The study involved Italian patients consulting our sleep center for respiratory sleep disorders. This study was conducted in accordance with the amended Declaration of Helsinki. The Institutional Ethical Committee, Azienda Ospedaliera Universitaria Policlinico “Paolo Giaccone”, Palermo 1, approved the protocol. All patients gave their written informed consent for personal data processing.

Recruitment was conducted between January 2015 and June 2019. The study recruited 2622 patients (male, 2011; female, 611). Italian-speaking patients referred to our sleep clinic because of OSA suspicion and at least 18 years of age were eligible for the study. Patients with uncontrolled psychiatric illness, neurocognitive impairment, usage of sedative or hypnotic medications, sleep time test duration shorter than 6 h, or “absence of answer” answer to at least one of the SANReSP items were not included in the study. The minimum sample size required for a screening study, with the null hypothesis for sensitivity and specificity at 0.5, a target significance level of 0.05, and a power of 80%, ranges from 22 (prevalence = 90%) to 980 (prevalence = 5%) [10]. The prevalence of OSA in patients at a sleep clinic is estimated to be approximately 80%. Thus, the study’s sample size of 2622 patients was consistent.

A portable, computerised system type III (Embla, Natus Inc., Middleton, USA) was employed to perform nocturnal monitoring. The sleep studies were performed at the sleep lab and at patients’ home. A sleep technician has set sleep studies device. The following signals were recorded: airflow by nasal cannula pressure, snoring, thoracic and abdominal efforts, body position, limb movements, arterial oxygen saturation, pulse rate, and pulse waveform. The recording duration was at least 6 h [11].

Type III Sleep Studies uses portable monitors that allow sleep studies to be done at the patient’s home or elsewhere. This option was introduced as a more accessible and less expensive alternative to in-laboratory polysomnography. Unlike Type I, Type III testing cannot measure the duration of sleep, the number of arousals or sleep stages, nor can it detect non respiratory sleep disorders. The American Academy of Sleep Medicine and Canadian Sleep Society/Canadian Thoracic Society guidelines recommend that portable sleep studies be provided under the direction of health professionals with accreditation in sleep medicine and as part of a comprehensive assessment [11].

Apneas and hypopneas were visually scored and OSA severity was defined according to the American Academy of Sleep Medicine 2012 standard criteria [12]. Apneas were identified on the airflow signal and were classified as obstructive, central, or mixed according to the behavior of thoraco-abdominal movements. Hypopneas were defined as discernible reductions in airflow or thoraco-abdominal movements for at least 10 s, followed by an arterial oxygen saturation fall of ≥3%. The AHI was calculated as the number of apneas and hypopneas per hour of estimated total sleep time. The percentage of the night with O₂ saturation <90% (TSat₉₀) was assessed. A sleep medicine expert, blinded to the questionnaire, scored all sleep studies.

To keep the questionnaire concise, easy-to-use, and quick to fill out, the questions were designed in a yes/no format. Based on a literature review, questions were designed using the most common clinical domains of OSA and hypertension. The SANReSP questionnaire includes six dichotomous (yes/no) questions: S: Do they tell you that you Snore? [Le dicono che russa?], A: Do they tell you that sometimes you stop breathing or have sleep Apnea? [Le dicono che talvolta smette di respirare o ha apnee durante il sonno?], N: Do you wake up during the night with an urge to urinate ‘Nocturia’? [Si sveglia durante la notte con il bisogno urgente di urinare?], Re: Do you sometimes feel unsatisfied with how you slept ‘Rest’? [Le capita di non essere soddisfatto/a di come ha dormito?], S: Do you frequently feel the desire or need to sleep during the day except after lunch ‘Sleepy’? [Sente frequentemente il desiderio o il bisogno di dormire durante il giorno eccetto dopo pranzo?], and P: Do you take medications for high blood Pressure? [Assume farmaci per la pressione arteriosa alta?] (Fig 1). SANReSP is a self-administered questionnaire. The score ranges from 0 to 6, attributing one point for each affirmative response and 0 is assigned when no answer is given.

Statistical analyses were performed using the commercial 2020 MedCalc statistical software (MedCalc Software Ltd., Ostend, Belgium). The Kolmogorov–Smirnov test was used for evaluating the accordance of the parameters with a normal distribution. Normally distributed continuous variables were reported as means and standard deviations. Non-normally distributed continuous variables were reported as median and 25–75% interquartile range. Categorical variables were reported as frequencies. The non-parametric Wilcoxon’s test was used to assess the differences. Comparisons with a p-value of ≤0.05 were considered statistically significant.

