User talk:Richardfeinman

Richardfeinman (talk) 19:04, 4 September 2017 (UTC)==Intention-to-treat analysis== My entry was removed on the grounds that "we report what reliable sources say, not what YOU say." However, Intention-to-treat (ITT)is an interpretation and what THEY say is an opinion of one group of (influential but not necessarily reliable) sources.

I am a reliable source. I am a professor in a medical school, widely published, former editor-in-chief of biology journal and what I say is one side of a controversy and, most of all, the source I cited lays out the argument from many examples in the literature (Feinman RD: Intention-to-treat. What is the question? Nutr Metab (Lond) 2009, 6:1).

In addition, Jerry Dallal, http://hnrca.tufts.edu/people/scientists/a-j/gerard-e-dallal-ph-d/, is a widely respected statistician and his explanations are even stronger in criticizing ITT.

Many establishment researchers, primarily physicians, hold to the ITT as in POV originally posted but science does not run on majority opinion. It is not clear that that is even a majority so much as a politically influential group. In fact, while I presented the case as two sides of a controversy, ITT is foolish, and stands as an example of imperial nakedness. Probably few people want to get in an argument and find it simpler to report per-protocol in addition to ITT (although many refuse to report ITT). I personally think that it should never be used but I tried to present both sides of the argument.

Bottom line: ITT says that if a cohort are assigned to take a medication, and nobody in the cohort actually takes the medication, ITT says you must report "the effect of the pill" in terms of the behavior of the pill. As I pointed out if you say that you are studying "the effect of assigning people to a medication" then there is nothing wrong with an ITT analysis but we are rarely interested in that.

In a wider perspective, the medical literature is currently highly criticized for lack of reproducibility and reliability (e.g. Horton R: Offline: What is medicine's 5 sigma? The Lancet 2015, 385:1380.; Gigerenzer, G. Calculated risks: how to know when numbers deceive you. New York: Simon & Schuster; 2002). ITT may be part of that picture. — Preceding unsigned comment added by Richardfeinman (talk • contribs) 14:19, September 4, 2017 (UTC)


 * You may be reliable, but you are not a reliable source. Wikipedia is based on verifiable information. Readers must be able to easily check that information comes from secondary and tertiary sources published by a source with an established reputation for fact checking and accuracy. Please see the links above for further information.
 * If you would like to further discuss your edit, please do so on the article's talk page. Until then, I have reverted the edit per WP:BRD. The "bold-revert-discuss cycle" suggests that if you boldly make a change to an article and another editor reverts you (as happened here) that the best course of action is to discuss the issue on the article's talk page, leaving the status quo in place until a consensus is established. Thanks. - Sum mer PhD v2.0 18:38, 4 September 2017 (UTC)