Health Matters: Artificial intelligence is changing modern medicine
Imagine the day: you go to your local health care facility and wait for your turn. Eventually the nurse brings you back to the testing room where you enter a small plexiglass chamber, similar to the airport security scanners. You place your feet on the spots indicated and stand quietly. Whereupon the artificial intelligence running the analysis system promptly spits out a page with your diagnoses and even some specific recommendations for treatment. All without the services (or time) of a physician. And the accuracy is impressive and treatments appropriate. This is the foreseeable future of medicine’s synthesis with AI, artificial intelligence.
Artificial intelligence in medicine is currently in its infancy. But all those knowledgeable agree, it is only a matter of time. In 10 to 20 years, the capability of the technology will likely catch up to the science fiction films and their “predictions.” Most of the mundane tasks now performed by less experienced healthcare professionals may also lend themselves well to being replaced by intelligent computer technology.
In medicine, AI has two two primary types, virtual and physical. The virtual consists of information storage, manipulation and management. Many healthcare systems are heavily involved in the use of AI, especially electronic health records. The physical branch is exemplified by robots, in all their incarnations. Examples are plentiful, like assisting the elderly in a variety of activities, even performing duties under the direction of a physician. Also included in this category are nanorobots, able to function as a drug delivery system.
Medical imaging studies, and there are many types, from X-rays to CT scans, MRIs and echocardiograms, all generate huge amounts of high quality data. What makes it of quality is that all these studies are producing digital information, which lends itself well to computer analysis. AI excels at the identification of patterns in all this complex data and is poised to do so on a scale and at a speed beyond any human ability. The hope is that this technology can be harnessed to help doctors and patients make better health-care decisions.
Although the concept was proposed many years previously, Alan Turing is considered one of the founders of modern computer science, with his first computer created in 1956. He changed the definition of intelligence in a computer, claiming it’s the ability to perform at the level of a human in cognitive tasks. AI is described as the science and engineering of intelligent machines. The term is applicable to a broad range of items in medicine such as robotics, diagnostic technology, medical statistics, and human biology.
Apparently there are different technologies being developed for AI. Advanced Neural Networks, referred to as ANN, seems to be the most popular. ANN’s are a tool for computational analysis are inspired by the human nervous system. These are networks of highly interconnected computer processors called ‘neurons’ that are capable of performing parallel computations for data processing. They are able to make sense of reams of electronic data, such as that about patients, their age, medical history, health status, tests, medical imaging, even DNA sequencing.
Doctors make difficult decisions, using complex data, multiple times a day. They are educated through years of medical schooling, doing assignments, practice examinations, even learning from mistakes. Artificial intelligence uses algorithms to learn how to do their jobs better and better, an actually learning process. This ability means AI systems will learn to make these complex decisions, factoring complex information, arriving at accurate diagnoses and treatments.
Modern medicine is now capable of producing a tremendously detailed analysis of the human body, and the associated challenges to human health. One result: the difficulty of analyzing and organizing this abundance of information. How do we best utilize this large amount of knowledge, data necessary to solve complex clinical problems?
There is no doubt of potential benefits including aiding in the diagnosis of cancer or heart problems. There are many different imaging tests, but all produce huge amounts of high-quality data. This massive quantity of information requires the processing power of AI to be best utilized. Digital images provide complete data sets, without missing numbers, so critical to analysis by artificial intelligence.
What will AI do? These technologies will assist care providers in identifying signals hidden inside massive amounts of data that would otherwise never be seen. Are there limits to what artificial intelligence will achieve? Are there levels of understanding that computers will never replicate? There are too many questions and too few answers. We do know that great potential exists for advances in the technology, but what effects to our society?
The way in which medicine is practiced is going to change, and perhaps sooner than anticipated. You are going to see a computer long before seeing a doctor. Physicians may become interpreters of data more than collectors. The synthesis of AI and developing electronics like wearable devices, monitoring patients’ activity levels and vitals (breathing, pulse rate, etc), will be part of these changes.
Because of the advances in artificial intelligence, it may be possible to eliminate the missed diagnoses and improve our ability to treat the root of a disease and not just the symptoms. Will artificial intelligence bring us good health or a better bureaucracy? Can physicians do a better job and care for more people with the assistance of AI? That is the hope but only time will tell.
Editor’s note: Dr. Conway McLean is a physician practicing foot and ankle medicine in the Upper Peninsula, with a move of his Marquette office to the downtown area. McLean has lectured internationally on wound care and surgery, being double board certified in surgery, and also in wound care. He has a sub-specialty in foot-ankle orthotics. Dr. McLean welcomes questions or comments at firstname.lastname@example.org.