Unresolved long COVID tracked leaving work in England

Most long COVID research measures symptoms. Fewer studies measure what the symptoms cost, and the cost that matters to a household is usually a job.

A study in The Lancet Regional Health - Europe took 45,864 people who were employed when they joined the Real-time Assessment of Community Transmission (REACT) study in England, and looked at where they were roughly two years later. Long COVID can impair ability to maintain employment, however its relationship to workforce retention and working hours remains unclear, the authors write. Previously the symptoms had been counted far more carefully than the consequences.

The people whose long COVID had not resolved were substantially less likely to still be working.

What long COVID is, and what makes it hard to study

A person who has symptoms that persist for 4 or more weeks after a case of COVID-19 is said to have long COVID. This study used a stricter threshold: long COVID was defined as symptoms lasting 12 weeks or more post-infection.

Long COVID may affect multiple organs in the body, such as the lungs, kidneys, heart, brain, and skin, and the symptom lists run across brain and nerves, mental health, bones and joints, the digestive system, and the heart and blood.

That breadth is the study problem. A condition defined by a long and varied symptom list is one where any two patients may share a diagnosis and nothing else.

It is also not uniformly severe. Although many people have mild symptoms, long COVID can be disabling for some people, and most people who have symptoms at 4 weeks after a COVID-19 infection will gradually improve.

How the REACT study followed people into work

The analysis drew on data from a late-2022 follow-up survey involving 45,864 participants of the REACT Study, with a median follow-up of 23 months.

Everyone in the sample was employed at recruitment, which is what makes the comparison possible: the question is not who had a job, but who still had one.

Two outcomes were tracked. Whether people remained in paid work, and whether their hours had changed.

The comparison group is people with no or short symptoms, defined as under four weeks, rather than people who never caught COVID at all.

What happened to work and hours

Of 45,864 participants employed at recruitment, 86% (N = 39,341) remained in paid work at follow-up and 11% (N = 4877) changed work hours. Approximately 4% (N = 1967/45,864) had unresolved Long COVID.

That 4% fared distinctly worse. Compared with participants with no/short (<4 weeks) symptoms, those with unresolved Long COVID had lower odds of being in paid work at follow-up, by about 38%, and higher odds of changing work hours, by a factor of more than four.

The hours finding is the larger of the two and the easier to miss. Leaving work entirely is visible; cutting from five days to three is not, and it happened at four times the rate.

The three shapes long COVID took

The researchers let the symptom data sort itself into groups. Three clusters were identified: multisystem severe, fatigue-predominant and anosmia-predominant Long COVID. Anosmia is loss of smell.

Those groups did not fare alike. Compared with the fatigue-predominant cluster, participants with multisystem severe Long COVID had lower odds of paid work and higher odds of changing work hours.

So the employment damage concentrated in one presentation. A person whose long COVID is a persistent loss of smell and a person whose long COVID spans breathing, heart and cognition are both counted in the same 4%, and only one of them is losing work.

What this long COVID survey cannot show

This is observational. People who stop working differ from people who keep working in ways that symptoms do not capture: the physical demands of the job, whether it can be done from home, household finances, age and seniority.

Long COVID here is also self-reported, as are the symptoms that formed the clusters. There is no test for the condition, so that is unavoidable and it means the groups are defined by what people said.

Survey cohorts lose people over time, and the people who drop out of a health survey are often the ones doing worst, which would understate the effect rather than inflate it.

The full text was not reachable for this piece, so the authors’ own limitations section was not reviewed.

What follows for workplaces and long COVID

The authors’ recommendation is about treating the condition as plural rather than singular: symptom clusters highlighted the importance of considering heterogeneity in Long COVID when assessing workforce impacts and designing public health responses.

For an employer, that argues against a single long COVID policy and for asking which symptoms a particular person has. Fatigue and a lost sense of smell need different accommodations, and the multisystem group needs the most.

For an individual, the useful thing here is scale. Reduced hours after long COVID is a common outcome across tens of thousands of people, and reading it as a personal failure of effort misreads the data.

What heterogeneity does not settle is what to do about it. For the cognitive cluster, cognitive rehabilitation is the one approach with a randomized result behind it, and it works from goals a person chooses rather than from a symptom list.