Sleep apnea-hypopnea syndrome (OSA) is a chronic disease that causes upper airway (UA) collapse resulting in multiple episodes of complete (apnea) or partial (hypopnea) obstruction, causing poor sleep quality and intermittent hypoxemia with vascular impact (1).

OSA prevalence in the general population is between 6 and 10% (2), being higher with increasing age. In Spain, between 1,200,000 and 2,150,000 people suffer from OSA, and of them, between 24 and 26% (3) present a very severe picture, being, in addition, OSA a condition related to cardiovascular diseases and occupational and traffic accidents (4).

However, it is currently estimated that only 5–9% of people with OSA have been diagnosed, with lack of accessibility to a diagnosis being the main cause attributed (5, 6).

Conventional polysomnography (PSG) is the gold-standard method for diagnosing OSA. PSG consists of continuous recording during the sleep period of neurophysiological (electroencephalogram, electrooculogram, and mental electromyogram), respiratory (peripheral saturation of O₂ and oronasal airflow by nasal cannula and/or thermistor), and other parameters (snoring, thoracoabdominal musculature movements, electrocardiogram, leg movement, and position); however, the realization of PSG requires a costly hospital infrastructure that not all health systems can cover (7). An alternative to PSG is home respiratory polygraphy (HRP). It consists of recording respiratory variables (peripheral O₂ saturation and airflow) and other variables (snoring, thoracoabdominal muscle movements, and position) in the patient's home via portable equipment. This method of diagnosis of OSA has demonstrated good psychometric properties in different health care settings, especially in patients with suspected moderate or severe OSA (7–10).

The search to improve accessibility to the diagnosis of OSA highlights the need to implement new formulas for early detection, the approach, and management of this condition, mainly from the primary care setting (2, 11, 12). In this sense, the STOP-Bang questionnaire (13) has proved its validity and reliability in various healthcare settings (14). However, at present, there are few studies using the Spanish version that support and have demonstrated with the necessary evidence its validity and reliability in primary care patients (15); therefore, we understand pertinent to perform more studies that demonstrate the clinical utility of this questionnaire.

The final number of patients recruited was 255 people, of whom 138 (54.1%) were men and 117 (45.9%) were women. The mean ± standard deviation (SD) of age was 54.76 ± 10 years (95% CI: 53.53–55.59), BMI 31.17 ± 6.58 kg/m² (95% CI: 30.36–31.98), and neck circumference 38.33 ± 5.78 cm (95% CI: 37.62–39.05; 95% CI), with significant differences in neck circumference and BMI by gender (Table 1).

The internal consistency, measured with the Kuder and Richardson coefficient, was 0.623 (95% CI: 0.541–0.665). The interobserver agreement, measured with Cohen's Kappa coefficient, was 0.936 (95% CI: 0.813–1.00; p < 0.001).

Significant gender differences existed in all responses to the STOP-Bang questionnaire (Figure 1).

A total of 38.43% (95% CI: 32.42–44.44) presented no OSA, with significant differences by gender (p = 0.038). Mild OSA was found in 22.75% (95% CI: 17.57–57.92) of people, 22.7% (95% CI: 17.57–57.92) of people presented moderate OSA, and 16.08% (95% CI: 11.54–20.62) presented severe OSA, with significant gender differences in mild OSA (p = 0.038) and severe OSA (p < 0.001) (Figure 2). The AUC for detecting AHI > 15 was 0.769 (95% CI: 0.704–0.833), with an optimum cutoff point of 3, providing a sensitivity of 84.85% (95% CI: 77.28–92.42) and a specificity of 55.10% (95% CI: 44.74–65.46) (Figure 3).

No significant differences were found by gender in the ability to diagnose true positives and true negatives (Chi-square = 2.18; p = 0.10).

In our study, for scores higher than or equal to 3, the STOP-Bang questionnaire showed a 77% diagnostic power for moderate/severe sleep apneas (AHI > 15), with an ability to classify true positives of 85% and true negatives of 55%. For the detection of an AHI > 5, the STOP-Bang sensitivity was 73.25% and specificity was 55.10%, with a good classification ability of 69.4%, results that should be considered with caution because respiratory polygraphy, which is recommended only for moderate to severe apneas (AHI > 15), was the gold standard used for comparisons.

Although Oviedo and Campo (18) indicate an internal consistency above 0.70 as optimal, we consider that a Kuder and Richarson coefficient (17) of 0.623 indicates an acceptable internal consistency considering that we are evaluating the coherence of the questionnaire components to determine the extent to which this tool reflects the theory of the phenomenon to be measured and that this tool has been developed by experts and validated for other populations. According to Landis and Koch (20), the interobserver agreement range obtained was “almost perfect,” confirming the reproducibility of the STOP-Bang questionnaire.

