Today's find from medRxiv: "SARS-CoV-2 viral load peaks prior to symptom onset: a systematic review and individual-pooled analysis of coronavirus viral load from 66 studies"

(A #medlibs thread touching on #systematicreview searching) https://www.medrxiv.org/content/10.1101/2020.09.28.20202028v1

"On June 8th, 2020, the investigators performed a PubMed search with the phrase, “((corona*) OR (COVID*) OR (”SARS-CoV-2”) OR (SARS*) OR (MERS*)) AND ((”viral load”) OR (”viral shedding”) OR (serolog*))."

Y'all, they searched using "PubMed’s online advanced search algorithm".

To the surprise of no one who is familiar with my harping on search things... I don't consider this a systematic review. The search sucks. Not to mention it was only done in PubMed. At least throw in Google or Google Scholar if you don't have access to other resources!

Side note: I really miss showing truncation expansion from legacy. "corona*" hits that 600 variations barrier, which new PubMed doesn't have. But new PubMed doesn't show any translation of truncation - another reason to question its use, especially for "systematic reviews".

But you know what bothers me the most? This bit of info:

"This publication is based on research developed by the Institute for Disease Modeling prior to its affiliation with the Bill & Melinda Gates Foundation."

The implications. <shudder>

Why is it so hard for folks to ask #medlibs to be involved in these things?! It's not like we can't be had for cheap, or even *free*. Plenty of qualified, experienced searchers could've donated their time and expertise.

Crowdsourcing it on twitter could come up with better. Heck, in less time than it took me to do this thread, even half-assing it or dealing with a migraine, I could come up with a better search. I know plenty of others who could as well.

Anyway. So yeah.