Build a no-fault, fail-safe system
The purpose of analyzing diagnostic errors is not to assign blame or point a finger at physicians, but rather to find and fix flaws in the medical system.20 That approach has been used by patient safety advocates to address other types of errors following the publication of the Institute of Medicine’s landmark report on medical error 11 years ago.21
Since then, many physician leaders have looked to the airline industry—a field in which the consequences for not strictly adhering to a fail-safe system are likely to be fatal. That reality has led to the development of vital checklists, forcing functions (in which the user is prevented from moving to the next step until the current step is completed), and computerized reminders in an attempt to eliminate, to the extent possible, the chance of human error.
The same principle can be applied to misdiagnosis. Recommended steps—ordering diagnostic tests or referring to a specialist, for example—should be put into motion whenever a set of predetermined parameters are met, rather than relying on physician memory or choice.20
Similarly, checklists should specify questions to ask or criteria to be met under specified circumstances to prevent physicians from prematurely settling on a (possibly incorrect) diagnosis. To avoid a rush to judgment, some patient safety advocates1 stress the importance of assessing the urgency of a patient’s condition, rather than trying to arrive at a definitive diagnosis the first time he or she presents with a perplexing set of signs and symptoms. Other recommendations follow:
Mandate a second look. Develop and adhere to a set of criteria to determine when a referral to a specialist or a physician consultation is needed, rather than deciding on a case-by-case basis.
Plug the holes in your follow-up system. Develop a fail-safe system for reviewing diagnostic tests or laboratory findings and reporting them to patients without delay. This can be done with an electronic health record (EHR) system or by developing and adhering to parameters requiring, for instance, that no test result get filed until there are 2 signatures on it—that of the physician who ordered the test, indicating that he or she has seen it, and that of a staff member, indicating that the patient has been notified of the results. As an additional back-up, tell patients undergoing tests when to expect to get results, and stress the importance of calling the office if they do not receive such notification within a specified time frame.
Partner with patients. Engage patients in the pursuit of a definitive diagnosis. Discuss your preliminary findings, describe your treatment decision and what you expect to occur, and urge patients to contact you with evidence that confirms or refutes that expectation. Elicit additional feedback at each visit until either the symptoms have fully resolved or you have gathered enough information to arrive at a definitive diagnosis.
Develop “don’t-miss” checklists. One list should cover diagnostic red flags to be considered anytime you see a symptomatic patient to ensure that you don’t overlook important signs and symptoms, and include findings that warrant hospital admission, specified diagnostic tests, and immediate referral. (A patient who comes in with a “common pink eye,” but has consensual photophobia, is at risk for iritis and needs an urgent ophthalmology evaluation, for instance.) Another list you should develop is a “must-do” list for well visits, featuring clinical scenarios to address and screening tests to remember, such as an eye exam for patients with diabetes.
Question your initial diagnosis. Beware of “premature closure”—the tendency to stop looking for other signs and symptoms once you find a presumptive diagnosis—and “diagnostic inertia”—evaluating new signs and symptoms almost exclusively on the basis of past medical history. If aspects of a patient presentation do not fit your presumed diagnosis, use a decision support system, if available, to review other possibilities.
Head off hand-off errors. Develop a problem list for each patient to reduce the likelihood that crucial information will be overlooked when more than 1 clinician is involved in his or her care. Include chronic and acute conditions, unexplained signs and symptoms, medications, and allergies. Create a fail-safe system for other potential hand-off problems, as well—requiring confirmation that the findings in a preliminary radiology report are the same as those in the final report before you take action based on the preliminary report, for example, and ensuring that you receive prompt notification whenever that is not the case.