A nationwide questionnaire study of post-acute symptoms and health problems after SARS-CoV-2 infection in Denmark, 2022, Sørensen et al

This risk difference thing
Statistical notes for clinical researchers: Risk difference, risk ratio, and odds ratio

So fatigue % at 6-12 months in Covid positive was 11.1% and in Covid negative was 3.1%. Then you just subtract one from the other, giving a risk difference of fatigue of 8%. Figure 1 in this paper that is the subject of the thread gives a risk difference of 8.43% - this is presumably because the risk has been adjusted for differences between the two samples i.e.

It's worth noting that last sentence in the quote - some, maybe a lot, of the people testing negative may have been tested because they had symptoms suggestive of Covid-19 or because they had close contact who had been diagnosed with Covid-19. So, it's likely that the Covid-negative sample might not have been totally healthy. They might in fact have been Covid positive (it's just the test didn't show it). It's also possible that the sample might conceivably include people who had another infectious illness, and might even have developed post-infection symptoms due to that other infectious illness.

 

So yeah, where I was heading with the risk difference was that, say for depression, it looks as though having a Covid-19 infection is associated with an additional 1 person in one hundred getting a diagnosis of depression between 6 and 12 months after infection. I'm not sure that that is enough of an increase to make it likely that Covid-19 has caused depression in any medical way. For one thing, people with other post-infection symptoms might be incorrectly diagnosed with depression; people who had a Covid-19 diagnosis were probably more likely to be grieving lost family members; and some people who now have difficult medical symptoms and associated financial stress might have secondary depression or sadness. So, I don't think there is evidence that depression is a direct symptom of Long Covid.

The differences in rates of diagnosis of PTSD are very small.
It's interesting to see that the difference in rates of diagnoses of fibromyalgia is so small. It's not appearing that fibromyalgia is a post-infection outcome. Possibly though it may be a result of diagnostic preferences in Denmark, as risk differences for joint and muscle pain were fairly big at 3.5%.

 

It's a great number of respondents, but self-selection looks to be a major problem.

 

Worth noting the samples aren't of the same type of people - most of those differences are highly significant.

That's a really big difference in new onset of physical exhaustion (defined as a health condition) (risk difference 40% - 45.5% in the test-positive sample and 7.3% in the test-negative sample).
That's difficult to square with the results for the symptom of fatigue (11.1 versus 3.1%).

A couple of interesting symptoms are:
Reduced strength legs/arm - risk difference 4.7%, and
Sleeping legs/arms - risk difference 3.5%
Legs and arms going numb was quite a distinctive symptom for me and my children at onset. I don't think it happens as much now, but it still happens a bit, especially when my illness is worse. The frequency and the low threshold was very noticeable - my arms could go numb just from the holding them up to eat at a table or to drive. I think it's a symptom that isn't talked about much.

 

I interpreted 'self-reported health problems with new onset between the test date and until 6-12 months after' as a new health problem starting after the Covid-19 test date (and probably having an ongoing impact), rather than as symptoms of the acute infection and a few weeks after. But I see from the Discussion that you are right.

I don't think those figures of exhaustion at any time from Covid-19 test onwards are very useful - it's such a low threshold and covers the period of the acute illness.

 

[See my post below - I have understood full time and part time sick leave differently from what was intended - full time sick leave is taking a whole day off, not being on sick leave all of the time since the Covid-19 test.]

I think monitoring of sick leave is a good thing for studies to do, as it is an objective measure that has clear economic and social consequence. Probably it would be more use if the number of days of leave were recorded for individuals in the year preceding the infection and in the year after the infection (or, in surveys of respondents where that sort of precision is difficult, if the sick leave taken was days, or weeks, or months).

I find the sick-leave results from this study hard to interpret. The period they ask about is from one month after the Covid test until the time of the survey, on average about 8 months later.

Full-time sick leave was reported by 9.4% of the test-positives and 6.5% of the test-negatives. That seems like quite a big percentage in both groups but might make sense if a lot of the people in the sample were being Covid-19 tested prior to admission to hospital with a range of ailments.

Weirdly, part-time sick leave was reported by 4.2% of test-positive and 1.7% of test-negatives. That seems incredibly low - I would have thought at least half of people of working age would take at least one day sick leave in an 8 month period.

 

agreed
Yes, I edited my post about the 'health problems' reporting.

I see that there were a lot of detailed questions about sick leave in the survey. It would have been good to see some of the results from those in the paper.

'Full-time sick leave' seems to refer to taking a whole day/week off, and 'part-time sick leave' refers to taking part of the day/week off - perhaps because the person only had part-time employment?


Questions about sick leave from the survey