OBJECTIVES: The electronic databases (ED) has been increasing used for clinical research since it reflects a routine care with large and heterogeneous samples. However, few studies report on its validity for such purposes, especially the ED from community hospitals. This study aims to assess the validity of EDs based on data from patients with atrial fibrillation (AF) receiving care from community hospitals in Phitsanulok. METHODS: The validity of ED was determined using out-patient medical records (OPMRs) as a gold standard.  A total of 200 AF patients were randomly sampled from a pool of patients with ICD-10 of AF (I48) from 2 community hospitals during Aug 2007-July 2008. For each patient, data of a randomly selected visit from the ED was matched to data of the same visit from OPMRs, abstracted by a standardized data collection form. The ED was cross-validated with OPMRs based on patient’s co-morbidities (risk factors for stroke) and bleeding events. All data were tabulated in a 2x2 format to calculate sensitivity, specificity and the Cohen’s Kappa adjusted for chance agreement. RESULTS: Out of 200 AF patients, 176 were documented as having diagnosis of AF in OPMRs (88%). The ED data on risk factors of stroke showed moderate to high sensitivity (range  66.67 – 100%) and high specificity (range 99.36 –100%). These results suggest that the risk factors of stroke were coded accurately. Based on these data, the agreement between two databases was considered good to very good (calculated kappa range 0.7940 – 0.9680). The specificity based on major bleeding was 100%, however; sensitivity and the Cohen’s Kappa could not be determined because the major bleeding diagnosis was found neither in the EMRs nor the OPMRs. CONCLUSIONS: The electronic database of AF patients from community hospitals was valid and in good agreement with the OPMRs.