Artificial intelligence (AI) is being used in medical settings. Initially, it began to be used in the first stage of examination: the medical interview. One company’s AI-powered medical interview system changes questions sequentially based on the patient’s symptoms to explore their condition. This company built an algorithm that analyzes the relationship between symptoms and diseases based on over 50,000 medical papers. This algorithm creates a system that doctors are provided with comprehensive information as soon as a patient enters the examination room. AI has long been prevalent in the area of examinations, which follow the medical interview. This includes using AI to analyze X-ray and endoscopic images, or electrocardiogram waveforms. Technology that uses AI to analyze electrocardiogram waveforms to detect diseases early has been put into practical use. Once the results of the medical interview and examinations are complete, a diagnosis is made, and the process moves on to selecting a treatment method. AI is also improving accuracy in this treatment selection process.
In Japan, since the 1990s, clinical guidelines summarizing the optimal examinations and treatments for each disease have been published. Clinical practice guidelines summarizing optimal examinations and treatments have been created and published by various academic societies. This has been a factor in raising the standard of medical care in Japan. Furthermore, doctors have increased their experience and knowledge, and those with more experience and knowledge tend to be called “master physicians.” Clinical practice guidelines have become the “textbook for medical professionals.” AI has learned from these guidelines and can now learn and analyze a vast amount of cases that a human doctor could not experience in a lifetime. AI-based diagnosis is based on a huge number of cases. It seems that AI medicine has the potential to provide every patient with a diagnosis from a “master physician.”
In the medical world, there are also opinions that the focus should be on preventive medicine rather than curative medicine. Early detection of signs of illness leads to early treatment and extends healthy life expectancy. The environment for such early detection and early treatment is being put in place. Health data that can quantify the signs and progression of disease is called digital biomarkers. With the spread of smartphones and wearable devices, this health data (such as electrocardiograms) has become easy to obtain. Electrocardiograms can also be measured 24 hours a day, and the results can be recorded in the cloud. The system can analyze respiratory rate, heart rate, and other data stored in the cloud, and if an abnormality is detected, it can notify the individual concerned. Upon receiving the notification, the individual’s electrocardiogram data will be further analyzed by a collaborating physician using the AI system, and treatment will be provided. The system is designed so that if it finds waveforms similar to those of past patients in the accumulated data, the AI will issue an alert. AI systems that support diagnosis and medical treatment appear to be shifting from a nascent stage to a widespread adoption stage.
