Possible Clinical Interpretations of Statistical Results

When a randomized clinical trial is conducted, a subject sample is chosen from the population of all possible subjects, and this sample is then randomized to the treatment groups. Analysis of the trial s data provides a precise result for that particular sample. However, importantly, while the sample can contain several thousand subjects, this is quite likely to be a very small percentage of the population from which that sample was chosen. Had a different sample of subjects been chosen, the chances of the data obtained being identical is so infinitesimally small that we can safely say that they would be different. The question of interest here is How different would they likely be Ideally, we would like them to be extremely similar, thus providing a result that is extremely similar to the result of the original trial the more similar the results from a second trial, the more confidence we could reasonably place in the results from the original trial. 

While the word confidence in the previous sentence occurs in its everyday use, the term is also used in Statistics in a precise manner, analogously to the statistical terms Normal and significant. Confidence intervals constitute a range of values that are defined by the lower limit and the upper limit of the interval. These limits are symmetrically placed on either side of the sample mean. A commonly used Cl is the 95% Cl. A commonly expressed view of a 95% Cl is that one can be 95% certain that 

New Drug Development Design, Methodology, and Analysis. By J. Rick Turner Copyright 2007 John Wiley Sons, Inc. 

---