Skip to main content

Kyna Fong in KevinMD: What if AI focused on respecting the craft of medicine?

Published Date
Author Kyna Fong

The following article is a version of a piece that was recently published on KevinMD. In it, Elation's CEO and co-founder, Kyna Fong, comments on the importance of developing AI tools that support's the clinician's ability to practice quality medicine with their patients.

Health care technology has a long history of getting it wrong. The electronic health record (EHR) boom of the late 2000s was a remarkable transformation — taking our sector from paper charts to computers in a matter of just a few years. Unfortunately, most of the technology that was so feverishly adopted was developed without clinical input and with a focus on medical reimbursement and policy-compliance checklists over medical care. Two decades later, many hospitals and practices today remain saddled with extensive “technical debt” — the industry term for old, difficult-to-maintain and update systems — as well as failed promises to save time, improve care, and enable coordination.

As a result, the health technology sector now finds itself in a full-blown usability crisis. Legacy EHR systems have left physicians feeling frustrated and burned out, with the CDC reporting that 9 out of 10 doctors have to spend time outside normal office hours on documentation every day. “Burdensome EHR systems are a leading contributing factor in the physician burnout crisis and demand urgent action,” said Christine Sinsky, MD, the AMA’s vice president of professional satisfaction.

With the advent of artificial intelligence (AI), health technology is entering a new era. We have a fresh chance to get it right…or to exacerbate the familiar issues. It is easy to imagine AI being applied in the same misguided, billing-and-compliance-over-clinical-quality ways that have previously frustrated physicians: More alert notification “noise” from low-quality data? More bias built into clinical decision support? More arbitrary upcoding suggestions? We’re seeing nurses at Kaiser Permanente already protesting these kinds of AI applications.

Usability must be our most urgent priority in applying AI to EHR technology. It is time to repay the “technical debt,” end this crisis, and focus on clinicians’ needs—giving them time back to focus on the craft of quality medicine and supporting their work in innovative new ways.

A recent physician survey showed that it was this kind of usability that stood out as exciting to clinicians among the many possible applications of AI. There is an eagerness for relatively simple tools such as AI-powered scribes that can ease the documentation burden for clinicians — more than a third of Elation’s clinicians have already tried one, with 90 percent reporting expecting it to reduce administrative time. The dreams of physicians don’t stop there. Physicians also want:

  • Systems where AI is trained based on the work of the individual clinician, creating templates and drafts that are easily edited and personal to the way each individual practices;
  • To use AI as a powerful translation tool, both in patient language and preference, helping physicians to support care and clarity for an increasingly diverse clinical and patient population; and,
  • For AI to help them make more room for the doctor-patient relationship by sorting through the large volumes of data, orders, referrals, and notifications they receive every day, organizing information into what should be read or signed first based on their own workload and patient needs.

Health technology has an opportunity to seize the power of AI for usability and transform our sector. Will we take it? It’s an important moment for both health technology leaders and users, a chance to speak up and get it right. Tactically, it is critical that physician usability and satisfaction are valued as a key metric in every step of the AI development and implementation process. Reduction in documentation time and extra time outside of office hours is one important data point, but we should be thinking beyond that about physicians reporting that AI is supporting their ability to practice quality medicine with patients.

To truly undo the harm of health care’s usability crisis, we need to use AI to transform clinicians’ relationships with technology—developing more trust, more helpfulness, and more support for the professional expertise that patients want at the point of care. I hope we are welcoming the EHR usability era.