Medical training oftentimes relies on pattern recognition, which is necessary to develop an excellent clinician who is both efficient and precise. There’s a flip side, though. An over-reliance on pattern recognition can miss outlying cases or entrench misguided practices. In addition, the knowledge base of medicine is vast, covering different specialties that are—in their own individual right—constantly evolving. The shifting landscape creates “potholes” that are not always readily apparent. It’s inevitable that we’ll step in these potholes, but may not even recognize them or, even if we do, we may find it difficult to acknowledge them. Overconfidence is one difficulty in seeing these potholes or acknowledging them.

Balancing confidence with self-cognizance can make us better clinicians who evolve with the medical landscape. There is a power in recognizing there are—as Donald Rumsfeld would term “unknown unknowns.” It relies on being cognizant of medicine’s and our own limitations, aware of our own egos, and then implementing concrete strategies, such as cognitive pauses, case reviews, and seeking feedback.

Overconfidence in Medicine

Medical knowledge evolves at a rapid pace. Peter Densen, MD, estimated that the medical knowledge “doubling” rate was 50 years in 1950, seven years in 1980, and just three and a half years in 2010.¹ It’s hard to imagine a physician being masterful in every diagnostic and therapeutic technique for every patient. It is even sometimes difficult for physicians to be aware of these knowledge gaps. Studies on overconfidence are present in physician imaging interpretation and diagnosis.^2,3 In one study in the intensive care unit, clinicians who were “completely certain” of a clinical diagnosis for 126 patients’ causes of death were actually incorrect 40 percent of the time, confirmed by post-mortem autopsy.⁴

Another interesting aspect of overconfidence is the Dunning-Kruger Effect, first described in 1999 in studies of participants’ self-perceptions in areas of logic, humor, and grammar. Its simplified findings were that the less proficient one was, the more likely one was to overestimate their proficiency. This has been similarly demonstrated in medicine. Residents’ confidence or self-perception of their knowledge in areas of diagnosis and communication were overinflated to their actual demonstration in these areas, compared to attending physicians. Furthermore, lower-performing physicians tended to rate themselves higher than their peers.^4–7

The landscape of medicine and its physician training have natural hurdles that make one prone to overconfidence, but sometimes for good reason. Physicians are trained in areas of pattern recognition, hearing a chief complaint, and coming to a hypothesis. Cognitive load and time spent are decreased as this pattern recognition of “fast thinking” leads a physician down a familiar pathway. To prevent overtesting and overconsultation, physicians must make a quick differential to focus diagnostics and treatment. Most of the time, the hypothesis is correct. However, relying on pattern recognition with blind confidence can lead to “early diagnostic closure,” the premature narrowing of diagnostic possibilities such that the patient’s true diagnosis is never considered. While testing may be done to confirm a diagnosis, there may also be confirmation bias or seeking data to confirm an inaccurate hypothesis.