Patients’ stories — what doctors call patient histories — are the bedrock of medicine. “Listen to your patient; they are telling you the diagnosis,” an aphorism attributed to Dr. William Osler, the founder of modern medicine, still holds true today. The disappearance of patients’ stories from electronic health records could be one reason that artificial intelligence and machine learning have so far failed to deliver their promised revolution of health care.
The medical industry’s fascination with artificial intelligence is understandable. Advancements in medicine have dramatically improved patient outcomes, and there is every reason to believe that machine learning, deep learning, artificial intelligence, and the like will do the same. But before we jump on the AI bandwagon, I offer this caution: consider the source of the data it is dependent on.
Health care AI companies currently harness data from electronic health records (EHRs) to build their products. EHRs are incomplete at best, dangerous at worst. They are so saturated with answers to questions required by insurance companies’ reimbursement rules and core measures from the Centers for Medicare and Medicaid Services that they end up having little to do with actual patient care.
Doctors enter a lot of information into EHRs using constrictive features like click-to-text templates, drop-down menus with pre-filled information, and macros. These records tell only part of the story of a doctor-patient interaction. More importantly, they almost never contain patients’ complaints, worries, and experiences in their own voices.
I believe these omissions are one reason for the growing disillusionment with AI, which a headline in Time magazine neatly summed up: “Artificial Intelligence Could Improve Health Care for All — Unless it Doesn’t.”
The journal Nature ran a far more critical piece, “The ‘Inconvenient Truth’ About AI in Healthcare,” which cautioned, “Simply adding AI applications to a fragmented system will not create sustainable change.”
That’s not to say AI can’t improve health care. It absolutely can — as long as the data AI is built upon includes patients’ and doctors’ voices. I believe in this so strongly that I am helping build a company that uses AI in chat-based doctor-patient communication. The goal is not to create the conversations, but to learn from them. Systems in which patients text their physicians and their physicians text back capture both voices in ways that electronic health records cannot.
Because chat collects the full conversation, the patient voice is the medical record. No check boxes, no drop-down menu autofill. Artificial intelligence built on chat records can fulfill both Osler’s mandate to “listen” and Silicon Valley’s promise to “learn.”
If we are going to build machines that use data to help diagnose and treat humans, then humans, not technology, should be at the center of the process. In much the same way that a glucometer uses a drop of blood to detect sugar levels or an electrocardiogram visually depicts a heart rhythm, AI can analyze humans’ words and help understand their conditions.
Once the medical industry finally ingests all the things patients say, the experiences they have, and the feelings they express, along with informed responses and further questions from doctors, we will be able to create the cluster maps and natural language processing engines that can actually make a difference to patient care.
Until the data become more meaningful, I will continue to raise an eyebrow any time an AI startup claims it can change the face of modern medicine by querying billions of medical records. Maybe someday that claim will be true. But it won’t happen until patients’ voices and experiences become fundamental parts of the artificial intelligence engine.
Blake McKinney, M.D., is an emergency medicine physician at Sutter Roseville Medical Center and co-founder and chief medical officer of CirrusMD.