A new study on Long COVID doesn't actually look at Long COVID

A new JAMA piece purports to suggest better outcomes in COVID positive patients, albeit with a study design not intended to actually measure Long COVID.

Edit: An additional phrase was added to one of the sentences below for a bit of additional context. The addition has been italicized so that readers can be made aware of this change. Also, the Subscribe button was included twice rather than the Share button. The buttons have been swapped as well.

Strangely, I’ve seen a few arguments pop up in recent months suggesting that Long COVID may not actually exist.

People are certainly within their right to make such arguments, and different perspectives always make for a more enriching discourse.

However, I’ve personally been critical of such arguments, and more broadly speaking I’m of the mindset that post-viral sequelae¹ has been a widely overlooked phenomena due to preconceived notions of these phenomena occurring predominately in women, which has led doctors to suggest that these events were caused by manic episodes and hysteria rather than something of viral etiology.

All this to say, the idea that long COVID or post-viral sequelae not being real has been a strange turn of events in recent months.

And a recent article published in JAMA² makes some very strange suggestions in regards to “Long COVID”:

Looking at the key points alone one may infer a strange conclusion, in that this study suggests that patients who tested positive for COVID actually improved in various measures:

Remember that, when reading studies, one has to be careful in looking at Key Points or Abstracts and inferring something from that part alone. These parts of a study are designed with the intention to grab your attention rather than provide you a brief summary of the study. Don’t be surprised to see a researcher’s perspective or spin come through the Abstract or Key Points.

Because when looking at the actual study I don’t find this to be a compelling piece for looking at Long COVID in particular, in that the study doesn’t appear to measure Long COVID more than just look at different measures at COVID infection and 3 months afterwards.

Remember that Long COVID itself is a very broad and often ambiguous concept.

When exactly would one be considered past the acute phase of the illness and be considered in the Long COVID phase?

In general, studies on Long COVID usually measure starting from the point of a negative test and 4 weeks afterward— any symptoms that persist 4 weeks post-negative test may be considered Long COVID symptoms.

However, the time at which the PCR/antigen test is conducted is important, since the researchers need to have some designated point in which the participants test negative for COVID in order to start moving into Long COVID territory, otherwise we won’t know if a participant included in the study had viral clearance at all.

For example, someone who tests positive for SARS-COV2 should then be tested routinely until considered negative, then a follow-up should occur at least 4 weeks after the viral clearance and an examination should be done to see whether the participant meets the criteria for PASC (post-acute viral sequelae of COVID). ONLY THEN would the included participant be considered to have had Long COVID, which can then be measured over longer periods to see alleviation or exacerbation of PASC.

To this point, the researchers don’t appear to have conducted any follow up PCR/antigen testing to ensure that the participants had cleared COVID, or when exactly the clearance occurred with respect to the 3-month follow up questionnaire:

This study included adult participants (aged ≥18 years) who were recently under clinical investigation for SARS-CoV-2 infection, were fluent in English or Spanish, had self-reported symptoms suggestive of acute SARS-CoV-2 infection (eAppendix in Supplement 1),13 and received testing for SARS-CoV-2 infection with any molecular or antigen-based assay approved or authorized by the US Food and Drug Administration within 42 days before enrollment. An individual was ineligible if the study team was unable to confirm the result of a diagnostic test for COVID-19 or if the individual was unable to provide informed consent, lacked access to an internet-enabled device or computer that would allow for participation, had a SARS-CoV-2 infection more than 42 days before enrolling in the study, or was imprisoned while participating in the study. The goal was to recruit individuals at a 3:1 ratio of those with positive results for COVID-19 to those with negative results for COVID-19.12

Without any routine testing we can’t figure out which participants had Long COVID, which had a persistent infection, or even if the included individuals would meet the 4 week post-viral clearance timepoint set by other studies.

So right off the bat, I’m hesitant to even surmise that this study is even measuring Long COVID if we don’t even know if the participants fit the criteria.

The researchers even make this comment suggesting that they may not be aware of the extent of actual Long COVID participants in their study (emphasis mine):

Notably, we recruited participants with acute symptoms suggestive of a first episode of COVID-19 illness, including those with and without a positive result on a COVID-19 test, which represented heterogeneous groups. The Centers for Disease Control and Prevention defines PCCs as new, ongoing, or recurrent health problems more than 4 weeks after acute SARS-CoV-2 infection; this definition is evolving and difficult to operationalize consistently across studies.13,20 The COVID-19–positive group in the current study contained both those who had mostly recovered after initial infection and those who might have had long COVID, for whom changes in well-being may have been different. Therefore, all results should be interpreted to reflect a more general burden of SARS-CoV-2 infection and the COVID-19 pandemic in addition to any specific burdens associated with long COVID.

So the researchers are suggesting that their study should be interpreted in a broader context than just Long COVID alone. Again, strange, but that just adds to the point that this study design could not inherently capture actual Long COVID participants.

Note that the questionnaire used in this study is also different than ones routinely used in Long COVID studies. Typical assessment of Long COVID use modifications from other questionnaires, and may broadly look for prespecified sequelae as a baseline indicator of Long COVID.

