Diabetic retinopathy is the leading cause of blindness in developed countries. The most effective way to prevent the evolution of the disease is its timely diagnosis, through population-based screening programmes. One of the main problems associated with this type of programmes is the low adherence rate. However, the individual decision to adhere to the screening was practically not explored in the simulation models found in the literature. Hence, the main objective of the present research is to demonstrate the relevance of computational simulation models in the study of this problem. To this end, an intelligent agent-based model was developed, supported by real data that, recognizing the importance of the individual features, effectively replicates reality. The use of the developed model to stage interventions outcomes, proved its usefulness in the discovery of knowledge and proposition of measures to the entities responsible for decision making.