Pharmacological strategies for schizophrenia have received increasing attention due to the development of new and costly drug therapies. Evaluating their relative costs and benefits in Canada requires modeling the natural course of schizophrenia. OBJECTIVES: To develop a Markov model with 1st-order Monte-Carlo simulations to simulate the natural course of newly diagnosed schizophrenic patients. METHODS: Six discrete disease states defined the Markov model: 1): first episode – FE; 2) low dependency state – LDS; 3) high dependency state – HDS; 4) Stable; 5) Well; and 6) Death. Patients’ movements between these disease states defined 17 probability transitions to be estimated. The model was based on data from the Régie de l’assurance maladie du Québec and Med-Echo databases. All individuals aged 0-60 years with a newly diagnosis of schizophrenia between 1998 to 2006 were first identified by ICD-9 codes. Using this data, 5 Cox proportional hazard models for competing risks were used to estimate the 17 probabilities of transition. Validation was conducted by comparing the model’s probability transitions’ predictions with the published literature.  RESULTS: A total of 12,754 individuals were identified as newly diagnosed patients with schizophrenia. After the FE of schizophrenia, 69.8% of patients passed in LDS, 11.2% in HDS, 1% in death state and 18% in Well state. The mean transition probabilities after one year of follow-up were: FE to Well at 0.28 (± SD=0.10), FE to HDS at 0.11 (±0.05), FE to LDS at 0.60 (±0.08) and respectively FE to Death at 0.01 (±0.01). The corresponding values were similar to those obtained from other published models. CONCLUSIONS: This model is the first Canadian model incorporating transition probabilities adjusted for individual risk factors profiles using Canadian data. Future applications will include pricing and cost-effectiveness of new therapies for schizophrenia.