Long Covid Can Be Both Psychosomatic And Real

We should clear some stuff up

I’m very pleased to be running an excerpt from The Age of Diagnosis: How Our Obsession with Medical Labels Is Making Us Sicker by Dr. Suzanne O’Sullivan. Released earlier this year by Thesis, an imprint of Penguin Random House¹, this is a great book and you should buy it. Featured below is the second half of Chapter Two, “Lyme Disease and Long Covid.” The excerpt is almost entirely self-contained — all you really need to know going in is that the “chronic” variant of lyme disease, or CLD, has kicked up a large controversy. Because it sweeps in such a wide range of symptoms, and because sometimes people who believe they have it have no evidence of a tick bite — or even live in a place where there aren’t any Lyme-carrying ticks — CLD is a moving target that is very difficult to pin down. There’s strong reason to believe that some cases, at least, are psychosomatic. This does not mean the patients are “faking” their symptoms; rather, it means that psychological causes are manifesting as physical symptoms, which is a very well-established phenomenon.

As Dr. O’Sullivan notes, some patients have built an identity and found community via believing themselves to have chronic Lyme disease. In the most disturbing cases, they’ve found themselves in the hands of cranks who have exposed them to dangerous, experimental treatments. (My former colleague Molly Fischer wrote an excellent article for New York magazine on this controversy back in 2019 that I highly recommend as well.)

Keep that all in mind as you read Dr. O’Sullivan’s views on long Covid. Premium subscribers can keep reading after the excerpt for a bit more from me. I’ve turned all the footnotes from this excerpt into links.

---

Usually, doctors and scientists hold all the power in developing new diagnostic concepts. However, like Lyme disease, a diagnosis that began with the observations of patients is long Covid.

The term was coined by Elisa Perego when, on May 20, 2020, she used it as a hashtag on Twitter to draw together multiple threads of concern about the long-term consequences of Covid‐­19. Perego lived in Italy, where the pandemic had taken off with particular ferocity after the first case was reported in mid-February 2020. She was infected soon after. Although her acute infection was not severe, she struggled to recover. She was one of many who turned to social media to report the possibility of persistent symptoms after an apparently mild infection. The traditional media’s focus at the time was on those who were dying, but social media was bringing to light a whole other population who were not in the hospital but who were also suffering. Perego invited others who had not recovered after a mild infection to share their experiences on social media using #longcovid. Her campaign took off very quickly. People came forward in droves to tell their stories. By June 2020, many were identifying as “#longhaulers,” predicting chronic disability for millions.

The reports of persistent symptoms after a mild infection alongside the term long Covid spread quickly from social media to news programs and newspapers and then to medical literature. In June 2020, in The Atlantic, science writer Ed Yong described young people attacked by rolling waves of symptoms from which they could not recover. In July, the British Medical Journal described an influx of long Covid to GP practices. Soon after, politicians began to present the public with this fresh threat as another reason as to why they must stay isolated to keep themselves safe.

Much of the early conversation about long Covid was driven by patient-activists and patient-scientists. New Yorker Fiona Lowenstein contracted Covid‐­19 in March 2020. Although her acute illness was short-lived and fairly typical, it later morphed into symptoms not previously associated with Covid, such as hives and sinus pain. Lowenstein went on to set up the Body Politic Covid‐­19 support group. Their mission was to facilitate patient-led and patient-involved research into Covid‐­19. They wanted a paradigm shift in which patients served as leaders in their own care. They referred to their movement as “health democratized.”

In the UK too, patients were at the forefront of the drive to have long Covid taken seriously. In an open letter published in the British Medical Journal in September 2020, Elisa Perego, who is a researcher in archaeology rather than a medical scientist, described long Covid as a “cyclical, multiphasic and multi-system condition.” Her co‐­author, Nisreen Alwan, a public health researcher and another long Covid patient, became a staple member of many public panels on the subject. Together, Perego and Alwan reinforced the message that long Covid was patient-created and that, going forward, patient experts should be central to research efforts.

