This study demonstrates that the BQ has a high sensitivity but low specificity and low positive predictive value. Moreover, BQ has a high misclassification rate and its diagnostic accuracy is no different than a random chance. Our findings corroborate the results of Netzer et al. (1999) concerning the sensitivity of BQ, but not specificity, positive predictive value, and the likelihood ratio, all of which were greater in high risk patients in the study of those authors, amounting to 77%, 87%, and 3.2, respectively.

In general, BQ has expectedly high sensitivity, as this tool has been developed for identifying high risk patients at the primary care level. However, low specificity and high misclassification rate suggest that BQ has a low discriminatory power and its utility is no different than the judgement of clinicians (Cowan et al. 2014; Sert Kuniyoshi et al. 2011). The present findings also support earlier studies showing that the BQ is of limited utility in specialized clinics (Ahmadi et al. 2008). Currently, clinicians look for a simple questionnaire that may be used as a tool to determine the risk of OSA syndrome and to predict the possible perioperative respiratory complications. The latter may improve clinical outcome when anesthesia and surgery are required (Gokay et al. 2016). The data from recently published systematic review suggest that the BQ is a questionnaire that enables to risk stratify patients for peri- and postoperative complications. However, testing of BQ is still required with a focus on specific surgery types, adjusted for potentially confounding factors (Dimitrov and Macavei 2016). A higher score of BQ in specific groups of patients after stroke or transient ischemic attack indicates that this tool is but moderately predictive for OSA exclusion (Boulos et al. 2016). It patients suffering from type 2 diabetes, BQ fails to identify 31% of patients with moderate-to-severe OSA, preventing such patients from receiving correct diagnosis and treatment. However, BQ may be suboptimal when OSA screening is done with home sleep monitoring devices (Westlake et al. 2016). An evaluation of BQ in Iranian patients with AHI >5 shows its sensitivity and specificity for OSA diagnosis as 77.3% and 23.1%, respectively, PPV of 68.0%, and NPV of 22.0% (Khaledi-Paveh et al. 2016), which is akin to present findings in the Polish population. The BQ has also been tested in Portuguese patients in whom it shows an acceptable reliability, but after excluding the following two questions: ‘Has anyone noticed that you stop breathing during your sleep?’ and ‘Have you ever dozed off or fallen asleep while driving?’ (Silva et al. 2016). Arunsurat et al. (2016) have assumed that the BQ may be useful as an OSA screening tool for the Thai or Asian populations after some adjustments. In addition, there is an apparent paucity of BQ testing in population samples comprising women and individuals of a low educational level (Silva et al. 2016). Interestingly, Gupta et al. (2016), in view of the unavailability of any screening tool for OSA in Hindi, have undertaken an attempt to explore the validity of a Hindi version of BQ, irrespective of the literacy status of subjects. The results have demonstrated sensitivity of 89%, specificity of 58%, PPV of 87%, and NPV of 63%, which supports the role of BQ as a valid tool for OSA screening OSA in that population. There is also a need to use simple tools for OSA screening in Africa, where the awareness of OSA is poor and its incidence is underreported, despite a high prevalence of symptoms (Desalu et al. 2016).