ABSTRACT
Background
Diagnosing obstructive sleep apnea (OSA) in children is challenging, with long wait times for polysomnography (PSG). This study assessed the diagnostic accuracy of home‐recorded video clips for OSA compared to PSG.
Methods
Children (2−18 years) referred for PSG for suspected OSA were enrolled. Parents recorded video clips of their child sleeping and completed the Pediatric Sleep Questionnaire (PSQ). Blinded clinicians scored videos using the Monash Obstructive Sleep Apnea score (MS). Participants underwent PSG, and outcomes included obstructive apnea‐hypopnea index (OAHI) and oximetry metrics (i.e., McGill Oximetry Score [MOS]; 3% Oxygen Desaturation Index [ODI3]). Diagnostic characteristics of MS, PSQ, MOS, and ODI3 were compared for detection of any (OAHI ≥ 1.5 events/h) and moderate‐severe OSA (OAHI ≥ 5 events/h).
Results
Forty‐one children (age 7.0 years, 49% female) participated. Median OAHI was 0.6 events/h (IQR 0.3, 3.1); 16 (39%) had OAHI ≥ 1.5 events/h, 5 (12%) had OAHI ≥ 5 events/h. PSQ identified 36 (88%) participants with a score ≥ 0.33. One child had MOS ≥ 2; ODI3 was ≥ 4.3 in 8 (20%) and > 7 in 6 (15%). Mean MS was 3.6 (SD 2.1). MS had 81.2% sensitivity and 52.0% specificity for any OSA and 100% sensitivity and 44.4% specificity for moderate‐severe OSA. A combination of MS and ODI3 improved diagnostic accuracy with an AUC of 98.3.
Conclusion
MS demonstrated high sensitivity but low specificity for the detection of moderate‐severe OSA. Video scores outperformed PSQ but were less accurate than oximetry. Combining MS and ODI3 yielded the strongest diagnostic characteristics. Video scores may aid in pediatric OSA screening.
Keywords: Monash score, obstructive sleep apnea, polysomnography, sleep, video clips
1. Introduction
Obstructive sleep apnea (OSA) is defined by partial or complete airway obstruction during sleep which affects 1%−4% of children [1, 2]. Long‐term consequences of untreated OSA include pulmonary and systemic hypertension, insulin resistance, impaired neurocognitive and behavioral functioning, and reduced quality of life [3, 4, 5]. Early identification of OSA is critical for timely diagnosis and treatment to potentially reverse associated negative outcomes.
Polysomnography (PSG) is the gold standard test for the diagnosis of OSA [1, 2]. However, there are significant barriers that limit access for children. Referral for PSG is associated with high rates of loss to follow‐up, with approximately half of children not completing the evaluation [6, 7]. In the United States of America, socioeconomic status and geographic barriers increase wait times for children from underserviced communities [6, 8, 9, 10]. In Canada, pediatric PSG wait time can be up to 2 years due to a lack of PSG resources, and this time has been further exacerbated by the COVID‐19 pandemic [11, 12].
Other screening tools for the detection of OSA include home polygraphy, oximetry, and patient questionnaires. While home polygraphy allows for the measurement of respiratory events, airflow, and oxygen levels, it is unable to detect arousals or hypercapnia [13], often leading to an underestimation of OSA severity [14]. This limitation, along with the fact that it remains incompletely validated [15], resulted in a lack of endorsement from the American Academy of Sleep Medicine [15, 16], although it is used in resource‐limited settings in other countries [13, 17]. Oximetry is less costly than sleep studies but is not uniformly funded across Canada, so accessibility remains a barrier. Additionally, while overnight oximetry has a high positive predictive value (PPV) for OSA, a normal oximetry study does not rule out OSA [18]. Patient questionnaires have consistently been reported to have limited sensitivity and specificity in the diagnosis of OSA [1]. Overall, these methods have restricted accessibility and/or only correlate weakly with PSG findings [19, 20, 21]. Alternate accessible and reliable screening tools are needed to accurately identify and prioritize children most likely to have OSA for definitive evaluation with PSG.
Smartphone video recording may be a feasible screening tool to identify children at risk of OSA. Home video recordings have been used to identify other clinical conditions such as autism and seizures with high sensitivity and specificity [22, 23]. In a pilot study, smartphone‐recorded video clips were analyzed with a novel standardized scoring system and found to have 100% sensitivity and 36% specificity for OSA, but further work is needed to evaluate its use in different populations [24].
In this study, we evaluated the accuracy of smartphone‐recorded video clips for diagnosis of OSA in an otherwise‐healthy pediatric population, by estimating the sensitivity and specificity of home‐recorded video clips in identifying the presence of OSA, compared to PSG. Additionally, we examined the predictive performance of video clips compared to questionnaire and oximetry screening tools for OSA, independently and in combination. We hypothesized that the video clips would exhibit sensitivity and specificity comparable to conventional screening tools, highlighting their strong potential as an alternative screening method for diagnosing OSA.
