2. Appraising the relevance of results

2.3 The eight steps

7. What are the results?

Reliance on the outcome: confidence interval

Clinical research is most often conducted on samples for the obvious sake of greater feasibility. Published results report the values observed in the samples. The true value (from the effect of a treatment, for example) that truly corresponds to a population, is somewhere close to this observed value. The role of statistics is to quantitatively define an range of values that contains the true value 95 times out of 100, so that it can be said that the sample represents the general population. The total values in this range are included in the the 95% confidence interval.

The reviewer must know that the true value lies between the two end points of the interval. As the interval becomes narrower, the true value becomes better defined, because the observed value is closer to the true value. And as the sample size becomes broader, the interval becomes narrower, and the number of subjects being studied is greater.