In today’s world, AI has been able to transcribe clinical notes, draft MyChart messages, and even interpret X-rays. The terms “ambient-scribe” and “in-basket automated response” have ascended with AI’s rise in popularity in the healthcare industry. AI has immensely impacted patients, physicians, and the healthcare systems. However, these aren’t the only prominent AI applications in healthcare. These two stories of health tech leaders uncover the recent unconventional AI tools that solve industry problems.
Dr. Jude Kong grew up in Shiy, a tiny village in the northwestern region of Cameroon, with the nearest hospital being 4 hours away with one doctor on staff. Kong explained that if someone were sick in the village, they would resort to self-medication. If someone were dying, they would need to be carried to the hospital as there were only two cars in the community. Due to the support from the women in Kong’s community, he became the first in his village to make it through his schooling years, including university. He had his first doctor’s appointment while studying in Italy.
Currently, Kong is a professor and the director of the AI and mathematical modeling lab at the University of Toronto. He also leads the Africa-Canada AI & Data Innovation Consortium and the Global South AI for Pandemic & Epidemic Preparedness & Response Network. Kong uses AI to customize proactive tools to improve public health access in communities like Shiy. His work has aided communities around the world. For example, an AI model in Ethiopia can analyze photos to discover if a patient’s paralysis indicates polio. In Peru, an AI-powered breathalyzer can diagnose respiratory illnesses. These tools are built and trained to target specific problems in a particular group of people that health systems commonly neglect. Kong’s organizations include local communities and researchers from the concept phase to implementation, which fosters trust in the new technology.
In the healthcare AI market, radiology plays a key role, accounting for 76 percent of FDA-approved AI and machine learning offerings. These tools can read and interpret scans, but they have other applications as well, according to Dr. Elisabeth Garwood, associate chief medical information officer and vice chair of AI and clinical innovation at UMass Memorial Health.
At UMass, at least 40 AI algorithms are run in its clinical workflows, making the imaging process more efficient for doctors and patients. One prioritization algorithm flags exams that seem unusual, so critical patients get results quicker. Another AI-powered billing software automatically codes easier cases so medical coders – who are short in supply – can prioritize more challenging cases.
Thanks to the healthcare system’s acceleration algorithms, patients have noticed a decrease in time spent in the MRI machine. It takes time to acquire sufficient data to create a clear image of the anatomy, and the picture can be grainy if the machine is forced to go faster. However, machine-learning algorithms can process accelerated data and reconstruct high-quality diagnostic images in less time.
MRIs are known to be extremely useful in medical decision-making but can be uncomfortable and loud. Approximately 1 in 10 MRI scans have to be terminated since up to 15 percent of patients experience severe forms of anxiety. AI is helping to make the process significantly more tolerable for patients so doctors can gather the results they need. “The acceleration algorithms at UMass are making our MRIs 25 percent faster, and that hacks the patient experience that they’re in the MRI for less time,” says Dr. Garwood.
Leave a Reply