Objective: To assess the prevalence of adverse events (AE) and to investigate its association with factors related to the patient and to hospital admission. Design: Cross-sectional study. Setting: Four general hospitals located in the southeastern region of Brazil. Participants: All patients admitted to the participating hospitals at the time of the study were surveyed. Intervention: The methodology was based on the Iberoamerican study of adverse events, a twostage medical record review. Main Outcome Measure: Medical records were screened for AE only in the day (24-h) immediately before the review process, independently of the admission date. Results: A total of 695 admissions were examined. Prevalence was 12.8%. Almost 43% of AE were preventable. More than 60% of patients with an event prolonged hospital stay. In final regression model, urgent admission (OR: 2.68; Confidence Interval (CI) 95%: 1.53–4.69), submission to a procedure (odds ratio (OR): 2.41; CI 95%: 1.33–4.39), presence of central venous catheter (OR: 2.25; CI 95%: 1.14–4.41) and immunosuppressive therapy (OR: 3.41; CI 95%: 1.57–7.40) were statistically associated with AE. Conclusions: Our results indicate that around 1.3 AE happen in each 10 hospital admissions in Brazil. As patient safety continues to be a Public Health concern worldwide and mainly in developing countries, this would indicate the potential use of prevalence measures for monitoring patient safety in Brazilian context.