John Booher

home

authors

john-booher

The explosive growth in the availability of big data for healthcare applications and advances in machine learning offer significant opportunities to alleviate some of the most challenging problems ailing the healthcare system. The question, though, is how far will these technologies-loosely grouped together under the rubric of artificial intelligence or AI-reshape the delivery and quality of healthcare in the future? 

Like other revolutionary advances in medicine (such as the development of modern sanitation practices in the 19

 century), AI is a tool to be integrated into healthcare practices, rather than a panacea unto itself.  Viewing AI in this manner, a few key opportunities, and limitations, are apparent.      

Opportunity #1: Early detection and triage

The ability of AI to cut through the noise of information flooding medical practitioners and early-detect potentially acute medical issues offers some of the most immediate and compelling benefits from the technology.  In February, Viz.ai, an artificial intelligence company, announced that it had received FDA clearance through de novo review of its Computer-Aided Triage and Notification Platform to identify Large Vessel Occlusion (LVO) strokes in CT angiography imaging.  In so doing, the FDA created a new regulatory classification for similar subsequent computer-aided triage software devices. 

In order for these technologies to reach their full potential, more resources will need to be allocated to the FDA for hiring knowledgeable examiners capable of timely reviewing new applications for these types of technologies, and new expedited pathways for approval will be required so that fast changing technologies do not leapfrog what is in the regulatory approval pipeline. 

Opportunity #2: Diagnostics and personalized medicine

The ability of AI to explore and exploit big data sets offers exciting opportunities to develop novel diagnostics through the identification of markers of disease (and disease risk) among disparate patient populations.  These same statistical processes can be applied to develop a personalized risk-benefit analysis in the application of therapeutic treatments. 

Whether AI will be truly transformative for diagnostics and personalized medicine depends on whether traditionally under-represented populations are sufficiently represented so as to avoid the age-old statistical problem of sample bias. Whether through the exploitation of the ubiquity of mobile devices or through other outreach approaches, fulfilling the potential of AI in diagnostics and personalized medicine will require a commitment to reaching these under-represented populations.        

Perhaps the most controversial application of AI in healthcare is in the recommendation or determination of a particular course of treatment.  In surveys, most individuals express significant concern with the idea of a version of HAL 9000 making life or death decisions.  But this is more the province of science fiction than science for the foreseeable future.  When viewed as an integrated aspect of medical protocol supporting clinical decision-making by professional human caregivers, it is instead a powerful tool for the clinician, rather than a substitute for the clinician. 

As with any such tool, the challenge for the human user is to not become overly reliant upon its use, and to maintain his or her own critical judgment and ability to independently assess a particular course of care.  In order to mitigate this risk, appropriate training will be necessary for medical personnel interacting with these tools.       

All of these applications offer the prospect of greatly reducing unnecessary and ineffective treatments and testing, improving access to care and reducing healthcare costs, but whether AI becomes as ingrained in the healthcare system as a surgeon scrubbing down before operating will depend on the extent to which the key participants in the healthcare system-patients, providers, payors and regulators-are willing to embrace these technologies.     

John Booher is a partner in the Silicon Valley office of the law firm Sheppard Mullin Hampton & Richter LLP, where his practice focuses on advising healthcare and technology companies, investors, and entrepreneurs on a range of matters.

Articles

Rapid advances in the availability of big data and in machine learning offer significant opportunities to improve the healthcare system.

Author | John Booher

Articles

Top 3 Opportunities for Artificial Intelligence in Healthcare