized, according to a recent Bloomberg Law analysis. Coverage depends on indisputable evidence in the medical records, which is difficult to achieve for a new, misunderstood condition like long Covid.

Identifying Long Covid

“What our paper shows is that the objective markers exist. They’re just sitting in the wrong layer of the medical record,” said Hossein Estiri, associate professor at Harvard Medical School and co-author of the study.

Insurers can now use the algorithm, which is public, to better identify long Covid patients, Estiri said. If they continue denying legitimate claims, “they’re basically ignoring the clear longitudinal data patterns that our AI identifies.”

“The legal and insurance systems are using the dimmest possible lens and then citing the dimness as evidence the disease isn’t there,” he said.

Insurance lobbying group AHIP previously collaborated with the Health Care Cost Institute on a similar 2023 study that found 1 in 4 Covid patients suffered long Covid symptoms, and that spending on those patients was five times higher.

“As this condition becomes better understood clinically, plans are constantly assessing best practices regarding coverage using evidence-based criteria,” the group said in a statement.

Julia Moore Vogel, who leads long Covid trials as senior director at the Scripps Research Translational Institute—and has long Covid herself—said the patient community often discusses how difficult it is to prove disability.

“It’s easier in a lot of cases to get the right to die approval than it is to get the disability support,” she added. “And it’s just like, what are we even doing as a society?”

Providers also don’t use the long Covid diagnosis code because there are no treatments for which to bill. Researchers are working to change that.

“The day that we have one of those, then it’s going to become very important to have it in your chart,” Vogel said.

Access Hurdles

Patients can wait years for a diagnosis as they navigate medical theories and skepticism from doctors with varying experience with the condition.

Julie Lam, a writer, marketing professional, and community representative for the National Institutes of Health’s Covid-19 RECOVER initiative, suffered from long Covid for three years before receiving her diagnosis in January.

“I finally can lift up my head and walk into my doctor’s office and say, I am clinically diagnosed with long Covid,” said Lam, whose symptoms include POTS and fibromyalgia-like pain that her doctors initially attributed to stress. “And so let’s talk about how I can get better instead of still going around with this diagnosis of exclusion.”

It took six years for Vogel to receive a long Covid diagnosis for her migraines and chronic fatigue, partly because she didn’t test positive for Covid-19 until she was reinfected. The study’s methodology was smart, she said, but could be limited based on how many people lacked a positive test result from the early days of the pandemic.

It could also be conservative because it excluded pre-existing conditions. Evidence exists that the virus can trigger conditions that a person was already genetically at risk for, Kaplan said. He pointed to his wife, who developed Alzheimer’s after her bout with Covid-19.

“The bug may have been the last thing in the game of Jenga,” he said, “where you pulled the last brick out and the whole thing toppled.”

The virus reactivates pre-existing conditions too, Lam said. Both her father and his wife died, she said, after Covid resuscitated cancer from which they had been in longtime remission. Lam was blocked from a study because she has a kidney condition that long Covid aggravated.

Excluding long Covid from medical records only worsens the fragmentation of care, Kaplan said.

“We’re too busy treating the symptoms and not understanding the underlying path of physiology,” he said. “That leaves people sick, that leaves people undertreated, and that leaves people disabled.”

The cumulative number of long Covid cases is also increasing, according to the study, which Vogel said is “exactly what we expect.”

“Yeah, we know you had three Covid infections and you’re OK, but you don’t know what’s going to happen on the next one.”