2. Methods
2.1. Design
This cross‐sectional study evaluated the accuracy of home‐recorded video clips in identifying the presence of any OSA and moderate‐severe OSA compared to the gold standard, PSG, in children referred for PSG evaluation for suspected OSA at the Children's Hospital of Eastern Ontario (CHEO). Video clips were reviewed by two pediatric sleep physicians who were blinded to the child's history and symptoms. A standardized scoring tool was used to assess the videos for presence and severity of OSA [24].
2.2. Study Population
The study population consisted of children referred for evaluation of suspected OSA at a pediatric tertiary care center, the CHEO. Eligibility was based on the following criteria: (1) children aged 2−18 years; (2) currently on the waiting list for PSG at CHEO; (3) their caregiver had access to mobile technology. Patients were excluded if: (1) they had a previous diagnosis of sleep‐disordered breathing based on prior PSG within the last 5 years; (2) had an underlying genetic or congenital syndrome; (3) had obesity (BMI > 95%ile for age and sex); (4) were unable to cooperate for PSG; (5) parent/caregiver did not speak English or French.
2.3. Recruitment
Children were identified from the PSG wait list and initially contacted by a member of the circle of care to obtain consent to contact for research. Those who were interested were then contacted by the Research Coordinator for consent and assent when appropriate. The ability of the child to provide consent and/or assent was based on capacity rather than age. The study protocol was approved by the Research Ethics Board from the Children's Hospital of Eastern Ontario Research Institute (REB Protocol #: 22/112X).
2.4. Questionnaire
Parents were asked to complete the Pediatric Sleep Questionnaire (PSQ) at baseline. The PSQ consists of 22 symptom items that ask about snoring frequency, loudness of snoring, observed apneas, difficulty breathing during sleep, daytime sleepiness, inattentive or hyperactive behavior, and other features of pediatric OSA. PSQ scores have been shown to correlate with OSA in referred populations of children [25]. Subscales include a 4‐item sleepiness scale, a 4‐item snoring scale, and a 6‐item inattention/hyperactivity scale derived from the DSM‐IV criteria for attention‐deficit/hyperactivity disorder (ADHD). PSQ scores ≥ 0.33 are considered positive and high risk for sleep‐disordered breathing [25].
2.5. Video Clips
Parents were given instructions to record a 3 min video of their child during sleep using their personal smartphone on three nights within 1 week. It was recommended to begin recording 2 h after sleep onset or at 5:00 a.m. to increase the likelihood of capturing the child during REM sleep, when OSA is most likely to occur [26]. Parents were instructed to record with adequate light and proximity to observe chest and abdominal movements, and with audio recording. Parents uploaded the videos onto a secure, electronic database (REDCap) before the PSG [27].
One of two blinded pediatric sleep physicians reviewed the video clips and scored for the presence and severity of OSA using a standardized scoring tool, called the Monash score, as previously described [24]. Briefly, Monash scoring is based on five components: inspiratory obstructive noises (1–4), presence of obstructive events (0–1), increased work of breathing (0–1), mouth breathing (0–1), and neck extension (0–1), to obtain a total score with a maximum value of 8. A Monash score of 3 or greater is considered indicative of the presence of OSA [24]. A minimum of two video clips of sufficient quality were reviewed per participant [28].
2.6. Sleep Study
Within 6 weeks of recording the video clips, PSGs were performed and scored in accordance with the standards of the American Academy of Sleep Medicine [29]. Severity of OSA was determined using the obstructive apnea‐hypopnea index (OAHI, events/h), with an OAHI of < 1.5 considered normal, an OAHI of ≥ 1.5 to < 5 consistent with mild OSA, ≥ 5 to < 10 as moderate OSA, and an OAHI of ≥ 10 as severe OSA [30, 31, 32]. Oximetry was scored using the Oxygen Desaturation (≥ 3%) Index (ODI3) and the McGill Oximetry Score (MOS). The ODI3 is scored by the number of ≥ 3% arterial oxygen desaturation events per hour of sleep, with threshold values of ≥ 4.3 and > 7 events/h [31, 32]. The MOS is a 4‐level severity score based on the frequency and depth of oxygen desaturation events [33]. In short, patients were assigned a score of 1 if there were 3 clusters of desaturations and < 3 desaturation events below 90%; 2 if there were ≥ 3 events below 90%, but ≤ 3 events under 85%; 3 if there were > 3 events below 85% but ≤ 3 events below 80%; and 4 if there were > 3 events below 80% [33].
2.7. Data Collection
Patient medical charts were reviewed, and age, sex, height, weight, medical comorbidities, sleep study results, and family history were collected.
2.8. Statistical Analysis
Study population characteristics were summarized with mean (±SD) and median (IQR) for continuous variables, and frequencies and percentages for categorical variables. Participants with missing video clip scores were excluded.
