I wish I had a bigger follow to promote this paper because I think this kind of research is so important. I just joined twitter, so I don't, but I'll try to do my part anyway. Have a look at: https://www.medrxiv.org/content/10.1101/2020.06.12.20129619v2

They found that HCWs were identified as cases of COVID at a rate 5.5x nonHCWs (range 2-6 mirroring the epidemic curve), and that 9.8% of HCWs likely transmitted to a household contact. Lots of other stuff in here but these were the findings I found most interesting.

I don't think the available data allowed them to adjust for a differential rate of testing, which was almost certainly higher among HCWs, though I doubt 5.5x higher. On the other hand, the denominator used to calculate rates of identified cases in HCWs was probably over-inclusive

and may include audiologists, chiropractors, psychologists, etc, for many of whom workplace exposures were probably not taking place during the lockdown. An unknown number of physicians and nurses who maintain their licences but are not in active practice may also be included.

All of these denominator issues would cause underestimation of the true difference in rates of identified cases among HCW as compared with nonHCW. When this paper goes for peer review, inclusion of the derivation of this number would increase confidence in the result.

I also have a different take on the rate of nosocomial transmission. The available data for this field, added about half way through the data set, were hopelessly incomplete with a definite conclusion in only 23.2% (unknown or missing in the remainder).

Thankfully, we can guess at the rate of nosocomial transmission from more reliable data in the study: 74.1% of HCW cases were outbreak-associated or a close contact of a confirmed case, while only 3.6% of HCW cases were acquired from a household contact.

Because 96.2% of study cases occurred during the lockdown, household contacts would have been the only other likely source of infection, yielding an estimate of 70.5% of HCW cases acquired nosocomially.

Last thought: 9.8% is probably an underestimate for HCW to household contact transmission because some PHUs (like mine) are not testing asymptomatic household contacts. Might be worth including this caveat in the text for context and comparison with estimates from other studies.

@DrKevinSchwartz call me out if you think I'm way off with any of this. Great stuff - hope to see more of this fellow 0T7's work as the pandemic evolves!

@threadreaderapp unroll