Diagnostic Time-Out: A Randomized Clinical Trial of a Checklist to Improve Diagnostic Accuracy

Diagnostic errors are common. They are more common than medication errors and they are the second leading cause of malpractice claims. They are more likely to harm patients and more likely to be preventable than other kinds of errors. Yet they have been largely ignored by patient safety groups, which have focused more on system problems than thinking problems. Diagnostic errors are often traced to physicians' cognitive biases and failed heuristics (mental shortcuts). We know how these faulty thinking processes lead to diagnostic errors, but we know little about how to resist them. Faulty thinking has plagued other high-risk, high-reliability professionals, such as airline pilots and nuclear plant operators. These professions have learned from their mistakes and have developed checklists to help prevent them. The medical profession has started to use checklists and time-out periods in the operating room and intensive care unit, but these strategies have not been used to reduce diagnostic errors. The most common reason that physicians fail to make the correct diagnosis is that they never consider it. This failure could potentially be prevented if the physician took a time-out to review a checklist. Our broad long-term goal is to reduce diagnostic errors by developing interventions that help counter faulty diagnostic thinking. The specific aims of this project are to (1) determine the feasibility of taking a diagnostic time-out in the acute outpatient setting (urgent care clinic and emergency department), (2) determine if new diagnostic possibilities are seriously considered as a result of the time-out and checklist, and (3) compare the initial differential diagnosis with the new differential diagnosis following the time-out, and with the discharge diagnosis documented in the medical record, and with the "final" diagnosis based on a one-month follow-up. To achieve these aims, the investigators will ask 5 urgent-care physicians to complete a time-out procedure for 10 diagnostically challenging adult patients and 5 physicians will serve as controls (no time out) for 10 diagnostically challenging patients (total of 100 patients). The investigator will ask the intervention physicians to take a 2-minute time-out to review a complaint-specific differential-diagnosis checklist, which includes the differential diagnosis for 60 common presenting complaints, such as dyspnea and chest pain. The time-out will occur at the conclusion of the history and physical exam. We will use descriptive statistics and qualitative methods to characterize physicians' reactions to the time-out and checklists. We will use this pilot project to plan a larger study that will determine the risks and benefits of diagnostic time-outs and checklists.