The sensitivity and specificity were determined for the identification of OSA with OAHI as the gold standard for each of the PSQ, ODI3, MOS, and the Monash video score. Sensitivity was defined as the proportion of true positives identified by each method, and specificity was defined as the proportion of true negatives. The thresholds for each method represent cut‐off points for mild‐to‐severe OSA (OAHI ≥ 1.5 events/h) and moderate to severe OSA (OAHI ≥ 5 events/h). Diagnostic accuracy was further evaluated using the PPV, negative predictive value (NPV), positive likelihood ratio (LR+), and negative likelihood ratio (LR−) which provide additional insight into the clinical performance of video scoring. To compare the diagnostic accuracy of each tool, as well as combinations of various tools, receiver operating characteristic (ROC) curves were generated, and the area under the curve (AUC) and confidence intervals were determined.
3. Results
A summary of the demographic and sleep study variables is shown in Table 1. A total of 68 patients were recruited, and 41 submitted video clips. The median age of participants was 7.0 (IQR 5.3, 8.5) years, 20 (49%) were female, and the median OAHI was 0.6 (IQR 0.3, 3.1) events/h. Sixteen individuals (39%) had OAHI ≥ 1.5 events/h and 5 (12%) had OAHI ≥ 5 events/h. Thirty‐six (88%) had a PSQ score ≥ 0.33, 1 had a MOS ≥ 2, ODI3 was ≥ 4.3 in 8 (20%), and > 7 in 6 (15%). Mean Monash video score was 3.6 (SD 2.1).
Table 1.
Demographics table (N = 41).
| Variable | N (%) | Median (IQR) |
|---|---|---|
| Age | — | 7.0 (5.3, 8.5) |
| Sex, female | 20 (48.8) | — |
| OAHIa | — | 0.6 (0.3, 3.1) |
| Mild to severe OSA (OAHI ≥ 1.5) | 16 (39.0) | — |
| Moderate to severe OSA (OAHI ≥ 5) | 5 (12.2) | — |
| Total AHIb | — | 2.4 (1.2, 4.6) |
| Total recording time (min) | — | 431.5 (407.0, 464.5) |
| % total recording time in REM | — | 15.9 (12.1, 18.2) |
| Mean saturation (%) | — | 97.8 (97.4, 98.0) |
| Lowest saturation (%) | — | 93.0 (92.0, 94.0) |
| Mean (SD) | ||
| ODI3c | — | 2.7 (3.8) |
| ODI3 ≥ 4.3 | 8 (19.5) | — |
| ODI3 > 7 | 6 (14.6) | — |
| MOSd | — | 1.0 (0.2) |
| MOS ≥ 2 | 1 (2.4) | — |
| PSQ e | — | 0.5 (0.2) |
| PSQ ≥ 0.33 | 36 (87.8) | — |
| Monash score | — | 3.6 (2.1) |
| Monash score ≥ 3 | 25 (61.0) | — |
Obstructive apnea‐hypopnea index (events/h).
Total apnea‐hypopnea index (events/h).
Oxygen desaturation ≥ 3% index.
McGill oximetry score.
Pediatric sleep questionnaire.
To assess the feasibility of this approach, we examined participant engagement and technical challenges encountered during this study. Of the 68 total families recruited, 41 (60%) submitted video clips. Among those who submitted videos, the majority (N = 31, 75%) of families did not note any problems with capturing the videos. Technical problems during video upload were relatively infrequent, affecting (N = 5, 12%) participants, who tended to be located in rural or remote areas. Finally, (N = 12, 29%) of parents reported that recording the video clips disrupted their sleep schedule, while the majority did not experience sleep disruption.
We determined the sensitivity and specificity of the different diagnostic tests for detection of moderate to severe OSA (OAHI ≥ 5 events/h) (Table 2). The sensitivity and specificity of the video clips for moderate‐severe OSA were 100.0% (95% CI 47.8, 100.0) and 44.4% (95% CI 27.9, 61.9), respectively. We also determined the PPV and NPV which were 20.0 (6.8, 40.7) and 100.0 (79.4, 100.0), respectively. The LR+ was 1.80 (1.34, 2.41) and the LR− was 0.00 (0.00, NE). We conducted ROC curve analysis to compare the different screening tools for diagnosis of moderate‐severe OSA (Figure 1). The AUC for the Monash score was 88.3 (95% CI 77.0, 99.6). We also evaluated the performance of combinations of the various screening tools with the video clips and found that the combination of Monash score with ODI3 yielded the highest AUC of 98.3 (95% CI 93.9, 100) (Figure 2).
Table 2.
