Validation of the Alliance Sleep Questionnaire for Screening Obstructive Sleep Apnea Compared to Gold Standard In-Lab Polysomnography Study

Background: Obstructive sleep apnea (OSA) is the most common type of sleep-disordered breathing disorder and is associated with various health repercussions. Given the high prevalence of OSA, diagnosis needs to be better streamlined to get individuals the care they need to improve their overall wellbeing. The current gold standard for diagnosing sleep disordered breathing requires patients to have an overnight in-lab polysomnography (PSG). The purpose of this validation study was to evaluate which questions from the self-reported Alliance Sleep Questionnaire (ASQ) have diagnostic utility for screening sleep apnea.
Methods: The sample used a research study cohort of individuals who completed a PSG and filled out the ASQ (n = 735 females, n = 631 males) of who were n = 439 no apnea, n = 443 mild apnea, n = 262 moderate apnea and n = 222 severe sleep apnea. Logistic regression models were used to identify relevant ASQ questions associated with apnea to find correlations, create regression models, classification trees and the area under the curve (AUC) for the final models (backward selection).
Results: The results of this study found that sex, BMI, and the primary complaint of loud snoring were the most predictive variables for sleep apnea. For all severities, (mild, moderate, and severe) being male was a strong predictor with odds ratios (OR) of 2.31, 3.92 and 4.69 respectively. BMI was split into three groups: normal weight, overweight and obese. Across all severities the odds ratios increased with weight category. The overweight group had OR’s of 1.44 for mild, 1.75 for moderate and 2.87 for severe. Similarly, the obese group had OR’s of 1.60 for mild, 2.79 for moderate and 8.68 for severe. Lastly, the primary complaint of snoring had OR’s of 1.89, 3.73 and 7.93 for mild,
moderate, and severe apnea. AUC for the final models were 0.69, 0.79 and 0.88 (mild, moderate, and severe).
Conclusions: Using the ASQ questionnaire to screen for sleep apnea and predict severity is possible using a specific subset of questions. There is need to replicate these findings in a community-based sample before it can be implemented into clinical practice.