Abstract
PURPOSE: Convergence insufficiency, the most common binocular vision anomaly, is characterised by a receded near point of convergence and an exophoria which is at least 4 prism dioptres (Δ) larger at near than at distance. However, the repeatability of standard heterophoria measures are poorly understood. This study assessed the ability of four common heterophoria tests to detect differences of 4Δ by evaluating the inter- and intra-examiner variability of the selected techniques. METHODS: Distance and near horizontal heterophorias of 20 visually-normal adults were measured with the alternating prism cover test, von Graefe prism dissociation, Howell Card and Maddox Rod by two examiners at two separate visits using standardised instructions and techniques. We investigated inter- and intra-examiner variability using repeatability and reproducibility indices, as well as Bland-Altman analysis with acceptable limits of agreement defined as ±2Δ. RESULTS: The Howell card test had the lowest intra-examiner variability at both distance and near, as well as the best 95% limits of agreement (±1.6Δ for distance and ±3.7Δ for near). Inter-examiner reproducibility results were similar, although at near the alternating prism cover test had better repeatability (1.1Δ, 95% confidence intervals -1.1Δ to 4.0Δ) than the Howell card (1.4Δ, 95% confidence intervals -1.9Δ to 5.9Δ). CONCLUSION: The low repeatability of many standard clinical heterophoria tests limits the ability to reliably detect a 4Δ difference. The Howell Card provided the most repeatable and reproducible results indicating that this technique should be used to detect small changes in heterophoria magnitude and direction.