A randomized experimental study to assess the effect of language on medical students' anxiety due to uncertainty

Background Communication and handoff failures are common causes of diagnostic errors in hospital. Human quest for certainty can increase the likelihood of premature closure in decision-making, the most common phenomenon in misdiagnosis. Little research exists on whether language choice in handoffs affects physicians' sense of uncertainty. Methods Medical students from a large US medical school were randomized to receive one of four language variations describing a presumed diagnosis in hypothetical handoffs from emergency department (ED) to inpatient ward. The control language arm used the word 'diagnosis'; experimental arms replaced this word with either 'hypothesis', 'probability of 60%', or 'working diagnosis' with a short differential. Outcome measures were students' anxiety due to uncertainty (range 5-30; higher scores indicating higher stress from uncertainty) and clinical uncertainty about the ED provider's presumed diagnosis. Results Mean anxiety due to uncertainty was significantly higher in subjects receiving the 'hypothesis' language arm compared to those receiving the control 'diagnosis' language [19.2 (4.6) vs. 15.5 (3.4); p<0.008]. Differences between subjects who received the probability language [17.2 (5.8) vs. 15.5 (3.4); p=0.26] and 'working diagnosis' language [16 (5) vs. 15.5 (3.4); p=0.69] were not statistically significant. There was no difference in items assessing clinical uncertainty after each scenario. Conclusions The word 'hypothesis' increased anxiety due to uncertainty compared to the word 'diagnosis', but did not change assessments of clinical uncertainty. Further research is needed to assess how use of language in clinical handoffs may influence perceptions and anxiety related to uncertainty and whether optimal language can be identified that leads to recognition of uncertainty without maladaptive stress or anxiety due to uncertainty.