Dedicated Olfaction and Taste Items do not Improve Psychometric Performance of the SNOT-22

Abstract

Objective: Previous work has shown the chemosensory dysfunction item of the 22-item Sinonasal Outcome Test (SNOT-22) that assesses problems with "taste/smell" has poor psychometric performance compared with other items on the SNOT-22, which we have hypothesized is due to the simultaneous assessment of two different senses. Our aim was to determine whether distinct smell and taste items in the SNOT-22 would improve psychometric performance.

Methods: One hundred and eighty-one CRS patients were recruited and completed the SNOT-22. Additional items querying problems with the senses of "smell" and "taste," using the same response scale and recall period were given to study participants. Item response theory (IRT) was used to determine IRT parameters, including item discrimination, difficulty, and information provided by each SNOT-22 item.

Results: Confirming previous studies, the chemosensory item of the SNOT-22 (reflecting "taste/smell") had poor psychometric performance. Use of a distinct smell or taste item instead of the combined "taste/smell" item did not improve psychometric performance. However, a dedicated smell question resulted in a left shift of threshold parameters, showing that the dedicated smell item better captures moderate CRS disease burden than the original taste/smell item of the SNOT-22, which by virtue of near-identical IRT parameters appears to more greatly reflect problems with taste.

Conclusions: A dedicated smell- or taste-specific item, rather than the combined "taste/smell" item currently in the SNOT-22 does not provide significantly greater psychometric performance. However, a dedicated smell item may better capture moderate CRS disease burden compared with the current chemosensory item on the SNOT-22. Laryngoscope, 132:1644-1651, 2022.

Keywords: CRS; IRT; SNOT-22; chronic rhinosinusitis; graded response model; item response theory; olfaction; quality of life; smell; taste.

Fig. 1

Item response category characteristic curves (ICCs) for the combined smell/taste, the distinct smell, and the distinct taste item. The curves represent each answer category, and the peaks denote the item discriminability [Color figure can be viewed in the online issue, which is available at www.laryngoscope.com.]