Creation of an abbreviated version of the 22-item Sino-Nasal Outcome Test (SNOT-22) in Patients with Cystic Fibrosis using Item Response Theory

Haidee Chen, BA

Daniel M. Beswick, MD

Tristan Tibbe, MA, MS

Myung-Shin Sim, DrPH

Ahmad R. Sedaghat, MD, PhD

Ryan Cottner

Ashleigh Halderman, MD

Naweed Chowdhury, MD

Anne Getz, MD

Jeremy P. Tervo 

Jonathan B. Overdevest, MD

Ethan Han, BS

Area of concentration: Basic, Clinical, and Translational Research

Specialty: Otolaryngology
Keywords: IRT, Chronic Rhinosinusitis, SNOT-22

Background: The Sinonasal Outcome Test-22 (SNOT-22) is widely utilized as a comprehensive patient-reported measure for assessing the impact of chronic rhinosinusitis (CRS) on quality of life. Given the frequency of CRS in patients with cystic fibrosis (PwCF), a population for whom survey burden can lead to decreased treatment compliance, a tailored and efficient assessment tool is immensely beneficial. This study seeks to apply Item Response Theory (IRT) to individual SNOT-22 questions in a population of PwCF with the objective of identifying the most informative questions of the survey and reducing survey burden in PwCF. 

Objective: Implementation of IRT using SNOT-22 questions to create a reliable, valid, shorter, abbreviated version of SNOT-22 to assess symptoms in PwCF and decrease survey burden.

Methods: Demographics and SNOT-22 data was collected from 180 PwCF across five institutions. IRT, chosen due to its potential to reveal the most informative items within a given instrument, was then used to analyze individual SNOT-22 questions to identify the most discriminating questions that provide the most information about CRS symptoms in each SNOT-22 subdomain. Based on this information, abbreviated 12-item and 7-item scales were created.

Results: Using IRT, we shortened the SNOT-22 questionnaire to surveys that contained 12 and 7 items. The abbreviated surveys’ item information functions demonstrate that they effectively encompass the spectrum of symptom severity for CRS in PwCF, with both scales demonstrating strong internal reliability (Cronbach’s alpha≥ 0.83) and are highly correlated with the original 22-item scale (r ≥ 0.93).

Conclusion: Identification of SNOT-22 survey questions that provided the most information concerning symptom severity for PwCF using IRT provides a potential opportunity to shorten the current SNOT-22 survey to decrease survey burden in PwCF while accurately assessing CRS symptom severity.