OBJECTIVES: To examine the association between continuity of care and  risk of potentially inappropriate concomitant medication (PICM) in continuous non-steroidal anti-inflammatory drug (NSAID) users.  METHODS: Longitudinal Health Insurance Database 2005 (LHID2005) which contains one million of randomly sampled beneficiaries from National Health Insurance (NHI) in Taiwan was used to identify continuous NSAID users who were prescribed with NSAID  more than 30 days in 2005. NSAID’s PICM was defined as if there are significant drug-drug interactions (DDI) event based on Drug Interaction Facts. Continuity of care (COC) index was used to measure the degree of continuous  care for  these NSAID users. Multiple logistic regression analysis was then applied to estimate the association between continuity of care and PICM. RESULTS: There were 52,010(5.2% of LHID2005) beneficiaries using NSAID more than 30 days in 2005, and 20.2% of continuous NSAID users had PICM. Of which, Beta blocking agents was the most frequently prescribed and accounted for 8.79% of total PICM. Compared with lower level of COC, the odds ratios of PICM were 0.89(95% confidence interval 0.84-0.94) and 0.76(95% confidence interval 0.71-0.80) respectively for medium and high  levels of continuity of care when personal, physician and institutional characteristics were controlled in logistic regression.  CONCLUSIONS: PICM was common for continuous NSAID users in Taiwan, especially in those who had lower level of COC. While the consequence of PICM has to be investigated further, health policy directed to improvethe continuity of caremay reduce the risk of drug interactions in NSAID users.