Artificial intelligence: German Medical Association on the status quo and plans

The German Medical Association (BÄK) has analyzed the opportunities and challenges of artificial intelligence in medicine in its statement "Artificial Intelligence in Medicine", as first reported by the Ärzteblatt. According to the statement, artificial intelligence (AI) plays a decisive role in overcoming the shortage of specialists and the increasing number of multimorbid patients. Cost savings can also be expected.

AI systems can be used in various areas of medicine and have already been used for years in imaging procedures, sometimes even as a second opinion. Through machine learning, AI algorithms can recognize patterns in large data sets and help to diagnose diseases more quickly. In diagnostics, the German Medical Association hopes that AI will help improve diagnosis in areas such as pathology, radiology, ophthalmology and dermatology. Personalized medicine is also possible with the help of gene data, for example. Virtual reality is already being used in operating theaters, as are care robots that support hospital staff.

"Smart hospitals" and interoperable data transmission

Hospitals could also be developed into "intelligent" facilities with more efficient processes in the areas of diagnostics, intervention planning and process management. The Medical Association refers to an existing project at Essen University Hospital. Data collected there, for example from medical histories, laboratory tests or imaging, is collected in a "data protection-compliant and interoperable manner" and compiled in a "digital patient file on the Smart Hospital Information Platform (SHIP)" and is then used for the further development of AI systems, for example. AI can also automate repetitive tasks in the healthcare sector, such as documentation.

"At Bonn University Hospital, the Innovative Secure Medical Campus is developing a digitalization and AI concept that can also be transferred to other clinics," the statement also states. Overall, AI could ensure better communication and coordination between the outpatient and inpatient areas and support telemedicine. Thanks to telemonitoring, for example, long-term monitoring based on vital data is possible. In addition, chatbots can support patient education, as can apps for disease management.

Another relevant point mentioned by the Medical Association is the use of AI in research and teaching. With the help of AI, such as Google DeepMind's AlphaFold, it is possible to predict protein structures, for example. AI also plays a role in blood glucose monitoring and can automatically control insulin delivery. The German Medical Association refers to the EU research project MELISSA ('Mobile Artificial Intelligence Solution for Diabetes Adapted Care') for patients with type 1 diabetes, which aims to provide fully automated diabetes management. MELISSA integrates data from multiple sources – including information on blood glucose levels, the amount of insulin injected, laboratory data, medical history data and lifestyle data.

The German Medical Association is also considering smart implants: "AI applications can be used to collect movement data and measured values on the bone implant and create a favorable healing situation for a fracture via sensor-controlled variation of stiffness. Such so-called 'smart' implants combine sensors and actuators, i.e. the perception of their environment and the reaction to it". The "currently developed implants" can be used for "any fracture healing". In the future, the BÄK also hopes that implants will be able to be adjusted after surgery with the help of AI. Emergency situations could also be simulated with the help of AI.

Ethical and legal aspects

However, the introduction of AI also raises ethical and legal questions that are currently being discussed. However, the responsibility for diagnoses and treatment decisions will presumably always remain with doctors. Currently, AI is primarily used to support doctors' work. According to the BÄK, medical staff should be transparently informed about how AI systems work so that they can critically scrutinize them. To this end, according to the German Medical Association, AI should provide answers as to "which combination of input parameters" no longer allows a reliable statement to be made and "which potentially medically relevant aspects" are no longer taken into account by the model. Data protection and compliance with confidentiality are also key issues that must be taken into account, as AI systems usually process personal data.

(mack)