Those of us that have been around a while can likely recall special events that captured the imagination of everyone. Those of us that visited Walt Disney World (WDW) in Orlando, Florida and were fortunate to see the Carousel of Progress,™ originally part of the Worlds’ Fair in 1964, and was also in WDW from 1975 to the present. For those that never saw it, it was originally sponsored by General Electric however, it is now all Disney owned and is tucked away at Tomorrowland. Essentially a darkened theater where the main “carousel” stage spins around to show off four different scenes depicting the future as envisioned 20 years prior. Each of the scenes takes place in a different decade of the 20th century, with a patented Disney Audio Animatronic, named John Progress, enthusiastically explaining the technological innovations made in the era — be that electric lights in the 1900s, appliances in the kitchen during the 1920s, or dishwashers in the 1940s.
Some of you may also remember the much-touted matches between world chess champion Bobby Fisher and the Goldblatt computer in 1977. Even though the computer with specific mathematical algorithms designed for the chess game Bobby Fisher still won three out of three games. One must wonder how well he would do against ChatGPT-4?
The reason Carousel of Progress and Bobby Fisher’s chess match came to mind is that the future is moving so fast it is hard to keep it all in frame. Not only did we mention in the last blog about the people who are excited by its development, but also those that are either frightened by the prospects or are predicting doom and gloom. As we continue development of our own applications for AI in healthcare, I read hundreds of articles about this space, this recollection made me think about what parts of society will be changed by AI, you ready?
As we have noted, there are amazing possibilities for AI, especially in healthcare. The one people connect to the quickest seems to be in decision support for physicians; some people even think they can replace physicians. This is complicated because given the relatively small amount of data physicians are provided by their patients, with the augmentation of simple diagnostic tools, providers of care they do amazingly well. Training, practice, and flow of patients keep their skill well sharpened. However, with the further use of technology and more information AI can prove to be a game changer. Let me elaborate.
The average patient (pre-Covid) waited about 45 minutes to see a physician, and in general, they spent less than 10 minutes with the provider. In part, some of that 10 minutes were spent “entering” data in to the EMR system, and I am now convinced this largely a waste the time. The more information the doctor had, the more they had to review and assess. This resulted in less time with the patient, and the more quickly a patient would get moved through the gauntlet of waiting room, intake and triage, doctor visit and out the door.
In an AI and Technology-driven healthcare model, including the use of Remote Patient Monitoring (RPM), the software augments the nurse or clinician that normal does patient in-take with vitals and questions regarding the reason for the visit. In this rudimentary model the interconnected devices and AI system could be collecting data for days or weeks before the visit. Allowing the AI platform to access the patients’ medical record, an advanced AI engine could even summarize the patient’s co-morbidities and align the patient complaint with their history and formulate some questions ad hoc regarding their condition.
This information is then quickly summarized by the AI-engine and can be given to the doctor as a quick snapshot preparing the doctor before they walk into the exam room with the patient. The AI program has essentially codified the patient’s relevant medical data and aligned the salient parts with the immediate complaint and then provided the physician with the vital signs, summary of complaint, and most likely diagnosis.
This is truly rudimentary AI, and while new to healthcare, these concepts and capabilities have been a part of many technological systems from auto repair to aviation. Engine vibration sensors on jet engines are constantly monitoring vibration and assessing condition to warn of a pending failure so the engine can be shut down before catastrophic damage occurs; and, this information is sent real-time back to the repair facility and manufacturer, not just the pilots! How amazing to have such monitoring on us to tell of an impending heart attack?
Many modern cars tell you what is wrong, even give you a diagnostic code (anyone thinking ICD-10?), and then set up the appointment in real-time with the repair facility to fix the problem that the mechanic can get directly by downloading when the car shows up. Why can’t we do that with patients? Well, we can do that! We just need to think better and smarter. I gave a speech at a university in 2008 in which I noted that my Lexus had better care than I did because computers were constantly monitoring performance and maintenance. Today with IoT, APIs and AI, we can do the same with patients.
However, there are significant challenges that must be addressed to implement AI in healthcare effectively. The industry is built on profiting from a lack of communication, duplication of effort, and 10-minute office visits. To realize the full potential of AI in healthcare, the industry must shift to a more patient-centric model that empowers consumers to control their data and use it to improve their care. This shift will require significant investment and a willingness on the part of the healthcare establishment to embrace these changes.