Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Name | Class |
|---|---|
| West China Hospital | OTHER |
| Second Affiliated Hospital of Soochow University | OTHER |
| Dalian Municipal Central Hospital | OTHER |
| Third Hospital of Inner Mongolia Autonomous Region |
Not provided
Not provided
Not provided
Not provided
This study will evaluate and compare the diagnostic accuracy and effectiveness of the Five Seasons Sleep Tracking Mat with the gold standard, polysomnography.
Given that obstructive sleep apnea (OSA) imposes a heavy disease burden, conducting effective public screening is regarded as a crucial step in health management and chronic disease prevention. Currently, the diagnosis of OSA relies on polysomnography (PSG) and manual scoring, which are constrained by equipment availability and personnel shortages, resulting in low efficiency, high costs, and scheduling difficulties. Additionally, PSG monitoring requires numerous contact-based sensors, leading to poor patient compliance. The lack of large-scale screening and long-term follow-up further restricts the implementation of standardized and evidence-based treatments. Developing new, effective, low- or zero-burden sleep monitoring devices to minimize sleep disruption while ensuring monitoring accuracy is, therefore, a key direction for future research.
The Five Seasons Sleep Tracking Mat (5S Sleep Tracking Mat) detects heartbeats, respiration, body movements, and snoring using ballistocardiogram (BCG) signals. The device is composed primarily of a control unit, monitoring mat, temperature/humidity sensor, power adapter, and light sensor. During operation, the device accurately captures the impact force of bodily movements at a high sampling rate of 2 kHz. It then uses proprietary signal processing and pre-trained AI models to extract heartbeat and respiratory waveforms, as well as body movements. Additionally, it automatically detects respiratory events, calculates the apnea-hypopnea index (AHI), and determines sleep stages. Since the monitoring mat only needs to be placed under the pillow without requiring direct contact with the body or wearable accessories, it is particularly suitable for home-based, long-term daily sleep monitoring.
In this study, researchers will use the current gold standard, PSG, as a reference to validate the Five Seasons Sleep Tracking Mat for sleep monitoring and OSA-assisted diagnosis.
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Name | Type | Description | Arm Group Labels | Other Names |
|---|---|---|---|---|
| Five Seasons [5S] Sleep Tracking Mat | Diagnostic Test | The Five Seasons [5S] Sleep Tracking Mat consists of a main control box, monitoring mat, temperature/humidity sensor, power adapter, and light sensor. It utilizes Ballistocardiogram (BCG) technology to detect heartbeat, respiration, body movements, and snoring. The collected data is analyzed by a proprietary pre-trained AI model, enabling automatic detection of respiratory events, calculation of the Apnea-Hypopnea Index (AHI), and sleep stage interpretation. Compared to other sleep monitoring devices, it is non-contact and requires no wearable components, making it particularly suitable for long-term, daily home use. |
| Measure | Description | Time Frame |
|---|---|---|
| Diagnostic sensitivity (AHI = 5) of 5S Sleep Tracking Mat compared to PSG | The proportion of participants with Apnea Hypopnea Index (AHI) ≥ 5 events per hour according to 5S Sleep Tracking Mat in participants with AHI ≥ 5 events per hour according to PSG scored with American Academy of Sleep Medicine (AASM) v2.6 Guidelines. The point estimation of diagnostic sensitivity of 5S Sleep Tracking Mat will be compared to the threshold of 0.8, with the half width of 95% confidence interval being no more than 0.1. | Day 1 |
| Diagnostic specificity (AHI = 5) of 5S Sleep Tracking Mat compared to PSG | The proportion of participants with Apnea Hypopnea Index (AHI) ≥ 5 events per hour according to 5S Sleep Tracking Mat in participants with AHI ≥ 5 events per hour according to PSG scored with American Academy of Sleep Medicine (AASM) v2.6 Guidelines. The point estimation of diagnostic specificity of 5S Sleep Tracking Mat will be compared to the threshold of 0.8, with the half width of 95% confidence interval being no more than 0.1. | Day 1 |
| Measure | Description | Time Frame |
|---|---|---|
| Diagnostic sensitivity (AHI = 15) of 5S Sleep Tracking Mat compared to PSG | The proportion of participants with Apnea Hypopnea Index (AHI) ≥ 15 events per hour according to 5S Sleep Tracking Mat in participants with AHI ≥ 15 events per hour according to PSG scored with American Academy of Sleep Medicine (AASM) v2.6 Guidelines. The point estimation of diagnostic sensitivity of 5S Sleep Tracking Mat will be compared to the threshold of 0.8, with the half width of 95% confidence interval being no more than 0.1. |
Not provided
Inclusion Criteria:
Exclusion Criteria:
Not provided
Not provided
Not provided
Age between 18 and 70 years, with potential symptoms of snoring at night, daytime sleepiness, or self-reported poor sleep quality, suspected of having OSA, recruited by department of otorhinolaryncology in collaborative hospitals.
