Seeing Around the Corners: Artificial Intelligence in Clinical Medicine

The powerful analytical systems in place today are very capable in showing providers where they have been.  The fact is that this data use presents the proverbial problem of driving the car while looking through the rear-view mirror.  While more real-time data would be helpful, it is still a retrospective view of events.  Even fresh clinical data gathered from the medical record is still a  look at “what was” or to some extent a “what is” point of view.  Today with the  financial  and outcome risk shifting toward the providers combined with a strong professional desire to do the best for the patient, physicians want to be able to look out into the future and prevent as many serious problems as possible.  In other words, physicians need the capacity to “look around the corner” to see what is coming. This concept is especially important with patients who have already developed chronic problems that will almost inevitably lead to more serious and complicated health issues in the future.  AI-powered analytics, appropriately applied, can also help solve acute real-time care problems.

Most analytical systems generate static reports that grind out consistent data into dashboards.  Some dashboard-based systems allow more experienced users to drill deeper into the information, partially satisfying their clinical curiosity.  On the other hand, there are a few systems, such as Salient Interactive Miner (SIM), that already assist the sophisticated subject matter experts (SME) to build and reuse the parameters that will identify the beneficiaries that could develop more serious issues in the future.  This then creates the opportunity for the provider to intervene with closer clinical tracking or medical and surgical therapies to improve the overall health outcome.  Yet even this advanced, sophisticated system will benefit from AI integration.  Nonetheless, if the SME does not ask the right questions, the answer may still fall short of what is really needed.

Never forget that any Artificial Intelligence solution is only as smart as the real intelligence that put it together in the first place.  In a clinical data analytical situation, AI can expand the sophisticated SME’s training and experience to others in the data analytical chain.  AI should be able to automate an expert’s clinical thinking for the wider application.  A note of caution is warranted:  AI is not a substitute for clear, rational, well-educated thinking.  Among other things, AI helps to manage the variables and opens the data to be explored in a reasoned, consistent manner.  AI can help anticipate clinical problems with complicated patients.

Every level of analytics–descriptive, diagnostic, predictive, and prescriptive–can add value to the health care experience.  The data points toward a significant change in AI adoption in the health care work place.  In fact, Gartner predicts that one in five workers focused on non-routine tasks will depend on AI to do the job.[i]  In actual patient care AI can push a marginal or distracted practitioner toward the correct diagnosis faster.  For example the diagnosis of sepsis should already be in the mind of a clinically-alert provider after a thorough history and physical examination and before all the laboratory information is available.  Nonetheless, AI can already alert the clinician to the sepsis diagnosis, define the laboratory limits of normal and abnormal (complete blood count, lactate, cultures, blood clotting studies, and others) and suggest a treatment process in advance of a confirmed diagnosis.  The value of AI is often the speed of diagnosis and intervention where it matters, such as in sepsis.  Specifically in this disorder, delay can be fatal even if the correct diagnosis is finally confirmed.  In both acute and chronic conditions AI-powered analytics is already showing active utility.

In health care it is not a matter of using AI to “put yourself ahead of any competitor.”  The objective of “unlocking the potential of AI-powered analytics” is to accomplish the purpose of the triple aim: better care, lower cost, and higher patient satisfaction.[ii]


[ii] ibid

Craigan Gray

About the Author

Craigan Gray, MD, MBA, JD

Dr. Craigan Gray, Salient Healthcare’s Chief Medical Officer, brings rich experience from private practice, hospital leadership, and governmental health-benefit programs. Prior to joining Salient, Dr. Gray was director of North Carolina’s $12 billion Medicaid program. His time as VPMA at Bon Secours Our Lady of Bellefonte Hospital in Kentucky was distinguished by moving the facility into the top-quality performance tier for Health Grades and CMS health quality indicators. Dr. Gray is a Stanford University trained Obstetrician/Gynecologist. In addition to an MD degree, Dr. Gray holds an MBA degree and a JD degree. He is a Certified Physician Executive and is published in various medical journals.

Leave a Reply

Your email address will not be published. Required fields are marked *