Sensitivity and specificity of diagnostic tools for detection of moderate‐severe OSA (OAHI ≥ 5).
| Tool | Threshold | Sensitivity | Specificity | PPV | NPV | LR+ | LR− |
|---|---|---|---|---|---|---|---|
| MS | ≥ 3 | 100.0 (47.8, 100.0) | 44.4 (27.9, 61.9) | 20.0 (6.8, 40.7) | 100.0 (79.4, 100.0) | 1.80 (1.34, 2.41) | 0.00 (0.00, NE) |
| MOS | ≥ 2 | 20.0 (0.5, 71.6) | 100.0 (90.3, 100.0) | 100.0 (2.5, 100.0) | 90.0 (76.3, 97.2) | NE | 0.80 (0.52, 1.24) |
| ODI3 | ≥ 4.3 | 80.0 (28.4, 99.5) | 88.9 (73.9, 96.9) | 50.0 (15.7, 84.3) | 97.0 (84.2, 99.9) | 7.20 (2.59, 20.02) | 0.22 (0.04, 1.30) |
| ODI3 | > 7 | 80.0 (28.4, 99.5) | 94.4 (81.3, 99.3) | 66.7 (22.3, 95.7) | 97.1 (85.1, 99.9) | 14.40 (3.49, 59.36) | 0.21 (0.04, 1.22) |
| PSQ | ≥ 0.33 | 100.0 (47.8, 100.0) | 13.9 (4.7, 29.5) | 13.9 (4.7, 29.5) | 100.0 (47.8, 100.0) | 1.16 (1.02, 1.32) | 0.00 (0.00, NE) |
Abbreviation: NE, not estimable.
Figure 1.
Receiver operating characteristic (ROC) curve comparing diagnostic tests for the detection of moderate to severe OSA (OAHI ≥ 5). Clinical thresholds are labeled on the curves from which the associated sensitivity and specificity from Table 2 can be viewed. AUC, area under the curve; MOS, McGill oximetry score; MS, Monash score; ODI3, oxygen desaturation (≥ 3%) index; PSQ, pediatric sleep questionnaire. [Color figure can be viewed at wileyonlinelibrary.com]
Figure 2.
Receiver operating characteristic (ROC) curve comparing combinations of diagnostic tests with the Monash score (MS) for the detection of moderate to severe OSA. AUC, area under the curve; MOS, McGill oximetry score; MS, Monash score; ODI3, oxygen desaturation (≥ 3%) index; PSQ, pediatric sleep questionnaire. [Color figure can be viewed at wileyonlinelibrary.com]
We also determined the sensitivity and specificity of the different diagnostic tests for detection of any OSA (OAHI ≥ 1.5 events/h) (Table 3). The sensitivity and specificity of the Monash score for the presence of OSA were 81.2% (95% CI 54.4, 96.0) and 52.0% (95% CI 31.3, 72.2), respectively. The PPV was 52.0 (31.3, 72.2) and NPV was 81.2 (54.4, 96.0). The LR+ was 1.69 (1.06, 2.71) and the LR− was 0.36 (0.12, 1.07). Monash score specificity was greater than that of PSQ, but not as high as that of oximetry metrics. We conducted ROC curve analysis to compare the different screening tools for diagnosis of any OSA (Figure 3). The AUC for the Monash score was 75.0 (95% CI 59.1, 90.9). We also evaluated the performance of combinations of the various screening tools with the video clips, which was highest when Monash score and ODI3 were combined (AUC 88.1 [95% CI 76.5, 96.8]; Figure 4).
Table 3.
Sensitivity and specificity of diagnostic tools for the detection of any OSA (OAHI ≥ 1.5).
| Measure | Threshold | Sensitivity | Specificity | PPV | NPV | LR+ | LR− |
|---|---|---|---|---|---|---|---|
| MS | ≥ 3 | 81.2 (54.4, 96.0) | 52.0 (31.3, 72.2) | 52.0 (31.3, 72.2) | 81.2 (54.4, 96.0) | 1.69 (1.06, 2.71) | 0.36 (0.12, 1.07) |
| MOS | ≥ 2 | 6.2 (0.2, 30.2) | 100.0 (86.3, 100.0) | 100.0 (2.5, 100.0) | 62.5 (45.8, 77.3) | NE | 0.94 (0.83, 1.06) |
| ODI3 | ≥ 4.3 | 43.8 (19.8, 70.1) | 96.0 (79.6, 99.9) | 87.5 (47.3, 99.7) | 72.7 (54.5, 86.7) | 10.94 (1.48, 80.75) | 0.59 (0.38, 0.91) |
| ODI3 | > 7 | 37.5 (15.2, 64.6) | 100.0 (86.3, 100.0) | 100.0 (54.1, 100.0) | 71.4 (53.7, 85.4) | NE | 0.62 (0.43, 0.91) |
| PSQ | ≥ 0.33 | 81.2 (54.4, 96.0) | 8.0 (1.0, 26.0) | 36.1 (20.8, 53.8) | 40.0 (5.3, 85.3) | 0.88 (0.68, 1.15) | 2.34 (0.44, 12.52) |
Abbreviation: NE, not estimable.
