A team of researchers from China has developed a computer program that learns, patients data, to evaluate and to detect children’s diseases from a harmless cold about Asthma to life-threatening inflammation of the brain. The program can be used to distinguish in emergency departments quickly urgent from the less urgent cases to support Doctors in training, or to help in the diagnosis of diseases with complex symptoms, write the researchers Huiying Liang from the center for children and women’s medical University hospital in Guangzhou. However, experts criticize the study because of various defects.
For their study in the current issue of the Scientific journal Nature Medicine (Liang et al., 2019) appears to be fed a computer program with a data set of about 800,000 of its electronic patient records of children. This is not included the medical history, laboratory results, and abnormalities that had been shown in the examination by a doctor but the diagnosis. Next, the information from the medical records were translated using a method called Natural Language Processing. This technology processes natural written or spoken language by machine and brings you in a structured, Algorithms processable Form. In a further step, an algorithm associated with the data, then finer and finer classes: first, the affected organ systems, groups of related diseases, to a final diagnosis.
was calculated As the diagnosis remains unclear
Subsequently, the scientists compared to scientists, the accuracy of the computer-aided diagnosis with the human children’s doctors – they used another set of about 400,000 data. The result sounds impressive: The Computer calculated diagnosis, agreed, depending on the disease in 79 to 98 percent of the cases with the attending physician.
According to expert opinion, the programme has, however, large defects to be already in clinics or doctor’s offices. According to the Medical computer scientist Thomas Neumuth, University of Leipzig, the applied Algorithms are not traceable. “Thus, it is not possible to find out how the decision is reached,” said the specialist for computer-based medical technology and surgery, the journalists of the German Science Media centre (SMC).
Science Media centre (SMC), The German Science Media centre
The Science Media centre in Germany, in short, SMC is a Service for quality assurance in science journalism. It is funded by universities, research institutes, companies and the media. Journalists of the SMC all editors for free of professionals tested information, scientific assessments, studies and quotes from researchers to the current messages available to subjects from the fields of medicine and life science, the environment and the climate, energy and mobility.
shareholders of SMC, the German Science journalists (WPK) and the Klaus Tschira Foundation in Heidelberg. Was founded in the SMC in the year 2015, according to the British model.