| Name | Role | Phone | Extension | |
|---|---|---|---|---|
| Weijun Huang, Dr. | Contact | 86 18930174480 | hellohuangwj@126.com |
Not provided
| Facility | Status | City | State | ZIP | Country | Contacts |
|---|---|---|---|---|---|---|
| Shanghai Sixth People's Hospital | Recruiting | Shanghai | Shanghai Municipality | 200025 | China |
All IPD that underlie results in a publication will be shared.
Not provided
Not provided
Not provided
Not provided
Not provided
| ID | Term |
|---|---|
| D020181 | Sleep Apnea, Obstructive |
| ID | Term |
|---|---|
| D012891 | Sleep Apnea Syndromes |
| D001049 | Apnea |
| D012120 | Respiration Disorders |
| D012140 | Respiratory Tract Diseases |
Not provided
Not provided
| UNKNOWN |
| Beijing HuiLongGuan Hospital | OTHER |
Not provided
Not provided
Not provided
| Day 1 |
| Diagnostic specificity (AHI = 30) of 5S Sleep Tracking Mat compared to PSG | The proportion of participants with Apnea Hypopnea Index (AHI) ≥ 30 events per hour according to 5S Sleep Tracking Mat in participants with AHI ≥ 30 events per hour according to PSG scored with American Academy of Sleep Medicine (AASM) v2.6 Guidelines. The point estimation of diagnostic specificity of 5S Sleep Tracking Mat will be compared to the threshold of 0.8, with the half width of 95% confidence interval being no more than 0.1. | Day 1 |
| Diagnostic sensitivity (AHI = 30) of 5S Sleep Tracking Mat compared to PSG | The proportion of participants with Apnea Hypopnea Index (AHI) ≥ 30 events per hour according to 5S Sleep Tracking Mat in participants with AHI ≥ 30 events per hour according to PSG scored with American Academy of Sleep Medicine (AASM) v2.6 Guidelines. The point estimation of diagnostic sensitivity of 5S Sleep Tracking Mat will be compared to the threshold of 0.8, with the half width of 95% confidence interval being no more than 0.1. | Day 1 |
| Overall accuracy of 5S Sleep Tracking Mat at AHI = 5 | Using an AHI threshold of 5 for true positive and true negative classification, the overall accuracy is the percentage of total cases correctly classified. | Day 1 |
| Overall accuracy of 5S Sleep Tracking Mat at AHI = 15 | Using an AHI threshold of 15 for true positive and true negative classification, the overall accuracy is the percentage of total cases correctly classified. | Day 1 |
| Overall accuracy of 5S Sleep Tracking Mat at AHI = 30 | Using an AHI threshold of 30 for true positive and true negative classification, the overall accuracy is the percentage of total cases correctly classified. | Day 1 |
| Cohen's kappa of 5S Sleep Tracking Mat at AHI = 5 | Using an AHI threshold of 5 for determining true positive or true negative results, Cohen's kappa (κ) is the ratio of the difference between the observed agreement (Pa) and the expected agreement (Pe) to the maximum possible difference. When κ=1, the sleep mat's diagnosis perfectly matches the PSG diagnosis. When κ=0, the agreement is entirely due to random effects. When κ<0, the sleep mat's diagnosis is completely inconsistent with the PSG diagnosis. | Day 1 |
| Cohen's kappa of 5S Sleep Tracking Mat at AHI = 15 | Using an AHI threshold of 15 for determining true positive or true negative results, Cohen's kappa (κ) is the ratio of the difference between the observed agreement (Pa) and the expected agreement (Pe) to the maximum possible difference. When κ=1, the sleep mat's diagnosis perfectly matches the PSG diagnosis. When κ=0, the agreement is entirely due to random effects. When κ<0, the sleep mat's diagnosis is completely inconsistent with the PSG diagnosis. | Day 1 |
