OBJECTIVE: To determine whether the use of celecoxib instead of a conventional non-steroidal antiinflammatory drug (NSAID) combined with either histamine-2 receptor antagonists (H2RA) or proton pump inhibitors (PPI) is cost-effectiveness in reducing gastro-intestinal (GI) side effects. METHODS: Expected values of effectiveness in terms of number of GI events avoided and health care cost in Thai Baht (40 Baht = 1 USD) were estimated for a six-month time horizon using a decision analysis model. An incremental cost-effectiveness ratio (ICER) for celecoxib was determined based on health care provider perspective. Data on the likelihood of GI disturbance for celecoxib and NSAID were obtained from Celecoxib Outcomes Measurement Evaluation Tools (COMET) study. Components of direct health care costs associated with the use of celecoxib and NSAID comparators were based on Thai health care context using a gastro-enterology expert opinion. Unit prices for the competing drugs were abstracted from electronic drug dispensing databases of 18 provincial hospitals located in all four regions of Thailand. RESULTS: For a reference case analysis, use of celecoxib instead of NSAID (diclofenac) plus H2RA (ranitidine) in 1000 hypothetical patients would incur an incremental cost of approximately 4.6 million Baht and reduce a total of 40.2 events of any GI side effects. This is equivalent to the ICER of 114,975 Baht per GI event avoided for celecoxib. Compared with NSAID plus PPI (omeprazole), ICER for celecoxib was as much as 948,611 Baht for an event avoided, equivalent to an incremental cost of 3.4 million Baht for lowering 3.6 GI events in 1,000 patients. These cost-effectiveness ratios were very sensitive to the unit price of celecoxib and GI-event risk categories. CONCLUSIONS: Use of celecoxib in Thailand was deemed not cost-effective unless concerned parties were willing to pay such a high amount for avoiding any GI events caused by the conventional NSAID.