Artificial intelligence as a diagnostic aid, patient app and telemedicine: Doctor Ludger Killich wants to revolutionize patient care. The first test is promising.
The first thing I tell an AI is that my nose is running, my head hurts and I feel sick. Immediately after I make an online appointment with my family doctor Ludger Killich, she asks me to describe all my complaints to her. The rules are simple: I write – and she asks politely. Am I breathing quickly? Do I have pain or chills? And how bad is the pounding in my head? The artificial intelligence is even interested in the color of my snot. The whole question-and-answer game takes a few minutes. Then it seems to know enough.
She sends my answers directly to the doctor along with her assessment that there is a 47.42 percent probability that I am suffering from bacterial sinusitis. According to Killich’s AI, other possibilities for my symptoms include viral sinusitis, an adenovirus infection, tonsillitis or a middle ear infection. Is this what the future of medicine looks like?
Artificial intelligence is a valuable co-thinker
I have to admit that my symptoms were all made up, and I only find out about the AI’s considerations because Dr. Killich shows me what it predicted. There is a reason for the fib: I want to find out how intelligent algorithms can improve care at the GP’s. The hybrid care center “i.Med MVZ”, which Killich set up six months ago in Halle, North Rhine-Westphalia, together with a partner, actually feels a little like looking into a crystal ball to me.
This is not about AI replacing doctors. “The AI cannot make the diagnosis for me.” However, the AI’s preliminary anamnesis contains important information that he can use to decide whether a Patient should come to the practice for certain examinations or whether he should refer him directly to a specialist or even to the clinic after a personal consultation. The AI is like a co-thinker. With all the medical knowledge in mind, it can remind him that some symptom combinations could also be a rare disease with perhaps a five percent probability.
The results of the blood pressure measurement are immediately entered into the app
In most cases, however, it is about giving the all-clear. “That is just as important,” says Killich: “A third of patients no longer have to come to the practice because I can discuss with them in a video appointment what can be done for a simple infection.” The discussions on site are also now much more focused when the patient does not have to start from scratch.
New technologies like these also save time – for the patient, the doctor and his staff. The same applies to Killich’s telephone assistant, who sorts the callers’ concerns in advance. Prescription request? Sent digitally. New appointment? Arrives by email straight away. “I’ve always been interested in how digitization can simplify and improve work processes and structures,” says Killich. “That’s why we also have an app in which patients can independently enter data such as blood sugar or blood pressure measurements, view parts of their patient files or chat with our practice staff.”
The practice concept could also be one of the urgently needed solutions for rural areas. There are currently over 5,000 general practitioner positions vacant in Germany. “Many existing practices lack successors,” says Killich. “With an offer like the video consultation, I can also take care of patients who live further away. My employees can do some tasks from home and decide more freely about their appointments because they could also offer a video consultation at eight in the evening.”
Used correctly, AI could improve health outcomes by up to 40 percent
I’m thinking of my family doctor’s practice in Hamburg’s Schanzenviertel.
Just a few weeks ago I was standing in line at the reception desk and could see how overwhelmed the medical staff was becoming. They were searching for notes, prescriptions were being thrown into thick piles, and the phone was ringing so loudly in the background that you wished someone would finally answer the phone, even if it meant waiting longer. Maybe it was an emergency.
How much better could we serve all these patients with a little technical support? Why do people still have to sit in a crowded waiting room for hours with a cold? And does a doctor really have to do everything herself? In Killich’s practice, for example, a trained care assistant takes over some of the home visits. Digitally upgraded, with an ECG device on the go, she can even determine initial blood values on the go and contact the doctor via tablet if necessary – this is called telemedicine.
The prestigious Harvard University estimates that AI alone, when used correctly, can improve health outcomes by up to 40 percent and reduce treatment costs by up to 50 percent.
Today, the digital eyes of an AI can already extract far more information from an X-ray scan than human experts. They often calculate the probability of sepsis from blood values better and reliably distinguish between unusual moles and skin cancer. A few weeks ago, a clinical study showed for the first time in a direct comparison: The probability of death was lower for patients when an AI warned the doctor about unusual ECG patterns.
Doctors are still needed
Doctors will still be needed, if only to classify the information from the AI. The best treatment depends not only on the measured blood values, but also on the overall situation: Does the elderly patient have help at home? Is the patient mentally stable? What conclusions can be drawn from his previous medical history? In the best case, the family doctor knows all of these things about his patients, but the AI does not. In addition, there is a completely different medical quality: the encouraging word, the listening ear or the hand on the arm. We know from countless studies how important these factors are for the success of treatment.
The German Ethics Council warns that the use of AI must not lead to a further devaluation of talking medicine or to staff cuts. “A complete replacement of medical professionals by an AI system endangers patient well-being.”
Finding the ideal combination of human experience and intelligent technical support will be one of the most important tasks of medicine in the future.
Will my family doctor also use the new options? The Health Foundation recently asked around 500 practicing physicians what they thought about the development of AI-supported contact points for initial assessment. Around 21 percent thought this was sensible and feasible. Almost 40 percent, on the other hand, did not think so at all. I can imagine which group my doctor belongs to – the second. What a pity.