Figure 3.
Receiver operating characteristic (ROC) curve comparing diagnostic tests for the detection of any OSA (OAHI ≥ 1.5 events/h). Clinical thresholds are labeled on the curves from which the associated sensitivity and specificity from Table 3 can be viewed. AUC, area under the curve; MOS, McGill oximetry score; MS, Monash score; ODI3, oxygen desaturation (≥ 3%) index; PSQ, pediatric sleep questionnaire. [Color figure can be viewed at wileyonlinelibrary.com]
Figure 4.
Receiver operating characteristic (ROC) curve comparing combinations of diagnostic tests with the Monash score (MS) for the detection of any OSA (OAHI ≥ 1.5 events/h). AUC, area under the curve; MOS, McGill oximetry score; MS, Monash score; ODI3, oxygen desaturation (≥ 3%) index; PSQ, pediatric sleep questionnaire. [Color figure can be viewed at wileyonlinelibrary.com]
4. Discussion
In an effort to improve existing OSA screening options, this study explored a novel tool to enhance the identification and triage of pediatric OSA patients. In this study, we performed an evaluation of the diagnostic accuracy of short video clips for the detection of OSA in children. We determined that in the identification of any OSA, or moderate‐severe OSA, the Monash video score was very sensitive, but less specific. These results underscore the strength of video clips as a screening tool to rule out children unlikely to have clinically significant OSA (OAHI ≥ 5)—those who may not require urgent PSG and may be better candidates for watchful waiting. When compared to other OSA screening tools, video scores performed better than questionnaire (PSQ) for the detection of moderate‐severe OSA, but less well than oximetry metrics (ODI3, MOS). To enhance accuracy, we combined the different screening tools with video clip scoring. The combination of video clip score and ODI3 showed excellent diagnostic characteristics for detection of any OSA (AUC 88.1 [95% CI 76.5, 96.8]) as well as moderate‐severe OSA (AUC 98.3 [95% CI 93.9, 100.0]). These findings highlight the potential of the Monash score—especially when combined with oximetry metrics—as a clinically useful triage tool, not only to identify children at high risk but to reliably exclude those at low risk, thereby streamlining access to diagnostic services and guiding more targeted treatment decisions.
This marks the first external validation of the Monash score in comparison to PSG beyond the original publication. In this new study population, we identified several practical considerations in collecting the video clips that will be important for future implementation. Parents were instructed to record their children 2 h after sleep onset or at 5 o'clock in the morning, which differed from the instructions in the original study, where parents were asked to record the breathing pattern they were most concerned about [24]. While this study attempted a more standardized approach, ultimately, many parents submitted videos that were not recorded during these times; instead, parents used their best judgment to record when their child was experiencing the most symptoms. Upon video upload, parents were asked about any issues they experienced at any point during the study. Overall, the home‐recorded video protocol proved feasible, with most families able to record and submit clips without major difficulty. A minority experienced challenges related to recording the video without disturbing the child or technical difficulties, primarily among those in rural or remote communities. While some parents did note disruption to their own sleep schedule, the approach was generally well tolerated.
The diagnostic characteristics described here align with previously reported findings. A study using 30 min professional video recordings to identify OSA, as defined by PSG, also found high sensitivity but low specificity [34]. Similarly, the sensitivity and specificity reported in the development of the Monash score are closely comparable to values reported here [24]. These studies demonstrate video clips perform consistently across different cohorts in the detection of OSA. This current study also evaluated the diagnostic performance of other conventional screening tools, allowing for a comparative assessment of video scoring with established screening measures for OSA. This broader scope provides a more in‐depth assessment of the diagnostic value of video clips and enhances the understanding of their reliability and utility in clinical settings. Further, the integration of multiple screening tools in our assessment is a unique aspect of this study, which is of clinical significance, especially in the context of institutions with limited resources, as the combination of measures improved diagnostic performance.
The video clips demonstrated higher predictive accuracy for OSA than conventional PSQ evaluations. In fact, the combination of PSQ and video clips did not change the AUC in the identification of moderate‐severe OSA (Figure 2). While PSQ continues to be a useful low‐resource clinical tool for sleep‐disordered screening, in particular for prediction of post‐adenotonsillectomy improvements in behavior outcomes [35], video clips can provide added value. While video clips were found to be less effective than oximetry metrics for accurate identification of OSA, their combination was found to be more predictive than the use of oximetry alone. However, certain accessibility challenges make oximetry a less practical choice of diagnostic test in some scenarios. For instance, the high upfront cost of oximetry technology remains a significant barrier, which explains why these resources are not uniformly available, even in developed nations like Canada [36]. In addition, studies have shown that pulse oximeters may provide less accurate readings in patients with darker skin pigmentation, leading to disparities in oximetry accuracy among patients of color [37, 38]. These inaccuracies lead to delays in diagnosis and treatment which put children at risk of adverse health outcomes [39]. While ODI3 displayed the highest accuracy out of all of the screening tools, it would be interesting to investigate whether that would still be true across all skin tones. Given the variability in resource availability within sleep medicine, it is important to focus on complementing existing tools rather than replacing them, ultimately expanding the toolkit for OSA detection.
