Setting and Problem
Diagnostic errors are estimated to occur in 10% to 15% of patient encounters. Cognitive errors contribute to over half of diagnostic errors and are associated with significant morbidity. Despite this, given the sharp-ended nature of discussing cognitive errors, educational initiatives tend to focus on system issues and fail to address the equally important cognitive component. However, addressing all contributing factors to diagnostic errors is crucial to optimizing patient safety, especially in cognitive fields such as internal medicine. Developing curricula to address cognitive errors through highlighting cognitive biases and teaching clinical reasoning and metacognitive strategies is crucial to a robust graduate medical education system.
Intervention
Our educational innovation utilizes a variation of the “cognitive autopsy” in an interactive case-based conference. A cognitive autopsy, typically performed individually, is a cognitive root-cause analysis where the analysis of medical outcomes occurs from a cognitive viewpoint (eg, thinking errors and biases: anchoring, premature closure, commission) as opposed to a system viewpoint (eg, staffing, policy, protocol). Residents use this reflective strategy in group settings to shed light on the myriad unconscious cognitive biases that affect their diagnostic reasoning.
Our biweekly clinical reasoning conference was implemented in January 2016 at a large internal medicine residency program.
The clinical reasoning conference team consists of a faculty director, a chief resident, and 15 senior residents.
The clinical reasoning conference starts with a brief clinical vignette presented by a resident. Cases are carefully chosen to illustrate a diagnostic error that highlights several cognitive biases.
Residents work in small groups (6 to 8) to evaluate and manage the patient in a real-time fashion. A clinical reasoning conference team member facilitates each small group by responding to the group's interventions in a simulated format.
The actual patient course with the diagnostic error is revealed. Thinking aloud, residents reflect on the case by performing a cognitive autopsy and generating a list of potential cognitive biases that might have led to the error in the actual patient course and in their own small group discussions (Did the physician(s) allow framing effect to influence the ordering of diagnostic tests? Did the physician(s) anchor to a specific element in the patient's presentation?).
We hypothesize that by demonstrating where and how cognitive biases can hide and lead to cognitive errors, residents can become equipped with metacognitive strategies that will help them mitigate bias, improve clinical reasoning, and ultimately decrease diagnostic errors.
Outcomes to Date
Our educational innovation has been very well received by the residents and is the most popular conference of the program. Formal evaluation is currently underway employing a mixed methods design (survey, reflections, and focus groups). The survey component with 102 responses reveals promising results (Table), with 90% of respondents indicated gaining new perspectives and reporting they are more aware of their cognitive biases. Ninety-five percent believed their clinical reasoning has improved, with 70% noticing their clinical decisions changing as a result of this intervention. Over 96% reported that the conference helps make them better clinicians. In addition, preliminary sampling of resident reflections points toward shifting perspectives indicative of transformational learning. Focus group data are still being collected.
Table.
Select Items From the Intervention Evaluation Questionnaire (N = 102)
We believe this conference has changed the culture of our program by normalizing attention to cognitive errors, integrating bias discussions into clinical rounds, and emphasizing metacognitive strategies in the face of uncertainty. We have observed that trainees embrace curricula that prepare them to tackle cognitive errors. Utilizing the cognitive autopsy in a safe, collaborative conference setting has proved to be an effective approach to deliver this curriculum. By providing a safe space to discuss cognitive errors, we are creating clinicians who are better equipped to tackle diagnostic uncertainty and ultimately provide safer care to patients.