Our study will enroll adults 20 years or older diagnosed with moderate to severe OSA (Apnea-Hypopnea Index ≥15) confirmed by polysomnography who are CPAP therapy naive at the time of enrollment.

We will exclude those with previous use of positive pressure ventilators, a diagnosis of central sleep apnea, a presence of uncontrolled acute mental illness, or a terminal illness.

The sample size was determined based on a power analysis conducted for the CPAP adherence rate. Effect sizes were estimated to be 0.32 based on our previous findings, which indicated that 32.25% of CPAP-related side effects were associated with the CPAP interface [17]. A power of 0.85 was used to detect a significant difference at P=.05 (2-sided). Given the highly variable dropout rate for CPAP therapy, we aimed to accommodate a 20% dropout rate, which necessitated recruiting 80 participants, with 40 participants in each group.

Figure 1 illustrates the proposed system architecture, consisting of a camera module, an Edge-AI computing module, and a display module. The camera module captures the user’s facial and hand images, while the Edge-AI computing module processes these images to detect relevant features and perform calculations to ensure a proper nasal mask fit and verify the accuracy of the wearing procedure. The display module presents the measurement results and offers real-time feedback on the recommended nasal mask size, following the mask sizing guidelines provided by ResMed [18].

Participants assigned to the standard care group will have their mask size visually checked by a case manager, who will also teach them how to use the CPAP machine.

CPAP adherence will be monitored daily using AirView, a cloud-based remote monitoring system. The research team will explain the study objectives to participants and obtain their written consent to authorize ResMed to grant access to their CPAP cloud data. This approval will enable the research team to retrieve and analyze adherence data with company authorization.

Data from CPAP devices (eg, AirSense 10 AutoSet) will be collected at T1, T2, and T3, tracking the percentage of nights meeting adherence criteria. Adherence refers to the patient’s adherence to CPAP therapy, which means using CPAP for more than 70% of the total sleep time during the night, with a minimum duration of 4 hours. The calculation method involves dividing the number of days with CPAP use exceeding 4 hours by the total number of days and multiplying by 100 to obtain the average percentage of CPAP use [19].

We developed a self-assessed CPAP adherence and comfort levels questionnaire to allow participants to track their daily adherence and comfort levels, providing a comprehensive overview of their subjective CPAP experience.

This structured daily assessment enables a detailed evaluation of patient experiences, helping to identify potential barriers to CPAP use and providing valuable insights into factors influencing adherence and comfort over time.

Common discomforts of using CPAP were collected and categorized based on symptoms identified by Australian scholars Ghadiri and Grunstein [6].

At the end of the study, participants will fill out a system usability scale (Multimedia Appendix 2) to evaluate system usability. This scale is a questionnaire assessing the usability of emerging technologies or systems, consisting of 10 items scored from 0 to 4 (from strongly disagree to strongly agree). The overall score is then multiplied by 2.5, with a maximum score of 100. According to Sauro [20], based on nearly 500 studies, a score of 68 is used as the benchmark for the system usability assessment: if the score is greater than 68, it indicates that the system’s usability is above 50%; if it is less than 68, the opposite is true.

The AR system’s recommended nasal size compared to the case manager’s estimation shows an agreement rate approaching 95% at T0.

We will calculate the time during the initial education session, hands-on training, patient questions and answers, and troubleshooting guidance at T1.

We will collect participants’ gender, height, weight, waist circumference, blood pressure values, education level, presence of other comorbidities (eg, hypertension, diabetes, cardiovascular diseases, chronic obstructive pulmonary disease, asthma, cancer), smoking habits (none, once a week, once a month, once a year), drinking habits (none, once a week, once a month, once a year), exercise habits (none, once a week, once a month, once a year), history of being diagnosed with mental health disorders, napping habits, and the approximate duration of naps.

The Apnea-Hypopnea Index, average blood oxygen saturation, and minimum blood oxygen saturation will be measured through polysomnography and a CPAP machine.

This study was approved by the Institutional Review Board of National Cheng Kung University on March 1, 2023 (approval A-ER-111-531). The study adheres to all institutional guidelines and ethical principles outlined in the Declaration of Helsinki.

All participants will provide written informed consent before enrollment. They will receive detailed explanations regarding the study objectives, procedures, and potential risks and benefits. Participants will have the right to withdraw from the study at any time without consequences.

All collected data will be anonymized and deidentified to protect participant privacy. Identifiable information will be securely stored and accessed only by authorized research personnel. Data will be reported in an aggregate format, ensuring that individual participants cannot be identified.

Participants will receive financial compensation (US $6) for T2 and T3 participation. Besides that, they will be provided with comprehensive CPAP education and follow-up support as part of the study. No additional costs will be incurred by participants for their involvement.

We confirm that no identifiable information of individual participants or users is present in any images included in the manuscript or supplementary materials.