Validation of the efficacy of the GPSS questionnaire for screening obstructive sleep apnea

This study aimed to compare the screening performance of a newly developed questionnaire for obstructive sleep apnea (OSA), the General Practice Sleep Scale (GPSS), with four commonly used screening questionnaires (NoSAS, Berlin, STOP, and STOP-Bang) across different levels of OSA severity, in order to assess their applicability in clinical practice.

The study retrospectively included 2,169 patients from the Sleep Medicine Center of the First Affiliated Hospital of Guangzhou Medical University (January 2012 to June 2017) as the first group, and 310 patients from the Sleep Center of the Second Affiliated Hospital of Guangdong Medical University (January 2020 to June 2025) as the second group, all of whom were assessed for OSA. The sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve (AUC) of each OSA screening questionnaire were calculated to evaluate their diagnostic performance.

The prevalence of OSA was 69.3% in the first group and 79.5% in the second group. In the first group, neck circumference, waist circumference, pulse rate, systolic blood pressure, diastolic blood pressure, age, height, weight, BMI, and sex showed significant differences between patients with OSA and those without OSA. In the comparison of sensitivity, specificity, PPV, and NPV of the five questionnaires among patients in the first group: for all OSA patients, the GPSS questionnaire showed higher specificity and positive predictive value compared to the other questionnaires; however, for moderate/severe OSA patients, the GPSS questionnaire demonstrated lower sensitivity, specificity, PPV, and NPV than the other four questionnaires. In the second group, among all OSA patients, the GPSS questionnaire demonstrated higher sensitivity, positive predictive value, and negative predictive value compared to the other questionnaires; among moderate/severe OSA patients, the GPSS questionnaire showed higher specificity and positive predictive value compared to the other questionnaires. ROC analysis results showed that, in the first group, the GPSS questionnaire demonstrated superior sensitivity, specificity, positive predictive value, and negative predictive value compared to the other questionnaires among all OSA patients, moderate OSA patients, and severe OSA patients. The corresponding AUC values were 0.75 (95% CI: 0.73-0.77), 0.73 (95% CI: 0.71-0.75), and 0.73 (95% CI: 0.71-0.75), respectively. In the second group, the GPSS questionnaire outperformed the other four scales in all OSA patients and moderate OSA patients, with AUC values of 0.77 (95% CI: 0.72-0.83) and 0.75 (95% CI: 0.70-0.80), respectively. However, it still had some reference value in patients with severe OSA, with a value of 0.69 (95% CI: 0.62-0.75).

According to the ROC results, the GPSS questionnaire was superior to the other four scales in screening for mild-to-moderate OSA, though there remains room for improvement in detecting severe OSA. Therefore, we recommend the GPSS for early screening of OSA, especially in primary healthcare settings with limited time. For patients at high risk of severe OSA, a combination of GPSS and other screening tools could be considered.