This study investigated how an initial diagnosis influences delivery of second opinions concerning skin cancers, including the possibility of an incorrect diagnosis.
Investigators are calling for a reexamination and refinement of the process in which second opinions for skin cancer–related diagnoses are delivered by dermatopathologists, after their study findings published in JAMA Dermatology showed that prior results likely influenced incorrect interpretation of study cases.
“Medical second opinions are common, although little is known about the best processes for obtaining them,” the study authors wrote, adding that they wanted “to measure the extent to which dermatopathologists’ diagnoses are influenced by prior diagnostic information from another dermatopathologist.”
The study cohort comprised 149 dermatopathologists (68%, male; median age, 47 [range, 34-76] years; 41%, Southern United States; 99%, board certified in dermatapathology) randomly assigned in 2 phases to interpret slide sets of melanocytic skin biopsy specimens from the Melanoma Pathology Study; these slides were organized by class, ranging from class 1 (benign/mildly dysplastic nevi) to class V (malignant invasive melanoma).
Phase 1 did not include prior diagnostic information from another dermatopathologist. Phase 2 involved reexamining the same 5 slide sets (90 cases) but with prior diagnostic information that indicated a less or more severe diagnosis compared with what each clinician produced in phase 1. All prior diagnostic information “were actually diagnoses from board-certified and or fellowship-trained dermatopathologists.”
Overall, when handing down their second opinion, the dermatopathologists were 52% more likely (relative risk [RR], 1.52; 95% CI, 1.34-1.73) to give a more severe diagnosis following their phase 2 slide review vs their phase 1 diagnosis—when prior diagnostic information was not provided—and if the prior diagnosis provided in phase 2 indicated more severe disease. In addition, although a slightly lower finding, they were also more likely, at 38% (RR, 1.38; 95% CI, 1.19-1.59), to hand down a second opinion of a less severe case of disease if their phase 2–provided diagnostic information indicated less severe disease.
There were 5322 overall interpretations, and most physicians had been interpreting melanocyte skin lesions for 5 to 9 years (26%) or 10 to 19 years (42%). Forty-eight percent spent 10% to 24% of their caseload performing these interpretations.
The study authors also noted trends of influenced opinions even among those who claimed not to be swayed by another diagnosis. Thirty-one percent of dermatopathologists who originally indicated their diagnostic ability was not affected by a prior diagnosis showed evidence of influence in this regard, according to a baseline survey. In addition, study findings showed that of those who consulted on cases, 21% were swayed when giving a decision and 47% handed down an incorrect phase 2 diagnosis even when their phase 1 diagnosis was correct.
“While medical practice has long incorporated second opinions, the potential influence of antecedent diagnostic information has never been systematically studied to our knowledge,” the authors wrote. “A lack of research on second opinions may reflect the challenges in studying the topic.”
They added that their results may actually be an underestimation of influence, however, because notes saved from phase 1 slide observations may have led the dermatopathologists to ignore the initial diagnostic information in phase 2. The possibility of being less influenced in a testing situation and training (board certification and fellowship) were both cited as potential reasons for this.
“The adage ‘2 heads are better than 1’ should be carefully considered in the context of medical decision-making,” the authors concluded. “We need to consider and refine the methods of how these 2 heads come together.”
Elmore JG, Eguchi MM, Barnhill RL, et al. Effect of prior diagnoses on dermatopathologists’ interpretations of melanocytic lesions: a randomized controlled trial. JAMA Dermatol. Published online August 10, 2022. doi:10.1001/jamadermatol.2022.2932