Validation of a clinical assessment score for pediatric sleep-disordered breathing

Abstract

Objectives/hypothesis: To validate a clinical assessment score for pediatric sleep-disordered breathing.

Study design: Prospective instrument validation.

Methods: One hundred children scheduled for overnight polysomnography were evaluated by a standardized history and physical examination and assigned a clinical assessment score. Parents completed the Obstructive Sleep Apnea (OSA)-18, the Pediatric Quality of Life Inventory (PedsQL) 4.0, and the Child Behavior Checklist questionnaires. Children with positive polysomnography underwent adenotonsillectomy or adenoidectomy. The identical assessments were performed at a mean follow-up of 8 months.

Results: Item reduction yielded a score of 15 items (Clinical Assessment Score-15 [CAS-15]) that demonstrated the best internal consistency and predictive utility (Cronbach α = .80). Intraclass correlation (ICC) demonstrated good intrarater (ICC, 0.78; 95% confidence interval [CI], 0.58 to 0.89) and inter-rater agreement (ICC, 0.65; 95% CI, 0.26 to 0.84). All change scores were significantly improved after surgery. Effect sizes were large for the CAS-15 (2.6), OSA-18 (2.4), and apnea-hypopnea index (1.4), and moderate for the Child Behavior Checklist (0.7) and PedsQL 4.0 (-0.5). Moderate to strong correlation was found between the initial CAS-15 scores and the external measures (|r| between 0.32 and 0.65). Receiver operating characteristic curves were constructed to determine the optimal initial CAS-15 score for predicting positive polysomnography. The area under the curve was 0.77 (95% CI, 0.67 to 0.87); and a score ≥32 yielded a sensitivity of 77.3% (95% CI, 65.3 to 86.7) and a specificity of 60.7% (95% CI, 40.6 to 78.5).

Conclusions: The CAS-15 proved useful in an office setting and correctly diagnosed 72% of referred children when compared to polysomnography. It correlated well with external measures and demonstrated a good response to clinical change.