Participation in population-based eye health surveys is lower but more gender-balanced in high- compared to low- and middle-income countries

<ns5:p>Background Population-based eye health surveys are an important source of evidence describing the eye health needs of a population, although differences in participation rates between population groups may lead to response bias. In some high-income settings, participation in surveys has decreased over time, although whether this has occurred for eye health surveys is unclear. The aim of this study was to compare participation in population-based eye health surveys conducted over the previous ~20 years between country income levels and different population sub-groups. Methods Participation rates were extracted from an existing database of all population-based eye health surveys undertaken anywhere in the world between 2000–2023 that estimated the prevalence of vision impairment. Overall participation and participation rates disaggregated between population groups (identified using the PROGRESS+ framework) were extracted and compared between country income levels. Results Participation was reported by 374 (86.6%) of the 432 included eye health surveys, was generally high (median=89.5%, interquartile range [IQR]: 81.1-94.6%) and did not change over the study period. Participation was lowest in high-income countries (79.2%, IQR 70.4-90.0%). Just over one-quarter of studies (n=125, 28.9%) disaggregated participation by gender/sex. In these studies, women were more likely to participate than men in low income (women: 90.5%, IQR 75.3-96.2 vs men: 86.8%, IQR 59.1-90.1%), lower-middle income (92.2%, IQR 86.1-95.5% vs 88.8%, IQR 80.1-93.4%), and upper-middle income (91.2%, IQR 85.8-94.6% vs 86.3%, IQR 77.4-90.1%) countries, but not in high-income countries where participation was similar (87.9%, IQR 80.8-95.4% vs 89.3%, IQR 79.1-92.9%). Participation rates disaggregated by age were reported by 68 surveys (15.7%) and participation was not significantly associated with age. Participation disaggregated between other population sub-groups was infrequently reported (&lt;10% of surveys) and could not be meaningfully compared. Conclusions Researchers could consider alternative recruitment strategies that achieve better participation in high income countries and for men in low- and middle-income countries.</ns5:p>