OBJECTIVES: This study investigated gastrointestinal (GI) risk level and compared treatment patterns according to GI risk level of arthritis patients. METHODS: This was a cross-sectional, observational study of arthritis patients taking non-steroidal anti-inflammatory drugs (NSAIDs) at least 1 month. Patients were recruited at 20 nationwide hospitals from December 2012 to September 2013. Data were collected through a self-administered questionnaire and medical chart review. Standardized calculator of Risk for Events (SCORE) was employed to estimate GI risk level. SCORE consists of six predictors; age, rheumatoid arthritis, current health status, steroid use, GI side effect history, and GI complication history instead of GI hospitalization history. SCORE ranges 0-39, and GI risk levels were defined as low(≤10), moderate(11-15), high(16-20) and very high(≥20). RESULTS: Study included a total of 1,896 patients. Mean SCORE was 19.21±5.80, and a considerable portion of patients had high (28.9%) and very high risk (41.0%). COX-2 selective inhibitor was prescribed in 56.5% of high risk patients and 53.3% of very high risk. GI risk level and COX-2 selective inhibitor prescription rate were analyzed by age stratification of 10 years. In patients aged ≥40, more than 60% of the patients were categorized in high or very high risk, and their prescription rates of COX-2 selective inhibitor were remained below 40% in high risk. In very high risk, the prescription rates of patients aged ≥40 for COX-2 selective inhibitor were found to be 30% in 40-49 and 39.8% in 50-59 and about 50% in 60-79. CONCLUSIONS: A majority of patients were at high or very high risk for NSAID-induced GI complications and undertreated in terms of COX-2 selective inhibitor use. Low prescription rate of COX-2 selective inhibitor could be explained by patients’ characteristics, GI protective agent use and limited reimbursement guideline. Nevertheless, GI risk should be sought more in pain control to minimize NSAID-induced GI complications.