For centuries, patients have sought medical help for their ailments. 

Just as in the past, however, there are still many illnesses – both wellknown, widespread diseases and rare conditions – that initially cause  few or inconclusive symptoms, and many patients leave the doctor’s  office with an incorrect diagnosis. In addition, diseases may progress  slowly or quickly depending on the individual. Sometimes it is best to  “watch and wait”, while in other cases patients should receive intensive  treatment from an early stage. In the context of ever-increasing  scientific knowledge, such complex medical decisions present physicians  with great challenges. We are inundated with digital information. As  digitization of healthcare systems progresses, new and previously  unimagined opportunities to improve patient care come to light.

For example, machine learning and artificial intelligence technologies  are making it possible to elicit important information from routinely  collected health data: Who is most likely suffering from a certain disease  but remains undiagnosed? Which patient is at a high risk of progression –  and which patients are most likely to benefit from a new treatment?

In this white paper, we present practical case studies for artificial  intelligence applications in medical care. We describe important  methodological approaches for the development of what are known  as predictive algorithms and show ways in which these algorithms are  already being employed successfully in healthcare practice today. We  also discuss the future outlook for the widespread use of predictive  algorithms in medical practice as well as potential obstacles that may  impede it.