News: Next-generation AI should shift from diagnostic predicting to “wayfinding”

Artificial intelligence (AI) in the healthcare setting provides adequate solutions for well-characterized, standalone diagnostic questions but aren’t yet capable of “interpreting context and providing cues that guide the diagnostician,” according to an article in the Journal of the American Medical Association (JAMA).  

Because AI tools aim only to predict a label (diagnosis) at the end of the process, they “overlook the upstream work of navigating the decision nodes along the diagnostic pathway, and therefore are unlikely to garner the trust of clinicians.” Clinicians encounter critical challenges interpreting complex patient data and determining what steps to take to care for that patient. “A new generation of AI is needed that considers the dynamism of the diagnostic process,” and helps to answer “what should be done next,”  according to the article. Future diagnostic AI solutions should focus less on the end diagnosis and more on  the “wayfinding” process of diagnostics. According to article authors, in a clinician’s perspective the “diagnostic journey begins with assessment of the patient’s signs and symptoms, which in turn triggers information gathering. […] As the clinician integrates and interprets accumulating data, next steps are planned, often with a general discretion or potential destinations in mind.”

As next steps are determined, new information is generated, prompting another round of collecting data, creating a cycle that repeats until the diagnosis is found. AI tools that support the diagnostic refinement process are more likely to help clinicians understand where they are on the diagnostic pathway and help them select the paths more likely to reduce uncertainty. This, in turn, is more likely to garner clinician buy in of AI technologies.

The article notes that development of wayfinding AI tools still needs to be generated. Observing digital records of clinicians’ current wayfinding activities can be a first step to revealing where AI can optimize information navigation and decision-making.

“A shift to wayfinding AI could help achieve a synergy of human intelligence and AI to achieve diagnostic excellence,” authors say.

Editor’s note: The JAMA published study can be found here.