Long Covid: A Disease in Search of a Diagnosis

It is now more than three years since I first described Long Covid as a “skivers’ charter”. At the time, this provoked indignation among those who believed that questioning the existence of Long Covid was tantamount to denying illness itself. It was not.

Then, as now, my concern was not whether people were unwell – plainly many were – but whether we had created a new disease entity out of a constellation of vague, overlapping and poorly defined symptoms that medicine had long recognised under other names. To understand how we arrived here, it is worth tracing the history of the Long Covid narrative as it unfolded.

2020: From Pandemic to Postscript

Almost as soon as COVID-19 was declared a pandemic in March 2020, stories began to circulate of people who ‘weren’t getting better’. Media outlets, notably the BBC, featured individuals reporting fatigue, brain fog, palpitations and breathlessness weeks after infection. There was never, to my recollection, a balancing interview with someone saying: “I had Covid six weeks ago and I feel fine.”

Post-viral syndromes have been known for decades. Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) has long occupied a contested space between neurology, immunology and psychiatry. Yet suddenly, in the glare of the pandemic spotlight, post-viral sequelae were rebadged. ‘Long Covid’ was born, not through laboratory discovery, not through pathological differentiation, but through the development of a narrative.

From the outset, definitions varied. Was it symptoms persisting beyond four weeks? 12? Three months? Did it require laboratory confirmation of infection? Increasingly, the answer was no. If someone believed they had Covid, that sufficed. Already, the epistemological ground was shifting.

2021–2022: Expansion and Elasticity

By 2021, Long Covid had accumulated symptoms like the proverbial snowball rolling down a hill. Lists expanded into the dozens. Fatigue, dizziness, palpitations, anxiety, depression, loss of taste, shortness of breath, ‘pins and needles’ and ‘other symptoms’. If one were minded to design a condition that almost anyone could have at some point in their lives, one could scarcely do better.

The Office for National Statistics began producing prevalence estimates. Headlines spoke of “millions affected”. Funding followed. In the UK alone, tens of millions of pounds were allocated to Long Covid research, dwarfing sums historically devoted to ME/CFS.

Meanwhile, studies emerged suggesting that a significant proportion of adolescents reporting Long Covid had never tested positive for Covid at all. In one study, roughly half. This was not a trivial footnote. If a condition is causally linked to a viral infection, but half those reporting it never had the infection, one must begin to wonder.

It was at this stage that I began to argue more explicitly that what we were witnessing might be less a novel disease and more a diagnostic umbrella or a charter under which all manner of non-specific malaise could shelter.

The Vaccine Pivot

As the vaccines were rolled out, a curious shift occurred. Long Covid, initially an argument for caution, became an argument for vaccination. Media reports proclaimed that the vaccinated were less likely to develop Long Covid.

Yet the studies underpinning this claim were often observational, comparing vaccinated and unvaccinated individuals who differed in many ways beyond vaccine status. Researchers themselves acknowledged confounding factors. Long Covid was defined inconsistently across studies. Symptoms were self-reported. Vaccination status was known to participants.

If a condition has a perceptual or psychosomatic component, and post-viral syndromes historically have been debated in such terms, then belief in protection may influence reporting. This is not conspiracy; it is the placebo effect.

At the same time, the possibility that some Long Covid-type symptoms might arise following vaccination was largely unexamined in mainstream discussion. Later research describing post-vaccination syndromes complicated the picture further. Were we disentangling infection effects from vaccination effects? Or simply attributing all roads to SARS-CoV-2 infection? The confounding was rarely addressed systematically.

2023: Narrowing, But Not Quite

By 2023, some studies began trimming the symptom list. Instead of 40 or 50 symptoms, perhaps seven predominated: fatigue; dyspnoea; palpitations; chest pain; hair loss and joint pain. Even so, many of these are common to other conditions: anaemia; deconditioning; menopause; anxiety disorders and thyroid dysfunction.