From the outset, long Covid — as a diagnostic concept, at least — was problematic. Alwan, as one of the pioneer doctors involved in supporting people with the condition, defined it as “not recovering several weeks or months after you had symptoms suggestive of Covid whether you were tested or not.” There was no disease definition. There were no specific symptoms and no diagnostic test results to set limits to the diagnosis. It was usually a self-diagnosis and no proof of infection was required. A negative Covid‐­19 test did not count against it. In fact, a substantial number of people with long Covid then and now either tested negative for the virus or had not been tested. When polled early in the pandemic, 70% of the members of one UK‐­based long Covid support group had tested negative for the virus. In an early survey run by Lowenstein’s Body Politic, as few as 15.9% of those with long Covid had tested positive for the virus.

That there was no disease definition saw the list of symptoms attributed to long Covid grow to more than 200 within days. Many of the symptoms were so non-specific that it would be hard to find a person who did not have at least one — such as dizziness, tiredness, depression, anxiety, muscle pain, headaches, nausea, joint pain, difficulty sleeping. One support group included “loneliness,” “feeling scared,” and “skin aging” on their symptom list. At one public seminar I attended, Alwan asked people to consider consulting their doctor even if they had no symptoms, for fear there could be people who had long Covid without realizing it. If a person didn’t need symptoms or a positive Covid test, that meant anyone or everyone could have it. With no diagnostic criteria to define it, long Covid could now feasibly be the cause of all sorts of physical and psychological suffering during the pandemic.

A great deal about long Covid as a single discrete condition caused by a viral infection was counterintuitive. With most infectious illnesses, or any disease really, the sicker a person is at the peak of their illness, the less likely they are to recover and the more likely they are to have lingering medical problems. Yet long Covid’s relationship to an acute Covid‐­19 infection was the opposite to that. Long Covid was more common after a mild infection than in those hospitalized with severe disease. Sicker people who survived the acute infection were making more complete recoveries than those with a mild infection. Some studies showed that non-hospitalized long Covid patients had a wider range of more severe symptoms than hospitalized patients.

The demographics of those with long Covid were also different to those hospitalized. Patients admitted to intensive care units and who died from Covid were more often the elderly, male, and those with comorbidities like diabetes. But long Covid patients were more likely to be younger and female, and diabetes was not a risk factor. Long Covid patients in the media often described good health and exceptional fitness pre-pandemic — the opposite of hospitalized patients, who were often older and frail. And, while acute Covid had some quite well-circumscribed symptoms, the manifestations of long Covid were highly variable and prone to change. One sufferer described them as “protean — imagine a mischief of mice moving through the walls of your house and laying waste to different bits of circuitry and infrastructure as they go.”

Years later, under the name “post Covid‐­19,” long Covid is defined by the WHO as the continuation or development of new symptoms three months after the initial SARS-CoV‐­2 infection, with these symptoms lasting for at least two months with no other explanation. The WHO estimates that 17 million people might have been affected in Europe in the first two years of the pandemic alone. Approximately 1.9 million in the UK had self-assessed with the disorder by March 2023. In March 2024 nearly 18 million American adults reported ongoing long Covid symptoms.

A variety of medical problems have undoubtedly been drawn under the single label of long Covid. They fall into four main groups.

Firstly, persistent symptoms would not be surprising for anybody sick enough to be hospitalized with Covid‐­19. I spent some time volunteering in the intensive care unit during the pandemic and I saw the toll the virus took. In the hospitalized, especially those who required support for their breathing, persistent symptoms could arise as a result of organ damage caused by the virus, side effects of treatment, and the complications of a long hospital stay.

Secondly, a percentage of those not hospitalized with Covid must certainly have developed persistent symptoms as part of a post-viral fatigue syndrome. This is a well-documented but poorly understood consequence after many viral infections. Most post-viral fatigue syndromes recover spontaneously, although they can be very severe and disabling while they last.

Thirdly, the lack of face‐­to‐­face appointments and overwhelmed healthcare facilities will certainly have seen some people misdiagnosed as either having Covid‐­19 or long Covid at the peak of the outbreak. In the middle of a pandemic, it is very easy to blame everything on the virus and to miss ordinary medical problems. Case reports are already emerging of people who were told they had Covid only to later discover that they had another medical problem, including cancer. However, these three explanations are unlikely to account for the largest group diagnosed with long Covid — those with a mild infection, often with a negative test for Covid or self-diagnosed, who have a wide range of persistent symptoms beyond those typically associated with the acute infection. The symptoms that this fourth group describe are very difficult to explain through pathological mechanisms related directly to a viral infection and are best explained by a psychosomatic illness.

