Determining the variability associated with visual acuity and refractive error measurements: a systematic review

The goal of this systematic review was to evaluate the measurement variability of distance visual acuity (VA) and refractive error (RE). Accurately measuring VA and RE is fundamental to eyecare, however, no scientific consensus has been reached on the measurement variability. Scientific literature suggests limits of variability of ±0.15 Logarithm of the Minimum Angle of Resolution (LogMAR) for VA and ±0.5 Diopters (D) for RE. This lack of consensus on variability affects our clinical decision-making, regulations, and research. This systematic review was performed in accordance with PRISMA guidelines and was registered at PROSPERO (ID CRD42024554663). We searched the Medline, PubMed, and Embase databases. The Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool was then used to assess the risk of bias (RoB). We included studies of adults that directly compared distance VA or RE and reported the Limits of Agreement (LoA) and mean difference. An overall summarized mean is reported in LoA and the certainty of evidence was assessed using GRADE. After applying the eligibility criteria, 12 studies reporting on VA and 6 studies reporting on RE were included. Six out of the 18 studies (33%) scored a low RoB in at least three out of four QUADAS-2 domains. In 25 out of 36 subgroups (69%), the LoA exceeded the suggested clinical accepted variability ranges for VA (±0.15LogMAR) or RE (±0.5D). The overall summarized mean is ±0.20LogMAR (95%CI 0.17;0.23) for VA and ±0.70D (95%CI 0.50;0.89) for RE. The GRADE certainty of evidence was very low. This study demonstrates that distance VA and RE measurements have a high measurement variability in adults, exceeding the suggested limits of variability (VA: ±0.15LogMAR, RE: ±0.5D). The evidence either fails to support the suggested limits of variability or is based on studies with methodological weaknesses, with a very low certainty of evidence. Methodological rigorous studies are therefore needed to accurately estimate the suggested limits of variability. For now, we propose using the overall summarized mean LoA obtained in our study as a provisional frame of reference (i.e. ±0.20LogMAR and ±0.70D).