Jonathan D. Linkous
CEO, Partnership for Artificial Intelligence and Automation in Health Care, PATH
Although lagging many other industries, the use of advanced technology in health care is nothing new. Primary care physicians use apps to identify potential drug interactions when writing a new prescription. Patient records are electronic in almost all hospitals and most doctors’ offices. The coming decade will witness an even more dramatic transformation in the delivery of health care through artificial intelligence (AI), automation and robotics.
Bertalan Meskó, M.D., Ph.D., the director of the Medical Futurist Institute, has identified nine areas of medicine that are already experiencing significant automation-related disruption:
- Mining medical records
- Designing treatment plans
- Assisting repetitive jobs
- Getting the most out of in-person and online consultations
- Health assistance and medication management
- Precision medicine
- Drug creation
- Open AI helping people make healthier choices and decisions
- Analyzing a health care system
Here are four more in-depth explanations about how AI is affecting health care:
1. Image analysis
Medical imaging was the first medical specialty to witness transformation resulting from AI. The days of holding an image up on a light board have long passed. Computers can often analyze and help interpret today’s digital images much more accurately and incorporate vast amounts of related data, such as lab tests, medical history, prior images of the patient, similar images of thousands of other patients and research findings from millions of relevant clinical studies. Philips, Siemens AGFA and IBM are already integrating AI into their medical imaging software systems. Years of studies in automated interpretation of radiology and pathology images and cardiac scans have shown that computers are becoming (or have become) superior to humans in identifying abnormalities. The visual examination of PAP smears is time consuming and expensive, while successful efforts to automate the analysis have existed for more than 60 years. The results of many electrocardiograms are read by the software first and only then reviewed by the provider. Today, very few doctors would dare to read an electrocardiogram without having it first read by a computer, if one is available.
Medical robots are already used to perform repetitive tasks and those that require intense dexterity. Robotic surgery has been in use in minimally invasive surgeries for over 30 years, and by 2025, surgeons could perform as many as 170,000 new robotic procedures annually. InTouch Health, manufacturer of telemedical robots, enables providers to remotely move around a hospital. The robots are in over 2,000 health facilities, boasting 750,000 documented clinical encounters.
3. Clinical decision making
Science fiction has long depicted how computers can provide medical care, but multiple projects using AI to help make medical decisions and provide treatment are here today. Many such applications use deep learning — a technique similar to our brain’s neural network — by sifting through vast, unstructured, seemingly unrelated data in order to teach themselves, identify relationships, make classifications and formulate predictions. It’s similar to the intellectual process used by skilled experts to make instant judgements based on years of experience and knowledge from a variety of sources. In medicine, deep learning is being used to predict patient outcomes and identify treatment plans, including procedures and medications tailored to the patient and based on volumes of historical data and research. Such exciting applications are starting to be used for everything from cancer to depression. Even apps that help patients diagnose themselves are in experimentation.
4. Effects on the medical profession
AI, automation and robotic applications in medicine are called “decision support,” implying that the doctor is still in control. However, these “virtual assistants” are learning fast, and some medical professions could see a time when such advanced technologies are more reliable, productive and trusted by the patient. That could result in a redistribution of some types of providers. But, like other industries, change has been constant in medicine and such change is getting faster.
It’s no surprise that the use of advanced technologies is accelerating in health care. Transportation, finance and entertainment already widely use AI. With the cost of health care spiraling and the population aging, it’s time to discover and deploy the benefits of technology to help us take better care of ourselves.