The evidence that a significant proportion of long Covid has a psychosomatic cause has built slowly. Multiple studies have shown that anxiety, depression, and perceived stress are consistent risk factors for long Covid.

One Norwegian study found the presence of loneliness or negative life events in the year prior to contracting Covid was a better predictor of who would develop long Covid than a positive test for the virus. A UK-based study of more than 30,000 children and young people also identified loneliness as closely associated with the development of long Covid. A German study that followed healthcare workers through the pandemic showed that psychosocial burden and expectations of symptom severity were risk factors for long Covid. A French study demonstrated that self-reported Covid‐­19 infection was more likely to lead to long Covid than laboratory confirmed infection. Those who expected to get sick, did.

Long Covid behaves just as psychosomatic illnesses do, with a flitting myriad of symptoms that defy anatomical explanation. Because there is no defining pathology in any single organ, it involves different bodily systems in different combinations in different people. Non-hospitalized long Covid patients often had a wider range of more severe symptoms than those recovering from hospitalization. It consistently contradicts biology. For example, sufferers who are the most short of breath also have the most normal lung function tests and medical investigations do not explain and are often at odds with the symptoms.

Long Covid as a psychosomatic disorder could have arisen through a variety of mechanisms. As we’ve seen, the nocebo effect is a potent generator of physical symptoms through the power of belief. Excess attention paid to the body during the pandemic changed how people experienced and used their body. Searching the body for evidence of infection may then have brought existing symptoms that might previously have been dismissed to the fore. Adrenaline, a change in diet and/or alcohol intake, and different activity levels negatively impacted general health and increased bodily white noise. Predictive coding — the means by which our brains process bodily sensations through the framework of our expectations — used our vivid internal models of the havoc Covid‐­19 could wreak and flooded our bodies with false signals. Some people’s expectation of long-term illness may have come to outweigh the reality of their healthy bodies as horror stories bore down on them from all sides. Long Covid spread through fear as alarming news and misinformation flooded mainstream and social media. Rather than being a single illness, long Covid was the array of consequences one can expect when people are caught in a global pandemic and have nowhere to turn.

Attempts to explain all long Covid as a consequence of viral pathology have so far fallen short. The presence of small clots in the blood, known as “microclots,” are often mooted as a cause and a potential route to treatment. That theory has been investigated by a Cochrane review, a systematic assessment of all available research results on a particular topic, and no evidence was found to support this. Studies found that microclots are equally found in other diseases and in healthy people.

Pathology that potentially explains persistent symptoms has been found in hospitalized patients who had severe infections. The virus really took a physical toll on those hit hardest and that has been reflected in the inflammation and ongoing infection seen in some of those who were hospitalized and in some of those who have, sadly, died. But these findings are often assumed to apply to all the other long Covid groups equally and are used to argue against a psychosomatic cause — when actually there is no evidence to say that they should apply to all groups. Normal tests with no evidence for infection or inflammation are the consistent finding in the group in whom psychosomatic illness makes best clinical sense. The most defining characteristic of those with long Covid after a mild or self-diagnosed infection is the lack of association between symptoms and tests and the lack of evidence for infection and inflammation. One study showed a greater correlation between psychiatric symptoms and disordered breathing than with lung pathology.

It is also worth noting that raised inflammatory markers are seen in association with stress and mental health conditions; so, if they did exist, they could also be explained in ways other than as a consequence of infection. In fact, the rush to explain long Covid saw patients who probably had multiple different problems amassed as a group which has made the likelihood of finding a commonality between them very unlikely. Many early studies that showed a wide range of persistent symptoms after an acute Covid‐­19 infection were of low quality and did not even include a control group. Later studies showed just as many symptoms in control subjects as were seen in people with long Covid. This suggests that the symptoms of long Covid were not exclusive to those infected and were more likely to be caused by some aspect of the social restrictions during the pandemic rather than by the virus itself.

It’s worth noting that the manifestations of long Covid and chronic Lyme disease are very similar. A person might be given either diagnosis depending on when and where they got sick. Both commonly come with a lot of non-specific symptoms such as fatigue, brain fog, depression, aches and pains, sleep problems, dizziness. Both disorders are so ill-defined that the number of symptoms associated with them is far more than those associated with either proven acute Covid‐­19 or true Lyme disease.

