Comparative validation of the Sleep Disordered Breathing scale of the Sleep Disorders Questionnaire 2nd edition and four other sleep apnea questionnaires against polysomnography

Abstract

Abstract Background The original Sleep Disorders Questionnaire (SDQ, 1994) contained a 12-item sleep apnea subscale (SDQ-SA). SDQ was shortened by factor analysis in 2024 (SDQ-2) and contains a 14-item subscale “Sleep Disordered Breathing” (SDQ2-SDB). This is the first external validation study of the new scale using polysomnography, with a concurrent validation against four commonly used sleep-disordered breathing questionnaires. Methods Members of a large urban police force (n = 733) completed the SDQ2-SDB, Berlin, STOP-Bang, NoSAS, and GOAL questionnaires. A subgroup (n = 101) also volunteered for a laboratory polysomnogram (PSG). SDQ2-SDB scores were analyzed by ANOVA in both samples. For the NPSG subgroup, Receiver Operating Characteristics (ROC) of the five apnea scales were calculated at a criterion of AHI ≥ 15. A correlation matrix among the five scales was also calculated. Results The NPSG subgroup with higher measured AHI showed significantly higher SDQ2-SDB scores. In the whole sample, SDQ2-SDB scores were significantly higher in males compared with females, in older compared to younger individuals, and in those with higher BMI. Age x Sex, Age x BMI and BMI x Sex effects were also significant. The areas under the ROC curve (AUC) were: SDQ2-SDB = 70%, Berlin = 65%, STOP-Bang = 73%, GOAL = 81%, NoSAS = 85%. The NoSAS AUC was significantly higher (p < 0.03) than those of the SDQ2-SDB, STOP-Bang, and Berlin scales but the latter three scales did not differ significantly from each other. Positive (PPV) and negative (NPV) predictive values of the SDQ2-SDB were 46% and 84% respectively, with the other scales ranging between 35 to 52% PPV and 74 to 89% NPV. The SDQ2-SDB correlated with the other four questionnaires with rho ranging from 0.558 to 0.855. Conclusions The SDQ2-SDB scale demonstrates adequate validity in a non-clinical sample. Its performance is similar to questionnaires commonly used for sleep apnea, like the Berlin and STOP-Bang scales.