These findings contribute to the growing body of evidence supporting the use of video clips as a promising screening tool for OSA. The Monash video scoring method has demonstrated excellent inter‐rater reliability, performing with high consistency ratings between two independent sleep physicians [28]. This analysis has been conducted in various population subgroups, including children with obesity, autism, and Down syndrome, as well as the same otherwise healthy population examined in the present study. Additionally, video clips have also been examined in combination with alternative diagnostic methods, compared to polygraphy, and demonstrated high sensitivity for identification of moderate to severe OSA, reinforcing their value in resource‐limited settings [40].
Building on these findings, there are several opportunities to refine and expand upon the use of home‐recorded video clips as a screening tool for pediatric OSA. As alternative screening tools become more widely accepted to address resource limitations in pediatric sleep medicine, it will be important to understand how the Monash score complements these approaches. Home‐based oximetry is a frequently used tool, and its ability to be performed at home makes it an appealing and accessible alternative for evaluating sleep‐disordered breathing. It would be valuable to examine the combination of the Monash score and home‐monitored oximetry, to compare how this differs from the PSG oximetry metrics. Additionally, it would be very interesting to know how well the Monash score performs for children with and without tonsillar hypertrophy or if they had adenotonsillectomy. Studies like this one contribute to opening up several promising lines of future research in the role of video‐based assessments within pediatric OSA screening strategies.
These findings provide promising evidence for the use of video clips as a screening tool for OSA, but there are some limitations to be considered. This is a relatively small sample size; it will be important to evaluate the use of video clips in a larger population before clinical practice is adapted. The night‐to‐night variability is another limitation that could be examined, although this has been mitigated by using three video clips. Furthermore, this study only examined otherwise healthy children. OSA is more prevalent in children with obesity or conditions such as Down syndrome, so it will be of clinical interest to explore the performance of video clips in these populations as well.
In conclusion, our findings provide preliminary evidence that home‐recorded video clips, especially when combined with oximetry metrics like ODI3, could become a critical tool in the early identification of pediatric OSA. As the demand for diagnostic testing continues to outpace availability, integrating these tools into clinical practice could dramatically reduce the time to diagnosis and intervention, potentially preventing the long‐term morbidities associated with delayed treatment.
Author Contributions
Sherri L. Katz: conceptualization, methodology, funding acquisition, writing – review and editing. Taylor Barwell: investigation, writing – original draft, review and editing, project administration. Vid Bijelić: methodology, data curation, formal data analysis. Nicholas Barrowman: methodology, data curation, formal data analysis. Henrietta Blinder: investigation, project administration. Naomi Dussah: investigation, project administration. Roya Shamsi: investigation, project administration. Alexa Leitman: investigation. Refika Ersu: conceptualization, methodology, funding acquisition, writing – review and editing, supervision.
Ethics Statement
Ethical approval was obtained from the Children's Hospital of Eastern Ontario Research Ethics Board (Approval Number: 22/112X). Informed consent was obtained from all participants (or legal guardians) before their participation. For individuals unable to provide consent, assent was obtained in addition to consent.
Conflicts of Interest
The authors declare no conflicts of interest.
Acknowledgments
This research was supported by a CHEST Foundation Grant in Sleep Medicine (#:17970).
