CES-EES Clinical Score and a Wearable Ring Sensor for Obstructive Sleep Apnea: Individual and Combined Validation Against In-Laboratory Polysomnography

Obstructive sleep apnea (OSA) is a common sleep-related breathing disorder associated with cardiovascular, cerebrovascular, and other long-term health consequences if untreated. In-lab polysomnography (PSG) is the diagnostic reference standard but is resource-intensive, costly, and associated with long waiting times. Existing screening tools such as the STOP-BANG questionnaire and the Epworth Sleepiness Scale (ESS) have recognized sensitivity/specificity trade-offs. The Clinical Examination Score (CES) combined with the ENT Examination Score (EES) is a previously described clinical scoring tool for OSA screening. The Belun Ring (Belun Sleep System BLS-100) is an FDA 510(k)-cleared (Class II), Malaysia Medical Device Authority-registered, AI/deep-learning-facilitated wearable capable of estimating AHI, sleep staging, and OSA severity.

This is a prospective, comparative, single-centre study conducted at Universiti Malaya Medical Centre (UMMC). Adults referred to the Otorhinolaryngology sleep clinic for suspected OSA, with high ESS and/or STOP-BANG scores, are screened with the CES-EES; those scoring 4 or more and otherwise meeting eligibility criteria are enrolled. After informed consent, participants undergo a standardised history, demographic data collection, and physical examination (oral cavity examination, flexible nasopharyngeal endoscopy, dental assessment), followed by a single overnight Level 1 in-lab PSG (Compumedics, 7-channel) at the UMMC Neurology Sleep Lab. The Belun Ring is fitted to the participant's index finger concurrently with standard PSG electrodes, so both devices record simultaneously during the same overnight study. A minimum of 4 hours total sleep time (TST) is required; participants with less than 4 hours TST are excluded from analysis. Both PSG and Belun Ring AHI are scored using AASM 1A criteria (hypopnea defined by desaturation of 3% or more). PSG is manually scored by a sleep technician; Belun Ring output is processed by the device's AI algorithm and independently cross-checked by two investigators.

The primary comparisons are: (1) CES-EES score against PSG AHI; (2) Belun Ring AHI against PSG AHI; and (3) CES-EES combined with Belun Ring AHI against PSG AHI. Planned statistical analyses include the paired t-test or Wilcoxon signed-rank test for comparative analysis, Bland-Altman plots for agreement/bias analysis, the intraclass correlation coefficient (ICC) for inter-scorer reliability, and descriptive statistics (mean, median, SD, range).