| Cohen's kappa of 5S Sleep Tracking Mat at AHI = 30 | Using an AHI threshold of 30 for determining true positive or true negative results, Cohen's kappa (κ) is the ratio of the difference between the observed agreement (Pa) and the expected agreement (Pe) to the maximum possible difference. When κ=1, the sleep mat's diagnosis perfectly matches the PSG diagnosis. When κ=0, the agreement is entirely due to random effects. When κ<0, the sleep mat's diagnosis is completely inconsistent with the PSG diagnosis. | Day 1 |
| Area under the Receiver Operating Characteristic (ROC) Curve (AUC) at AHI = 5. | The ROC of the 5S Sleep Tracking Mat will be drawn based on PSG diagnostic results. The 95% confidence interval of the AUC will be calculated using model-based estimation, with true positive/negative results determined with an AHI threshold of 5. | Day 1 |
| Area under the Receiver Operating Characteristic (ROC) Curve (AUC) at AHI = 15. | The ROC of the 5S Sleep Tracking Mat will be drawn based on PSG diagnostic results. The 95% confidence interval of the AUC will be calculated using model-based estimation, with true positive/negative results determined with an AHI threshold of 15. | Day 1 |
| Area under the Receiver Operating Characteristic (ROC) Curve (AUC) at AHI = 30. | The ROC of the 5S Sleep Tracking Mat will be drawn based on PSG diagnostic results. The 95% confidence interval of the AUC will be calculated using model-based estimation, with true positive/negative results determined with an AHI threshold of 30. | Day 1 |
| Intra-Class Correlation Coefficient (ICC) of Apnea Hypopnea Index (AHI) | ICC between the AHI according to 5S Sleep Tracking Mat and the AHI according to Polysomnography (PSG), using one-way random effects model. | Day 1 |
| Intra-Class Correlation Coefficient (ICC) of Total Sleep Time (TST) | ICC between the TST according to 5S Sleep Tracking Mat and the TST according to Polysomnography (PSG), using one-way random effects model. | Day 1 |
| Intra-Class Correlation Coefficient (ICC) of Sleep Latency (SL) | ICC between the SL according to 5S Sleep Tracking Mat and the SL according to Polysomnography (PSG), using one-way random effects model. | Day 1 |
| Intra-Class Correlation Coefficient (ICC) of Wake After Sleep Onset (WASO) | ICC between the WASO according to 5S Sleep Tracking Mat and the WASO according to Polysomnography (PSG), using one-way random effects model. | Day 1 |
| Intra-Class Correlation Coefficient (ICC) of the proportion of the stage 1 and 2 of Non Rapid Eye Movement (NREM) sleep | ICC between the proportion of the stage 1 and 2 of NREM sleep according to 5S Sleep Tracking Mat and the proportion of the stage 1 and 2 of NREM sleep according to Polysomnography (PSG), using one-way random effects model. | Day 1 |
| Intra-Class Correlation Coefficient (ICC) of the proportion of the stage 3 of Non Rapid Eye Movement (NREM) sleep | ICC between the proportion of the stage 3 of NREM sleep according to 5S Sleep Tracking Mat and the proportion of the stage 3 of NREM sleep according to Polysomnography (PSG), using one-way random effects model. | Day 1 |
| Intra-Class Correlation Coefficient (ICC) of the proportion of Rapid Eye Movement (REM) sleep | ICC between the proportion of REM sleep according to 5S Sleep Tracking Mat and the proportion of REM sleep according to Polysomnography (PSG), using one-way random effects model. | Day 1 |
