There is much fear around artificial intelligence (AI) and what it will mean to industries -- in particular the healthcare industry. Discussions on AI often center around how it will impact people's lives through job loss. The fact that it's being applied in healthcare, the natural assumption is that its greatest impact might be on physicians and clinicians.
AI has already been utilized to better enable medical diagnoses. There is an AI application via Infervision, a Beijing-based company, that is being used to read CT scans and x-rays. Stanford researchers claim to have created an AI algorithm that can accurately identify and diagnose skin cancers through the utilization of known images of cancerous lesions. According to a study by Frost & Sullivan, the market for AI in healthcare is projected to reach $6.6 billion by 2021.
The greatest value of AI in healthcare exists through solving minor human issues and inconveniences. In a study recently published in the Annals of Internal Medicine, it was concluded that for every hour physicians were seeing patients, they were spending nearly two additional hours on paperwork. The real direction of breakthroughs in AI, and the realization of the potential of the technologies grouped into AI, will not be in replacing people in the performance of their work, or the execution of their reasoning, but in its ability to allow for greater productivity. Through machine learning, we can guide physicians in doing their jobs, by giving them the gift of time so they can use it to connect the dots for improved patient care. The system can present them with fewer dots, guided by what has been deduced as being the best set of information to present. Clinical trials can be found based on disparate data, so physicians can match patients to protocols about which they didn't know, and really don't have the time to avail themselves of if it is at the cost of seeing more patients.
As we assess the impact of these technologies on the healthcare industry, we have to take care to not overstate the capabilities, by oversimplifying the complexity of the decision making process of physicians. One of my favorite studies published by the Harvard Medical School, found that medical students who took art appreciation courses were 38 percent more successful in making accurate medical diagnoses when compared to the control group. They spent time examining various works of art to hone their observational, analytical and communication skills, and this had a direct impact on their capabilities as bedside physicians.
To me, this signifies that there will always have to be a human element in the system. The system will best serve us and the industry by automating that which doesn't require art; the complex decision trees, thereby making certain the right information is always available to the physician at the right time. There will always have to be a human side to the machine, and that means us.
Crucial time and tremendous amounts of resources are lost every day in the world’s healthcare systems. Misdiagnoses cost unnecessary additional tests, result in delayed treatment plans and diminished survival or remission rates from what would have transpired had it been caught and identified correctly earlier.