Artificial intelligence could revive the art of medicine

Doctors practice medicine to deliver care, not do data entry.

Yet in the era of electronic medical records (EMRs), for every hour spent with a patient, physicians spend nearly two hours on paperwork. What if technology could take care of the paperwork for us?

Record-keeping systems in health care were built for back-office functions, not bedside medicine. Most EMR vendors started out building products to collect payments and schedule appointments. When these vendors got around to building interfaces for doctors, they built EMRs that also primarily serve administrative purposes: they create a paper trail for billing audits, regulatory reporting, and medical malpractice suits.

Imagine if technology could collect and organize data to satisfy those administrative requirements. Then doctors, instead of being data entry clerks, could focus on practicing medicine.

Artificial intelligence (AI) is a tool that can help us realize that vision. Despite the frenzy about computers that can read a CT scan or choose a patient’s cancer therapy, the role of AI in medicine is not to replace doctors. AI will be applied most powerfully when used in products that integrate seamlessly with doctors’ workflow, and relieve us of administrative tasks that take us away from direct patient care.

AI-powered products will assist doctors. They will not supersede doctors by out-performing us at diagnosis. Nor should they tell doctors how to do their jobs. Technology, when implemented right, should help you and never get in your way.

How can technology facilitate doctors’ workflow? One application that I’m excited about is our nascent ability to listen to a conversation between a doctor and patient, and create a clinical note out of that interaction. This is part of what we’re building at the company I’m a part of, but AI is only one piece of the puzzle.

AI can be a useful tool because it can learn from patterns, so the longer we work with a doctor, the better we get at generating accurate notes. But we also care about simple, user-centric design, scalable cloud architecture, and introducing a voice interface that will be familiar to fans of Amazon’s Alexa. We see these elements everywhere in consumer technology, but they’ve been slow to appear in medical applications.

Imagine if you could ask your intelligent assistant to pull up your patient’s blood pressure trend since starting lisinopril. Or, look up treatment recommendations for an unusual condition -- I had a patient recently who swallowed both of her hearing aids, batteries and all (she had them in a medicine cup and mistook them for pills). A Google search returned a clinical algorithm from the CDC that helped me think through this unusual twist on battery ingestion. If you’re searching Google in unusual cases, you’re already using AI in medicine. Wouldn’t it be nice if data in your EMR were as easily accessible as all the world’s data indexed by Google?

One of the sad consequences of our clunky EMR systems is how much data we lose because the interface is so limited. A natural way of documenting clinical encounters would not only make doctors’ lives easier, it would improve the quality of data by capturing richer information about clinical decision-making.

As products evolve, someday, an AI tool that has learned my practice patterns will be able to anticipate my next steps, like a good PA or resident would. As a trauma surgeon, the next time I see a restrained driver who was T-boned on the driver’s side, if he is tender over the left lower chest wall, my AI assistant might remember that I like to CT the abdomen to look for splenic injury (the spleen lives just under those left-sided ribs). When I’m rounding on a patient with ileus, my AI assistant can figure out how long my patient has been unable to eat, and remember that I typically start parenteral nutrition after seven days.

Most of the medicine I practice wasn’t learned from a textbook. It doesn’t come from top-down quality protocols. I learned it from my mentors, and from experience. If AI has any hope of cataloguing and disseminating the practices that make up good medicine, it will be through participating in patient care. If we succeed on our journey, products that utilize AI may one day help restore the humanity of our profession.

By doing what computing does well -- collecting and organizing information -- AI can allow doctors to focus on caring for the sick and injured. It can provide us the support we badly need to practice the art of medicine.

Erin Palm, MD, MBA, is the director of product at Suki and a practicing general surgeon.

Erin Palm, M.D., M.B.A.
Director of Product, Suki
@erinpalmmd

Dr. Erin Palm, a Stanford-trained trauma surgeon leads product at Suki, an artificial intelligence-powered, voice-enabled digital assistant for doctors. Suki works by understanding voice commands from a physician and uses them to create a clinically accurate note that is then pushed to an electronic health record (EHR) system.

Before Suki, Dr. Palm wore many different hats in the health care community, including completing a fellowship in surgical critical care at LA County + USC Medical Center. She joined Suki because she has never experienced technologists so clearly interested in the doctor’s point of view.

Suki is like an Alexa for doctors’ offices that lifts the burden of medical documentation and difficult-to-use EHRs, enabling physicians to focus on what they love: taking care of patients. Suki is personalized to each doctor, gets smart as they use it, and is rapidly and inexpensively scalable. Moreover, it is easy to implement and similar to having a chief resident in the exam room who knows how a doctor practices and, as a result, makes the work day easier.

A member of Suki’s executive team, Dr. Palm is joined by co-founders Punit Soni and Karthik Rajan, as well as Dr. Nathan Gunn, a UCSF-trained internist, and technologists from Apple, Google, Salesforce, and 23andMe. Suki is backed by investors such as Venrock, First Round, Social Capital, and Marc Benioff.

Before becoming a surgeon, Dr. Palm was a dual degree M.D./M.B.A. student at Stanford. She also attended Stanford University for her undergraduate degree in environmental science and health policy. Along her journey through the health care sector, Dr. Palm worked as an EHR product manager in India and venture capital analyst. She remains actively involved in patient care, practicing trauma and acute care surgery at Kaiser’s level II trauma center in Vacaville, Calif.

© Copyright ASC COMMUNICATIONS 2018. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months