| Bland-Altman Limits of Agreement (LOA) of Apnea Hypopnea Index (AHI) | The mean and difference of AHI according to both 5S Sleep Tracking Mat and PSG are analyzed from the raw data. The Bland-Altman plot is created with the mean on the x-axis and the difference on the y-axis. The mean difference and the 95% limits of agreement (LOA) are calculated based on the plot which must fall within the clinically accepted threshold range. | Day 1 |
| Bland-Altman Limits of Agreement (LOA) of Total Sleep Time (TST) | The mean and difference of TST according to both 5S Sleep Tracking Mat and PSG are analyzed from the raw data. The Bland-Altman plot is created with the mean on the x-axis and the difference on the y-axis. The mean difference and the 95% limits of agreement (LOA) are calculated based on the plot which must fall within the clinically accepted threshold range. | Day 1 |
| Bland-Altman Limits of Agreement (LOA) of Sleep Latency (SL) | The mean and difference of SL according to both 5S Sleep Tracking Mat and PSG are analyzed from the raw data. The Bland-Altman plot is created with the mean on the x-axis and the difference on the y-axis. The mean difference and the 95% limits of agreement (LOA) are calculated based on the plot which must fall within the clinically accepted threshold range. | Day 1 |
| Bland-Altman Limits of Agreement (LOA) of Wake After Sleep Onset (WASO) | The mean and difference of WASO according to both 5S Sleep Tracking Mat and PSG are analyzed from the raw data. The Bland-Altman plot is created with the mean on the x-axis and the difference on the y-axis. The mean difference and the 95% limits of agreement (LOA) are calculated based on the plot which must fall within the clinically accepted threshold range. | Day 1 |
| Bland-Altman Limits of Agreement (LOA) of the proportion of the stage 1 and 2 of Non Rapid Eye Movement (NREM) sleep | The mean and difference of the proportion of the stage 1 and 2 of NREM sleep according to both 5S Sleep Tracking Mat and PSG are analyzed from the raw data. The Bland-Altman plot is created with the mean on the x-axis and the difference on the y-axis. The mean difference and the 95% limits of agreement (LOA) are calculated based on the plot which must fall within the clinically accepted threshold range. | Day 1 |
| Bland-Altman Limits of Agreement (LOA) of the proportion of the stage 3 of Non Rapid Eye Movement (NREM) sleep | The mean and difference of the proportion of the stage 3 of NREM sleep according to both 5S Sleep Tracking Mat and PSG are analyzed from the raw data. The Bland-Altman plot is created with the mean on the x-axis and the difference on the y-axis. The mean difference and the 95% limits of agreement (LOA) are calculated based on the plot which must fall within the clinically accepted threshold range. | Day 1 |
| Bland-Altman Limits of Agreement (LOA) of the proportion of Rapid Eye Movement (REM) sleep | The mean and difference of the proportion of REM sleep according to both 5S Sleep Tracking Mat and PSG are analyzed from the raw data. The Bland-Altman plot is created with the mean on the x-axis and the difference on the y-axis. The mean difference and the 95% limits of agreement (LOA) are calculated based on the plot which must fall within the clinically accepted threshold range. | Day 1 |
| D020919 |
| Sleep Disorders, Intrinsic |
| D020920 | Dyssomnias |
| D012893 | Sleep Wake Disorders |
| D009422 | Nervous System Diseases |