Long COVID symptoms and loneliness: findings from the World Trade Center Health Registry, 2025, Sisti et al

[forestglip]

Long COVID symptoms and loneliness: findings from the World Trade Center Health Registry

Sisti, Julia S.; Packard, Samuel E.; Metzler, Janna

[Line breaks added]

Background
Symptoms of long COVID can profoundly impact affected individuals’ functioning, including their ability to participate in social activities. While individuals experiencing long COVID symptoms frequently report loneliness, few studies to date have investigated whether loneliness is more common among those with symptoms compared to those without.

We examined associations between long COVID symptoms and loneliness among World Trade Center Health Registry (WTCHR) enrollees.

Methods
Among WTCHR enrollees who reported an acute COVID-19 infection in 2022-23 on a self-administered survey, we used modified Poisson regression to estimate multivariable-adjusted prevalence ratios (PR) and 95% confidence intervals (95% CI) for associations of self-reported long COVID symptoms (any vs. none, selected from a predefined symptom list) with loneliness.

Overall loneliness was assessed with the 6-item de Jong Gierveld loneliness scale (range: 0–6); social and emotional loneliness were assessed with their respective subscales (range: 0–3). We also assessed whether level of social support prior to COVID-19 infection modified associations of long COVID symptoms with loneliness.

Results
Analyses included 5,692 enrollees (mean age: 62 years); prevalence of loneliness was 61%. In fully adjusted models, enrollees who reported any long COVID symptoms had higher prevalence of loneliness compared to those without symptoms (PR = 1.19, 95% CI:1.13, 1.25).

Associations were somewhat stronger for emotional loneliness than for social loneliness (PR = 1.22, 95% CI:1.15, 1.29 and PR = 1.12, 95% CI:1.07, 1.18, respectively). Effect modification by social support was not observed on either the additive or multiplicative scale.

Conclusion
Long COVID symptoms were associated with prevalence of loneliness in a sample of primarily older adults. As loneliness itself is associated with subsequent adverse health outcomes, addressing loneliness among people living with long COVID may help prevent further reductions in quality of life.

Web | DOI | PDF | BMC Public Health | Open Access

 

[Utsikt]

So a cross-sectional study found that sick people are more lonely? Isn’t that what you’d expect?

How does «addressing loneliness» help the situation if people are lonely because they are sick?

 

[forestglip]

How does «addressing loneliness» help the situation if people are lonely because they are sick?

Why not if one considers loneliness a distressing long COVID-associated symptom that might be relieved by something like an online support group, in a similar way to how pain might be relieved by a painkiller?

 

[Utsikt]

Why not if one considers loneliness a distressing long COVID-associated symptom that might be relieved by something like an online support group, in a similar way to how pain might be relieved by a painkiller?

Because:

Effect modification by social support was not observed on either the additive or multiplicative scale.

 

[forestglip]

Because:

I think that's referring to this:

we also assessed whether associations between long COVID symptoms and loneliness
differed by pre-infection social support.

We did not find evidence that associations between long COVID symptoms and
loneliness were modified by level of pre-COVID social support on the multiplicative scale

So it's looking at pre-COVID social support's effect on the association between LC symptoms and lonelineless post-COVID.

I also wouldn't necessarily expect pre-COVID advil intake to modify the association between long COVID and pain symptoms post-COVID, and wouldn't base any claims of advil efficacy on negative findings from that.

 

[Utsikt]

I think that's referring to this:

So it's looking at pre-COVID social support's effect on the association between LC symptoms and lonelineless post-COVID.

Oh, that could have been made clearer in the abstract. Thanks.

 

[rvallee]

I can't really see the usefulness of this study because the way they assess loneliness is too superficial and generic, it does not ask why. The proliferation of bad questionnaires is seriously hindering the progress of medical research. Even the good ones have significant limitations and most of them are terrible.

Loneliness was measured on both the COVID-19 baseline and follow-up surveys with the 6-item De Jong Gierveld Loneliness scale [14]. This scale comprises two 3-item subscales measuring social loneliness and emotional loneliness, separately. The scale includes three negatively worded statements (e.g., “I feel rejected”) and three positively worded questions (e.g., “There are enough people I feel close to”). Respondents were asked to rate the level to which each statement described their feelings, with positive (“yes”), neutral (“more or less”), and negative (“no”) response options.

This is actually weird:

Several sensitivity analyses were performed. First, to ensure that associations between long COVID and loneliness were not affected by symptoms

But, why else would this be happening? I guess they put it that way in a "we're trying to falsify the hypothesis", but, really, why else? Just say it, it's the only plausible explanation. Plus, the questions they ask have nothing to do with being illness, are far too generic for that. They did find that symptoms are the primary reason, but the conclusion of addressing it doesn't make sense on its own, and it's pretty ironic that it's always put this way because medicine has nothing to offer to improve symptoms, and yet there is the whole fiction about this being a thing.

Statistics are very useful mathematical tools but doing that on qualitative numbers just produces too much garbage. It reminds me of the growing problem where politicians do everything the same way, and end up producing garbage as well, holding no positions or convictions unless they poll well in some setting. Except things aren't any better when they try to freestyle with open-ended discussions, because they never actually are open-ended, so this is all very screwed up.

 

[forestglip]

This is actually weird:

Several sensitivity analyses were performed. First, to ensure that associations between long COVID and loneliness were not affected by symptoms

But, why else would this be happening? I guess they put it that way in a "we're trying to falsify the hypothesis", but, really, why else?

I think that's specifically about excluding people that recently had COVID to make sure the association they're seeing is related to long COVID and not acute:

First, to ensure that associations between long COVID and loneliness were not affected by symptoms or isolation due to a recent COVID-19 infection, we excluded enrollees who reported that their most recent COVID-19 infection occurred within 3 months of their COVID-19 follow-up survey.