A University of Minnesota doctor is recommending reforms to medical education to address the problem of misdiagnosis — arguing that the existing apprenticeship approach isn’t preparing U.S. physicians for the tough calls they have to make.

In an editorial published last week in the medical journal Diagnosis, the U’s Dr. Andrew Olson and colleagues called for cognitive training that helps current and future doctors avoid the psychological pitfalls that can lead them to incorrect conclusions about illnesses. That training is lacking in the current system, which loads up medical students with biological and clinical facts, and has them learn decisionmaking by watching and emulating their mentors, they wrote.

“Trainees learn the basic steps of the diagnostic process and can arrive at a diagnosis but may have little or no appreciation for the underlying cognitive psychology, or on the many ways that the process can be degraded in the actual contexts of providing care,” they wrote.

The problem of misdiagnosis gained attention in 2015 after a report by the U.S. Institute of Medicine called it a “blind spot” in health care that causes an unknown number of problems and injuries. By some estimates, as many as 5 to 15 percent of diagnoses are off.

The University of Minnesota has been trying to address the problem. Olson has developed training in diagnostic reasoning for U medical students.

The school also holds an annual lecture on the topic in the name of a 15-year-old who died in 2005 following diagnostic mistakes and assumptions in her care.

Doctors still must master the biological and clinical information about various diseases and conditions to diagnose them, but Olson and colleagues argued that they also must learn about the pitfalls of the decisionmaking process, and about what to do if they suffer fear or doubt in assessing complicated patients.

Training needs to “introduce the concept of metacognition, an implicit skill required to recognize one’s own limitations, the need for including team members and consultants in the diagnostic process, and the importance of lifelong learning to improve diagnostic performance,” they wrote.

A report in the journal Academic Medicine listed the mental tricks that can lead to misdiagnoses. They include “anchoring,” which is the tendency to lock in on one theory based on initial symptoms, and “confirmation bias,” which is the tendency to look for diagnostic information to support a hunch.

Doctors also can suffer from the “Gambler’s fallacy” when doubting a diagnosis because a recent string of other patients had it, and the “visceral bias” of allowing feelings toward patients to cloud their judgment.