Professor Johner, which AI methods are currently being applied in healthcare?
The best known methods are machine learning methods that detect patterns in existing data pools and thus are able to find solutions autonomously. One of these methods is Deep Learning which uses neuronal networks. Today we use it for example in imaging. More recently, boosting procedures such as XGBoost, which is also a machine learning approach, have been quickly gaining ground.
Moreover, AI is used to detect depression by way of analysing language and movement patterns. Other areas of application are production, selection and dosage of pharmaceuticals, error detection in patient records and signal detection in ECG in order to recognize arrhythmias. These are diagnostic methods which have been approved and are commercially available. In addition, AI methods are being used in therapy, for example in triage where they support decision-making processes.
Which types of AI innovations do you think will have the most significant impact on the healthcare system?
Hard to tell. Any prognosis has to take the hype curve into consideration. In imaging AI has reached peak hype. In other areas, I am sure, the hype is still to come. AI will sweep over us in waves at different points. The next focus will be on false diagnoses and treatments which will generate considerable attention. These AI applications will become part of medical practice sooner or later without the patient ever noticing their existence, a bit like anti-lock brakes in the car.
Which regulatory requirements do manufacturers of AI applications for healthcare have to fulfil?
The legislator neither can nor wants to adopt a separate set of laws for each medical device category. The Medical Device Directive, in the near future the Medical Device Regulation, defines the requirements. This legal framework has to be adjusted for the current AI procedures with a focus on verification and validation of the systems, stability and reproducibility as well as fitness for use – issues that are already regulated. The law demands that benefits of the systems be proven quantitatively and that they outweigh the inherent risks of the systems.
This regulatory framework provides good guidance. We are currently working on specific standards for AI. In cooperation with the notified bodies, such as TÜV, I drafted a guideline covering these requirements. This does not only refer to the devices themselves but also to organisations – they inter alia have to define and prove staff competencies.
Which limitations does or should AI have?
AI should not cement bias – something that happens quite easily depending on the data pool the developers use to train their algorithms. Furthermore AI methods must not be used when there is no clear evidence of their benefit and when the risks outweigh the benefits.
We also have to discuss whether we need to regulate the economic framework together with the AI methods. Case in point: should health insurers be allowed to use AI for reimbursement decisions?
There is a fear that AI will replace physicians, particularly in imaging. Is this going to happen?
Already today, there is a shortage of physicians and healthcare will not become easier in the future. I hope that AI will ease the workload of physicians – after all, these systems are programmed to support them in routine tasks. In some areas AI is more powerful than the human brain but that’s not a new experience. A truck, we all agree, is better at transporting things than a human. This is how we ought to look at AI: it is a tool which can perform certain tasks better than we humans. AI will give the physicians more time to actually deal with the patients.
This means we need physicians who can use the information culled from AI in a meaningful way in the diagnostic workup or who can assess and implement therapy suggestions made by an AI system. These skills create an added value, so to speak, for human intelligence.
Professor Dr Christian Johner is founder and owner of the Johner Institut GmbH, a consulting company for manufacturers of medical devices. A physician by training, Professor Johner teaches software architecture, software engineering, software quality assurance and medical IT at the University of Constance, Germany. In 2010 and 2011 he was research associate at Stanford University where he contributed to the development of the ontology editor “Protégé” and taught “Modelling of Biomedical Systems”.