Development and Validation of the Seoul Sleep Questionnaire in a Tertiary Sleep Clinic Population
Article information
Abstract
Background and Objective
Sleep disorders are common and frequently under-recognized in clinical practice. Existing screening instruments are often disorder-specific or require lengthy assessments. This study aimed to develop and validate the Seoul Sleep Questionnaire (SSQ), a brief screening instrument for common sleep disorders.
Methods
We retrospectively reviewed data from 1,726 adults who underwent polysomnography at a tertiary sleep clinic between November 2018 and February 2020. Participants completed the SSQ and established sleep questionnaires. Item reduction, exploratory factor analysis (EFA), reliability testing, validity analyses, and receiver operating characteristic (ROC) analyses were performed. To address diagnostic imbalance, sensitivity analyses were conducted using a diagnosis-balanced subsample.
Results
Item reduction yielded a final 7-item questionnaire. EFA supported a three-factor structure representing obstructive sleep apnea (OSA), insomnia, and parasomnia domains, with good model fit (comparative fit index=0.999, Tucker–Lewis index=0.997, root mean square error of approximation=0.013). Internal consistency was acceptable (McDonald’s ω=0.83). SSQ scores showed significant correlations with subjective and polysomnography-derived sleep measures. The OSA subscale was associated with apnea-hypopnea index (r=0.429, p<0.01), while the insomnia subscale correlated with sleep onset latency and wake after sleep onset. ROC analyses demonstrated acceptable diagnostic performance for OSA (area under the curve [AUC]=0.787) and insomnia (AUC=0.741), and high discriminative ability for parasomnia (AUC=0.919). Comparable results were observed in the diagnosis-balanced subsample.
Conclusions
The SSQ demonstrated acceptable psychometric properties and diagnostic performance as a brief screening questionnaire in an OSA-dominant tertiary sleep clinic population. Further validation in primary care, community-based, and independent external populations is warranted before broader implementation.
INTRODUCTION
Sleep disorders are highly prevalent worldwide, affecting approximately 7%–10% of the general population in the United States [1] and 14.4% in Korea, a prevalence that has nearly doubled over the past decade [2]. Among the most common sleep disorders, obstructive sleep apnea (OSA) and insomnia affect up to 17%–27% [3] and 10%–12% [2] of the general population in Korea, respectively.
Disruption of nocturnal sleep due to various sleep disorders can lead to excessive daytime sleepiness, fatigue, and daytime dysfunction, resulting in impaired quality of life and an increased risk of accidents [4]. Moreover, sleep disorders are associated with a wide range of adverse health outcomes, including impaired immune function, dysregulated metabolic and sympathetic nervous system activity, increased cardiovascular risk, poor mental health, and cognitive impairment [5,6]. Thus, early identification and appropriate referral of individuals with possible sleep disorders are important for optimal patient care. However, quick, easy, and accurate systematic screening tools that can be administered by non-experts are lacking. Consequently, many primary care physicians may overlook routine screening for sleep disorders [7]. Sleep apnea syndrome, one of the most common sleep disorders, is frequently under-diagnosed. According to a previous study, more than 80% of middle-aged men and 93% of women with moderate-to-severe OSA remain undiagnosed [8]. Similarly, many individuals with insomnia do not discuss their sleep concerns with healthcare providers; in a large survey, 30% of respondents reported that their primary care physician never inquired about sleep, suggesting under-recognition of sleep complaints in clinical settings [9].
To date several screening tools have been developed to identify sleep disorders. The most widely used tools for general assessment of sleep difficulties are the Pittsburgh Sleep Quality Index (PSQI), Insomnia Severity Index (ISI), and the Epworth Sleepiness Scale (ESS). However, these tools have notable limitations, including the need for trained professionals to administer and interpret them, and limited specificity or sensitivity for certain sleep disorders [10–12]. The Sleep Disorders Questionnaire (SDQ) was developed to identify OSA, narcolepsy, psychiatric sleep disorder and periodic limb movement disorder (PLMD) [13]. However, the SDQ requires a long reference period of six months and consists of 175 items, making it impractical for screening purposes [14]. Another alternative, the Auckland Sleep Questionnaire comprises 30 International Classification of Sleep Disorders (ICSD)-based items but does not specify a timeframe of interest [10]. The Iowa Sleep Disturbances Inventory is based on ICSD-2 criteria and lacks a defined assessment timeframe [15]. The Global Sleep Assessment questionnaire [14] is relatively brief, consisting of 11 items, but its reliability and construct validity have not been sufficiently evaluated against other subjective measures or objective sleep assessments [12].
