Reliability of the Turkish version of the European Obstructive Sleep Apnea Screening (EUROSAS) questionnaire for drivers

Motor vehicle accidents (MVA) are one of the most important causes of death and injury in Turkey [1] and worldwide [2, 3]. According to accident statistics in Turkey, 95% of MVAs are caused by drivers, mainly due to high speed, alcohol, substance use, and sleepiness [1]. The most common medical disorder causing sleepiness is obstructive sleep apnea (OSA). According to a questionnaire-based study, the Turkish Adult Population Epidemiology of Sleep (TAPES) [4], including a nationwide representative sample of 5021 adults, the OSA prevalence in Turkey is around 14%, based on the Berlin questionnaire [5], and excessive daytime sleepiness (EDS) is 5.4% based on the Epworth Sleepiness Scale (ESS) scores [6, 7].

During the last decades, several attempts have been intensively debated regarding the development of national and international strategies to screen high-risk drivers for sleepiness and OSA. The increasing awareness of OSA as a risk factor for MVAs, which is reversed by effective treatment with continuous positive airway pressure (CPAP), has led to a revision of annex III of the European Union (EU) directive on driving licenses that is subject to mandatory implementation by all member states from December 31, 2015 [8]. This directive was based on the recommendations from a working group established by the Transport and Mobility Directorate of the European Commission in 2012 [9]. Accordingly, Turkish legislation has also been revised for regulation of driving license issues for individuals with OSA. For instance, a full-night polysomnography (PSG) is recommended and treatment should be initiated as soon as possible when the OSA diagnosis is confirmed [10]. However, given the high prevalence of OSA in general population and the limited resources for PSG, there are many controversies and difficulties in implementation of these rules in practice.

In this context, we aimed to use a questionnaire that was developed by the European Union Driver License Committee as a screening tool for OSA (Supplement 1) [9]. We named the questionnaire the European Obstructive Sleep Apnea Screening (EUROSAS) questionnaire for drivers and sought to test the reliability of its Turkish version.

As shown in Table 1, most of the participants were older than 30 years, non-obese, and were married. Few had comorbidities. Almost half of the entire study population were university graduates with the highest proportion (90%) among the NPFDs, and the lowest (10.0%) among the PMDs (p < 0.001).

The normality assumption of the total EUROSAS scores was violated (W = 0.94, p < 0.001 for the first test, and W = 0.92, p < 0.001 for the retest, respectively). As shown in Fig. 1, the female drivers had significantly lower median scores than the male drivers; in the first test 8.0 (IQR 6.8–12.0) in PMD, 8.0 (IQR 6.0–11.0) in NPMD, and 5.0 (IQR 4.0–8.0) in NPFD (p < 0.001). Corresponding values in the retest were 9.5 (IQR 7.0–13.0), 8.0 (IQR 6.0–10.0), and 5.0 (IQR 4.0–8.0), respectively (p < 0.001).

As demonstrated in Fig. 2, the proportion of participants with EUROSAS score ≥ 10 increased from 34% in the first test to 50% in the retest among the PMD (ns). The corresponding values were 34% vs 24% in NPMD (ns), and 8% vs 16% in NPFD (ns). There was a strong positive correlation between two measurements (r = 0.864; p < 0.001), indicating a good reproducibility of the EUROSAS scores (Fig. 3). Cronbach’s alpha coefficients of the EUROSAS for the entire population were calculated as 0.463 for the first test, and 0.562 for the retest.

As presented in Table 2, range of mean item scores was relatively wide on both measurements. The highest mean score was yielded on the item 1 (gender), which was calculated as 1.67. The lowest mean score was yielded on the item 8 (Have you been told that your breathing stops during sleep?) which was 0.09 for the first test and 0.11 for the retest. Likewise, the range of the item total correlation coefficients was also considerably broad (Table 2). The lowest item total correlation was yielded on the item 10 (Do you suffer arterial hypertension?), which was 0.031 for the first test, and 0.048 for the retest, respectively. The item 11 (ESS) showed the highest item total correlation, 0.346 on the first test, and 0.491 on the retest, respectively (Table 2).

The main conclusion of the current study is that the EUROSAS questionnaire is easy to understand and is applicable with good reproducibility among Turkish professional and non-professional drivers. However, the test-retest reliability level seems to be poor.

