Bias

Author:Prof. Dr. med. Peter Altmeyer

All authors of this article

Last updated on: 29.10.2020

Dieser Artikel auf Deutsch

Requires free registration (medical professionals only)

Please login to access all articles, images, and functions.

Our content is available exclusively to medical professionals. If you have already registered, please login. If you haven't, you can register for free (medical professionals only).


Requires free registration (medical professionals only)

Please complete your registration to access all articles and images.

To gain access, you must complete your registration. You either haven't confirmed your e-mail address or we still need proof that you are a member of the medical profession.

Finish your registration now

DefinitionThis section has been translated automatically.

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.

Authors

Last updated on: 29.10.2020