Statistical overview of the Sniffin’ sticks olfactory test from the perspectives of anosmia and hyposmia

Abstract

The Sniffin’ Sticks test is evaluated by summing the scores of threshold, discrimination, and identification subtests to establish an olfactory diagnosis (anosmia , hyposmia , normosmia). However, variations in thresholds, ranges and inconsistencies have been observed. Statistical analyses (distributions, quantiles), protocol simulation, and implementations of the Sniffin’ Sticks test were conducted. This study contributes a statistical revision: determining a cut-off point for the total TDI score, optimising the number of alternatives in the identification subtest, and equalising correct responses (frequency, balanced in time, subsequent pairs). The upper score threshold used to diagnose anosmia is lower than the typical first-order error thresholds applied in comparable assessments ( anosmia ≤ 16.75 points, hyposmia ≥ 17 points at 90% confidence level). In the culturally adapted 16-pen Sniffin’ Sticks identification tests, the frequency of correct responses was aligned with that of the original test. However, the equalisation of the frequency of adjacent correct response pairs and the temporal occurrence of correct responses was not met in any test. The order of all correct answers that simultaneously satisfied all three matching conditions was determined. The practical significance of the results is that the recommended protocol of the Sniffin’ Sticks test system can be implemented in clinical practice with minimal modification.