OBJECTIVES: The risk of subsequent events after a first cardiovascular event in persons with type 2 diabetes has received less attention to date. Simulation models, including risk engines and health-economic cost-effectiveness models, have thus relied primarily on estimations of the risk of first events and assumed constant transition probabilities for subsequent events. The aim of the current study is to analyze the differences in risk of having a first and a second myocardial infarction (MI) for persons with type 2 diabetes. METHODS: Observational data from the Swedish National Diabetes Register (NDR) for 35,238 persons with type 2 diabetes aged 30-74 years at diagnosis from January 1, 2004 to December 31, 2008 were analyzed using the conditional non-frailty Weibull model. To not underestimate the effect of BMI, two specifications of the model were estimated. Age at diagnosis, sex, hypoglycaemic treatment, diabetes duration, microalbuminuria and smoking were common covariates in both models.   RESULTS:   A total of 1409 patients had one MI event and 200 experienced two events. The results showed that the risk of a second MI differ from the risk of having a first MI. In addition, the effects of covariates were not constant between multiple events. Women had a lower risk for developing a first event compared to men, but a higher risk for a second event conditional on the first MI. Preliminary results indicate four times higher hazard of developing a MI conditional on a first MI during the follow up. CONCLUSIONS: The findings show the need for an update of simulation models including health-economic models and risk engines to include separate transition probabilities for first and subsequent events for correct predictions of costs and quality of life gains. Using recurrent event risk equations may reduce the bias from the previous assumption of constant transition probabilities for consecutive events in health economic models.