Self-driving cars, voice powered digital assistants, and even the latest driver designed by Calloway Golf—what do these and a myriad of other items have in common? The claim that they are built using some type of artificial intelligence (AI). With the ubiquity of claims related to AI and machine learning, it can be hard to separate myths from reality; and ultimately, develop an understanding as to how AI-based technologies can deliver benefits at both individual and societal levels.
The hype-cycle around AI is clearly overheated, yet it’s fair to say that some AI-based technologies are delivering results. Although still at a very early stage in day-to-day application, we are seeing examples of how machine learning and predictive analytics can enhance solutions and services including navigation and travel; digital and streaming entertainment; security and risk assessment; banking and customer support — to name just a few.
Healthcare has seen significant impact. A recent Morgan Stanley research report on medical devices and AI identified several areas where significant progress is being made, including:
- Chronic care – user monitoring and alerts
- Digital dentistry – optimize treatment plan from intra oral digital scans
- Diagnostic imaging – analyze and interpret images for diagnosis
- Dialysis – optimize dosing, treatment and follow-up
- Orthopedics – assess pre-op data to create best practice treatment plan
- Radiation therapy – create and replicate best practice treatment plans
At Medial EarlySign, our machine learning-based disease risk predictors (“AlgoMarkers”) help healthcare systems identify patients at high-risk for either already having or being on rapid trajectories for developing life-altering medical conditions, sometimes even before they appear symptomatic. Our priority is to help healthcare organizations find early intervention opportunities for their efforts with early detection and prevention of high-burden diseases.
With EarlySign, medical and care management prfessionals can identify and notify those individuals at high risk and get them in for further evaluation. We also devote special attention to make sure that our predictive risk models are integrated within care teams’ clinical workflows.
Clients benefit from our helping them using data they already have—and squeeze every bit of value out of that data to deliver insights that inform better decision making. This concept can best be understood by paraphrasing our Chief Medical Officer, Jeremy Orr, MD, MPH. In a recent informational webinar with SLUcare, he explained how our AlgoMarker for lower GI disorders can help providers with their early detection and efforts to prevent high-burden diseases:
When clinicians look at a lab tests, they usually look for values outside the normal ranges while the machine looks at things a bit differently. Imagine the best hematologist you’ve ever met, except this one is a bit unusual:
- He pays attention to every subtle variation of CBC indices, even within the normal ranges
- He’s able to discern patterns involving multiple indices (or even all indices)
- He has seen hundreds of thousands, or even millions of CBC results
- He remembers every one of them
- And he can tie those to outcomes like patients that have had bleeding, CRC, adenomas, or other lower GI disorders
Ultimately, one can now imagine that this “machine learning-based” hematologist can spot potential abnormalities better than a clinician who looks at lab tests one at a time.
Looking forward, the nature of risk detection for a variety of diseases will change dramatically, as more sophisticated markers continue to emerge and healthcare systems “cross the chasm” from being early adopters and innovators to mainstream use. We see the widespread adoption of predictive analytics addressing a broader range of high-burden diseases. This combination of novel biomarkers with advanced machine learning and predictive markers will make great impact in terms of quality of care, affordability, and patient experience.