The present meta-analysis included factors associated with the purchase of a CPAP machine in patients with OSA. The types of studies conducted in adult patients with OSA included: Randomized controlled trials, observational studies or descriptive studies comparing factors between those who purchased CPAP and those who did not. The diagnosis of OSA was made using polysomnography upon evidence of an apnea-hypopnea index (AHI) or a respiratory disturbance index (RDI) of five or more events/h. After being diagnosed with OSA, CPAP trial or CPAP titration was initiated either in a sleep laboratory or at home. CPAP titration is performed with the aim of identifying the appropriate CPAP pressure for each patient. A decision to purchase a CPAP machine is made after CPAP titration. Studies conducted in adult patients with OSA that had received CPAP titration either in a sleep laboratory or at home were included.

In the present meta-analysis, five databases were searched: PubMed (https://pubmed.ncbi.nlm.nih.gov), Central database (www.cochranelibrary.com/central), Scopus (www.scopus.com), CINAHL Plus (https://web.s.ebscohost.com/) and Web of Science (www.webofknowledge.com). The search terms used were ‘obstructive sleep apnea’, ‘sleep apnea syndrome’, ‘predict^*’, ‘independent’, ‘factor^*’, ‘variable^*’, ‘purchase^*’, ‘buy’, ‘bought’, ‘pay’, ‘paid’, ‘expend^*’ and ‘spend^*’, where ‘^*’ was used to search for all terms which began with the preceding letters. The full list of search terms is shown in Table SI, Table SII, Table SIII, Table SIV and Table V. The final search was performed on February 8, 2021. After the duplicates were removed, initial screening was performed for non-relevant articles. Studies were considered relevant if they had been conducted to evaluate different factors between patients who purchased a CPAP machine and those who did not. Data extraction and full text reports were reviewed by two independent authors (BS and KS). Of these, the articles that met the study criteria were included in the final analysis (14).

The studied variables included both marketing and clinical factors. The definitions of the studied variables were as follows: Average to high income determined in each study, which may vary among countries; health insurance covering CPAP machine costs, which indicated cost reimbursement by the insurer; and Epworth Sleepiness Scale (ESS) score, which is a subjective sleepiness evaluation method with a value range of 0-24(15). The primary outcome assessed were factors that varied between patients who purchased a CPAP machine and those who did not. At least two studies were required to calculate the differences of the studied variables between the two groups. The mean differences were calculated between the two groups with a 95% CI for numerical factors and the odds ratio with 95% CI for categorical outcomes. Heterogeneity was computed and reported as I squared (I²). A forest plot of each comparison was created based on I². If I² was ≤75%, fixed effect was used. For factors displaying an I² value of >75%, the random effect was used.

The Newcastle-Ottawa Scale, adapted for cross-sectional studies, was used to evaluate the study quality of observational studies (16). The scale comprised three categories: Selection process, comparability and outcome measurement, with a score of 5, 2 and 3 points, respectively. The total score was 10 points and classified as very good (9-10 points), good (7-8 points), satisfactory (5-6 points) and non-satisfactory (0-4 points). Study quality was evaluated independently by two authors (BS and CN). Disagreements between these two authors were discussed and a final decision was made by a third reviewer (KS). All analyses were performed using Review Manager 5.4 (Copenhagen, The Nordic Cochrane Centre, The Cochrane Collaboration, Denmark).

There were 598 articles from five databases, which met the search criteria (Fig. 1). After duplicate article removal, 390 articles were included in the screening process. There were 12 articles found to be eligible for full text evaluation. A total of four studies were excluded: Two studies were excluded as CPAP compliance was evaluated, not purchasing of CPAP machines; one study was excluded as the rate of CPAP purchasing did not include variable evaluation; and one study was excluded as only the abstract was available, which was from a conference. In total, there were eight eligible studies involving 1,605 patients from four countries (6-10,17): Five studies from Israel (11,16-20), one from Mexico (12), one from Poland (17) and one from Thailand (10). The most recent study was published in 2018 by Sawunyavisuth (10) from Thailand (Table I).

The most common study design was cross-sectional and was used in three of the studies (10,13,18). The diagnosis of OSA was made using polysomnography; however, the inclusion criteria for OSA were variable, with the highest AHI of 30 events/h (11). The duration of CPAP titration or trial, prior to CPAP purchase, was a maximum of 2 weeks (11,13,18). Not all patients with OSA, in three of the studies, underwent CPAP titration [Brin et al (18), 183/400 patients; Shahrabani et al (19), 150/194 patients; Simon-Tuval et al (13), 132/162 patients]. In these three studies, data for analysis were obtained from the total population in the study. Most studies used self-administered questionnaires or telephone interviews to record patient information. The numbers of participating patients varied between 50(20) and 400(18), with an average CPAP purchasing rate of 49.8%. Only one study evaluated marketing strategies on CPAP purchasing (10).

There were 11 variables available for comparison between patients who purchased a CPAP machine and those who did not (Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Fig. 10, Fig. 11 and Fig. 12). These factors were age (Fig. 2), sex (Fig. 3), years of education (Fig. 4), living with a partner (Fig. 5), income (Fig. 6), insurance (Fig. 7), smoking (Fig. 8), hypertension/cardiovascular disease (Fig. 9), ESS (Fig. 10), body mass index (Fig. 11) and AHI/RDI (Fig. 12). A total of six factors were found to be significantly different between both groups: Age (Fig. 2), years of education (Fig. 4), income (Fig. 6), smoking (Fig. 8), ESS (Fig. 10) and AHI/RDI (Fig. 12). AHI/RDI was significantly different between the two groups with the highest mean difference of 10.40 events/h (95% CI, 4.95-15.86) as shown in Fig. 12. Patients who purchased a CPAP machine were older (by 1.11 years), had more years of education (0.93 years more), were smoking more (by 1.15 pack/year), and had higher ESS (by 0.61) and AHI/RDI (by 10.40) scores than those who did not purchase a CPAP machine, as shown in Figs. 2, 4, 8, 10 and 12, respectively. Additionally, those who purchased a CPAP machine had a 1.47 times higher income than those who did not (Fig. 6). The quality of most of the studies was evaluated as satisfactory, except for the longitudinal study by Tarasiuk et al (11), which was considered as good, with a score of 7/10 (Table II). Furthermore, the study by Byśkiniewicz (17) was not scored as it was published in Polish.