This international, ambispective, real-world, cohort study is being conducted in 4 European countries (France, Belgium, Spain, and Portugal). It includes a retrospective analysis with the reuse of home support provider data routinely collected for PAP-related services and a prospective evaluation with the collection of patient data using a web-based patient questionnaire.

The main objective of the study is to compare PAP adherence and the PAP therapy persistence rate between countries and between different levels of home support provider service.

The study’s other objectives are to (1) identify service components that influence PAP adherence and persistence; (2) identify factors associated with PAP adherence and persistence; (3) describe in each country sociodemographic and OSA-related clinical characteristics, home support provider PAP-related service, OSA and PAP literacy of participants (perceived and actionable competence), patient experience and satisfaction with PAP therapy and home support provider support, patient-reported OSA symptoms, and quality of life (determined using the EQ-5D-5L questionnaire [32]); and (4) describe patterns of PAP use.

The study protocol was developed in collaboration with a scientific committee composed of experts in OSA and methodologists from the 4 countries involved. The regulatory process and patient consent for a study with the reuse of data and a questionnaire were different for the 4 countries.

In France, the study followed the methodology MR-004 of the French Data Protection Authority (Commission Nationale Informatique et Liberté) concerning research that reuses data that have already been collected [33] and was approved by the local ethics committee of Angers University Hospital in November 2023 (2023-159). In Portugal, the study was considered observational and required approval by a central ethics committee; the approval was obtained from the ethics committee of Lisboa in May 2024 (24/24). In Spain, the study was considered observational and required approval by the ethics committee at each participating hospital; approval was obtained from the CEIm Hospital Universitario Ramón y Cajal of Madrid in July 2024 (095/24) and from the CEIm Hospital Universitario Doctor Peset of Valencia in December 2024 (127/24) and is pending for other regions. In Belgium, the study was considered interventional and required approval by a central ethics committee (Commissie voor Medische Ethiek Ziekenhuisnetwerk Antwerpen Institutional Review Board, Ziekenhuis aan de Stroom Middelheim) and by the local ethics committee at each participating hospital. Ethics approvals from the belgian central ethics committee are still pending.

In France, potential participants have the option to decline the reuse of their personal data through an “opt-out” process. Conversely, in Belgium, Portugal, and Spain, an “opt-in” process has been established, requiring explicit consent from participants to join the study.

No compensation is provided to participants.

Study outcomes, time points, and data sources are summarized in Table 1. An ePRO tool will be used to administer the patient questionnaire. The ePRO tool was designed to assess various aspects of a patient’s status, condition, and experience at 2 time points: the time of ePRO completion and the beginning of PAP therapy. This included sociodemographic characteristics (age, relationship status, employment, education level, and social support), health status (BMI, smoking status, alcohol intake, comorbidities, sleep duration, and use medications to aid sleep), OSA-related factors (when PAP therapy was started, apnea-hypopnea index before starting PAP, and OSA symptoms), PAP therapy experience and satisfaction (tolerability of PAP, attitude to PAP, device noise, and level of motivation to use PAP), patient experience and satisfaction with home support providers, quality of life (including current health status), and OSA-PAP literacy. The full ePRO questionnaire is provided in Multimedia Appendix 1.

Home support provider service will be assessed using 4 components: frequency of interactions between the home support provider and a patient; content of support (logistical, educational, supportive, or behavioral); personalization of support according to patient profile or needs; and telemonitoring (telemetry, data visualization, and alerts triggering home support provider interventions).

Potential participants will be identified by applying the inclusion criteria to the data sourced from the ALH Enterprise Resource Planning (ERP) systems. Initially, potential participants will receive an email or SMS invitation to connect to a secure platform for the consent process. Those who provide consent will be granted access to the ePRO tool and will be invited to complete the questionnaire, which takes about 10 minutes.

The study database will be created from 3 data sources: the ALH ERP systems that contain data routinely collected for PAP-related services; the database containing information about PAP device use (collected remotely through the Internet of Things or by download from device memory cards during home provider support visits); and the ePRO tool. Database structure and content will be standardized across countries.

Because the ALH ERP and PAP device use databases are primarily designed for administrative, logistical, and patient follow-up purposes, rather than research, all data will undergo thorough review for completeness, accuracy, and reliability before statistical analysis. These reviews will be conducted in accordance with a detailed data management plan, involving multiple experts with complementary areas of expertise. All decisions made during the review process will be fully documented.

