Confidence intervals for clinical importance

In a collection of four placebo-controlled trials in hypertension a difference of 4mmHg in terms of mean fall in diastolic bp is to be considered of clinical importance anything less is unimportant. The results, are given in Table 9.2, where gUj and p.2 re the mean reductions in diastolic bp in the active and placebo groups respectively. 

Note the mathematical connection again with the first two trials giving significant p-values and the second two trials giving non-significant p-values. 

In comparing the results from trials 1 and 2 it is clear that the p-value does not tell the whole story. In terms of p-values they are indistinguishable but the first trial has demonstrated a clinically important difference while trial 2 has detected something that is clinically irrelevant. 

14mmHg. This is classically the trial that is too small with low power to detect even large differences. 

Again the p-value is not giving the whole story. There is clearly nothing of any clinical importance in trial 3 but in trial 4 there could be something worthwhile, it is just that the trial is too small. 

---