301
Views
28
CrossRef citations to date
0
Altmetric
Original Research

Obstructive Sleep Apnea Screening with a 4-Item Instrument, Named GOAL Questionnaire: Development, Validation and Comparative Study with No-Apnea, STOP-Bang, and NoSAS

ORCID Icon, , ORCID Icon, , ORCID Icon & ORCID Icon
Pages 57-67 | Published online: 23 Jan 2020
 

Abstract

Background

Obstructive sleep apnea (OSA) is a very prevalent disorder. Here, we aimed to develop and validate a practical questionnaire with yes-or-no answers, and to compare its performance with other well-validated instruments: No-Apnea, STOP-Bang, and NoSAS.

Methods

A cross-sectional study containing consecutively selected sleep-lab subjects underwent full polysomnography. A 4-item model, named GOAL questionnaire (gender, obesity, age, and loud snoring), was developed and subsequently validated, with item-scoring of 0–4 points (≥2 points indicating high risk for OSA). Discrimination was assessed by area under the curve (AUC), while predictive parameters were calculated using contingency tables. OSA severity was classified based on conventionally accepted apnea/hypopnea index thresholds: ≥5.0/h (OSA≥5), ≥15.0/h (OSA≥15), and ≥30.0/h (OSA≥30).

Results

Overall, 7377 adults were grouped into two large and independent cohorts: derivation (n = 3771) and validation (n = 3606). In the derivation cohort, screening of OSA≥5, OSA≥15, and OSA≥30 revealed that GOAL questionnaire achieved sensitivity ranging from 83.3% to 94.0% and specificity ranging from 62.4% to 38.5%. In the validation cohort, screening of OSA≥5, OSA≥15, and OSA≥30, corroborated validation steps with sensitivity ranging from 83.7% to 94.2% and specificity from 63.4% to 37.7%. In both cohorts, discriminatory ability of GOAL questionnaire for screening of OSA≥5, OSA≥15, and OSA≥30 was similar to No-Apnea, STOP-Bang or NoSAS.

Conclusion

All four instruments had similar performance, leading to a possible greater practical implementation of the GOAL questionnaire, a simple instrument with only four parameters easily obtained during clinical evaluation.

Disclosure

The authors declare no conflicts of interest.