It is very hard to talk openly about psychological mechanisms as being responsible for either of these conditions. I know that conversation will upset people, partly because it will be misunderstood. Psychosomatic disorders are often confused with malingering when the two are completely unrelated. Malingering is deliberately pretending to be ill. But psychosomatic disorders are disabling, unconsciously generated medical disorders. Someone diagnosed with a psychosomatic disorder may feel they are being dismissed by the medical establishment, as, in the popular imagination, psychosomatic disorders are seen as “lesser” than other diseases, which means the diagnosis is taken to diminish the person’s level of suffering. Many people still assume that they are imagined or purposeful.

None of that is true. Psychosomatic disorders produce real physical symptoms. Like palpitations that occur with a sudden fright, psychosomatic symptoms are genuinely experienced but are not due to a disease. A psychosomatic mechanism does not mean a person isn’t disabled, only that the problem has arisen through a complex mind–body interaction rather than being due to tissue pathology caused by a virus. I have spent much of my career making people aware of the

seriousness of psychosomatic conditions. I can assure you the pain or fatigue of a psychosomatic condition can be as debilitating as that of cancer and this explanation should not be seen to diminish the suffering of those with this form of long Covid.

It is precisely because it is such a difficult conversation that a psychosomatic explanation for people with long Covid has not featured nearly enough in public discourse. Even doctors and scientists who strongly suspect long Covid and CLD are psychosomatic don’t necessarily raise it. That fear has stymied research progress into both. When one scientist published the results of her research into long Covid, which showed that people who were under stress were more likely to

develop long Covid, she attached a disclaimer to the publication saying that this should not be “misinterpreted as supporting a hypothesis that post-Covid‐­19 conditions are psychosomatic.” She later expressed regret for that statement, saying that she meant only that people with long Covid were not pretending or duping their doctors. Her research did in fact lend support to long Covid being psychosomatic, based on the technical definition of the term, she said later. It seems she was afraid to say that in the first instance knowing the misunderstandings it could provoke.

Patients often reject this explanation for their suffering, as psychosomatic symptoms are all too commonly presented as less disabling than the symptoms of other medical problems. Psychosomatic, psychological, and social suffering are still highly stigmatized and neglected. Disorders like long Covid and CLD present more appealing alternatives to those people with multiple disabling symptoms who are looking for a diagnosis and treatment, as they allow people to access help while navigating a way around stigma. They also offer a support network, peace of mind through an explanation, and validation that the suffering is real.

Society has a general lack of caring institutions, except for medical facilities. This means that physical illness is always prioritized and so it is more straightforward when distress is expressed as a medical problem. No institution took responsibility for Sian as long as she seemed undiagnosable. There were no systems of support for people stuck at home during the pandemic who did not catch the virus. It wasn’t easy for a person with unexplained symptoms to visit their doctor, which meant people missed out on vital reassurance. As Elisa Perego’s campaign pointed out, quite rightly, it was not only those who were severely ill who were suffering. A medical diagnosis brings a person under the jurisdiction of one of the few institutions available to offer support in a crisis. Without long Covid, many of those feeling the social and psychological effects of the pandemic had no voice. Disorders like CLD and long Covid grow in prevalence because they bring help to people who need help. That is facilitated by the fact that the subjectivity of diagnosis makes it so available.

I am a highly specialized doctor who works in a state‐­of‐­the-art environment. I understand the value of tests and technology when they’re used well. I don’t want to minimize the importance of the technical diagnostic tools that I have to hand today that did not exist, or were not as reliable, when I qualified as a doctor. Back then, before magnetic resonance scanning (MRI) was widely available, it took months, or even years, of unpleasant, invasive tests to make a diagnosis of a

disease like multiple sclerosis. I remember so many people in comas, with seizures, paralysis, and other mystery illnesses that I couldn’t explain back then but could easily explain today through scientific advances.