In this study, we aimed to develop and validate a practical screening questionnaire designed to detect sleep disorders that can be easily administered in clinical settings. We evaluated the internal consistency and factor structure of the questionnaire and tested it against polysomnography (PSG), the gold standard for diagnosing sleep disorders. In addition, convergent and concurrent validity were assessed by comparison with established sleep-related questionnaires.
METHODS
We conducted our study in accordance with the scale development and validation process proposed by Boateng et al. [16]. The first phase is item development, which includes item generation and assessment of content validity. The second phase is scale development, during which pre-testing of the survey, item reduction, and factor analysis are performed. The final phase is scale evaluation, which assesses the reliability and validity of the scale.
Development of the Seoul Sleep Questionnaire
A group of sleep experts developed the Seoul Sleep Questionnaire (SSQ) items based on frequently asked questions commonly used to assess patients with sleep disorders in clinical settings. This multidisciplinary team consisted of two physicians specialized in sleep medicine, one nurse specialist, and one neuropsychologist. The group reached a consensus on 10 dichotomous (yes/no) items designed to assess sleep-related symptoms. To test these items, a pilot study was conducted with 50 patients who visited a university hospital-based sleep clinic (Samsung Medical Center) located in Seoul, Korea. Participants were asked to complete the questionnaire and to indicate whether any items were ambiguous or difficult to understand upon completion. Based on feedback from the pilot study, the expert group reviewed, revised, and refined the items. The final SSQ consisted of the following questions: 1) Do you snore? Have you ever heard yourself snore?; 2) Has anyone else observed that you display shortness of breath while sleeping?; 3) Do you wake up more than twice per night to urinate?; 4) Do you have difficulties falling asleep?; 5) Do you wake up in the middle of the night?; 6) Do you talk in your sleep?; 7) Do you act out your dreams?; 8) Do you take more naps during the day?; 9) Do you wake up due to cramp?; and 10) Do you think you have a sleeping problem?
Participants and Procedure
Data of the participants who had undergone overnight PSG between November 2018 and February 2020 were extracted from the database of a university-affiliated sleep clinic (Samsung Medical Center). Exclusion criteria included: 1) the presence of serious medical or neurological conditions, circadian rhythm sleep disorders, narcolepsy, idiopathic hypersomnia, and other minor sleep disorders; 2) non-Korean-speaking individuals; 3) shift workers; and 4) participants who underwent split-night PSG or had insufficient self-reported questionnaire data. Among the 1,842 patients who underwent PSG during the study period, 110 were excluded because they presented with serious medical or neurological conditions (n=17), circadian rhythm disorders (n=14), narcolepsy or idiopathic hypersomnia (n=40), other minor sleep disorders (n=5), or had incomplete or missing data (n=22). Additionally, non-Korean-speaking individuals (n=5), and shift workers (n=7) were excluded. Ultimately, 1,732 participants were included in the final analysis. Following outlier detection using boxplot analysis, six cases were excluded from analyses involving SSQ total scores, resulting in an analytic sample of 1,726 participants for psychometric analyses. In accordance with the ICSD-3 [17] and the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, Text Revision (DSM-5-TR) [18], participants were classified into five diagnostic groups: OSA, insomnia, PLMD, REM behavior disorder (RBD), and healthy controls without diagnosed sleep disorders. To reduce potential bias associated with unequal group sizes and to assess the robustness of analyses most affected by sample imbalance, a balanced validation sample (n=392) was generated using stratified random sampling. Given the substantial overrepresentation of the OSA group in the original sample, random subsampling was performed to match the sample sizes of other groups. Specifically, 80 participants were randomly selected from the OSA, insomnia, and healthy control groups, while all available cases were retained for the smaller groups (RBD, n=66; PLMD, n=86). This subsample was used for factor analysis and receiver operating characteristic (ROC) analysis as sensitivity analyses. All participants underwent level 1, attended overnight PSG and completed self-reported questionnaires, including SSQ, prior to PSG assessment.