The results of our study were similar to those reported by others performed in Spanish-speaking primary care populations. Cruces-Artero et al. (15) conducted a validation study of the STOP-Bang questionnaire for the identification of moderate/severe sleep apnea (AHI ≥ 15) in primary care in a Spanish population (Galicia-Spain). They recruited a sample of 178 people (57 women and 121 men) over 18 years old whose sociodemographic characteristics (age, gender, BMI) and selection criteria were similar to those of our study. They performed the diagnosis of OSA for the comparison of the STOP-Bang questionnaire score by using PSG and obtained for women a sensitivity of 93.8% (95% CI: 69.80–99.80) and a specificity of 63.4% (95% CI: 46.90–77.90), with an AUC of 0.816 (95% CI: 0.709–0.922) and an optimum cutoff point of 4, and for men a sensitivity of 55.2% (95% CI: 41.50–68.30) and a specificity of 73% (95% CI: 60.30–83.40), with an AUC of 0.686 (95% CI: 0.594–0.778) and an optimum cutoff of 6. Compared with the present study, the study conducted by Cruces-Artero et al. (15) showed better results in women both in sensitivity and specificity; however, in men, the sensitivity obtained by our study was higher and the specificity lower. This fact could be explained because Cruces-Artero et al. (15) used PSG to establish the diagnosis of sleep apnea and not HRP, developing their study for an AHI ≥ 15. In contrast, we used HRP to establish the diagnosis of sleep apnea in our study because it was performed entirely in the primary care setting, developing the study for an AHI > 15.

Saldias Penafiel et al. (21) recruited a sample of 205 Spanish-speaking people (95 men and 110 women) from a metropolitan area of Chile, with a mean age of 47.8 ± 20 years, who presented with clinical symptoms of sleep-related breathing disorders (habitual snoring and/or observed episodes of breathing pauses). The diagnosis of OSA was performed by using HRP. The STOP-Bang questionnaire achieved a sensitivity of 89% and a specificity of 36%, with an AUC of 0.67 for detecting moderate/severe apnea (AHI > 15). These results are consistent with those obtained in our study.

Similar results were also reported in the validation studies of the STOP-Bang questionnaire in the field of primary care in the non-Spanish language (22–26), with a mean sensitivity of 80.11% (range: 77.3–83.9) and a mean specificity of 61.46% (range: 52.8–66.35%).

Comparing the screening ability for moderate/severe apnea of the Berlin questionnaire, a widely used questionnaire whose sensitivity is 76% and its specificity is 59%, we observed that the STOP-Bang questionnaire (27) has a better ability to detect true positives.

The STOP-Bang questionnaire, which consists of dichotomous questions, that is, its questions have only two possible answer options (Yes or No), is very useful as a screening tool. It allows people to populate it in ~1–2 min. It would be an easier tool than the Berlin questionnaire because the questions in the Berlin questionnaire are more difficult, given that they present polyatomic answers from 5 categories. The STOP-Bang questionnaire also shows acceptable internal consistency and good reproducibility.

The demonstrated good psychometric properties, and the ease of use make the STOP-Bang questionnaire an effective tool for screening moderate/severe OSA. Its use in primary care centers could contribute decisively to reducing the underdiagnosis that this disease presents today (3, 28), with the consequent impact on the complications inherent in the lack of treatment.

This study and the existing ones and the SEPAR recommendations that indicate the need to address OSA in primary care suggest that it is necessary to sensitize the primary care professionals to the use of measuring tools for the management of OSA. The STOP-Bang questionnaire is a recommendable tool for screening and diagnostic support.

Although the results support the existing evidence, we believe that it would be advisable to conduct new studies demonstrating the usefulness of the STOP-Bang questionnaire in different population groups in the primary care setting. This fact becomes even more necessary in the Spanish language version.

The development and validation of rapid and inexpensive tools for screening for sleep-related breathing disorders could facilitate their detection under the limited availability of time per patient of the primary health care professional.

Although we would have preferred to obtain a larger sample size, the situation produced by the coronavirus disease (COVID-19) pandemic significantly hampered the process of recruiting the study people, having to limit the fieldwork. Consequently, the precision of the parameters analyzed has been lower than desired, although above the estimated values in the sample size calculation.

Selection bias may have occurred because of the non-use of probabilistic sampling techniques. However, convenience or opportunistic sampling is commonly used in this type of validation study. As some authors have pointed out (29), it is necessary for validation studies that the selected sample covers a broad clinical spectrum of individuals, from the asymptomatic patient to the patient with symptoms specific to the condition being studied, to avoid overestimating the validity of the measuring tool. Given that the recruitment of patients was conducted in a population setting through the primary care services, we consider that this fact has been achieved, leading us to think that the possible selection bias was unimportant.

It is worth noting that, in our study, all the development of the study (recruitment of participants, completion of the STOP-Bang questionnaire, and completion of respiratory polygraphy) has been conducted in the primary care settings by medical and nursing health professionals.

Likewise, we understand the blinded process used as a strength of the study. We prevented the researcher who conducted the polygraphic analyzes from having prior knowledge of the results of the STOP-bang questionnaire, ensuring that the equality of the groups was maintained during the execution of the study. Thus, we reduced the risk of an information bias that could arise from the psychological influence of the knowledge of the interventions received in the groups among the study participants.

It is our intention, when the situation produced by the coronavirus disease (COVID-19) pandemic allows it, to increase the sample to obtain more precise results.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by Comité de ética de la investigación de Córdoba-España. The patients/participants provided their written informed consent to participate in this study.

JS-M, LP-d-T, MV-A, RM-G, and EN-M contributed to conception and design of the study. JS-M organized the database and performed the statistical analysis. RM-G wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.