Healthcare organizations are increasingly embracing machine learning and other forms of artificial intelligence (AI) to improve patient experiences and care outcomes. AI can be thought of as any task performed by a human that can be performed as well as or better by a computer.
The recently concluded HIMSS Annual Conference and Exhibition in Las Vegas offered nearly countless examples of how potential applications for healthcare are coming into the marketplace (including EHR uses), some promising to have exponential impacts on practice efficiency.
Now, the focus of many AI-based applications is shifting beyond operational concerns and into the clinical side. The following are just a sampling of how AI is being used to transform healthcare:
This expanding list of applications has pushed the amount of venture capital spending on AI-enabled healthcare startups from less than $200 million in 2015 to nearly $800 million last year, according to PitchBook.
In a recent article published in Forbes, Dr. Robert Pearl discusses three applications of AI:
Algorithmic solutions: Evidence-based approaches programmed by researchers and clinicians such as treatment alternatives for cancer and chemotherapy protocols
Visual tools: Applications to recognize patterns in medical care where the human eye can fail, such as cancer screening and detection from storing thousands of images for comparisons
Medical practice solutions: Applications that reduce errors and improve compliance by making sure patients take their medications correctly and on time
AI can also be used to help patients and office staff interact with patients reminding them of their upcoming appointments and make sure they complete any necessary procedures prior to or after their office visit.
A March 13 MGMA Stat poll with 1,257 responses asked if respondents would consider receiving medical advice through AI in a healthcare setting: 35% of them said yes, 40% said no and 25% said they were unsure. “Yes” respondents’ reasons included a willingness to try AI once or twice to compare it to current processes and the belief that it might be convenient or a better alternative for minor health issues/concerns. Those who responded “no” said that they were cautiously optimistic in the technology but noted that complex issues need discussion between providers and patients, a belief that you can’t replace hands-on medicine. They also noted concern that AI can’t understand human emotions, stress and pain. For those who said they were unsure, many respondents who didn’t know what AI is or its application in the healthcare setting.
Healthcare is complex, and the severity of illness and the type of information needed plays a big part in the role of AI. However, with the desire for convenience and immediate attention, patients will tolerate and choose a certain amount of AI if they perceive that it provides value to them based on their needs.
But as with any new technology, finding partners to create solutions depends largely upon the tech side’s understanding of how the American healthcare system works. More importantly, practice leaders must understand how these partners intend to provide value and improvements to the processes they have in place and must master for successful operations.
Another key element in embracing AI and other technological solutions is trust. To leverage your practice’s data into an AI solution, you must be able to capture accurate and complete data from your EHR and other practice management systems.
Additionally, practice physicians need to trust that the proposed solutions will not simply improve the overall level of care and workflow of the practice, but also show a return on investment.
As noted in the MGMA 2017 DataDive Better Performers data, better-performing, physician-owned practices managed to spend less on information technology (IT) than their peers, signaling that being judicious with IT spending leads to more efficient solutions and better outcomes over time. This is certainly something to keep in mind considering how to use AI in your practice.