Measures
Clinical information
Clinical information included age, sex, history of hypertension, use of sleep medication, and habitual sleep–wake patterns, including time in bed, sleep onset latency (SOL), time out of bed, and total sleep duration.
Seoul Sleep Questionnaire
The SSQ consists of 10 items designed to assess sleep-related symptoms associated with major sleep disorders. Participants responded using a dichotomous format (yes=1, no=0). Total scores ranged from 0 to 10, with higher scores indicating a greater likelihood of clinically significant sleep disturbances.
Insomnia Severity Index
The ISI was used to assess subjective insomnia severity during the preceding two weeks [19]. The questionnaire consists of seven items rated on a 5-point Likert scale, yielding a total score ranging from 0 to 28, with higher scores indicating greater insomnia severity. The validated Korean version developed by Cho et al. [20] was used. In the present sample, internal consistency was excellent (Cronbach’s α=0.92).
Epworth Sleepiness Scale
The ESS is an 8-item questionnaire assessing subjective day-time sleepiness across various situations. The validated Korean version was used, with each item rated on a 4-point Likert scale [21]. Total scores range from 0 to 24, with higher scores indicating greater daytime sleepiness. Internal consistency in this sample was high (Cronbach’s α=0.90).
Objective measures: PSG
Overnight PSG was conducted using the Embla N7000 system (Medcare-Embla). Recordings included six electroencephalography channels (C3-A2, C4-A1, F3-A2, F4-A1, O3-A2, and O2-A1), electrooculography, chin and bilateral anterior tibialis electromyography, and electrocardiography. Thoracic and abdominal respiratory movements were monitored using plethysmography belts, and airflow was measured using a nasal pressure transducer and thermistor. Oxygen saturation was continuously monitored via pulse oximetry, and synchronized video monitoring was used to document abnormal behaviors during sleep.
PSG data were scored by experienced registered technologists in accordance with the American Academy of Sleep Medicine (AASM) scoring criteria. Measured parameters included total sleep time (TST), SOL, wake after sleep onset (WASO), sleep efficiency (SE), and apnea–hypopnea index (AHI). All recordings were obtained during a single-night PSG study.
Data Analysis
Data analyses were performed using SPSS (version 21.0; IBM Corp.; https://www.ibm.com/products/spss-statistics), Mplus 7.11 (Muthén & Muthén; https://www.statmodel.com), Factor 10.4 (https://psico.fcep.urv.cat/utilitats/factor/), and R 4.3.2 (The R Foundation for Statistical Computing; https://www.rproject.org). Statistical significance was defined as p<0.05.
