## Critical appraisal checklist and calculator for an article on therapy or prevention. This calculator works off-line.

####
First of all, for the results of a single trial to be valid

1. The assignment of patients to treatment options must be randomised and concealed.

2. all the patients entered must be accounted for at the conclusion, and analysed in the groups to which they were randomised.

Ideally the subjects and clinicians should be blind to the treatment group, and all other treatment should be identical. The groups must be shown to be similar at the start.

Programme written by DJR Hutchon (copyright)
This calculator is for educational use. It is believed accurate but no responsibility for accuracy of the results is accepted by the author. David J R Hutchon BSc, MB, ChB, FRCOG Consultant Gynaecologist, Memorial Hospital, Darlington, England.

Comments and suggestions are welcome and will be included in the comments section. E-mail to me at DJRHutchon@hotmail.co.uk

1/1/1999

Algorithm used

function calculate(form) {

var CER=form.contyes.value/form.contotal.value

var EER=form.expyes.value/form.exptotal.value

form.RRR.value=(CER-EER)/CER

form.ARR.value=(CER-EER)

form.NNT.value=1/(CER-EER)

var arrconfid =(1.96*Math.sqrt(((CER*(1-CER*1))/form.contotal.value*1)+((EER*(1-EER*1))/form.exptotal.value*1)))
var confid=1
var fromconfid="NNT(benefit) "+Math.round((1/(form.ARR.value*1 + arrconfid*1)))
if (Math.round((1/(form.ARR.value*1 + arrconfid*1)))<0) {
fromconfid="NNT(harm) "+Math.abs(Math.round((1/(form.ARR.value*1 + arrconfid*1))))
Reference
Altman D G. Confidence intervals for the number needed to treat. BMJ 1998;317:1309-12