All identifiable personal information will be protected and stored in the study database in pseudonymized form so that it cannot be viewed by the study sponsor or any external vendors that are processing the study data. A summary of data sources and data flow is provided in Figure 1.

While the association between PAP adherence and costs has been demonstrated in the United States [21,23,37-42], data from Europe remain limited [43,44]. To further determine the benefits of home support providers, an ancillary study, named APAIR (Association Between Positive Airway Pressure Adherence, Morbi-Mortality and Costs Among Patients With Obstructive Sleep Apnea in France: Real World Study Using Claims and Telemonitoring Data), will be conducted in the subset of AWAIR study participants from France. APAIR will investigate the association between PAP adherence and persistence, and key outcomes (mortality, morbidity, and health-related costs). AWAIR study participants from France will be identified in the French Healthcare Claims Database (Système National des Données de Santé), which provides longitudinal patient-level data for approximately 99% of the French population [45]. Data on hospitalizations and sick leave related to the complications of OSA will be retrieved for these individuals. The aim of APAIR is to obtain data relating to new and more relevant thresholds for PAP adherence and the personalization of PAP therapy support (eg, for those with higher health risks or higher health care resource use).

The high and increasing number of patients with OSA treated with PAP therapy drives concerns among health systems regarding the associated health care resource consumption and costs. This has led some countries to implement significant changes in the patient pathway, notably toward more home-centered models [30,46]. Regardless of the care setting, PAP adherence and persistence are essential because they are associated with health benefits [19], whereas poor adherence and PAP therapy terminations have been associated with increased disability, health care resource consumption, and direct or indirect costs [47-49].

Despite a large body of evidence regarding the factors associated with PAP adherence [50] and persistence [25,51], and the reimbursement of PAP-related services in many European Union countries, there is currently a lack of evidence regarding the impact of home support provider services on these outcomes. In addition, although defined by country or local regulations, the level of home support provider service is poorly described. Differences exist in service frequency, content, and communication channels between home support providers and patients, the use of PAP telemonitoring, and the qualifications of home support provider personnel. Some of these components have been shown to impact PAP adherence [29,30,52,53].

Our working hypothesis is that a higher level of home support provider service would be associated with better PAP adherence and persistence. This study will also investigate between-country differences in terms of patient characteristics, OSA-related characteristics, and services, and identify which components of service are most effective at improving adherence and persistence.

We anticipate that this research will provide valuable evidence to inform decision-making by authorities and payers regarding the regular reassessment of services and associated tariffs. It could also support guidelines relating to the follow-up of individuals receiving PAP therapy.

The comparative, real-world design with between-country comparisons was the most feasible approach for this study. A randomized controlled trial was not viable because home support provider services are implemented for all patients prescribed PAP. The need to include a control group without services in a randomized trial would have markedly reduced ethical approval and acceptance of the study and would have introduced important bias related to behavioral factors for those accepting home support provider management. Furthermore, the controlled conditions of clinical trials do not adequately capture the complexity of factors influencing adherence in real life. Furthermore, there was a need for evaluation of the real-world impact of home support providers on a representative, unselected population. Real-world studies provide information about outcomes with less rigid treatment patterns compared with a controlled trial [54] and bridge the gap between clinical research and everyday practice. We followed the Haute Autorité de Santé methodological recommendation for real-world studies, proposing a design consistent with identified questions, using preexisting high-quality data, integrating PROs and PRE, and ensuring transparency [55].

One of the major strengths of this study is its large sample size, which enhances the representativeness of the population and the generalizability of the findings. The retrospective design allows for long-term evaluation of PAP adherence and persistence and ensures that there is no influence on participants’ behavior because data were collected prior to their inclusion in the study. In addition, the AWAIR study will describe between-country differences in home provider services and patient characteristics for the first time, and report on patient satisfaction and experience with PAP therapy.