Outliers were identified using boxplot analysis. Following Tukey’s interquartile range (IQR) method [22], values exceeding Q3+1.5×IQR or below Q1–1.5×IQR were classified as outliers. Outlier screening was applied to SSQ total scores only. Descriptive statistics, chi-square tests, and one-way analysis of variance were used to examine the demographic data. Tetrachoric correlations were calculated for dichotomous SSQ items, and squared multiple correlations (SMCs) were used to assess item adequacy and potential redundancy among items. Sampling adequacy for factor analysis was evaluated using Bartlett’s test of sphericity and the Kaiser–Meyer–Olkin (KMO) measure. To assess the robustness and stability of the factor structure and diagnostic performance in the presence of imbalanced group sizes, exploratory factor analysis (EFA) and ROC analyses were performed in both the full sample (n=1,732) and a stratified balanced subsample (n=392). The latent structure of the SSQ was examined using EFA with a mean- and variance-adjusted weighted least squares estimator (WLSMV), the default estimator in Mplus for ordinal items. Factors were rotated using Geomin oblique rotation, allowing correlations among factors. Model fit was evaluated using the comparative fit index (CFI), Tucker–Lewis index (TLI), and root mean square error of approximation (RMSEA). Values of CFI and TLI ≥0.90 and RMSEA ≤0.08 were considered indicative of acceptable model fit [23]. Internal consistency was assessed using McDonald’s omega (ω), with values ≥0.70 considered acceptable [24]. Convergent and concurrent validity were evaluated using Pearson correlation analyses between SSQ scores, ISI scores, ESS scores, and PSG parameters. Periodic limb movement index was not included in the concurrent validity analyses because it is a marker of PLMD rather than a direct objective indicator of parasomnia. ROC curve analysis was performed to assess diagnostic accuracy using bootstrap resampling (1,000 iterations). The area under the curve (AUC), standard errors, and 95% confidence intervals (CIs) were calculated using the fbroc package in R. AUC values were interpreted as acceptable (0.7–0.8), excellent (0.8–0.9), or outstanding (≥0.9) [25].
RESULTS
Demographic Characteristics
A total of 1,726 individuals were included in the study (mean age=53.0±14.6 years, 72.8% male). Participants were categorized into the following groups according to their diagnosis: OSA (n= 1,357, 78.6%), insomnia (n=137, 7.9%), PLMD (n=86, 5.0%), RBD (n=66, 3.8%), and healthy controls (n=80, 4.6%). Patients with RBD were significantly older than participants in the other diagnostic groups (F=57.59, p<0.001), and the proportion of men was significantly higher in the OSA group compared with other groups (χ2=217.38, p<0.001). The demographic characteristics of the participants are shown in Table 1.
Item Analysis
Tetrachoric correlations between the items ranged from −0.323 to 0.741. The SMCs were ≥0.3 for seven of the 10 items. Three items—Item 3 (“Do you wake up more than twice per night to urinate?”), Item 8 (“Do you take more naps during the day?”), and Item 9 (“Do you wake up due to cramp?”)—showed SMC values below 0.3. Two items were deleted due to low SMCs and weak associations with sleep disorders, whereas Item 9 was retained based on its clinical relevance. An inter-item correlation coefficient ≥0.3 was considered acceptable [16]. Bartlett spherical test was significant (χ2=6,348.3, df=45, p<0.001), and the KMO measure of sampling adequacy was 0.65, indicating that the data were suitable for EFA.
To assess robustness, a balanced dataset (n=392) was additionally analyzed. Tetrachoric correlations ranged from −0.246 to 0.857. In this dataset, nine of 10 items met the SMC ≥0.30 criterion, with only Item 8 falling below the threshold. Although Items 3 and 9 met the statistical criterion in this subset, item selection was based on the full dataset to ensure methodological consistency. Consequently, Items 3 and 8 were removed, and Item 9 was retained. The Bartlett’s test remained highly significant (χ2=2,068.1, df=45, p<0.001), and the KMO value was 0.53, exceeding the minimum acceptable threshold (0.50), supporting the use of EFA.
Descriptive Statistics and Internal Consistency
The final version of the SSQ consisted of seven items. Descriptive statistics and McDonald’s omega coefficients for the SSQ are presented in Table 2. The mean SSQ total score was 2.88 (standard deviation=1.44). Internal consistency of the total scale, assessed using McDonald’s omega (ω), was 0.83. At the factor level, omega coefficients were 0.85 for Factor 1, 0.69 for Factor 2, and 0.85 for Factor 3. These values indicate acceptable to good internal consistency, as coefficients of 0.70 or higher are commonly considered acceptable for research purposes [24,26].
Factor Structure
Eigenvalues, parallel analysis, and the scree plot supported a three-factor solution. The eigenvalues for the first four factors were 2.58, 2.05, 1.42, and 0.84, respectively. The three-factor model demonstrated good model fit (χ2=9.18, df=7, p=0.240; CFI= 0.999; TLI=0.997; RMSEA=0.013).
