Mastodawn

OK, I came across an interesting paper that presents a set of 62 papers that reported race or ethnicity data. I was thinking it would be interesting to see if the papers are easily discoverable by would-be readers looking for race/ethnicity data
- are the population subgroups mentioned in the title/abstract/author keywords?
- are the population subgroups mentioned in the Medline or Embase subject indexing?
- if the population subgroup info is hidden away in the the full text, does the publisher allow full text to be searched outside of its own platforms?
#medlibs #HealthDisparities

Show thread

Kate Nyhan Mar 25

But... then I realized that the original paper is, well, not how I would have done it ¯\_(ツ)_/¯

Show thread

Kate Nyhan Mar 25

They used Claude as part of their screening and data extraction workflow.
They used a filter designed to retrieve studies *from* the US to retrieve studies *about* the US.
They don't actually report their search strategy in enough detail that it could be repeated.

Show thread

Kate Nyhan

So I don't think their list is a comprehensive one.
OTOH I don't necessarily need a comprehensive list of *all* observational studies of these diseases with race/ethnicity data. I just need a list of *some* studies with race/ethnicity data, and for the list to have been compiled by people who didn't start by searching for race/ethnicity search terms. And this, as far as I can tell, fits the bill.