I am not talking about utility, but if I had: Assuming that people truly understand what the test can tell them (is a high bar), then: - this T2D PRS is unlikely useful for anyone - this T2D PRS is not useful for indivs that **DON'T** resemble 23andme customer base
-
This Tweet is unavailable.
-
-
(Repost) Let me attempt to restate your stance then on this particular score. The T2D PRS is not useful if you are aware that you have high environmental risk. The T2D PRS is probably not useful for individuals that don't resemble the majority of the 23andMe userbase.
1 reply 0 retweets 0 likes -
Replying to @ATorkamani @cecilejanssens and
Well - I guess we disagree on "unlikely useful for ***anyone***" If you are healthy, (near)European and you fall in the highest risk tier - 2X increase risk even with the relatively low incidence in healthy individuals is meaningful and worth investigating.
2 replies 0 retweets 1 like -
Replying to @ATorkamani @cecilejanssens and
So, presumably, the test should only ever be sold and marketed to people who meet those fairly exacting criteria, if at all?
1 reply 0 retweets 0 likes -
Replying to @GidMK @cecilejanssens and
No - that is the group that the score is *most* useful for. It's a personal value judgement for the individual. Although yes - I would say on the marketing side it should be clear how well the test would perform across various groups - which is exactly what was done.
1 reply 0 retweets 0 likes -
Replying to @ATorkamani @cecilejanssens and
But what you're saying is that in the best case scenario the test - for people who are low on every other risk, resemble 23andMe userbase, and score in the highest risk tier - will inform people of a risk increase of, what, about 0.05%? Roughly?
2 replies 0 retweets 0 likes -
Replying to @GidMK @cecilejanssens and
According to the white paper its >2X increase in lifetime risk in the top 5%. Risk increase depends on how you define "healthy." But the lifetime risk for BMI (18-25) is >10%. Seems like a fine definition for how most people perceive their health. 10% -> 20% (roughly)
2 replies 0 retweets 0 likes -
Replying to @ATorkamani @cecilejanssens and
But lifetime risk is usually pretty useless to the individual. If you look at the modelling curve in the white paper, the increased 5-year risk looks to be almost 0 and 20-year increase is about 0.5%
3 replies 0 retweets 0 likes -
Replying to @GidMK @ATorkamani and
agree. and let's not forget .... 23andme didn't study lifetime risk ...
1 reply 0 retweets 1 like
They did some modelling based on OR and some comparisons with another model built on real data, but that's certainly not the same as building their own model based on lifetime assessments or even specific diagnoses rather than self-report
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.