The OSA probability determined by SANReSP was compared to the diagnostic sleep study results. Receiver operating characteristic (ROC) curve analysis was performed. The specificity, sensitivity, positive predictive value (PPV), negative predictive value (NPV), likelihood ratio (LR) for a positive test result (LR+), LR for a negative test result (LR-), accuracy, and odds ratio (LR+/LR-) (OR) were calculated. All estimates were reported with their respective 95% confidence intervals (CIs). The CIs for sensitivity, specificity, and accuracy were ‘exact’ Clopper–Pearson CIs. The CIs for the LRs was calculated using the ‘Log method’ of Altman et al. 2000 (as given on page 109) [13]. The CIs for the predictive values were the standard logit CIs given by Mercaldo et al. 2007 [14]. The CIs for OR were calculated according to Altman DG 1991 [15].

The LR+ and LR- indicate if a patient is more or less likely to have a disease, respectively. Since LRs are not affected by disease prevalence, they might be more useful parameters than sensitivity, specificity, and predictive values. The higher the LR+ (>1), the stronger is the prediction of a disease. Conversely, the lower the LR- (<0.5), the better is the prediction of the absence of disease.

The optimal cut-off that maximises true positives and minimises false positives was determined, and it corresponds to the highest sensitivity + specificity. The response was dichotomised by AHI ≥ 5/h for all OSA patients, AHI ≥ 15/h for moderate–severe OSA patients, and AHI ≥ 30/h for only severe OSA patients. The predictive probability (post-test probability) was calculated using logistic regression. The statistical analyses were conducted separately for men and women. The statistical significance of the difference between the areas under ROC curves (AUCs) of men and women was evaluated using the DeLong method [16].

Descriptive statistics of the analysed sample are presented in Table 1. Body mass index (BMI) and age were greater in women than in men, while neck circumference and AHI severity were greater in men than in women. Of the 2622 patients, 265 (10.1%) patients did not have OSA, 2357 (89.9%) had AHI ≥ 5/h, 1801 (68.7%) had AHI ≥ 15/h, and 1264 (48.2%) had AHI ≥ 30/h.

The median SANReSP score was 4 for men and 5 for women. In Table 2, we can see the positive responses for SANReSP, from the total population and between males and females. Fig 2 shows the distribution of the SANReSP scores for men and women. A score of 5 was most frequent in both sexes. Figs 3 and 4 show the ROC curves for all patients and men/women, respectively, at AHI severity levels of ≥ 5/h, ≥ 15/h, and ≥ 30/h. The comparison between the ROC curves of men and women were statistically different at AHI ≥ 15/h. Additionally, the AUCs for women were larger than that of men.

Evaluation of the questionnaire’s validity for AHI severity (≥ 5/h, ≥ 15/h, ≥ 30/h) is presented in Table 3. For AHI ≥ 5/h, the optimal SANReSP cut-off was >3 (sensitivity, 74.76%; specificity, 67.92%; accuracy, 74.07%; OR, 6.30; AUC, 0.76). For moderate–severe OSA (AHI ≥ 15/h), the optimal cut-off was >3 (sensitivity, 78.18%; specificity, 46.53%; accuracy, 68.27%; OR, 3.11; AUC, 0.66). For severe OSA (AHI ≥ 30/h), the optimal cut-off was >4 (sensitivity, 59.10%; specificity 64.73%; accuracy, 62.01%; OR, 2.67; AUC, 0.65).

Fig 5 shows the relationship between the predictive probability of sleep apnea (AHI ≥ 5/h) versus the SANReSP questionnaire score; the dashed line indicates the 95% CIs. The probability of OSA increased as the SANReSP score increased, with the same trend in both women and men. The probability of AHI ≥ 5/h was 86.8% in men and 68.7% in women for a score of 3, while it was 98.7% in men and 97.9% in women for a score of 6.

The SANReSP questionnaire is a low-cost, easy, short, and self-administered questionnaire, which exclusively consists of clinical parameters and does not require patient measurements. Thus, it is excellent for screening patients with suspected OSA. The sample size of the study is among the largest sample sizes used to investigate the predictivity of OSA screening questionnaires. A limitation of the study is the use of home polygraphic studies for diagnostics rather than complete laboratory-based polysomnography. PSG is widely accepted as the gold standard test for diagnosis of OSA. However, within the appropriate context, the use of Level 3 sleep study is considered an acceptable method for diagnosing OSA by the American Academy of Sleep Medicine [27]. As reported by AASM Rules, Level 3 sleep study were performed after clinical evaluation. Additionally, sleep assessment was performed under the supervision of certified sleep medicine physician. Furthermore, the study has a selection bias since the results apply to a sleep clinic population, i.e. patients with a high probability of pre-test disease, which is not representative of the general adult population. Additionally, the sensitivity of this tool is moderate and some of the patients referred to sleep clinic can be missed. Furthermore, the questionnaire development was based on a literature review and questions were related to the most common clinical domain of OSA. No factor analysis was performed.