Objective: To evaluate the accuracy of the French health administrative database to describe patients' medication and primary care visits, in the context of a transitional care intervention including an in-hospital medication reconciliation followed by a structured community follow-up by the patient's general practitioner and pharmacist. Design: A retrospective cohort study of older persons enrolled in the transitional care intervention between January 1st, 2015 and December 31st, 2018. Results: Only 46.1% of the community follow-up were timely billed, in the 3 months after the patient discharge. The sensitivity of the health administrative database to identify medications was 90.0%. Its positive predictive value was 50.1%. Conclusion: This study reveals that the French health administrative database was poorly reliable to identify both community follow-up and chronic medications.