Whoop recently reported a "Novel Algorithm Capable of Identifying 80% of Symptomatic COVID-19 Cases" using #Wearables.
Interesting study. Glad they're exploring & sharing findings.
But what caught my eye were the low Sensitivity values reported in the study preprint...
1/
I'd love to get input/perspective from folks w/ experience/expertise in diagnostic & screening tests.
Thoughts? Comments? Implications of this level of Sensitivity?
Basically: Is the algorithm promising? Why or why not?
Here's link to preprint:
medrxiv.org/content/10.110…
2/
For those unfamiliar with Sensitivity and Specificity of diagnostic or screening tests, here's a nice 3-minute video introduction:
3/
Also here's a diagnostic test calculator. You can plug in Sensitivity & Specificity values (e.g. from preprint Table) with sample size & prevalence of disease. Helps you to visualize false positives & false negatives, or do quick though experiments:
araw.mede.uic.edu/cgi-bin/testca…
4/
Example: Assume 1000 people w/ Covid prevalence of 3%. Imagine a test has 25% Sensitivity & 95% Specificity. Test would correctly flag 8 (of 30) people with Covid, but falsely flag 48 people who were healthy. So out of 56 positive tests, only 14% would actually have Covid.
5/5
Share this Scrolly Tale with your friends.
A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.