Two-field non-mydriatic fundus photography for diabetic retinopathy screening: a protocol for a systematic review and meta-analysis
Dongjing Yu,Xiaoyan Dou,Jiao Chen,Ying Lu,Baikang Ye,Xiaojun Wu,Zijing Wu,Qi Li,Xiaohe Tian,Bo Zhou,Ying Deng,Wei Li,Xinglin Hu,Lisha Mou,Zuhui Pu
DOI: https://doi.org/10.1136/bmjopen-2021-051761
IF: 3.006
2021-10-01
BMJ Open
Abstract:Introduction Diabetic retinopathy (DR) is one of the most prevalent microvascular complications of diabetes mellitus. Guidelines for DR screening in different countries vary greatly, including fundus photography, slit-lamp biomicroscopy, indirect ophthalmoscopy, Optical Coherence Tomography (OCT), OCT-A and Fundus Fluorescein Angiography (FFA). Two-field non-mydriatic fundus photography (NMFP) is an effective screening method due to its low cost and less time-consuming process. However, it is controversial due to the sensitivity and specificity of two-field NMFP. This review intends to evaluate the performance of the two-field NMFP in diagnosing DR and helps clinicians determine the most optimal screening method. Methods and analysis Two reviewers will independently search on the Medline, Embase, Cochrane databases, ProQuest, Opengrey, Chinese National Knowledge Infrastructure, Wanfang Data, VIP China Science and Technology Journal Database, Chinese BioMedical Literature Database, ISRCTN, ClinicalTrials.gov and the WHO ICTRP to identify relevant studies. There is no restriction posed on the language of the study. Included studies focus on the performance of two-field NMFP in detecting DR in diabetes patients. Analysis and evaluation of the studies will be examined by two reviewers independently using the Quality Assessment for Diagnostic Accuracy Studies-2 tool and later evaluated using the Population, Intervention, Comparison, Outcome, Study design criteria. A random-effect model will calculate the diagnostic indicators, including the sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, diagnostic OR, area under the curve and 95% CIs. We will also develop a summary receiver operating characteristic curve. We anticipate analysing subgroups according to the factors, which may lead to heterogeneity, including DR levels of patients, the reference standards, camera models, the interpretation criteria. The data will be analysed by STATA software. This study was registered with PROSPERO. Ethics and dissemination This review will analyse the published data. Patients/the public were not involved in this research. The results of this study will be published in peer-reviewed journals. PROSPERO registration number CRD42020203608.
medicine, general & internal