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Marcus C. L., Brooks L. J., Ward S. D., et al., “Diagnosis and Management of Childhood Obstructive Sleep Apnea Syndrome,” Pediatrics 130 (2012): e714–e755. [DOI] [PubMed] [Google Scholar]
- 2. Gileles‐Hillel A., Bhattacharjee R., Gorelik M., and Narang I., “Advances in Sleep‐Disordered Breathing in Children,” Clinics in Chest Medicine 45 (2024): 651–662. [DOI] [PubMed] [Google Scholar]
- 3. Decuzzi J., Redline S., and Isaiah A., “Secondary Analyses of the Childhood Adenotonsillectomy Trial: A Narrative Review,” JAMA Otolaryngology–Head & Neck Surgery 148 (2022): 779–784. [DOI] [PubMed] [Google Scholar]
- 4. Menzies B., Teng A., Burns M., and Lah S., “Neurocognitive Outcomes of Children With Sleep Disordered Breathing: A Systematic Review With Meta‐Analysis,” Sleep Medicine Reviews 63 (2022): 101629. [DOI] [PubMed] [Google Scholar]
- 5. Smith D. F. and Amin R. S., “OSA and Cardiovascular Risk in Pediatrics,” Chest 156 (2019): 402–413. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Boss E. F., Benke J. R., Tunkel D. E., Ishman S. L., Bridges J. F. P., and Kim J. M., “Public Insurance and Timing of Polysomnography and Surgical Care for Children With Sleep‐Disordered Breathing,” JAMA Otolaryngology–Head & Neck Surgery 141 (2015): 106–111. [DOI] [PubMed] [Google Scholar]
- 7. Harris V. C., Links A. R., Kim J. M., Walsh J., Tunkel D. E., and Boss E. F., “Follow‐Up and Time to Treatment in an Urban Cohort of Children With Sleep‐Disordered Breathing,” Otolaryngology–Head and Neck Surgery 159 (2018): 371–378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Bisgaier J. and Rhodes K. V., “Auditing Access to Specialty Care for Children With Public Insurance,” New England Journal of Medicine 364 (2011): 2324–2333. [DOI] [PubMed] [Google Scholar]
- 9. Boss E. F., Smith D. F., and Ishman S. L., “Racial/Ethnic and Socioeconomic Disparities in the Diagnosis and Treatment of Sleep‐Disordered Breathing in Children,” International Journal of Pediatric Otorhinolaryngology 75 (2011): 299–307. [DOI] [PubMed] [Google Scholar]
- 10. E. B. Penn, Jr. , French A., Bhushan B., and Schroeder J. W. Jr., “Access to Care for Children With Symptoms of Sleep Disordered Breathing,” International Journal of Pediatric Otorhinolaryngology 76 (2012): 1671–1673. [DOI] [PubMed] [Google Scholar]
- 11. Katz S. L., Witmans M., Barrowman N., et al., “Paediatric Sleep Resources in Canada: The Scope of the Problem,” Paediatrics & Child Health 19 (2014): 367–372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Gerdung C. A., Tsampalieros A., Barrowman N., Dibartolo M., and Katz S. L., “Tertiary Pediatric Sleep Services in Canada: A National Survey,” Canadian Journal of Respiratory, Critical Care, and Sleep Medicine 9 (2025): 14–20. [Google Scholar]
- 13. Kaditis A. G., Alonso Alvarez M. L., Boudewyns A., et al., “Obstructive Sleep Disordered Breathing in 2‐ to 18‐Year‐Old Children: Diagnosis and Management,” European Respiratory Journal 47 (2016): 69–94. [DOI] [PubMed] [Google Scholar]
- 14. Kissow Lildal T., Boudewyns A., Kamperis K., et al., “Validity of In‐Lab and Home Respiratory Polygraphy for Detecting Obstructive Sleep Apnea in Children,” Sleep Medicine 103 (2023): 195–203. [DOI] [PubMed] [Google Scholar]
- 15. Kirk V., Baughn J., D'Andrea L., et al., “American Academy of Sleep Medicine Position Paper for the Use of a Home Sleep Apnea Test for the Diagnosis of OSA in Children,” Journal of Clinical Sleep Medicine 13 (2017): 1199–1203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Tan H. L., Kheirandish‐Gozal L., and Gozal D., “Pediatric Home Sleep Apnea Testing,” Chest 148 (2015): 1382–1395. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Borsini E., Blanco M., Bosio M., Fernando D. T., Ernst G., and Salvado A., “‘Diagnosis of Sleep Apnea in Network’ Respiratory Polygraphy as a Decentralization Strategy,” Sleep Science 9 (2016): 244–248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Brouillette R. T., Morielli A., Leimanis A., Waters K. A., Luciano R., and Ducharme F. M., “Nocturnal Pulse Oximetry as an Abbreviated Testing Modality for Pediatric Obstructive Sleep Apnea,” Pediatrics 105 (2000): 405–412. [DOI] [PubMed] [Google Scholar]
- 19. Mitchell R. B., Garetz S., Moore R. H., et al., “The Use of Clinical Parameters to Predict Obstructive Sleep Apnea Syndrome Severity in Children: The Childhood Adenotonsillectomy (CHAT) Study Randomized Clinical Trial,” JAMA Otolaryngology–Head & Neck Surgery 141 (2015): 130–136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Patel A. P., Meghji S., and Phillips J. S., “Accuracy of Clinical Scoring Tools for the Diagnosis of Pediatric Obstructive Sleep Apnea,” Laryngoscope 130 (2020): 1034–1043. [DOI] [PubMed] [Google Scholar]
