Validation of Entropy Measurements Extracted From Portable Monitoring Devices for Diagnosing Obstructive Sleep Apnoea

Abstract

Obstructive sleep apnea (OSA) is a prevalent condition characterized by repeated episodes of partial or complete upper airway obstruction during sleep, significantly impacting sleep architecture leading to symptoms such as loud snoring, daytime sleepiness, and fatigue. OSA is also linked to diabetes, cardiovascular disease, road crashes, healthcare-related costs, and workplace productivity losses. Obesity is the primary risk factor for OSA. Timely diagnosis and treatment are essential to minimize the potential complications. Diagnosis primarily involves polysomnography (PSG), considered the gold standard, along with home based sleep tests. Furthermore, nonlinear methods such as entropy have gained popularity in diagnosing OSA due to their ability to assess the irregularities associated with this condition.

Aim: This study aimed to validate entropy measurements from two portable devices in individuals with OSA.

Methods: Sample entropy analysis was employed using the MATLAB R2024a (MathWorks, Natick, MA, USA) to assess the heart rate variability (HRV) and oxygen saturation (SpO2) measurements collected from a sample of 11 participants utilising both portable monitoring devices: WatchPAT (WP300; Itamar Medical Ltd., Caesarea, Israel) and Embletta (Embletta Gold, Natus Medical Incorporation). Bland-Altman analysis used to assess the agreement between sample entropy measurements.

Results: The Bland-Altman analysis presented in Figures 9 and 10 demonstrated a strong agreement in sample entropy values for HRV and SpO2 between the WatchPAT and Embletta devices. This suggests that sample entropy is a reliable method for evaluating and capturing important physiological data related to OSA. The consistent readings between the two devices not only validate sample entropy's effectiveness but also suggest that devices like WatchPAT, may offer greater patient satisfaction and reduce discomfort compared to the Embletta device.

Conclusion: This study demonstrates that sample entropy is a reliable tool for assessing HRV and SpO2, which are crucial for diagnosing OSA. Employing sample entropy allows for more streamlined and accessible screening processes, potentially making OSA diagnosis rapid and less invasive. The WatchPAT device may be preferred due to its simplicity and patient comfort. However, the study's limitations, including a small sample size, may affect the generalizability of the findings.