[Utility of medical devices: approaches to planning and conducting clinical trials].

Medicines and medical devices do not only differ in the approval process, but also in the aim and conduct of clinical trials. We first discuss important differences between medicinal products and medical devices. Emphasis is put on the differences in the framework for clinical trials. We point out that a different analysis set should be used in clinical trials of medical devices when compared with medicinal products and medical devices in the USA. Specifically, regulators generally ask for the full analysis set based on the intention-to-treat principle as proof of efficacy of medicines. A central aspect of clinical trials of medical devices is that they have to be tested under normal conditions of use according to the performance data. As a result, all data acquired while the medical device was not during normal conditions of use should be excluded from statistical analyses. We discuss statistical methodological particularities of medical devices, such as blinding and the control of placebo effects. Using the conservative treatment of anal incontinence as an example, we show that comprehensive technical and physical knowledge is required for assessing the utility of medical devices. Finally, we consider reporting of severe adverse events and of severe adverse device effects of medical devices.