Semantically Ambiguous Language in the Teaching Operating Room

Teaching and training surgeons work hard in the OR to understand each other, yet miscommunication is an important cause of preventable adverse events in surgery. Our objective was to perform a formal semantic analysis of language in authentic teaching surgical cases, identify the prevalence and typology of ambiguous or potentially ambiguous language, and describe their potential for miscommunication.

In this secondary analysis of qualitative data, we collaborated with a semanticist, categorizing linguistic phenomena often associated with miscommunication. We defined an ambiguous phenomenon as a string of language that could be reasonably interpreted in more than one way. We analyzed transcripts of 319 minutes of surgery, coding for 14 linguistic categories. Cohen's kappa was calculated. We determined the prevalence and rate of each linguistic category and chose illustrative examples.

Six surgical attendings, four fellows, and six residents, ranging from PGY1 to PGY4, at the University of Pittsburgh Medical Center, a tertiary medical center in Pittsburgh, Pennsylvania.

We found 3912 examples of potentially ambiguous language, 12.3 per minute. Percentage agreement between two expert raters was 76.3%. The most common phenomena were deixis (3.1 per minute), directional (2.6), anaphora (1.3), implicit instruction (1.3), and degree modifiers (0.7). Restatements/reframing occurred 1.4 times per minute. We identified 131 near misses associated with potentially ambiguous language. Cohen's kappa was 0.70 among expert semanticists.

Potentially ambiguous language is common and has the potential to jeopardize safe teaching surgery. We postulate that the high amount of potentially ambiguous language use in the operating room places a burden on the training surgeon to comprehend surgical instruction.