Narayana Nethralaya Postgraduate Institute of Ophthalmology, Bangalore, India.
Indian J Ophthalmol. 2014 Jan;62(1):41-9. doi: 10.4103/0301-4738.126178.
To report the Karnataka Internet Assisted Diagnosis of Retinopathy of Prematurity (KIDROP) program for retinopathy of prematurity (ROP) screening in underserved rural areas using an indigenously developed tele-ROP model.
KIDROP currently provides ROP screening and treatment services in three zones and 81 neonatal units in Karnataka, India. Technicians were trained to use a portable Retcam Shuttle (Clarity, USA) and validated against ROP experts performing indirect ophthalmoscopy. An indigenously developed 20-point score (STAT score) graded their ability (Level I to III) to image and decide follow-up based on a three-way algorithm. Images were also uploaded on a secure tele-ROP platform and accessed and reported by remote experts on their smart phones (iPhone, Apple).
6339 imaging sessions of 1601 infants were analyzed. A level III technician agreed with 94.3% of all expert decisions. The sensitivity, specificity, positive predictive value and negative predictive value for treatment grade disease were 95.7, 93.2, 81.5 and 98.6 respectively. The kappa for technicians to decide discharge of babies was 0.94 (P < 0.001). Only 0.4% of infants needing treatment were missed.The kappa agreement of experts reporting on the iPhone vs. Retcam for treatment requiring and mild ROP were 0.96 and 0.94 (P < 0.001) respectively.
This is the first and largest real-world program to employ accredited non-physicians to grade and report ROP. The KIDROP tele-ROP model demonstrates that ROP services can be delivered to the outreach despite lack of specialists and may be useful in other middle-income countries with similar demographics.
报告印度卡纳塔克邦互联网辅助早产儿视网膜病变（ROP）筛查计划（KIDROP），该计划利用本土开发的远程 ROP 模型，在服务不足的农村地区进行ROP 筛查。
KIDROP 目前在印度卡纳塔克邦的三个地区和 81 个新生儿单位提供 ROP 筛查和治疗服务。技术人员接受了使用便携式 Retcam Shuttle（美国 Clarity）的培训，并与进行间接检眼镜检查的 ROP 专家进行了验证。一个本土开发的 20 分评分（STAT 评分）根据三向算法评估他们进行成像和决定随访的能力（I 级到 III 级）。图像也被上传到一个安全的远程 ROP 平台，并由远程专家通过智能手机（iPhone，苹果）进行访问和报告。
对 1601 名婴儿的 6339 次成像进行了分析。三级技术员与所有专家决策的 94.3%一致。治疗级疾病的敏感性、特异性、阳性预测值和阴性预测值分别为 95.7%、93.2%、81.5%和 98.6%。技术员决定婴儿出院的kappa 值为 0.94（P<0.001）。只有 0.4%需要治疗的婴儿被漏诊。在 iPhone 上报告和 Retcam 上报告需要治疗和轻度 ROP 的专家之间的kappa 一致性分别为 0.96 和 0.94（P<0.001）。
这是第一个也是最大的在现实世界中使用经过认证的非医生来分级和报告 ROP 的计划。KIDROP 远程 ROP 模型表明，即使缺乏专家，ROP 服务也可以在缺乏专家的偏远地区提供，并且可能对其他具有类似人口统计学特征的中等收入国家有用。
