Sample size calculation
When planning a clinical study, sample size determination is one of the most important factors for the statistical analysis and hence for the whole study. If the sample size is too large, it can unnecessarily increase costs, resources, and time. If the sample size is too small, it will not provide valid results.
An insufficient number of cases results in low power and therefore the results are not very reliable. Low power can be recognized by very wide confidence intervals. These are also known as barn doors, i.e. the estimate is extremely imprecise and therefore not meaningful. As a result, there is a risk that all or parts of the study will have to be repeated.
If there are no regulatory specifications, sample size should rest upon statistical calculations. Therefor information on the assumed performance is needed. Depending on the result type of the IVD assay, different parameters must be provided by the sponsor and might be based on literature or former studies. Here, it is reasonable that calculations are based on rather conservative assumptions, i.e. the worst but still acceptable scenario.
