Talk:Attributable fraction among the exposed

Figure caption

I believe in the figure caption the text "(AFe = 1/4)" is wrong. It should be (AFe = 1/3) as I_e = 12 I_u = 8 so AFe = (I_e-I_u)/I_e = 1/3. Unfortunately, I con't find the place to fix it. Nafradi (talk) 15:24, 28 October 2019 (UTC)

Thanks! Fixed it. Psarka (talk) 16:02, 28 October 2019 (UTC)

AF and related formulas

I think there's an error here and suggest checking an epi text before using these interpretations of the AF and related formulas.

Please tell me your concerns - possibly with a reference - and we´ll work it out. Further, I couldn´t find any reference to "AF" you mentioned in the article - please explain. Also please don´t forget to sign your post ;) T.pienn (talk) 17:02, 31 March 2010 (UTC) Moved original post to new section for better readabilitiy T.pienn (talk) 14:19, 16 April 2018 (UTC)

Let's sort out some of the confusion

I finally got around to try to improve things at least a little bit. I did investigate a few seemingly authoritative sources (most notably A Dictionary of Epidemiology and Statistical Methods in Medical Research), and I am now confident that the usage of the combinatorial explosion of terms {Population, Exposed} x {Atributable, _} x {Risk, Factor, Fraction} x {Percentage, _} is a complete mess :(

My currently favored approach would be to follow the Dictionary Of Epidemiology, and settle on the longer but explicit names: "Attributable fraction among the exposed" for (1 - RR) / RR, and "Attributable fraction for the population" for the P_e(RR-1)/(1 + P_e(RR - 1)). This Attributable risk percent (current name) is slightly tricky, as it is formally not equal to (1 - RR) / RR but rather that times 100%, so it contradicts the table. If you'd rather keep the "Attributable risk something" for the sake of symmetry with "Attributable risk", then I would propose "Attributable risk among the exposed": This drops the problematic percent thing and is present in both Dictionary of Epidemiology and Statistical Methods. Psarka (talk) 11:58, 9 May 2018 (UTC)

Also, I made a (draft) page "Attributable fraction for the population": https://en.wiki.x.io/wiki/Draft:Attributable_fraction_for_the_population following the terms of the Dictionary of Epidemiology. What do you think about it? Psarka (talk) 13:37, 9 May 2018 (UTC)

You´re right, (1-R)/RR is not a percentage, so the naming itself is bad. Even worse, the formula in the worksheet https://en.wiki.x.io/wiki/Template:ARR_RRR_worksheet is probably wrong, too, I missed that one until now (in my defence, the line with ARP was introduced by https://en.wiki.x.io/wiki/User:Arcadian). If you have a reliable resource for this formula then please correct it, thank you. T.pienn (talk) 19:28, 16 May 2018 (UTC)