Generative AI, Health Informatics and the Future: A Chat with Michael Bice
Michael Bice served as a senior executive of academic medical centers and large healthcare systems for over 25 years. During that time, he was an adjunct professor at the University of Florida College of Public Health where the Dean of the College was Robert Frank, later appointed Provost at Kent State University. Frank hoped to establish a master’s degree in health administration at Kent and tapped Bice for the job of creating it. After much research and discussions with major health employers in Northeast Ohio, it was determined that the area was already well supplied with M.H.A. programs. But what it didn’t have was master’s degree program in Health Informatics.
CCI: What happened when you recommended an H.I. program?
MB: I took that recommendation back to the various university councils, and the H.I. master’s degree program was approved for the School of Library and Information Science [now the School of Information] in 2010. From the outset, it was an online program only. And they asked me to come out and get the program started. I was the only faculty member at that time. I recruited faculty and students, designed the curriculum and created an Advisory Council. I taught the first class (Introduction to HI) and recruited 25 students that very first class.
CCI: What was the state of health informatics then?
MB: Vastly different from what it is today. In 2010, the driving force for health informatics was the conversion from paper to electronic health records, and that was a very big deal. You had to have people who were trained in how to implement the EHR, how to make improvements as needed, how to use clinical analytics and how to overcome user design issues. There was real demand for health informatics graduate students to help institutions make the conversion. It was a very timely development, and Kent was uniquely positioned at the time. There were only three accredited programs in health informatics in the whole country.
CCI: Kent State’s health informatics program was just recently accredited. What’s the value in accreditation?
MB: If you were creating a program, would you make it out of thin air? The Commission on Accreditation of Health Informatics and Health Information was the resource, and the program was designed to be accredited. CAHIIM Accreditation means that your institution has met or surpassed the standards of your profession and that you are prepared to take your place as one of the prominent HI programs in the country.
CCI: What was appealing to you about the field of health informatics?
MB: Once you understand that the future of HI is linked with advancements in AI, then you ask yourself, what is the future of artificial intelligence in healthcare? Since 2010, we have witnessed the development of virtual reality, augmented reality, predictive analytics, care robots, voice activated devices and remote patient monitoring. That's one of the continuing attractions of H.I. You’re in a great space in healthcare, and it's growing rapidly.
CCI: How would you say this technology has changed medicine?
MB: There are basic parts of the delivery of healthcare which won't change, irrespective of technology. The relationship between the practitioner and the patient, standards of care and standards of excellence. They're the bedrock. But then you say, Okay, what is the promise of technology? Tech innovation is beginning to branch out beyond basic smart watches into care robots, service robots that do infection control, and virtual reality's use in therapies to minimize pain or reduce the effects of PTSD. We're only scratching the surface with generative AI (ChatGPD). A Journal of the American Medical Association study found that reviewers picked an AI chatbot’s replies nearly 80% of the time [over a human physician’s] because the replies were longer and more empathetic. Now, that's kind of amazing. We're at a point in which the generative AI will begin in non-clinical areas. But the long-term potential is in clinical care.
CCI: What kind of guardrails are in place to ensure that the technology is safe and used ethically?
MB: Well ethical issues in medicine were present prior to generative AI, but with generative AI, they have become extraordinarily important because this AI is so easy to use and misuse. So right now, we have Google, Microsoft and Meta driving this AI train, and they have conflicting objectives. They have to build consumer confidence in the safety of AI, but they're also in business selling a product in demand. We all await the development and promulgation of guardrails such as watermarking and strict requirements for AI accuracy.
CCI: What kind of changes do you see AI making to fields like health informatics?
MB: Well, there's reskilling because you will have to upgrade people's skills. There's risk management for AI, which is a whole new field. But then there's prompt engineering, and prompt engineering has to do with asking the right question. For example, if I say to ChatGPT “What are the elements of the social determinants of health?” Bang! Here are the five in a list, but that doesn't get me anywhere. So, the prompt engineer would say “Tell me why a primary care physician does not use social determinants of health when they do a history and physical?” And you get a lot of information which is very useful. Boston Children's Hospital has just hired a prompt engineer. Some H.I. grads from Kent are going to be prompt engineers. Your ability to ask the right question and do it in a way that elicits the most useful response is a very, very important skill set.
CCI: And what would you tell potential students who are looking at the HI program?
MB: If you want to be a practicing health informatics professional, Kent State is the place to be. It has history. It has connections with the provider community that goes back years. Its students and its graduates are all very competent people. It has a great faculty. In the long haul, the Kent State program in Health informatics is going to be an international leader.