Date Rishabh C, Shen Kevin L, Shah Beena M, Sigalos-Rivera Mara A, Chu Yvonne I, Weng Christina Y
Baylor College of Medicine, Department of Ophthalmology, Houston, Texas; Ben Taub General Hospital, Department of Ophthalmology, Harris Health System, Houston, Texas.
Baylor College of Medicine, School of Medicine, Houston, Texas.
Ophthalmol Retina. 2019 Apr;3(4):343-349. doi: 10.1016/j.oret.2018.12.003. Epub 2018 Dec 24.
To determine the accuracy of a county teleretinal screening program of detecting referable diabetic retinopathy (DR) and treatable diabetic macular edema (DME), as well as to evaluate patient compliance with clinic follow-up after referral from teleretinal screening.
Retrospective observational study.
Patients in the Harris Health System (HHS, Houston, TX) older than 18 years of age who underwent teleretinal screening between July 2014 and July 2016.
Teleretinal imaging (TRI) consisting of single-field 45-degree nonmydriatic color fundus photography with referral thresholds of severe nonproliferative DR, proliferative DR, and significant DME. Teleretinal imaging results for all referred subjects were obtained and cross-referenced with dilated fundus examination findings with regard to DR severity and the presence of DME. Follow-up status was also noted. Subjects underwent OCT if deemed necessary by the examining specialist. Agreement between TRI and dilated fundus examination (DFE) findings was determined by calculating the Cohen κ coefficient.
The primary outcome measure is agreement between TRI results and DFE findings with regard to DR severity and the presence of DME. The secondary outcome measure is compliance with follow-up.
Of 1767 patients who were screened and referred for clinical examination, 935 (52.9%) attended their clinic appointment. Overall agreement between DFE and TRI was moderate (weighted κ 0.45) in terms of DR severity. There was agreement within one DR severity level in 86.2% of patients. The positive predictive value for detecting referable disease was 71.3%. Of patients referred for DME, 30.4% were deemed to have treatable DME.
The HHS teleretinal screening program demonstrates a high level of accuracy in the detection and classification of referable DR, but a lesser degree of accuracy in the detection of treatable DME. Only slightly more than half of participants were compliant with follow-up after a TRI referral. This large-scale study provides insight into the utility of teleretinal screening in a county health care system.
确定县级远程视网膜筛查项目检测可转诊糖尿病视网膜病变(DR)和可治疗糖尿病黄斑水肿(DME)的准确性,并评估患者在远程视网膜筛查转诊后对临床随访的依从性。
回顾性观察研究。
哈里斯健康系统(HHS,得克萨斯州休斯顿)中2014年7月至2016年7月期间接受远程视网膜筛查的18岁以上患者。
远程视网膜成像(TRI)包括单视野45度非散瞳彩色眼底摄影,转诊阈值为重度非增殖性DR、增殖性DR和显著DME。获取所有转诊受试者的远程视网膜成像结果,并与散瞳眼底检查结果就DR严重程度和DME的存在情况进行交叉对照。同时记录随访状态。检查专家认为必要时,受试者接受光学相干断层扫描(OCT)。通过计算Cohen κ系数来确定TRI与散瞳眼底检查(DFE)结果之间的一致性。
主要观察指标是TRI结果与DFE结果在DR严重程度和DME存在情况方面的一致性。次要观察指标是随访依从性。
在1767名接受筛查并被转诊进行临床检查的患者中,935名(52.9%)参加了门诊预约。就DR严重程度而言,DFE与TRI之间的总体一致性为中等(加权κ 0.45)。86.2%的患者在一个DR严重程度级别内存在一致性。检测可转诊疾病的阳性预测值为71.3%。在转诊DME的患者中,30.4%被认为患有可治疗的DME。
HHS远程视网膜筛查项目在可转诊DR的检测和分类方面显示出较高的准确性,但在可治疗DME的检测方面准确性较低。TRI转诊后只有略多于一半的参与者依从随访。这项大规模研究为县级医疗保健系统中远程视网膜筛查的效用提供了见解。