1

The study Attributes and predictors of Long-COVID: analysis of COVID cases and their symptoms collected by the Covid Symptoms Study App is a prospective observational cohort study of COVID-19 symptoms in a subset of 4182 users of the COVID Symptom Study app meeting inclusion criteria.

Briefly, the subset comprised individuals

  • who had tested positive for SARS-CoV2 by PCR swab testing
  • who logged as “feeling physically normal” before the start of illness (up to 14d ays before testing) so that we could determine onset.

From Table 1. "Characteristics of COVID-19 by duration of symptoms" contains values in brackets() and i would like to know what they mean

            |   A       |    B      |     C     |     D     |
            | Overall   |   Short   |   LC28    |    LC56   | Remark
Number      |  4182     |   1591    |    558    |    189    | A-B-C-D = 1844 question 7
Age (years) | 42.8(13.4)| 40.1(13.6)| 48.9(12.7)| 50.9(12.5)| questions 1 to 3
BMI         | 27.3(5.9) | 26.8(5.7) | 27.5(5.7) | 27.5(5.8) | question 4

Questions 5 and 6 concerns line 111:

Cardiac symptoms (palpitations, tachycardia) were over-represented in the LC28 group (6.1%) compared to in short-COVID (0.5%) (p<0.0005)

My understanding (Q5, Q6)

  • I assume that p stand for p-value. According to exclusions of null hypothesis (often denoted H0) the null hypothesis is a default hypothesis that there is no difference between two measured phenomena.

  • If H0 is the hypothesis that there is no correlation between cardiac symptoms among short-covid compared to LC28 than the likelihood / probability of this hypothesis is very low (0.0005)?

Questions

  1. Row Age (years) What does 42.8 (13.4) mean - is 42.8 the median age and 13.4 Interquartile range (IQR) from 25% to 75%?
  2. If 13.4 is the IQR does that mean the lower boundary (25%) age is 29.4 years (42.8 - 13.4) or 36.1 years (42.8 - (13.4/2))?
  3. If 42.8 (13.4) means something different could you explain?
  4. Row BMI What does 27.5(5.8) mean regarding body mass index?
  5. Does (p<0.0005) mean the p-value (likelyhood) for no correlation?
  6. Is my understanding correct?
  7. Does anyone know what is up with the delta of 1844 persons? Are they without symptoms?

1 Answer 1

2

You're looking at a non-peer-reviewed study with some sloppy conventions. There is not enough information contained in the paper to answer your questions about it.

You can assume what you want for many of your questions, but they are better directed at the authors, who have not indicated when they present means or medians, standard deviations or IQR, etc.

It's really quite sloppy, but these mistakes happen. Ordinarily they get caught by careful peer reviewers before final publication (or journal staff, or, preferably, their own coauthors before they post a preprint).

2
  • Thanks this encourages me to ask the author.
    – surfmuggle
    Nov 24, 2020 at 0:35
  • Prof. Helen M Colhoun, Medical Informatics & Epidemiology on 22th of october 2020: [...] these results may be subject to serious bias [...] estimating risk based on a non-representative sampling frame [...] dropping around half of the 45839 persons who tested positive as being asymptomatic a further 14443 are dropped because of starting to use the app whilst already unhealthy [...] [more see second comment]
    – surfmuggle
    Nov 24, 2020 at 7:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.