I’d start with asking questions that a. matter and b. have some degree of ability to implement and impact the people who would benefit most.https://twitter.com/Alan_Watson_/status/1223224456677666817 …
-
Show this thread
-
Ps. It’s not easy to knock. It takes a lot of time and effort to constantly pick apart studies that are (most likely) done for CVs and for publishers to drive traffic. All under the remit of “evidence-based”. As Altman said. Less research, better research, for the right reasons
1 reply 1 retweet 4 likesShow this thread -
Replying to @dnunan79
It isn't challenging to churn out this sort of attack though, is it? And you picked up on this thread. You don't even need to read the paper (though if GidMK had done so then he would have known 1) they didn't just "ask people what they eat (once)" & 2) it was a 15 year study)pic.twitter.com/oGPax75GWO
1 reply 0 retweets 0 likes -
Replying to @Alan_Watson_
I imagine after doing like over 100 deep dives into these types of studies
@GidMK defaults to this stance as he knows it’s probably about right 95/100 (P = 0.05). These studies produce associations mixed up of exposure effect, confounding, bias & “interpret w caution” And I do1 reply 0 retweets 0 likes -
Replying to @dnunan79 @Alan_Watson_
Eh, I did read the study. Exposure assessment was conducted using a single FFQ conducted 5 years after the initiation of the original longitudinal cohort. They then linked that to routine medical data for mortality outcomes 10-15 years later
3 replies 0 retweets 4 likes -
Which is not what you described your tweet above. There were three questionnaires, the first in 1990, the second 1995 (for 14-28 days) & then a ten year version - these included seasonal adjustment and cross checking statistical analysis In total the study was 22 (not 10) yrspic.twitter.com/Y4rN7P0DKR
1 reply 0 retweets 0 likes -
Replying to @Alan_Watson_ @dnunan79
Read that again. They validated the FFQ using a subset who completed dietary diaries as well. They didn't use the 0 years or 10 year data, because the FFQS were different. The length of the entire longitudinal cohort is immaterial, they didn't use all the data
1 reply 0 retweets 0 likes -
You claimed that "they ask people what they eat (once)" - they did it 3 x, & checked over an extended period too then assessed the soy intake vs the 10 year questionnaire. The length of the cohort is relevant if only because you had claimed it was 10 years - it clearly wasn't.pic.twitter.com/h7HcSBVOXL
1 reply 0 retweets 0 likes -
Replying to @Alan_Watson_ @dnunan79
Sigh. They didn't use the 10-year data to calculate exposure, they checked correlation between the two. And 0.42 is very low, which actually supports my argument - obviously, the exposure changed a lot I said 10-15 years - from the 5 year to the 20th year, depending on death
1 reply 0 retweets 0 likes
They "checked it 3x" but only USED a single measure in the analysis, so it doesn't matter what other data they collected. Anyway, I remember why I muted you in the first place, farewell
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.