- 21. Weinstock T. G., Rosen C. L., Marcus C. L., et al., “Predictors of Obstructive Sleep Apnea Severity in Adenotonsillectomy Candidates,” Sleep 37 (2014): 261–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Chen D. K., Graber K. D., Anderson C. T., and Fisher R. S., “Sensitivity and Specificity of Video Alone Versus Electroencephalography Alone for the Diagnosis of Partial Seizures,” Epilepsy & Behavior: E&B 13 (2008): 115–118. [DOI] [PubMed] [Google Scholar]
- 23. Fusaro V. A., Daniels J., Duda M., et al., “The Potential of Accelerating Early Detection of Autism Through Content Analysis of YouTube Videos,” PLoS One 9 (2014): e93533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Thomas R. J., Dalton S., Harman K., et al., “Smartphone Videos to Predict the Severity of Obstructive Sleep Apnoea,” Archives of Disease in Childhood 107 (2022): 148–152. [DOI] [PubMed] [Google Scholar]
- 25. Chervin R. D., Hedger K., Dillon J. E., and Pituch K. J., “Pediatric Sleep Questionnaire (PSQ): Validity and Reliability of Scales for Sleep‐Disordered Breathing, Snoring, Sleepiness, and Behavioral Problems,” Sleep Medicine 1 (2000): 21–32. [DOI] [PubMed] [Google Scholar]
- 26. Bonsignore M. R., Mazzuca E., Baiamonte P., Bouckaert B., Verbeke W., and Pevernagie D. A., “REM Sleep Obstructive Sleep Apnoea,” European Respiratory Review 33 (2024): 230166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Harris P. A., Taylor R., Thielke R., Payne J., Gonzalez N., and Conde J. G., “Research Electronic Data Capture (REDCap): A Metadata‐Driven Methodology and Workflow Process for Providing Translational Research Informatics Support,” Journal of Biomedical Informatics 42 (2009): 377–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Katz S. L., Blinder H., Dussah N., et al., “Inter‐Rater Reliability of Video Clips to Assess for Obstructive Sleep Apnea in Children,” Sleep Medicine 131 (2025): 106483. [DOI] [PubMed] [Google Scholar]
- 29. Troester M. M., Quan S. F., Berry R. B., et al., The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications. Version 3 (American Academy of Sleep Medicine, 2023), 1. [Google Scholar]
- 30. Witmans M. B., Keens T. G., Ward S. L. D., and Marcus C. L., “Obstructive Hypopneas in Children and Adolescents: Normal Values,” American Journal of Respiratory and Critical Care Medicine 168 (2003): 1540. [DOI] [PubMed] [Google Scholar]
- 31. Evans H. J., Gibson N. A., Bennett J., et al., “British Thoracic Society Guideline for Diagnosing and Monitoring Paediatric Sleep‐Disordered Breathing,” Thorax 78 (2023): 1043–1054. [DOI] [PubMed] [Google Scholar]
- 32. Polytarchou A., Ohler A., Moudaki A., et al., “Nocturnal Oximetry Parameters as Predictors of Sleep Apnea Severity in Resource‐Limited Settings,” Journal of Sleep Research 32 (2023): e13638. [DOI] [PubMed] [Google Scholar]
- 33. Nixon G. M., Kermack A. S., Davis G. M., Manoukian J. J., Brown K. A., and Brouillette R. T., “Planning Adenotonsillectomy in Children With Obstructive Sleep Apnea: The Role of Overnight Oximetry,” Pediatrics 113 (2004): e19–e25. [DOI] [PubMed] [Google Scholar]
- 34. Sivan Y., Kornecki A., and Schonfeld T., “Screening Obstructive Sleep Apnoea Syndrome by Home Videotape Recording in Children,” European Respiratory Journal 9 (1996): 2127–2131. [DOI] [PubMed] [Google Scholar]
- 35. Isaiah A., Spanier A. J., Grattan L. M., Wang Y., and Pereira K. D., “Predictors of Behavioral Changes After Adenotonsillectomy in Pediatric Obstructive Sleep Apnea: A Secondary Analysis of a Randomized Clinical Trial,” JAMA Otolaryngology–Head & Neck Surgery 146 (2020): 900–908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Herbert L. J. and Wilson I. H., “Pulse Oximetry in Low‐Resource Settings,” Breathe 9 (2012): 90–98. [Google Scholar]
- 37. Sjoding M. W., Dickson R. P., Iwashyna T. J., Gay S. E., and Valley T. S., “Racial Bias in Pulse Oximetry Measurement,” New England Journal of Medicine 383 (2020): 2477–2478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Shi C., Goodall M., Dumville J., et al., “The Accuracy of Pulse Oximetry in Measuring Oxygen Saturation by Levels of Skin Pigmentation: A Systematic Review and Meta‐Analysis,” BMC Medicine 20 (2022): 267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Wong A. K. I., Charpignon M., Kim H., et al., “Analysis of Discrepancies Between Pulse Oximetry and Arterial Oxygen Saturation Measurements by Race and Ethnicity and Association With Organ Dysfunction and Mortality,” JAMA Network Open 4 (2021): e2131674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Ersu R., Shamsi R., Bjelic V., et al., “Screening for Obstructive Sleep Apnea in the Resource‐Limited Setting of the COVID‐19 Pandemic (SPARE Study),” Sleep Medicine 132 (2025): 106559. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.