Keyboard Liberation

Session: SGO Presidential Invited American Board of Obstetrics and Gynecology

Speaker: Eric Topol, MD, American cardiologist, scientist, and author

AI set to help physicians push back from the computer so they can regain valuable patient-to-doctor interaction

In an oversimplified definition, artificial intelligence, or just plain AI, ingests massive amounts of data, churns everything around until it all makes sense, and then spots patterns or conclusions that even the most skilled human eye misses. As mankind continues to design algorithms that can make different and even deeper decisions, many medical thought leaders have come out in favor of this still-evolving machine science, stating that AI is set to change the very delivery of healthcare.

“Artificial intelligence, what I call deep medicine, is the use of AI to improve humanity in medicine. This may sound counterintuitive—technology to enhance medicine. I’m here today to convince you otherwise.”

That’s how Eric Topol, MD, began his SGO 2021 discussion about the important role artificial intelligence plays in the delivery of holistic, compassionate, and efficient medical care.

For the record, Dr. Topol has multiple distinctions that establish him as an expert in the use of artificial intelligence, deep data, and smart technology in the practice of individualized medicine. He has authored three best-selling books on the topic, and he also founded Scripps Research Translational Institute, fully dedicated to individualized medicine.

Despite decades of experience and insight, Dr. Topol still appears amazed with what AI is capable of accomplishing—and also what he says AI will never accomplish. For example, he stated unequivocally that AI will never replace the physician. “The Robot Will See You Now” is only a clever 2019 Time magazine cover headline, he explained. But that doesn’t preclude a few personnel changes, which Dr. Topol succinctly explained by quoting professor and neurosurgeon Antonio Di Leva: “Machines will not replace physicians. But physicians using AI will soon replace those not using it.”

Quick Work

Looking at how AI will benefit medicine long-term is a tough call—since the science is still being developed and the possibilities are infinite at this point. However, Dr. Topol confidently shared several benefits already popping up. “The most striking advantage that can lead to improvement in medicine is liberation from the keyboard,” he said. “Doctors are looking at computer screens rather than looking at the patient. And the patient is disenchanted, having waited on average three weeks to get a primary care doctor appointment. That’s unacceptable.”

It’s a problem that continues to disrupt the physician/patient relationship, Dr. Topol said. “It’s not why any of us went into medicine.” And, he added, time spent focused on technology rather than a fellow human being may very well be a contributing factor behind the overall physician burnout rate in America.

Natural language processing and machine learning can take over much of a physician’s screen time, with a big payback. “Each minute AI gives back to us has a profound influence on how much more time we have for humanity in medicine. The gift of time will help us not only restore the essence of medicine but also make the most accurate diagnoses possible,” Dr. Topol said.

And accuracy, the doctor added, is no small issue, with an estimated 12 million adults in the United States experiencing a diagnostic error every year in outpatient settings alone.1

Here are several examples where Dr. Topol sees AI giving physicians the gift of time.

  • Pathologists are being armed with a new dimension of what can be seen on a whole slide image, with deep learning-based pattern recognition able to incorporate clinical, radiologic, and genomic data to strengthen and quicken an accurate disease diagnosis and predict patient prognoses.
  • In some oncological scenarios, a sample can be sent to pathology and when paired with AI, the turnaround time for an interpretation is quicker and more consistently accurate. “This could be a way we move forward and not have to wait with patients and operating room teams in the OR for an accurate answer,” Dr. Topol said.
  • In Japan, machine vision is being used to diagnose polyps during a colonoscopy and interpret the likelihood that a polyp is cancerous. “The doctors know in real time if a biopsy is indicated or not indicated. It’s exciting, advanced, and already being studied in randomized trials,” Dr. Topol said.
  • AI delivers the ability to take an MRI or PET scan with considerably higher resolution and greater speed. “We can also convert images that are not as crisp as we’d like,” Dr. Topol explained. “That gives us better interpretability, in less time, and with less chance of a misdiagnosis.”

As machine learning takes over diagnoses, it’s still and always will be the physician who advises the patient along their healthcare journey. “There’s not going to be any algorithm for empathy. That’s clear,” Dr. Topol said. “Empathy is a human quality, and it is indeed the essence of medical practice.”