STUDY DESIGN: This prospective diagnostic accuracy study was divided into two phases.

SETTING: The study was conducted at a quaternary teaching hospital.

METHODS: Polysomnograms performed from 1 April 2014 to 31 December 2021 were included. In Phase 1, the parameters of 95 oximetry trend graphs were evaluated, and a modified scoring system was constructed. In Phase 2, the modified scoring system was employed in 272 oximetry trend graphs, and its diagnostic accuracy was determined. A logistic regression model was used to assess the ability of each scoring system to predict paediatric OSAS.

RESULTS: A total of 367 patients were recruited. In Phase 1, a four-tier severity classification system was constructed. In Phase 2, its diagnostic accuracy was found to be 53.3% sensitive, 97% specific, with positive predictive value of 98.5% and negative predictive value of 34.6%. The lowest detectable apnoea-hypopnoea index (AHI) was 4.5. The inter-rater reliability calculated was 80%. Logistic regression was applied to assess associations of the modified McGill score (MMS) or McGill oximetry score (MOS) with OSAS. The area under the receiver operating characteristic curve was higher for the MMS than for MOS (0.78 [95% CI 0.73-0.84] vs. 0.59 [95% CI 0.51-0.66]).