In recent years, Artificial Intelligence is being increasingly prevalent in business and society. Since it has the potential to transform many aspects of patient care as well as administrative processes within providers, payers and pharmaceuticals it is being applied in healthcare more and more everyday. A number of research studies already show that AI can perform in key health-care tasks, such as diagnosing illness, as well as or better than humans. Algorithms already outperform radiologists today to spot malignant tumours and instruct researchers on how to develop cohorts for expensive clinical trials. For a number of reasons, however, we assume it will be several years before AI replaces humans with large medical process domains. This is shown through one of the studies in the Journal of American Medical Informatics Association that studied AI-assisted EHR documentation.
This being said, today whenever AI is used in the healthcare field, it always needs physician assistance. The study shows that while AI documentation assistants (or digital scribes) offer great potential in the primary care environment, they still need to be monitored by a person before there is good evidence for their autonomous potential. Researchers have consulted with primary care physicians who often use EHRs for reporting purposes to clarify their opinions on potential assistants in recording AI. Three main trends emerging from the discussions were identified: professional autonomy, human-AI collaboration and new care models. Researchers have collaborated with primary care physicians who often use EHRs for documentation purposes to clarify their opinions on potential assistants in recording AI. Three main trends emerging from the discussions were identified: professional autonomy, human-AI collaboration and new care models.
First, the doctors underlined the value of their ability to care for patients in their own way with the capabilities offered by AI technology. “If [patients] think we ‘re just getting suggestions from a machine, then maybe they should just get suggestions from a machine. I think it’s getting harder to persuade them that our ideas are more important than anything they can pick up on the internet,” one doctor said.With respect to human-AI cooperation, physicians have shared a range of views about what roles should be assigned to AI. Many assumed that activities such as reporting, referrals and other paperwork could be supported by an AI system. Most said empathy would be lacking in the AI systems. “GPs raised many concerns, including some possible biases in patient data and system design, the time required to correct the errors and train the system, difficulties in coping with complex cases, and AI auditing,” the researchers wrote.
The problem of reducing the burnout of clinicians affiliated with EHR has loomed large, with vendors and researchers seeking to find major triggers – and possibly remedies, in turn.
As one such solution, AI was introduced, with many major EHR vendors offering plans to integrate the technology into their workflows. But human feedback continues to be crucial, as noted by the new JAMIA study and other studies. AI could “bring back sense and intent in medicine practise while offering new standards of effectiveness and accuracy,” Stanford researchers wrote in a report by the American Medical Association Journal of 2017.However, they added, physicians must “proactively direct, supervise, and track the implementation of artificial intelligence as a patient care partner.” So even though technology have its advantages AI-assisted EHR documentation will need physician assistance.