Instead, the researchers used an assessment called the Patient-Reported Outcomes Measurement Information System, or PROMIS:

The baseline and 3-month surveys included questions from the PROMIS-29 and the PROMIS SF-CF 8a.15 The PROMIS instruments use T score measurement, in which a score of 50 represents the mean score of a reference population (ie, the US general population), with an SD of 10.11,16 For PROMIS measures, higher scores correspond to a greater degree of the outcome being measured (eg, greater fatigue). The T score was used to measure outcomes for 7 of the PROMIS-29 subscales (physical function, anxiety, depression, fatigue, social participation,1 sleep disturbance, and pain interference) and the PROMIS SF-CF 8a survey. A single-item numerical rating scale for pain intensity (“In the last 7 days, how would you rate your pain on average?”; score range, 0-10, with 0 indicating no pain and 10 indicating worst imaginable pain) is included in the PROMIS-29 survey.

The NIH website on PROMIS suggests that this questionnaire is given to those suffering chronic illness, so we can see how this can create problems when we’re not sure to what extent the participants in this study would be qualified as having Long COVID, and therefore chronic illness.

I should also note that the proceeding lines from the JAMA study state that researchers gave out the wrong questionnaire, and therefore had to alter their statistical analysis to account for the needed corrections, so that itself may have caused some issues with the assessment of the results.

Now, when we look at some of the variables measured there are some noteworthy points which, again, may allude to the fact that this study isn’t measuring Long COVID:

From Wisk, et al. Figure 2. The above shows 8 of the 29 measurements taken from the PROMIS29 questionnaire.

The above figure looks at different measurements given the PROMIS questionnaire. Note that blue corresponds to those who tested positive while orange corresponds with those that test negative.³ The square notes the initial score averages, while the triangles indicate the 3-month follow up score (essentially see the scores should make an arrow). Subscale measures on the left indicate improvement the higher the T score while measures on the right indicate improvement the lower the T score.

On first glance we may notice that there are differences in baseline measurements between the COVID positive tested group and the COVID negative tested group, which may be attributed to the COVID infection.

However, when looking at the data I may come up with a different conclusion.

Note that this questionnaire is intended for people who have chronic illness. The inclusion of the “social participation” should alert people that the intent is to measure how much one’s chronic illness may influence their ability to go out and socialize (the PROMIS questionnaire appears to be more directed to those who are given an intervention for their chronic illness to see if they improve).

Given this fact, I may unironically argue that the COVID illness itself may be the intervention here, in that those who were told that they had COVID may feel relieved and have their perception/well-being changed for the better compared to those who tested negative.

I can just as readily infer that the drastic increase in “social participation” as well as the “physical function” and decrease in “anxiety” may be owed to the fact that many people may have felt more free to engage in social activities and move around after having a COVID infection, possibly with the thought that they have natural immunity and therefore would be well-protected and therefore would not need to be hunkered down and isolated.

Note that the study was enrolled participants at the end of December 2020 up until September 2021, and therefore the perception of COVID, natural immunity, vaccines, and continuous lockdowns would be far different then than it would be now (emphasis mine):

In this interim analysis, we included the first 1000 participants who completed the PROMIS surveys12 at baseline and 3-month follow-up (Figure 1). Participants included in this analysis were enrolled from December 11, 2020, to September 10, 2021.

The questionnaire can’t control for this aspect unless it was created to intentionally measure people’s perception of COVID at baseline and at the follow up 3-month timepoint. And even with that being said, there’s likely to be that factor in which people who had COVID may feel more comfortable going out and regaining some sense of normalcy.

In summary, I don’t find this study to be an examination of Long COVID. The lack of follow-up testing to check for time of viral clearance, the inability to even figure out which participants meet the criteria for Long COVID, the results for the PROMIS questionnaire alluding to “social participation” which would raise some questions as to what is actually being measured here, all suggest that the study should not be used to infer anything with respect to Long COVID.

Note the researchers make such a remark in their study as well.

Now, as to why Long COVID has been criticized recently is a different matter, and that may be best included in a separate post. My sneaking suspicion points to a similarity in the emergence of virus skepticism, in that this line of thinking, well, removes some thinking and alleviates one of the mental burden of having to worry about these issues while simultaneously preventing government and pharmaceutical utilization of this issue.

Again, maybe that will be saved for another time. For now, just be aware that the study looking at Long COVID doesn’t appear to be looking at Long COVID.

If there are any questions related to this study please let me know in the comments!

---

If you enjoyed this post and other works please consider supporting me through a paid Substack subscription or through my Ko-fi. Any bit helps, and it encourages independent creators and journalists outside the mainstream.

Click on the image above or visit the website at https://ko-fi.com/moderndiscontent

1. Please note that the use of “post viral sequelae” may be entirely inaccurate. The literature uses the phrase “post viral syndrome”, and as it relates to COVID refers to long COVID as post-acute sequelae of COVID (PASC) to signify the time-lapse between infection, recovery, and persistence of symptoms. So always take my use of such a phrase with skepticism.

2. Wisk LE, Gottlieb MA, Spatz ES, et al. Association of Initial SARS-CoV-2 Test Positivity With Patient-Reported Well-being 3 Months After a Symptomatic Illness. JAMA Netw Open. 2022;5(12):e2244486. doi:10.1001/jamanetworkopen.2022.44486

3. It’s imperative that readers be aware that PCR/antigens tests are proxies for a SARS-COV2 infection, and that inclusion of participants doesn’t necessarily relate to test positivity. For example, participants who tested negative may have, in fact, been infected with SARS-COV2 but they just presented with a negative test which would also bias the results. So there’s a possibility that some of the participants may have been misclassified.

The researchers note this as well in their limitations:

Fourth, COVID-19 tests may yield false-negative or false-positive results25,26; therefore, we cannot exclude the possibility that participants may have been misclassified (as having either a positive or negative COVID-19 test result) based on their documented test result; this misclassification could explain part of the similar change in well-being observed between the 2 groups.