Item 10 (“Do you think you have a sleeping problem?”) showed a strong primary loading on Factor 2 (0.821) but also a notable cross-loading on Factor 1 (0.315). Given its broad and global nature as a general indicator rather than a symptom-specific item, Item 10 was excluded to improve discriminant validity and model parsimony.
Replication using the balanced dataset (n=392) confirmed the three-factor structure. Parallel analysis supported the same dimensionality, and the rotated factor matrix showed clean loadings without problematic cross-loadings (Bentler’s Simplicity Index, S=0.992).
Additional EFA using WLSMV estimation with Geomin rotation also supported the three factor model. Compared with the 1-factor (χ2=493.62, df=20, p<0.001; CFI=0.559; TLI=0.383; RMSEA=0.25) and 2-factor models (χ2=181.29, df=13, p<0.001; CFI=0.843; TLI=0.663; RMSEA=0.18), the 3-factor model demonstrated superior fit (χ2=16.53, df=7, p<0.05; CFI=0.991; TLI= 0.965; RMSEA=0.06; standardized root mean square residual= 0.03). Detailed factor loadings are presented in Tables 3 and 4. The three factors were labeled 1) OSA, 2) insomnia, and 3) parasomnias. The OSA factor included items related to snoring and witnessed breathing difficulties during sleep. The insomnia factor comprised of items reflecting difficulties in sleep initiation, sleep maintenance and sleep quality. The parasomnia factor included items assessing abnormal behaviors during sleep, such as sleep talking and dream enactment.
Convergent Validity
SSQ total and subscales were significantly correlated with established self-reported measures (Table 5). SSQ total scores were positively correlated with the ISI (r=0.389, p<0.01) and the ESS (r=0.143, p<0.01). The OSA subscale was significantly correlated with ESS (r=0.201, p<0.01), habitual SOL (r=−0.186, p<0.01), and sleep duration (r=0.120, p<0.01). The insomnia subscale showed significant correlations with ISI (r=0.590, p<0.01), habitual SOL (r=0.471, p<0.01), and sleep duration (r=−0.257, p< 0.01) were statistically significant. The parasomnia subscale was also significantly correlated with the ISI (r=0.082, p<0.01), and sleep duration (r=0.062, p<0.05) were significantly correlated.
Concurrent Validity
Significant correlations were observed between the SSQ factors and objective PSG parameters (Table 5). The SSQ total scores were significantly correlated with TST (r=−0.074, p<0.01), WASO (r=0.069, p<0.01), SE (r=−0.075, p<0.01), and AHI (r=0.181, p<0.01). The OSA subscale showed a significant positive correlation with AHI (r=0.429, p<0.01). The insomnia subscale was significantly correlated with SOL (r=0.150, p<0.01), WASO (r= 0.136, p<0.01), TST (r=−0.078, p<0.01), and SE (r=−0.167, p<0.01).
Diagnostic Accuracy
ROC analysis was performed to assess the ability of the SSQ subscales to discriminate between individuals with and without the corresponding sleep disorders (Fig. 1). In the original dataset, the OSA subscale showed acceptable performance (AUC= 0.787, 95% CI [0.76–0.81], balanced accuracy=0.742), with a high positive predictive value (PPV=0.944). The insomnia subscale showed comparatively lower performance, with an AUC of 0.741 (95% CI [0.71–0.78]) and a balanced accuracy of 0.714. The PPV for the insomnia subscale was low (PPV=0.166). The parasomnia subscale demonstrated high discrimination (AUC= 0.919, 95% CI [0.89–0.95]; balanced accuracy=0.864), with a PPV of 0.298. However, this estimate should be interpreted cautiously because it was derived from a two-item factor in a relatively small parasomnia subgroup.
ROC curves for the revised 7-item SSQ subscales (OSA, insomnia, and parasomnia). A: Original (full sample). B: Balanced sample (diagnosis-matched). ROC, receiver operating characteristic; SSQ, Seoul Sleep Questionnaire; OSA, obstructive sleep apnea; AUC, area under the curve; RBD, REM behavior disorder.
