OBJECTIVES: This study determined the prevalence of potentially inappropriate medication use in a nationally representative sample of the non-institutionalized elderly population using the updated Beers 2003 criteria. METHODS: Participants with age 65 years and older were selected from the Medical Expenditure Panel Survey (MEPS) panel five for the period of January, 2000 to December, 2001, which was constituted by a series of five-round interviews. The average duration within each round was 3.5, 5.0, 6.0, 6.0, and 3.5 months from round one to five, respectively. Potentially inappropriate prescription medications were identified using the NDC (National Drug Code). All the calculations of prevalence were adjusted by the weight variable in the MEPS. RESULTS: Among the 1161 elderly participants in panel five, 402 patients were recognized having at least one inappropriate prescription during a two-year period, with a prevalence of 34.6%. A total of patients (25.2%) had inappropriate medication use in 2000 and 314 patients (27.5%) in 2001. The prevalence was also calculated for five rounds, of which 178 participants (15.8%) out of 1161 received at least one inappropriate medication in round one, 201 participants (17.5%) in round two, 215 (18.5%) in round three, 230 (20.2%) in round four, and 184 (16.8%) in round five. The round-specific results indicated that the prevalence did not shift dramatically during the two-year period. Propoxyphene and amitriptyline were still the most prevalently used inappropriate agents. CONCLUSIONS: Since the updated Beers 2003 criteria changed with a net addition of 24 new agents compared to the 1997 list, the prevalence of inappropriate prescriptions in the nation did not decrease after four-years compared to the statistics reported by Zhan's study (2001), which analyzed a sample of 1996 MEPS. Necessary interventions or policy making are still imperative in order to effectively prevent the use of potentially inappropriate prescriptions in the elderly.