To our knowledge, this is the first reliability study of a screening questionnaire developed by the EU Driver License Committee, and the first definition of the acronym EUROSAS in Europe. The increasing consciousness about OSA as a risk factor for MVAs [2, 3], and the observed beneficial effect of CPAP treatment in reducing this risk in retrospective studies [11] has led to a revision of annex III of the EU directive on driving licenses that is a mandatory document for implementation by all member states from December 31, 2015 [8, 9]. As summarized in an editorial [12], this directive includes: “(a) applicants or drivers in whom a moderate or severe OSA syndrome is suspected shall be referred for further authorized medical advice before a driving license is issued or renewed. They may be advised not to drive until confirmation of the diagnosis; (b) driving licenses may be issued to applicants or drivers with moderate or severe OSA syndrome who show adequate control of their condition and compliance with appropriate treatment and improvement of sleepiness, if any, confirmed by authorized medical opinion; (c) applicants or drivers with moderate or severe OSA syndrome under treatment shall be subject to a periodic medical review, at intervals not exceeding 3 years for noncommercial drivers, and 1 year for commercial drivers, with a view to establish the level of compliance with the treatment, the need for continuing the treatment and continued good vigilance.” [12].

In accordance with the aforementioned EU directive [8, 9], the Turkish legislation has been revised as following [10]: (a) subjects with severe OSA (AHI > 30/h), and subjects with moderate OSA (AHI 15–30/h) with documented daytime sleepiness should not be given driving license unless this condition is treated; (b) the efficient treatment should be documented to include the severity of OSA at baseline, response to treatment, compliance with CPAP; (c) the applicants with BMI > 33 kg/m² should undergo full-night PSG regardless of symptoms; (d) the applicants with witnessed apneas and daytime sleepiness should undergo full-night PSG regardless of BMI. Given that access to OSA diagnosis and treatment is already limited for symptomatic patients, the current legislation has led to an even higher number of requests for specialist evaluation for drivers with BMI >33kg/m² regardless of symptoms, and lengthened the waiting lists for PSG. Thus, screening for OSA by using a full-night PSG on a large scale is not feasible in Turkey. On the other hand, though OSA increases the risk of traffic accidents, the disorder is associated with EDS in only one-fifth of the patients according to populationstudies [13]. The majority of evidence support that driving risk in OSA is closely related to the degree of daytime sleepiness rather than the severity of OSA in terms of AHI [14, 15]. There is yet no consensus on the definition of “a high-risk driver” other than a report from the American Thoracic Society [16]. The report suggests that a high-risk driver is the one who has moderate to severe daytime sleepiness, and a recent unintended MVA or a near-miss attributable to the sleepiness, fatigue, orinattention [16].

In the medical literature, there have been several screening questionnaires for OSA in general populations and sleep clinic cohorts [5, 6]. The current EUROSAS questionnaire for drivers may be a good preliminary screening tool to identify the high-risk drivers, especially in countries that have less access to full polysomnography for diagnosing OSA, and consequently initiating CPAP treatment. As expected, the probability of OSA was less common among the female drivers. In males, the risk of having OSA showed an increasing pattern among the professional drivers while there was a decreasing trend among the non-professionals drivers. Though this might be explained by the regression toward the mean, it might also reflect the initial legal concerns of the professional drivers. Nevertheless, there was a strong correlation between the first and second measurements in the entire cohort, suggesting a good reproducibility of the EUROSAS questionnaire. Of note, the threshold of 10 is arbitrary, and requires further validation.

As reflected by the Cronbach’s alpha coefficients, the reliability of the EUROSAS questionnaire was poor in the current study population. Expectedly, the highest mean score was found in item 1 (gender), and the lowest mean score was found in item 8 (Have you been told your breathing stops during your sleep?) while the item 11 (ESS) showed the highest item total correlation (Table 2). The lowest item total correlation was found in item 10 (Do you suffer arterial hypertension?), which might be related to the self-reported hypertension diagnosis, and may not be sufficiently reliable for evaluation of the EUROSAS scores. Thus, given the broad range of the item scores both in the first test and the retest, our results should be interpreted cautiously, and regarded as a pilot evaluation of the EUROSAS questionnaire for drivers.

We conclude that the EUROSAS–Turkish version is easy to understand and is reproducible among the drivers in Turkey. However, the test-retest reliability level is poor, and the probability of having OSA seems to be underestimated at the first test among the PMDs. This underestimation may reflect the drivers’ concerns for legal issues for losing their driving licenses. Further validation of the questionnaire by objective sleep studies and fitness-to-drive testing is required.