Prevalence and Patterns of Refractive Errors Among College Students in Low- and Middle-Income Countries: A Systematic Review and Meta-Analysis

Background Uncorrected refractive error is the leading cause of correctable visual impairment worldwide, and students in visually demanding curricula may be at particular risk, yet routine vision screening is inconsistent across low- and middle-income countries (LMIC). We estimated the pooled prevalence and pattern of refractive error among college and university students in LMIC. Methods We searched PubMed/MEDLINE and complementary regional and web-based sources for cross-sectional studies reporting refractive error prevalence among college or university students in low- or middle-income countries (World Bank classification); the review was registered in PROSPERO (CRD420261475059). Study-level proportions were logit-transformed and pooled using a random-effects model with restricted maximum likelihood estimation and the Knapp–Hartung adjustment. Robustness was assessed by sensitivity, leave-one-out and subgroup analyses and under two alternative model specifications. Results Fourteen cross-sectional studies from 12 low- and middle-income countries, comprising 5,431 students, were included; 2,663 had refractive error of any type, 1,947 myopia, 908 astigmatism and 254 hyperopia. The pooled prevalence of any refractive error was 50.7% (95% CI 35.0–66.3%; 95% prediction interval 8.0–92.5%). Myopia predominated (28.7%, 18.9–41.1%), followed by astigmatism (11.6%, 5.4–23.0%) and hyperopia (4.2%, 2.3–7.6%). Heterogeneity was extreme for every outcome (I² 93.7–98.9%), and the pooled estimate was stable across all robustness analyses (46.7–54.6%). In exploratory, post-hoc subgroup analysis, pooled myopia prevalence was 24.9% where categories were mutually exclusive versus 44.3% where they could overlap (P = .032, unadjusted for multiple comparisons), with no difference for any refractive error (P = .533). No clear funnel-plot asymmetry was detected for the primary outcome; subtype asymmetry tests were difficult to interpret because the logit standard error depends on the proportion itself. Conclusions Around half of college students in the LMIC settings sampled have refractive error, predominantly myopia, but extreme heterogeneity means the pooled figure should be read alongside its wide prediction interval rather than as a generalisable value. These findings support consideration of routine, low-cost vision screening at tertiary-education entry and greater methodological standardisation in future primary studies.