Duarte · Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine 2023 · cross-sectional diagnostic accuracy study · n=4,499

Screening for obstructive sleep apnea: comparing the American Academy of Sleep Medicine proposed criteria with the STOP-Bang, NoSAS, and GOAL instruments.

Cited 15 times in the scientific literature.

Level 3 - non-randomized controlled study

Cross-sectional diagnostic validation cohort compared against gold-standard polysomnography

PubMed 36872648 · doi:10.5664/jcsm.10546 · record verified 2026-08-29

What was done

This cross-sectional diagnostic accuracy study evaluated the performance of the 2017 American Academy of Sleep Medicine (AASM 2017) screening criteria compared with three validated instruments: NoSAS, STOP-Bang, and GOAL. The cohort included 4,499 adults undergoing overnight polysomnography at a single referral center between July 2019 and December 2021. High risk under AASM 2017 was defined as excessive daytime sleepiness plus at least two of three criteria: loud snoring; observed apnea, gasping, or choking; and hypertension. Obstructive sleep apnea (OSA) severity thresholds based on the apnea-hypopnea index were 5.0, 15.0, and 30.0 events/h. Predictive performance was evaluated using area under the receiver operating characteristic curve (AUC) and contingency tables.

What was found

For any OSA severity, AASM 2017 criteria showed a sensitivity of 31.0% to 40.6% and a specificity of 80.8% to 89.6%. AASM 2017 had higher specificity but markedly lower sensitivity across all severity cutoffs compared to GOAL, STOP-Bang, and NoSAS. GOAL, STOP-Bang, and NoSAS each achieved AUC values > 0.7 for any OSA severity, significantly outperforming AASM 2017 (all P < .001). Performance was similar among GOAL, STOP-Bang, and NoSAS across all OSA severity levels (all P > .05). Specific numerical sensitivities, specificities, and exact AUC values for the comparator tools were not reported in the abstract.

Why it matters

The AASM 2017 screening criteria miss a substantial proportion of patients with OSA due to low sensitivity, demonstrating that STOP-Bang, NoSAS, and GOAL are far more effective screening tools in clinical referral populations.

Limits

The study was conducted in a single referral center cohort, introducing referral bias and limiting generalizability to primary care or general community settings. Numerical diagnostic metrics (exact AUCs, sensitivity, specificity) for the comparator questionnaires were omitted from the abstract.

Cited by