In contrast with conventional clinical trials, the real-world setting ensures that the results will reflect the usual course of care [56]. This is particularly important for evaluating adherence because it minimizes the risk of bias that could arise from study participation. Furthermore, real-world evidence is particularly relevant for meeting health authority post-approval effectiveness requirements [57]. Well-conducted comparative effectiveness research can provide valid results similar to randomized trials for this measure of known confounding factors [56], and can use methods (such as stratification, matching, or regression analyses) to account for or adjust for such factors and yield valid results [58]. In the real-world comparative AWAIR study, adjustment for potential confounders identified through literature review and clinical expertise, combined with the use of hierarchical models accounting for contextual differences related to the international nature of the study, will help the data to provide an accurate assessment of the net effect of home provider services on PAP adherence and persistence while minimizing the potential biases inherent in real-world studies. Additionally, the use of an ePRO tool facilitates the collection of data on confounding factors that would not otherwise be available in the retrospective database analysis, along with PROs, PREs, and health literacy across a large and diverse patient population from various geographic regions [59].

Selection bias cannot be entirely ruled out because participants were required to have a valid email address or smartphone number to be eligible for the study. This may lead to underrepresentation of certain subgroups of PAP users, such as older adults. Patients who answer the questionnaire might also differ from those who do not answer it. In addition, this study will only include individuals who use PAP supplied by a home support provider affiliated with ALH and it is possible that the level or content of additional services may differ slightly from other home support providers in the same country. To assess the potential impact of such biases, we will compare the characteristics of our study population with published data on OSA, focusing on age, sex, and disease severity.

Data from the ePRO tool may be subject to recall bias and inaccuracy because participants will complete the web-based questionnaire at variable periods of time since the start of PAP therapy (up to 5 years later for those who started PAP in 2019) and medical data will directly be provided by the patient. Confounding bias other than the key confounding factors considered is another limitation, particularly with respect to medical follow-up (data that will not be collected in this study).

Finally, the period over which this study was conducted included times before and during the COVID-19 pandemic (including lockdowns), and the worldwide recall of a PAP device made by Philips Respironics [36]. Sensitivity analyses will be performed to determine whether these factors had any influence on the overall study findings.

The study faced several challenges, including the regulatory classification of an ambispective study and the heterogeneous regulatory processes between countries. In France, the reuse of data for clinical research is facilitated by a simplified procedure that does not exist in other countries. Although the decentralized approach used in this study (ie, e-consent + ePRO tool) has advantages, especially in terms of managing a large number of participants, proposing the study to potential participants remotely can decrease the rate of enrollment. This is why a video-based approach was chosen because using videos to explain the research aims and procedures to potential participants of a clinical trial was preferred by patients over paper-based systems and improved their understanding [60].

Another challenge is the quantity of data to be processed due to the large number of participants and the daily PAP data collected over up to 5 years, requiring robust data management and quality control systems. Moreover, data were collected for routine home support provider services, in different countries with different services and enterprise resource planning systems, with different data collection processes, which adds another layer of complexity to the study. Therefore, those involved in the study need to have a good understanding of the data collection processes, data structures, formats, contents, and constraints of each system to allow standardization of data to constitute a research-ready database.

Clinical trials are critical in the assessment of health products and services. However, managing and executing randomized controlled trials pose significant challenges in terms of inclusion, data collection, follow-up, comparator group, and costs [61]. Specifically in patients with OSA treated with PAP, it was shown that recruitment strategies in randomized trials could result in a lack of representativeness of the participants, because only up to 20% of typical patients with OSA meet the eligibility criteria, limiting the generalizability of the results [62]. The expanding availability of real-world data combined with advancements in operational tools and methodologies presents a significant opportunity to develop alternative study designs that leverage existing data. Furthermore, the European Health Data Space will create a common framework in Europe for the reuse of health data for research [63]. The design of the AWAIR study integrates multiple real-world data sources (including operational databases, device data, and an ePRO tool), the use of comparative real-world methods, and the implementation of e-consent. This innovative approach not only makes the study more robust but also sets a precedent for future research in similar fields, particularly regarding the number of participants who will consent and complete the ePRO tool. The design allows the study to be continuously enriched with data from new countries to measure the effect of other service levels and also to extend patient follow-up to assess longer-term persistence or include other patients in the countries of interest to assess the impact of changes in service levels.

The AWAIR study has a unique design, leveraging an unprecedented number of eligible participants, decentralized technologies, and a real-world comparative methodology across multiple countries. This approach will highlight intercountry differences in terms of patient characteristics, PAP adherence and persistence, PROs, PREs, and satisfaction with home support provider services. By assessing the added value of home support provider services, the results will support best practices for patient management and decision-making by health authorities and payers.