Bias

In clinical trials, a bias is a systematic error or bias that does not rectify itself on average when repeated. A bias causes a deviation from a high or low measured value, which can be caused, for example, by the incorrect calibration of devices.

The causes of a systematic error are usually to be found in the study design or in the conduct of the clinical study. For example, in a selection bias, groups with a non-existent or particularly pronounced risk profile may be incorrectly weighted. In the measurement bias, the endpoints are incorrectly weighted, whereas in the wear and tear bias, results from studies with a low risk profile are considered valid.

BiAS is also a biometric-statistical program package developed by the University of Frankfurt for the calculation and evaluation of clinical studies. Bias includes in particular methods of experimental design (sample size and power calculations) and additional programs for specific questions that are not included in other statistical programs such as SPSS (e.g. Fliess-Kappa).

The Clinical Research Organizations take over the case number calculation of studies and prepare plans and protocols for the conduct of clinical studies.