Artificial intelligence is ready for the health care industry, but is the health care industry ready for AI?
As AI progresses, it is incorporating more machine learning into its methodologies. For machine learning applications, a model is fed a large amount of data, like patient MRI images. Then the algorithm interprets these images to determine which contains a tumor and which does not. For someone in computer science, this seems straightforward. But for doctors, who have to close the loop by discussing outcomes and diagnoses with patients, reliance on technology to do previously manual work leaves many with doubts.
However, the health care industry has a bigger problem, one both AI and machine learning can solve: The medical field is woefully understaffed, particularly primary care doctors. And, like any other field, when humans need help performing more tasks with less staffing, AI is emerging as a possible solution. In mid-August, the American Medical Association, the Association of American Medical Colleges and the National Association of Community Health Centers, alongside a series of other high-profile medical associations, announced their support for the Human Diagnosis Project, which uses a blend of AI techniques to provide care for underserved patients. Currently in 10 health centers, the project aims to be in 10,000 facilities with in five years.
Projects like this one could expand the ways that AI and machine learning move the medical field into the future. There are already numerous applications for the technology — and the results are promising. Computer models can now interpret if a patient will have a heart attack with more accuracy than traditional medical guidelines. Google’s Deep Mind can automatically detect some of the leading causes of blindness in patients that still have their vision at the same rate as ophthalmologists. Some companies are even using the application to analyze the molecular structure of pharmaceuticals currently on the market to determine if the medication has another promising, but untested, use.
While there may still be resistance to AI in health care, the field requires mass data interpretation, and an increasingly small amount of doctors are left to interpret it. As medical associations warm up to using AI as a way to help diagnose patients in need, we at Merritt Group anticipate this trend to grow well beyond its current applications. In the near future, nearly every specialty could be transformed by AI, pushing down patient wait times, increasing accurate diagnoses and — yes — even gaining physician approval.