Validation study of the claims-based definition for age-related macular degeneration at a single university hospital in Japan.

PURPOSE To evaluate the efficacy of the claims'-based definition of age-related macular degeneration (AMD) in detecting clinically-diagnosed AMD. STUDY DESIGN A validation study using cross-sectional data. METHODS Seven hundred clinically-diagnosed AMD patients and seven hundred non-AMD individuals were randomly selected from patients at the Kyoto University Hospital's ophthalmology outpatient clinic between January 2011 and December 2017. We evaluated the sensitivity, specificity, and positive/negative likelihood ratio of eight different claims'-based definitions of AMD for detecting clinically-diagnosed AMD. These definitions consist of the diagnosis name (AMD) in combination with (1) fluorescent fundus angiography, and/or (2) treatment of AMD, and (3) the exclusion of patients who had a diagnosis of central serous chorioretinopathy (CSC) or myopic choroidal neovascularization. RESULTS Defining by the diagnosis name AMD in the claims' data showed the highest accuracy (sensitivity 94.9%, specificity 92.6%, accuracy 93.7%). Combining the diagnosis name AMD with fluorescence fundus angiography and/or anti-vascular endothelial growth factor (anti VEGF) treatment increased the specificity at the expense of sensitivity. Notably, the combination with AMD treatment achieved a specificity of 98.3%. CONCLUSION The current validation study elucidated the high accuracy of the disease name (AMD) in the claims' data for identifying clinically-diagnosed AMD at a single university hospital. Although drawing wider conclusions may be limited, the results of this study contribute to creating real-world evidence in ophthalmology, based on the National Database of Health Insurance Claims in Japan.