How do you use the answers to both questionnaires

The ideal trial will be designed to detect reliably the smallest difference identified by these summaries. However, discussion of the questionnaire results with potential collaborators should inform the decision as compromises will often have to be made; these are discussed towards the end of this chapter. Even so, using such an approach has a number of benefits. The questionnaires involve the potential participants much more closely in the determination of sample size, a process which often does not get as much thought as it deserves. Equally importantly, they enable the person determining the sample size to understand the degree of clinical uncertainty prevalent and where appropriate, build this into the sample size calculation. In the CHART example above, the median 'expected difference' was 10 per cent, the average range of equivalence ranged from 11 to 14 per cent; the trial was designed to detect an absolute increase in survival of 10 per cent. In this case, the clinicians demanded a substantial benefit to CHART before they would consider using it routinely, reflecting the 'costs' of implementing CHART in practice.

These questionnaires can be useful not only in designing trials, but also in monitoring their progress. Reference [10] shows how data such as these can be used during interim analyses to help determine the impact of the interim results on clinicians' opinions, and hence to guide when a trial should stop.

0 0

Post a comment