Symposium on promises, pitfalls of technology in medicine

Technology is transforming medicine. Today, computers can recognize skin cancer. Data maps plot locations where people are likely to have asthma attacks. Sensors track the handwashing hygiene of caregivers in a hospital.

In the midst of these ongoing changes, how do doctors retain a human connection with patients and ensure that new developments enhance, rather than impede, medical care?

Speakers discussed that challenge April 17 at the daylong Human Intelligence and Artificial Intelligence in Medicine Symposium, hosted by the Presence Center and co-sponsored by the Stanford Department of Medicine.

“We don’t want to go back to those days when there was very little that we could offer scientifically,” said Lloyd Minor, MD, dean of the School of Medicine. “But unfortunately, the growth of the technology of medicine — and the way that technology is being implemented today — has made it very difficult for physicians to establish the relationships, the bonds that have classically characterized the profession and that for centuries before have provided what was really at the heart and core, the meaningful aspect, of being a physician.”

Obligations, such as completing electronic medical records, have pulled physicians away from interpersonal interactions with patients and have left many feeling burned out, Minor said. However, other technological advances hold great promise for predicting or even preventing health problems — the essence of precision health.

In his talk, Abraham Verghese, MD, professor of medicine and the founding faculty director of Presence, said that ideally the data-driven, quantitative side of medicine would be balanced with the human aspect of a patient’s story and what a diagnosis means to that person’s life.

“It’s no use knowing the genome and proteome of a patient,” he said, “if you’ve missed the fact that they have the outline of a cigarette packet in their shirt pocket.”

The symposium’s keynote speaker Eric Topol, MD, founder of the Scripps Translational Science Institute, agreed that doctors are challenged today to practice patient-centered medicine in an era of high-technology and big business.

He noted that in contrast with other wealthy nations, Americans spend more on health while their life expectancy is declining and he listed other problems in U.S. medicine, such as diagnostic errors, false positives on cancer screenings and prescription drugs that are not working as intended.

“The terrible situation that we’re in today, in U.S. health care especially … it’s because we haven’t been using the data that we could,” he said, adding that artificial intelligence is on the cusp of changing that.

One issue is that although researchers have developed algorithms that can classify skin cancer lesions and identify eye disease in people with diabetes, overall few published studies to date provide clinical validation of a computer’s medical assessment, Topol said.

Additionally, doctors and researchers are just beginning to recognize the potential of patient-generated data. Wearable devices are now helping people monitor everything from potassium levels to asthma attacks to migraine triggers, Topol said. The ultimate result — a “virtual medical coach” — will empower patients with knowledge about their own health and how to improve it.

“AI will help patients. It's not just doctors,” Topol said.

Overall, he said, the hope is that artificial intelligence will give back to physicians a resource they desperately need: time.