ELISA was developed in the 1970s and Western blot in the 1980s [note: these are my links, not footnotes from the book’s text]. They are an essential part of a doctor’s diagnostic toolbox and a key research tool. They give confidence and speed to a diagnosis. That the tests used to help diagnose Lyme disease can be tricky to interpret should not be seen to denigrate the extreme usefulness of tests but should rather help people to understand that tests alone cannot themselves make a diagnosis. Used alone, they can be uninterpretable and even misleading. Diagnosis is an art and a science, but, surprising as it may sound to some, the art still takes precedence. All tests have a false positive and a false negative rate, so a test that is done as a kind of sweeping diagnostic survey in a person with vague symptoms has a high chance of producing confusing results. If a doctor doesn’t know all the confounding variables that affect the test they have requested, they can easily give too much weight to an abnormal, or a normal, result. Frankly, it’s impossible for any doctor to know all the confounding variables for every test, so those requesting tests they don’t encounter every week can easily find themselves out of their depth. And doctors working with tests they do order every week can find themselves out of their depth too! As a neurologist, I send patients for MRI brain scans countless times a month, but I wouldn’t claim to know the meaning of every single tentatively “abnormal” finding. I am regularly handed a scan that shows something that may or may not be relevant to my patient. When that happens, only a discussion with colleagues and radiologists in which the patient’s story is given central position will resolve the dilemma.

Nor do I want to imply that normal tests, as are seen in many people with CLD and long Covid, are the whole reason that a doctor says a disorder could be psychosomatic. They aren’t. There are lots of diseases that are hard to detect with tests. A large number of the patients I see with epilepsy have normal test results, but the diagnosis of epilepsy can still be made with some confidence based on the description of the seizures. The psychosomatic diagnosis is made in just the same way — based on how the symptoms present; the way they evolve; the way they move around the body, defying anatomy; and also on the contradiction between physical findings and the degree of symptoms.

Diagnosis is made through inferences, using tests in concert with all the clinical information. It is a clinical art that is highly qualitative and cannot be distilled to a simple analysis based on a list of common symptoms. Doctors are prepared to dismiss positive and negative tests equally, if the clinical diagnosis is strong. An excellent diagnostician is one with good clinical acumen and, more often than not, they are also the one with the most experience. Doctors continue to learn throughout their career through listening to their patient’s subjective

account of their symptoms. I find these conversations fascinating. They speak to how unlikely it is that feeding symptoms into a computer will ever produce a reliable diagnosis. If in talking about a patient’s headache I use the word pain, they might correct me and say no, it’s not a pain, it’s a soreness. Unpicking why one word is a better description than another, similar word is often the most important part of a diagnostic discussion. Medical investigations contribute to diagnosis, but they are not the objective thing they are perceived to be. They can be misused, misinterpreted, and manipulated. The quality of the doctor doing the test is as important, if not more important, than the quality of the test.

Developing new diagnoses is also a science and an art, but in this instance, the science must take precedence. In seeking help for people with long Covid, Fiona Lowenstein’s Body Politic movement called for the democratization of health. But scientific answers aren’t at the convenience of the majority opinion. Understanding patient experience is fundamental to setting research priorities, but the scientific process must still be systematic, methodical, rigorous, and open to any answer. Both Lyme disease and long Covid gained such traction in the public arena that they grew outside of the control of the researchers. The discussion around long Covid moved from social media to mainstream media to medical media without adequate scientific interrogation. Society’s demand for immediate answers didn’t allow enough time for scientists to stand back and ask the most basic question of long Covid: How do we reliably define this new disorder so we can study it methodically? Social pressure during the pandemic led scientists and doctors to skip over the essential work of dividing up different types of long Covid for individual study to make research and conversation around the disorder more meaningful. The public rejection of a psychosomatic formulation to explain long Covid shut down that important line of enquiry to the detriment of those affected.

A good doctor is one who is experienced enough to hear the nuance in the story, who understands the primacy of the clinical context, who doesn’t order tests every time, who doesn’t give a label to every ailment, who knows when watchful waiting is the best strategy but also knows when to act. A good scientist is one who is logical and careful, who is objective, who is creative and playful with ideas, but who is also skeptical and unafraid to question strongly held assumptions.

From THE AGE OF DIAGNOSIS: How Our Obsession with Medical Labels Is Making Us Sicker by Suzanne O'Sullivan, published by Thesis, an imprint of Penguin Publishing Group, a division of Penguin Random House LLC. Copyright (c) 2025 by Suzanne O'Sullivan.

---

Anyone who has been reading this newsletter for a while will probably understand why I sought to publish this excerpt.