Originally published 2021-02-01.

A large University of Oxford study found that about one in three people who had COVID-19 (34%) received a neurological or psychiatric diagnosis in the six months after infection. For around one in eight (13%), it was the first such diagnosis they had ever had. Anxiety and mood disorders were by far the most common; strokes, brain bleeds and dementia were much rarer but more frequent in people who had been seriously ill.
The results first appeared as a preprint in January 2021 and were then peer-reviewed and published in The Lancet Psychiatry on 7 April 2021. They became some of the most quoted numbers of the pandemic, and often some of the most misunderstood. Here is what the study actually measured, what it can and cannot tell us, and what it means for anyone recovering from the virus.
What the Oxford researchers studied
The team, led by Prof. Paul Harrison and Dr Max Taquet of Oxford's psychiatry department, used TriNetX, a US network of electronic health records covering more than 81 million people. They identified 236,379 patients over the age of 10 who had a confirmed COVID-19 infection after 20 January 2020 and were still alive on 13 December 2020.
They then looked at 14 neurological and mental health conditions recorded in the following six months. To see whether COVID-19 was different from other illnesses, they compared these patients with 105,579 people who had flu and 236,038 who had any respiratory tract infection, matching the groups for age, sex, ethnicity and existing health problems.
The key numbers, in plain language
Across all COVID-19 patients, 34% had at least one neurological or psychiatric diagnosis within six months. That figure includes people who already had such a condition before, such as a long-standing anxiety disorder recorded again. Counting only first-ever diagnoses, the figure was 13%.
The most common diagnoses were:
- Anxiety disorders: 17%
- Mood disorders, such as depression: 14%
- Substance misuse disorders: 7%
- Insomnia: 5%
Serious neurological conditions were less common: ischaemic stroke in 2.1% of patients, dementia in 0.7% and brain haemorrhage in 0.6%.
After adjustment, COVID-19 carried a 44% higher risk of neurological and mental health diagnoses than flu, and a 16% higher risk than respiratory infections in general.
Severity made a big difference
The sicker people were with COVID-19, the more likely they were to receive a later diagnosis. The overall rate rose from 34% across all patients to 38% among those admitted to hospital, 46% among those treated in intensive care and 62% among those who developed delirium (encephalopathy) during their illness.
The gap was starkest for serious brain conditions. Ischaemic stroke was recorded in 1.3% of patients who were never hospitalised, but in 6.9% of intensive care patients and 9.4% of those with encephalopathy. Dementia rose from 0.4% to 1.7% and 4.7% across the same groups. Psychotic disorders went from 0.9% to 2.8% and 7%.
Even so, people with milder illness were not untouched. Anxiety and mood disorders were common across all groups, which matters because most people who catch COVID-19 are never admitted to hospital.
What the study does not show
Headlines such as "COVID causes mental illness in one in three" went further than the evidence. The authors were careful to list the limits of their work:
- It shows an association, not proof of cause. An observational study can show that diagnoses were more common after COVID-19 than after flu, but it cannot prove the virus directly caused them. Stress, isolation, money worries and the experience of being seriously ill can all play a part.
- Health records are imperfect. The completeness and accuracy of the records are not known.
- The sample leans towards sicker patients. Many people with mild or no symptoms never saw a doctor, so they were not counted. The true rate across everyone infected is probably lower.
- Severity and duration are unknown. A diagnosis code says nothing about whether a condition was mild and brief or serious and lasting.
The researchers found no clear link between COVID-19 and parkinsonism or Guillain-Barré syndrome, two conditions that had worried doctors early in the pandemic. When asked by reporters how long the effects might last, Dr Taquet said that question could not yet be answered and that longer follow-up was needed.
How COVID-19 can affect the brain
The Oxford study did not examine mechanisms, but other research offers clues. In July 2020 a team from University College London described 43 patients in the journal Brain with neurological complications, including inflammation of the brain and spinal cord known as acute disseminated encephalomyelitis (ADEM), strokes, nerve damage and delirium. Some had only mild breathing symptoms.
Scientists have proposed several pathways. Severe infection can trigger widespread inflammation and blood clotting, which raises the risk of stroke. Low oxygen levels during serious illness can injure brain tissue. The immune system's response may sometimes attack the nervous system. And the psychological weight of a frightening illness, long hospital stays and lockdown isolation can take its own toll.
These findings also fed into growing research on long COVID, in which people report fatigue, "brain fog" and difficulty concentrating for months after the initial infection.
Why the findings mattered for health services
By late January 2021 the world had recorded more than 100 million confirmed COVID-19 cases and over two million deaths, and the true number of infections was far higher because testing was limited in the first wave. Even a small percentage of new psychiatric or neurological diagnoses translates into a large number of people needing care.
Prof. Harrison said the findings confirmed high rates of psychiatric diagnoses and showed that serious nervous system disorders also occur, though less often. Dr Jonathan Rogers of University College London, writing in a linked commentary, noted that many of these conditions are chronic or recur. Other psychiatrists said the findings would help researchers decide which conditions, such as dementia and psychosis, deserved closer study.
What this means if you have had COVID-19
Most people recover from COVID-19 without lasting neurological or mental health problems. But if, in the weeks and months after infection, you notice persistent low mood, anxiety, sleep problems, memory trouble, or new physical symptoms such as weakness or numbness, it is worth taking seriously rather than putting it down to tiredness.
Talk to a doctor about any symptoms that worry you. Sudden weakness on one side of the body, trouble speaking or a drooping face can be signs of a stroke and need emergency help straight away. For emotional difficulties, a GP or family doctor can point you towards support, and many countries now have services that specialise in post-COVID symptoms.
Sources: University of Oxford, 7 April 2021: Largest study to date suggests link between COVID-19 and subsequent mental health and neurological conditions; medRxiv, January 2021: preprint of the Oxford study. Study by Max Taquet, Paul Harrison and colleagues, The Lancet Psychiatry, 2021; earlier coverage by The Guardian.

COMMENTS