Differential diagnosis is the bread and butter of clinical medicine. Yet the Long Covid label often preceded exclusion. A Queensland study comparing individuals who had tested positive for Covid with those who had other viral illnesses found no significant difference in persistent symptoms between groups. Ongoing symptoms appeared indistinguishable from other post-viral syndromes. That should have been a turning point. It was not.

2024: Consensus Without Confirmation

In 2024, the US National Academies proposed defining Long Covid as an “infection-associated chronic condition”. Notably, the definition did not require laboratory confirmation of prior SARS-CoV-2 infection.

This was presented as inclusive and pragmatic. To the sceptical mind, it was revealing. A disease entity that does not require proof of its putative cause to be diagnosed sits uneasily within evidence-based medicine.

Symptom lists ballooned again; at one point reportedly reaching into the hundreds. A diagnosis defined by breadth rather than precision risks losing discriminatory power. If everyone qualifies, the category ceases to inform.

The overlap with ME/CFS remained conspicuous. Yet funding disparities persisted. Long Covid basked in the sanctity conferred by pandemic association. ME/CFS, long marginalised, did not enjoy the same glow.

2025: The Publication Industry

By 2025, it seemed that ‘Long Covid’ in a title almost guaranteed publication. Observational studies linking pre-infection fitness levels to post-infection fatigue reported that those less fit before Covid were less fit after Covid. Such findings were presented as insights into Long Covid rather than into baseline health disparities.

In other cases, Long Covid was identified entirely through self-report. Vaccination status was missing. Confounders were plentiful. Yet conclusions were couched in language implying causal inference.

None of this proves that Long Covid does not exist. It does suggest that the evidentiary bar has sometimes been set remarkably low.

The Core Problem: Definition and Distinction

The central issue remains unchanged since 2020: what is Long Covid, precisely? Is it a distinct pathophysiological entity with identifiable biomarkers? A subset of post-viral syndromes indistinguishable from ME/CFS? A heterogeneous cluster of symptoms with multiple causes? Or a sociocultural phenomenon amplified by media and policy focus?

Without agreed diagnostic criteria, laboratory markers or clear causal pathways, we risk medicalising uncertainty. This does not invalidate suffering. People are tired. People are breathless. People are depressed. The question is whether labelling these experiences as ‘Long Covid’ advances understanding or obscures it.

Incentives and Narrative

One cannot ignore institutional incentives. Research funding follows headlines. Public health bodies benefit from sustained relevance. Pharmaceutical companies benefit from extended narratives around prevention and treatment. None of this necessarily implies malice. It does suggest structural pressures that may favour expansion over contraction of diagnostic categories.

The ‘sanctity’ once attached to cancer research appeared, for a time, to transfer to Covid-related research. To question Long Covid became, in some circles, to question compassion itself. But science is not compassion. It is disciplined doubt.

Where We Stand

Five years on, Long Covid remains a term in flux. Some cases may represent genuine persistent post-infectious pathology. Others may reflect psychological stress, co-morbidities, vaccine effects or entirely unrelated conditions coincident in time.

Until Long Covid can be shown to be clinically distinct from established post-viral syndromes, diagnosable through objective criteria rather than self-report alone, and causally linked to SARS-CoV-2 independent of confounding factors, it remains, at best, a provisional construct.

The danger is not that we care too much – or too little – about people’s health. The danger is that we mistake fashion for fact and consensus for evidence. We have been here before in medicine. Diagnoses expand, contract, merge and disappear; for example, the descriptions of ‘hysteria’ (no longer used) and ‘neurasthenia’ (now subsumed under depression and anxiety).

If Long Covid is real, then it needs clear definition, objective markers and exclusion of confounders. Otherwise, it will be no more than a medical fashion that flourished in the spotlight before dissolving in the cold light of evidence.

Professor Roger Watson is Distinguished Professor of Nursing at Southwest Medical University, China. He has a PhD in biochemistry. He writes in a personal capacity.