Optimizing Pediatric Sleep-Disordered Breathing Screening: Key Predictive Questions Identified from CHAT and PATS Databases

Clinical Question

Can a small set of screening questions accurately predict or rule out severe pediatric obstructive sleep apnea (OSA)?

Bottom Line

Three key screening questions can predict severe pediatric OSA with a positive predictive value of 100% and a negative predictive value of approximately 95.97–96.64%, offering a practical alternative to extensive questionnaires.

Background: Screening for pediatric sleep-disordered breathing (SDB) is essential to guide polysomnography (PSG) referral and perioperative planning. Existing tools such as the Pediatric Sleep Questionnaire (PSQ), OSA-18 questionnaire, and Epworth Sleepiness Scale are often lengthy and show variable accuracy, particularly in distinguishing disease severity.

Study Design: Secondary analysis of prospectively collected data from two major pediatric trials (CHAT and PATS). All possible combinations of two to four screening questions were analyzed for predictive accuracy using PSG-defined OSA outcomes.

Setting: Multicenter pediatric clinical trial datasets from the National Sleep Research Resource

Synopsis: A total of 645 children aged five to nine years were analyzed. Severe OSA prevalence ranged from 10.7% to 13.3% depending on the PSG metric (RDI, OAHI, AHI). The study systematically evaluated all possible combinations of screening questions derived from PSQ, OSA-18, and ESS. Among 304 unique combinations, several sets of three to four positive screening responses consistently achieved 100% positive predictive value (PPV) with high negative predictive value (NPV). The most effective combinations incorporated breathing abnormalities during sleep (e.g., breath-holding, choking, or gasping) or behavioral concerns (hyperactivity, aggression, discipline problems), mouth breathing, and growth delay.

The top four combinations demonstrated NPVs of 95.97%-96.64%, effectively ruling out severe OSA when negative. Notably, snoring alone was not a strong predictor and was absent from the high-performing question combinations, highlighting the limitations of relying on snoring-based screening. Tonsil size was associated with OSA severity in univariate analysis but did not independently predict severe disease in multivariable models, suggesting limited utility as a standalone screening tool. The findings emphasize the importance of integrating respiratory symptoms, behavioral changes, and growth patterns in screening. Clinically, these simplified question sets can streamline decision-making regarding PSG referral, surgical planning, and post-operative monitoring, particularly in resource-limited settings.

Limitations include dataset heterogeneity, exclusion of obese children, and reliance on parent-reported symptoms, which may introduce reporting bias.

Citation: Sturm H, et al. Optimizing pediatric sleep-disordered breathing screening: key predictive questions identified from CHAT and PATS databases. Otolaryngol Head Neck Surg. 2026;174:265-271. doi:10.1002/ohn.70054.

Comment: This study looked at pooled data from the CHATs and PATs studies to assess which combination of elements from common questionnaires (PSQ, OSA-18, and ESS) were predictive of severe OSA. Six hundred forty-five patients were pooled, and 304 screening question combinations were analyzed; three question combinations came to have the highest PPV (100%) and NPV (~96%): 

1. Presence most of the time or all the time of breath-holding spells or pauses in breathing at night OR choking or gasping sounds while asleep OR aggressive or hyperactive behavior OR discipline problems;
2. Tendency to breathe through the mouth during the day; and
3. Growth delay at any time since birth.

–Eric Gantwerker, MD