North America

Shelly Nash

AI, machine learning, deep learning. You hear these terms tossed around in many disciplines these days. Transportation, finance, manufacturing, education, and of course healthcare. If we think about huge amount of data coming at end-users, analysts, executives, even consumers it’s no wonder the idea of machines, or rather computers, helping our overloaded brains sift through all of this and make sense of it is appealing. Artificial intelligence, defining it as the ability of computer systems to translate data and present it or trend it to alert a user of an important perhaps crucial fact is not only appealing but seems revolutionary. Focusing on healthcare, it has been estimated that the doubling time of medical knowledge (articles, studies, research outcomes) was 7 years in 1980, 3.5 years in 2010, and is now 0.2 years, or 73 days. This means that trying to keep up with new guidelines, recommendations, and clinical best practices is impossible for the human brain; but not for an updated well-run computer system. This is where AI can and has been playing a role.

Areas like sepsis prediction and alerting in electronic health systems, review of patterns in EKGs and sending notifications to clinicians, as well as pattern recognition and analysis to allow better diagnoses on radiology and pathology studies are examples where AI has been successful. Electronic health record vendors have also been using AI to aid clinicians in recognizing data from not just their current encounter or visit but from other visits or medical treatments. Meaning if you see a doctor in one state and don’t recall or mention a health problem you had (and often the healthcare providers use different record systems) AI can be used to comb not just your current chart or record but all of the data available about you in the electronic universe. Some computer systems can even review dictated or non-discrete scanned documents. This would take a human hours or days to do. Predictive models or applying these to individual patients is also where AI shines. Companies are working to compare your current information—things like your current health conditions, weight and height or blood pressure, current medications—to not only help clinicians know which treatment is right for you now, but to advise them, based on data from other patients with your similar characteristics, as to what might be the best treatment going forward. The whole field of precision medicine is based on data analysis where artificial intelligence plays a key role.

“Areas like sepsis prediction and alerting in electronic health systems, review of patterns in EKGs and sending notifications to clinicians, as well as pattern recognition and analysis to allow better diagnoses on radiology and pathology studies are examples where AI has been successful.”

Is AI the Holy Grail or just a flash in the pan? Most experts would agree it is more than a phase or craze and the everyday implications and possibilities continue to be developed. AI is not ready, as many fear, to replace humans. It helps to glean information quickly and accurately to people. It allows for the quick integration of new algorithms or knowledge which can be applied or built into everyday work. Just like any technology, we need to think of ourselves as the proprietors or gatekeepers of AI. It can suggest, alert, advise, but the end-user or human is still in charge. This is an exciting time to be working in healthcare technology. We should be open, not fearful yet diligent, in our reviews of technology as the “human intermediaries” in using AI as it develops.