Recording vital parameters such as blood oxygen saturation, pulse and temperature is essential for asessing the current medical status of patients and can even facilitate early diagnosis of diseases, before they break out severely. For such medical applications, the multisensor systems worn on the body for monitoring vital parameters have to function reliably and deliver high quality data. With the multimodal sensor data, machine learning methods can be used to detect features in the data sets that allow conclusions to be drawn about specific diseases. Spontaneous change in vital parameters over time can signalize diseases (such as sepsis or emergence of wounds) early on, and via a respective warning system facilitate their prevention. This shows that there is a very high potential in the continuous monitoring and evaluation of vital parameters in the diagnosis of diseases.