OBJECTIVES: The Syreon model was developed to predict the long term effects of screening, treatment and control of type 2 diabetes. After a successful internal validation the model’s outcomes needs to be compared to outcomes of cohorts that were not used for the modeling exercise. The objective of this study was to demonstrate the methods and results of the external validation. METHODS: As a first step, we identified the applicable clinical trials and cohort studies which had not been used to build the model and simulated the patient cohorts for each study according to the published demographic, epidemiologic characteristics and treatment patterns. The incidence rates of the predicted and observed outcomes were calculated for comparison and the results were evaluated using statistical methods and expert opinion. RESULTS: 92 validation analyses were performed. The differences between the estimated and measured values of the incidences were within the range of 44% and -100%. The slope of the fitted linear regression line was 0.5326 while the R2 value was 0.6956. The macular oedema submodel presented the best fit and the estimated values from the foot ulcer submodel had the lowest accuracy compared to values observed in other studies. CONCLUSIONS: In overall the model performed well, however it frequently underestimated the incidence of the outcomes observed in the studies. This is most likely due to the limited information about the patient characteristics from the studies under evaluation. In most cases the information published about the population characteristics, treatment patterns and effectiveness were not sufficiently detailed to precisely match the model’s input parameters. Without sufficient information average values were used as input parameters, and this way the model presumably simulated healthier patient cohorts than the ones participated in the studies.