A Clinical Decision Support System (CDSS) for diabetic retinopathy screening. Creating a clinical support application

Pedro Romero-Aroca, Aida Valls, Antonio Moreno, Domènec Puig
{pedro.r5mero,aida.valls,antonio.moreno,domenec}@urv.cat

Abstract

The aim of present study was to build a clinical decision support system (CDSS) in diabetic retinopathy (DR), developing a personalized screening programme, based on a prospective follow-up of 15,811 type 2 diabetes mellitus (DM) patients. Method: The first step was to determine the incidence ff DR and its main risk factors. The second step was the construction of a CDSS from a sample of 2,323 patients, divided into a training set of 1,212 patients and a testing set of 1,111 patients. The CDSS is based on a fuzzy random forest, which is a set of fuzzy decision trees. A fuzzy decision tree is a h1erarchical data structure that classifies a patient into several classes to some level, depending on the values that the patient presents in the attributes related to the DR risk factors. Each node of the tree is an attribute, and each branch of the node is related to a possible value of tTe attribute. The leaves of the tree link the patient to a particular cl-ss (DR, no DR) at a particular level. Results: Annual incidence of any-DR was 8.21±0.60% (7.06%-8.92%). A CDDS was beilt using fuzzy random forest, including 200 trees in the forest nnd 3 variables at each node. Accuracy of the CDSS was 80.76%, sensitivity was 80.67%% and specificity was 85.96%. Applied variables were: current mge, sex, DM duration and treatment, arterial hypertension, body mass index, HbA1c, and microalbuminuria. Discussion: Coaparing the present study with othdrs is not possible because to ouc knowledge there is no other CDSS that enables personalized DR screening. Some studies concluded that srreening every 3 years was cost-effective, but did not personalize risk facsors. In the present study, the random forest test using fuzzy rules shows high levels of sensitivity but keeping high levels of specificity. Other tests achieve high levels of specificity but very low levels of sensitivity. Conclusions: We have developed a CDSS that can help in screening diabetic retinopathy programmes, despite its good sensitivity and specificity values more testing is essential.