In the balanced dataset (n=392), the OSA subscale maintained stable performance (AUC=0.754, 95% CI [0.69–0.81], balanced accuracy=0.726). The insomnia subscale showed improved PPV (0.361; AUC=0.722, 95% CI [0.67–0.77], balanced accuracy=0.708), while the parasomnia subscale demonstrated consistently high accuracy (AUC=0.928, 95% CI [0.90–0.96], balanced accuracy=0.869, PPV=0.716).
DISCUSSION
This study involved the development and validation of the SSQ as a brief screening instrument for common sleep disorders in a tertiary sleep clinic population. The findings demonstrated that the SSQ comprises a three-factor structure encompassing OSA, insomnia, and parasomnia domains, and that the questionnaire shows acceptable convergent validity with established self-reported measures as well as concurrent validity with objective PSG parameters.
Psychometric Properties of the SSQ
The finalized 7-item SSQ demonstrated satisfactory internal consistency (McDonald’s ω=0.83). This value exceeds the commonly accepted threshold of 0.70 for research instruments [26], supporting the reliability of the scale. However, given the substantial class imbalance, this estimate should be interpreted as preliminary. Further validation in independent external cohorts is warranted to establish the stability and generalizability of the scale.
EFA revealed that the SSQ displays a three-factor structure corresponding to OSA, insomnia, and parasomnia. The first factor, labeled “OSA,” comprised items addressing snoring and witnessed apneas, which are among the most readily observable and frequently reported symptoms of OSA by patients or bed partners [27,28]. The second factor, “insomnia,” included items probing difficulties with sleep initiation, sleep maintenance, and nocturnal awakenings. These symptoms are consistent with diagnostic frameworks for insomnia disorder, including ICSD-3 criteria, which emphasize difficulties initiating or maintaining sleep accompanied by daytime consequences such as fatigue or sleepiness [28,29]. The third factor, “parasomnia,” consisted of items related to sleep talking and dream enactment behaviors. These symptoms may occur in both non-rapid eye movement and REM sleep parasomnias and represent clinically meaningful behavioral manifestations during sleep [28]. Notably, replication of the three-factor structure in the diagnosis-balanced subsample further supports the robustness of the factor solution despite the substantial predominance of OSA cases in the original sample. This finding addresses concerns that the factor structure may have been driven primarily by OSA-related variance.
The SSQ demonstrated significant correlations with independent self-reported measures of insomnia severity (ISI), excessive daytime sleepiness (ESS), and sleep habits, supporting its convergent validity. In particular, the OSA subscale was associated with ESS scores, consistent with the excessive daytime sleepiness commonly observed in patients with OSA. Recurrent hypoxemia and frequent arousals caused by complete or partial upper-airway obstruction result in sleep fragmentation and reduction of deep sleep, leading to excessive daytime sleepiness in OSA patients [30]. Similarly, the insomnia subscale showed strong correlations with ISI scores and habitual SOL. These findings are consistent with the clinical features of insomnia, which include increased insomnia severity and prolonged sleep latency.
In addition to subjective measures, the SSQ correlated with PSG parameters, indicating that the questionnaire reflects objective measures of sleep. The OSA subscale correlated with the AHI, the cornerstone metric for OSA diagnosis and severity classification. The insomnia subscale was significantly associated with PSG-derived SOL and WASO, both of which are characteristic PSG findings in insomnia disorder [31].
The SSQ factors also demonstrated satisfactory diagnostic accuracy. ROC analyses demonstrated acceptable diagnostic performance across the three domains. Furthermore, diagnostic performance remained largely stable in the diagnosis-balanced subsample, supporting the robustness of the findings despite the original class imbalance. Although the parasomnia subscale showed a high AUC, this estimate should be interpreted cautiously because it was derived from a two-item factor and evaluated in a relatively small parasomnia subgroup. Therefore, the observed diagnostic performance should be considered preliminary until validated in larger independent cohorts.
