Investigating the minimal clinically important difference for SNOT-22 symptom domains in surgically managed chronic rhinosinusitis

Background: Prior work has described 5 domains within the 22-item Sino-Nasal Outcomes Test (SNOT-22) that allow for stratification of symptoms into similar clusters and that can be used to direct therapy. Although the outcomes of various interventions on these symptom domains have been reported, minimal clinically important difference (MCID) values have not been investigated, which has limited clinical interpretation of these results. Methods: This study was designed as a secondary analysis of a prospective, multi-institutional, observational cohort. A total of 276 patients with medically refractory CRS who underwent surgical management were enrolled. Distribution-based methods (half-standard deviation, standard error of measurement, Cohen's d, and the minimum detectable change) were used to compute MCID values for both SNOT-22 total and domain scores. The Medical Outcomes Study Short Form 6D (SF-6D) health utility score was used to operationalize anchor-based associations using receiver-operating characteristic (ROC) curves. Results: The mean MCID of several distribution-based methods for total SNOT-22 scores was 9.0, in agreement with previously published metrics. Average MCID values for the rhinologic, extranasal rhinologic, ear/facial, psychological, and sleep domain scores were 3.8, 2.4, 3.2, 3.9, and 2.9, respectively. Anchor-based approaches with the SF-6D did not have strong predictive accuracy across total SNOT-22 scores or domains (ROC areas under-the-curve ≤ 0.71), indicating weak associations between improvement in SNOT-22 scores and health utility as measured by the SF-6D. Conclusion: This estimation of MCID values for the SNOT-22 symptom domains allows for improved clinical interpretation of results from past, present, and future rhinologic outcomes research.