Statistical Power


Health Economics

  • In health technology assessment, 'power' most frequently refers to the statistical power of a trial. In general, a trial ought to be big enough to have a high chance of detecting a statistically significant effect on cost or outcome and one that is also clinically or biologically significant. Sample size is therefore critically important. The researcher needs to decide the degree of difference between two groups being compared that would constitute a minimally clinically significant effect. How large the sample needs to be to deliver statistically significant results can be determined by using a statistical nomogram. The power of the study (moderate, high, or very high) is the chance of detecting a pre-specified true clinically relevant difference between the groups at a pre-specified p-value (usually p < 0.05).