A modified goldfrank sniffing bar improves accuracy in identifying simulated toxins in a case-based teaching model

Objective:  To determine if olfactory analysis of toxin proxies aid in the identification of poisons when compared to case histories alone.

Methods:  A convenience sample of 42 emergency medicine residents participated in a prospective single-blind observational study using a sniffing bar designed with the technique described by Goldfrank et.al. Each subject received 12 written cases describing a patient with a chemical exposure. Ten chemicals with odors corresponding to a specific case history and 2 chemical case controls with no odor comprised the sniffing bar. Each subject was asked to identify the toxin first by reading the case history alone, then again after smelling the corresponding odor. All subjects were asked to record their level of confidence for each case before and after the addition of the sniffing bar using a 5-point Likert scale. Percentages correct before and after the sniffing bar intervention were calculated and two-tailed p-values and 95% confidence intervals were measured. Continuous comparative variables such as the means of the Likert scale confidence measurements were analyzed with the two-tailed t-test.

Results:  Median percentage of toxins correctly identified was 64.3% (range 9.5-97.6) when based on case description alone and 86.9% (range 9.5-95.2) with addition of the sniffing bar. The percentage of initially incorrect identification reversed after additional data provided by the sniffing bar was 11.5% (range 0-14). Of the 10 actual toxins, correct identification of 6 achieved statistical significance after introduction of the sniffing bar (p=0.0017; 95% CI 4.6-11.4). There was no association between level of training and improved accuracy with use of the sniffing bar, and overall confidence did not improve significantly after addition of the sniffing bar.

Conclusion:  Diagnostic accuracy of poisons in case-based scenarios is improved with the use of a sniffing bar. The sniffing bar is a useful adjunct to a traditional case-based emergency medicine toxicology teaching curriculum.