The Clinical Divide: Implementing AI to Improve Patient Outcomes

Kevin Campbell, M.D.
JANUARY 21, 2019
As Healthcare Analytics News™ reported earlier this month, a recent study published in the Journal of the National Cancer Institute revealed that artificial intelligence (AI) outperformed humans in identifying cervical precancer lesions in a primate model. Now, another study published last week demonstrated that AI was able to effectively screen a human population for pre-cancerous lesions as well.  These results are definitely promising and have inspired many to push AI forward, much faster. How can we as physicians and healthcare executives leverage this technology to improve patient outcomes? And with the U.S. Food and Drug Administration (FDA) now defining ways to get AI in medicine approved as well, how are our hospitals going to keep up with these new technologies as they are rolled out?

Welcome to The Clinical Divide. I’m Dr. Kevin Campbell, a Duke-trained cardiologist and CEO of the health data startup PaceMate. Every week, this Healthcare Analytics News video series examines technology and medicine’s top news. I really try to bring the views that help physicians and healthcare executives bridge the clinical divide.



Let’s begin today with the latest National Cancer Institute study. Using deep learning algorithms, a form of AI, a research team trained AI to study digitized photos and distinguish cervical conditions that would and would not require treatment. The cervical photos were obtained from a review of over 9,000 women in Costa Rica where human papilloma virus (HPV) and early cervical cancers are relatively common and often go untreated.  Historically, it has been exceedingly difficult to accurately identify those precancers in populations like this where resources are lacking.  In the study, the AI algorithm successfully identified 127 of 228 precancers diagnosed in the whole population. This is more than 55 percent and is far greater rate than what’s been identified using standard methods due to lack of resources and lack of access to care in Costa Rica.

>> WATCH: The Clinical Divide: Leveraging Telemedicine to Treat Mental Health

As healthcare systems continue to be overwhelmed with patients and as healthcare costs continue to rise, we as doctors and CEOs must continue to push for new ways to improve the health status and outcomes of the patient populations that we are charged to serve. 

AI, I think, is critical to the development of these efficient cost-effective technologies and CEOs and doctors must work together to implement these technologies into every day patient care.  At PaceMate, for example, we are using algorithms to streamline the remote monitoring of cardiac devices, and this allows more patients to be served with even better diagnostic accuracy and efficiency than was ever possible via traditional methods.

The momentum for AI in medicine is picking up speed.

Even the government is starting to “get it.” The FDA has announced that it will be focused on AI and deep learning-type algorithms and how they may successfully be applied in medicine. Commissioner Scott Gottlieb, M.D., has committed to creating simple and safe pathways for expedited approval for these technologies. As patient advocates and the deliverers of care, both CEOs and M.D.s must prepare and seek out ways to implement these new technologies into the daily care of their patients. It’s essential to the success of our mission.

Thank you for joining me on this episode of The Clinical Divide. Until next week, I’m Dr. Kevin Campbell, for Healthcare Analytics News.

Get the best insights in healthcare analytics directly to your inbox.

Related
FDA Clears RAPID AI Software for Selecting Stroke Victims for Clot Removal
Objective Eye Tracking Tool for Concussion Granted FDA Approval
AI Diabetes Tech Continues FDA's Software Push

SHARE THIS SHARE THIS
15
Become a contributor