DETECTION OF VISUAL FIELD CHANGE IN CLINICAL TRIALS

用于检测青光眼视野丧失的自动视野检查已成为 在过去的十年中，大多数研究人员和临床医生都可以使用。 辨别视野患者的几种分析方法 已开发并测试了自动视野检查的缺陷 这一次。然而，用于识别现场缺陷是否 自纵向视觉以来，随着时间的推移，进展还没有得到很好的测试 有足够的跟踪调查和样本量的现场数据现在才出现 变得有空。国家眼科研究所目前赞助了三个 比较青光眼不同治疗方法的随机临床试验 病人。这些是晚期青光眼干预研究(AGIS)， 协作性初始青光眼治疗研究(CIGTS)和早期 临床青光眼试验(EMGT)试验。每项研究都开发了一种 确定进行性视野损失的不同方案。二 试验使用分级和计分领域的系统。比分是 用于识别在手术过程中视野恶化的患者 学习。第三种使用青光眼变化概率图 汉弗莱·斯塔帕克计划。然而，这三种不同的方法提出了 这些研究还没有通过将它们应用于大型 具有广泛随访的纵向数据集。我们建议评估 并比较了三种不同的视力评估方法 使用两个不同的纵向数据集的场级数 没有参与这些临床试验的患者。这将允许 辨析三者之间的异同 使用相同数据集的方法。

Automated perimetry to detect visual field loss in glaucoma has become available to most researchers and clinicians in the past decade. Several analytic methods for identifying patients with visual field defects on automated perimetry have been developed and tested during this time. However, methods for identifying whether field defects progress over time have not been well tested since longitudinal visual field data of sufficient follow-up and sample size are only now becoming available. The National Eye Institute currently sponsors three randomized clinical trials to compare different treatments for glaucoma patients. These are the Advanced Glaucoma Intervention Study (AGIS) , Collaborative Initial Glaucoma Treatment Study (CIGTS) and the Early Manifest Glaucoma Trial (EMGT) trials. Each study has developed a different protocol for determining progressive visual field loss. Two trials use a system of grading and scoring fields. The scores are to be used to identify patients whose visual field worsens during the study. The third uses the Glaucoma Change Probability maps of the Humphrey Statpac program. However, the three different methods proposed by the studies have not been evaluated by applying them to a large longitudinal data set with extensive follow-up. We propose to evaluate and compare the three different approaches to the assessment of visual field progression using two different longitudinal data sets of patients not involved in these clinical trials. This will allow for the identification of similarities and differences between the three approaches using the same data sets.