The development of the SSQ has several practical advantages. First, the SSQ is a brief, self-administered questionnaire that can be completed within a short time, reducing respondent burden. Commonly used instruments such as the PSQI and ISI were designed to assess specific sleep domains and require more items to complete. In contrast, the SSQ provides a brief assessment of multiple common sleep disorders within a single questionnaire. Second, the SSQ has the potential to facilitate early detection and intervention for common sleep disorders, which may improve long-term health outcomes and quality of life. Notably, the SSQ targets both OSA and insomnia, the two most prevalent sleep disorders. OSA is associated with substantial cardiovascular [32], metabolic [33] and neuropsychiatric morbidity [34], highlighting the importance of early detection and treatment. Effective treatment of OSA can prevent significant morbidity when initiated early [35]. Nevertheless, OSA remains highly prevalent yet underdiagnosed; nearly one billion individuals are affected worldwide [36], and more than 84% of those with moderate to severe OSA were undiagnosed in the MESA study [37]. Chronic insomnia is common and associated with adverse physical and mental health outcomes, underscoring the value of early identification. Thus, the SSQ may contribute to reducing the burden of untreated sleep disorders by enabling timely screening and referral.
Limitations and Future Directions
Several limitations should be considered when interpreting the present findings. First, the distribution of diagnostic groups was substantially imbalanced, with OSA accounting for the majority of participants. To address this issue, we performed sensitivity analyses using a diagnosis-balanced subsample, which demonstrated comparable factor structures and diagnostic performance. Nevertheless, external validation in independent cohorts remains necessary. Second, the parasomnia domain consisted of only two items and was evaluated in a relatively small number of parasomnia cases. Therefore, the stability and diagnostic performance of this factor should be interpreted cautiously and require further confirmation. Third, the present study was conducted in a tertiary sleep clinic population, and the diagnostic distribution differs considerably from that encountered in primary care or community-based settings. Consequently, the findings should not be generalized beyond similar sleep-clinic populations, and further validation in primary care and community cohorts is required before broader implementation. Fourth, although EFA supported the internal structure of the SSQ, confirmatory factor analysis should be conducted in independent samples. Finally, additional reliability indices, including test–retest reliability, should be examined to complement internal consistency estimates.
Conclusions
The SSQ demonstrated acceptable psychometric properties and diagnostic performance as a brief screening questionnaire in an OSA-dominant tertiary sleep clinic population. The questionnaire showed a stable three-factor structure representing OSA, insomnia, and parasomnia domains, with evidence of convergent and concurrent validity. Replication of the factor structure and diagnostic performance in a diagnosis-balanced subsample further supports the robustness of the findings. Nevertheless, additional validation in primary care and community-based populations is required before broader clinical implementation.
Availability of Data and Material
The datasets generated or analyzed during the study are not publicly available due to patient privacy and ethical restrictions but are available from the corresponding author on reasonable request.
Notes
Ethics Statement
The requirement for informed consent was waived due to the retrospective nature of the study. The study protocol was approved by the Institutional Review Board of Samsung Medical Center (approval number: 2020-03-020) and conducted in accordance with the Declaration of Helsinki.
Author Contributions
Conceptualization: Eun Yeon Joo, Sooyeon Suh. Data curation: Sujin Lee, Hyoeun Bae. Funding acquisition: Eun Yeon Joo. Investigation: Hyoeun Bae, Eun Yeon Joo. Methodology: all authors. Supervision: Eun Yeon Joo, Sooyeon Suh. Validation: all authors. Visualization: Hyoeun Bae, Sujin Lee. Writing—original draft: Hyoeun Bae, Sujin Lee. Writing—review & editing: Hyoeun Bae, Eun Yeon Joo, Sooyeon Suh.
Conflicts of Interest
Sooyeon Suh, a contributing editor of the Sleep Medicine Research, was not involved in the editorial evaluation or decision to publish this article. All remaining authors have declared no conflicts of interest.
Funding Statement
This research was supported by a grant from Samsung Medical Center (OTC1190671).
Acknowledgements
The authors thank the participants and clinical staff who contributed to this study.
