Talk:Number needed to harm

Attributable risk

let me amend my earlier comment today. I think the error is in using the term 'attributable risk' to calculate the NNH/NNT value. It should be the inverse of the absolute risk reduction. The example in this article does the calculation correctly but confuses AR with ARR. Patmalone51 (talk) 18:19, 2 November 2010 (UTC)

Y It is actually the inverse of the absolute risk increase. Psarka (talk) 23:37, 8 September 2018 (UTC)

Error in calculation?

This appears to have an error in the example of working out attributable risk. I believe it's Risk A minus Risk B divided by Risk A. the article leaves out that last part. This results in a major error in calculating the ivnerse of the AR to get to the NNH. Patmalone51 (talk) 16:57, 2 November 2010 (UTC)

Y It is now correct. Psarka (talk) 23:37, 8 September 2018 (UTC)

There is an error in the example cells. For example, 75 + 100 != 115. — Preceding unsigned comment added by 162.99.225.23 (talk) 14:27, 21 February 2020 (UTC)

NNT vs NNH?

Why does this stub switch suddenly to talking about two treatments? I mean, i'm going to figure this out in the next few days one way or another and will then be able to fix it... providing i don't totally forget to do it. Bakerstmd 05:09, 18 May 2007 (UTC)

Y It is no longer a stub, and it no longer switches. Psarka (talk) 23:38, 8 September 2018 (UTC)

References

"2." None of these URLs are extant (404 Errors). E Hassen (talk) 19:51, 22 April 2016 (UTC)

Y Replaced that by a proper reference. Psarka (talk) 23:39, 8 September 2018 (UTC)

Worked example

Worked example - where is it coming from? Any reference? I believe we either have to remove everything related to the follow-up period, or explain how it is relevant and how it can be used. In the absence of the model of the disease the follow-up period numbers are misleading. [original question was not signed]

The worked example explains - though in a rather mathematical way - how the relevant numbers are found. The disease model is not really relevant, but here an example of a cohort study (for more information, please see the article about it on WP) is given, which means (in short terms) that the same individuals are repetitively examined over a period of time and any changes noted. From the differences between the groups (one exposed to an influence, like smoking, versus the other unexposed group, like non-smokers) these numbers are derived.

Aim is to give an estimate how the reduction of or the exposure to an influence (e.g. smoking) would benefit or harm a population. T.pienn (talk) 14:05, 16 April 2018 (UTC)

Y The example was way overkill, replaced it by a simpler standard table. Psarka (talk) 23:40, 8 September 2018 (UTC)

Request of examples

Please give examples of actual medicines, that have very high NNTH numbers compared to NNT = Number needed to treat 128.214.78.170 (talk) 14:54, 14 December 2017 (UTC)

This article is about the general concept of NNT/NNH, as are the numbers in the example. For your requested list of drugs, please do your own research in medical literature about this. T.pienn (talk) 13:53, 16 April 2018 (UTC)

I second the request, and think that such an example would improve the article. Psarka (talk) 23:42, 8 September 2018 (UTC)