Assessment of a surgical pattern recognition examination

Recent cognitive research in medicine has demonstrated that expert diagnosticians solve many clinical problems through "pattern recognition" rather than deductive reasoning strategies. Development of the ability to recognize and diagnose common surgical problems is a critical objective of undergraduate surgical education. A surgical pattern recognition examination (PAT) was developed to assess surgical diagnosis recognition in clerkship students.

Surgical faculty selected 18 diagnoses for 20 possible presenting complaints (eg, leg pain, soft tissue mass). A distinctive patient description was written for each diagnosis. An examination of 200 items was compiled covering 20 presenting complaints.

The PAT was administered to clerkship students during the 1992-1993 academic year (n = 77). PAT scores ranged from 33% to 90% with a mean of 65% (SD = 12.7). PAT scores correlated highly with the National Board of Medical Examiners (NBME) Surgery Subject Examination (SSE) scores. The validity of the examination was assessed by a speeded administration of the exam to surgical house staff. Senior residents scored significantly higher than junior residents (mean 82% versus 63%, P = 0.004). Performance on the PAT was significantly affected by the order of the student rotation in the academic year (P = 0.02) while performance on the NBME SSE was not (P = 0.40), which is consistent with diagnostic ability improving with clinical experience.

The surgical PAT examination is a new, reliable, and valid method for assessing diagnostic ability in third-year students.