DRRisk is an educational tool that assesses the risk of current diabetic retinopathy in individuals who have diabetes. It uses fourteen risk factors to make a determination of an individual's current risk of retinopathy.
The diabetic retinopathy risk assessment tool was developed at the Center for Biomedical Informatics at Charles R. Drew University of Medicine and Science, using electronic health record data from 27, 223 Los Angeles County Department of Health Services (LACDHS) patients with Type 1 or Type 2 diabetes seen between 2015 and 2017. The average age of the individuals was 58 years, 57.5% of the individuals were women and 75.3% of the individuals were Latino. The tool is based on a deep neural network learned on the LACDHS data. Details can be found in the following publication:
Ogunyemi OI, Gandhi M, Lee M, Teklehaimanot S, Daskivich LP, Hindman D, Lopez K, Taira R. Detecting Diabetic Retinopathy through Machine Learning on Electronic Health Record Data from an Urban, Safety Net Healthcare System. JAMIA Open. 2021 August 19;4(3):1 - 10.[Click here to view the publication]Please note that this diabetic risk assessment tool may be not be as accurate for non-Latinos, as most of the data used to develop it came from Latinos.
The tool was evaluated on data from 13,408 individuals with diabetes seen at LACDHS between 2015 and 2017, with an AUC of 0.81, sensitivity of 73.6% and specificity of 72.8%. Data from the individuals evaluated was not used to build the tool.
The tool was also evaluated on data from 9,300 individuals with diabetes seen at LACDHS in 2018, with an AUC of 0.8, sensitivity of 72.2% and specificity of 74.2%. Data from the individuals evaluated was not used to build the tool.
The tool can be used by submitting just six pieces of information about a person with diabetes:
Work on developing the diabetic retinopathy risk assessment tool was funded by the National Library of Medicine under grant 1 R01 LM012309. The content presented on this site is solely the responsibility of the investigators on that grant and does not necessarily represent the official views of the National Institutes of Health.
Find out more about the funded project, Predicting Diabetic Retinopathy from Risk Factor Data and Digital Retinal Images here.