Exploring possibilities for shortening the 22‐item Sino‐Nasal Outcome Test (SNOT‐22) using item response theory. Issue 2 (26th August 2021)

• Record Type: Journal Article
• Authors: Liu, David T.; Phillips, Katie M.; Speth, Marlene M.; Besser, Gerold; Mueller, Christian A.; Sedaghat, Ahmad R.
• Is Part Of: International forum of allergy & rhinology. Volume 12:Issue 2(2022)
• Page Start: 191
• Page End: 199
• Publication Date: 2021-08-26
• Subjects: chronic rhinosinusitis -- CRS -- IRT -- item response theory -- quality of life -- SNOT‐22; 617.51005
• DOI: 10.1002/alr.22878
• Languages: English
• ISSNs: 2042-6976

Abstract

Abstract: Background: Even with a high‐quality instrument, such as the 22‐item Sino‐Nasal Outcome Test (SNOT‐22), patients can be overwhelmed by repeated completion of questionnaires, leading to nonadherence and reduced data quality. The aim of this study was to evaluate whether the SNOT‐22 could be made more concise without losing the valuable information that it provides. Methods: A modern psychometric approach, item response theory (IRT), was used on a sample of 800 patients with chronic rhinosinusitis (CRS). The SNOT‐22 was refined based on: (1) the item information criteria, retaining only items with above‐average information; and (2) the total test information, retaining only the most discriminating items to cover at least 30% of the information within each subdomain. The preliminary validity and reliability of these refined scales were assessed using Cronbach's alpha and Pearson's correlation. Results: Using an IRT approach, we find that it may be possible to shorten the SNOT‐22 to an 11‐item and six‐item version based on psychometric properties. Item information functions of the shortened 11‐item and six‐item scales demonstrate that both versions accurately covered the CRS symptom‐severity continuum. Preliminary reliability and validity analysis showed that both refined scales had good to excellent reliability (Cronbach's alpha ≥ 0.80) and were highly associated with the original full‐length scale ( r ≥ 0.90). Conclusion: IRT provides data‐driven opportunities for the continuous development and refinement of the SNOT‐22. While patient and provider input must be accounted for too, our results show that future revisions of the SNOT‐22 could include significantly fewer items.