Talk:Risk difference

Excess risk merge
I think that Excess risk is another synonym for ARR and RD in some circles. If anyone agrees, feel free to redirect it to ARR. Cmcnicoll (talk) 07:53, 2 February 2010 (UTC)
 * The formula given in Excess risk seems different but does turn out to be essentially the same. The problem with just doing a redirect is that the detailed context being dealt with is rather different, and context is important. It may be that all of the content of absolute risk reduction can be applied in both contexts, and so can be expanded to cover this. A follow-on problem is that many of the other articles wikilinked from here are using the same framework and so should be adjusted accordingly. Melcombe (talk) 14:01, 2 February 2010 (UTC)
 * The differences between the formulas is now resolved, and there was no context in the previous Excess risk page. I can't find any reasonable sources of the term which would distinguish it somehow from the risk difference, so I proceeded to turn it into a redirect. Psarka (talk) 19:45, 8 September 2018 (UTC)

Table is incorrect
The table is incorrect. ARR should = AR when exposure is preventing disease. The cited source is correct, the table is not. —Preceding unsigned comment added by 96.229.250.151 (talk) 20:17, 22 April 2010 (UTC)
 * Table is now correct, as the distinction between risk reduction, risk increase, and risk difference was made explicit. Psarka (talk) 19:46, 8 September 2018 (UTC)

Presenting Results
I may be wrong, but doesn't "reduces 2 cases of colon cancer to 1 case if you treat 6,000 people for five years" mean that the absolute risk reduction in this case is actually 1/6000 (i.e. 0.00017), rather than "0.003 fewer cases per person, using the colon cancer example above" as claimed under this heading? I have a sneaking suspicious that this article was written by a doctor.... ;-) 195.89.26.224 (talk) 12:48, 3 June 2010 (UTC)

Worked Example
I could be wrong, but wouldn't CER − EER = 40% - 10% = positive 30% ?

I believe the textbook Basic Epidemiology describes ARR as Risk Difference, and states "The risk difference, also called excess risk, is the difference in rates of occurrence between exposed and unexposed groups". This would make the correct equation EER- CER and therefore 10% - 40% = (-)30%.

This would make the answer equal the one found in the table.

The textbook can be found at http://whqlibdoc.who.int/publications/2006/9241547073_eng.pdf — Preceding unsigned comment added by 60.224.89.189 (talk) 13:14, 11 March 2012 (UTC)


 * You´re right - the problem is that someone edited the table and changed the original (mathematically correct) EER-CER to CER-EER, which messes up all the signs in the examples.
 * But as always in medicine there are (at least) two ways to look at the problem - starting from the treatment group or from the control group, accordingly the signs change...
 * I´ll edit the template to provide the (mathematically) correct examples. T.pienn (talk) 13:12, 5 April 2014 (UTC)


 * Basic Epidemiology 2nd ed. does not describe ARR as Risk Difference, and does not even contain the term. Dictionary of epidemiology 6th ed. describes it as the negative risk difference. The formulas are now modified accordingly. Psarka (talk) 19:55, 8 September 2018 (UTC)

NNT in the table
Why does the NNT in the work example table have "< 0: number needed to treat" when a number needed to treat is a positive number? — Preceding unsigned comment added by 72.161.242.83 (talk) 06:22, 10 April 2012 (UTC)


 * Because with the same formula you can express NNT and NNH - depending if the outcome is greater or less than zero. T.pienn (talk) 13:13, 5 April 2014 (UTC)


 * Table was replaced by two, which resolved the signs issue. Psarka (talk) 19:54, 8 September 2018 (UTC)