The objectives of the study are to develop a cost-effectiveness model comparing biological therapy (BT) with methotrexate (MTX) alone, in the treatment for rheumatoid arthritis (RA), combining clinical and quality-of-life data from international trials with local costs and local epidemiological data. We designed a six-month cycle Markov model with five functional states, based on Health Assessment Questionnaire, with patients initiating treatment in any of the predefined states, based on a sample of 150 local RA patients. Simulations ran for 10 and 20 years, and for the whole life span. Utilities, in quality-adjusted life years (QALY), were taken from international literature. Discount rate was 3 % for costs and utilities. We calculated direct and indirect costs using a combination of international and local data. Results are presented as incremental cost-effectiveness ratios (ICER). ICERs in euros per QALY were <euro>143,072 for 10 years; <euro>139,332 for 20 years; and <euro>137,712 for the whole life span. Total costs with MTX were lower than with BT, despite higher out of pocket, productivity, and complication costs. Under conventional thresholds, and for the "average" RA patient, BT would not be cost-effective in Colombia. BT compared to MTX provides more QALYs, but at a high cost. When ICERs were estimated for Colombia, BT would not be cost-effective. We propose different thresholds for different conditions, perhaps prioritizing chronic diseases that lead to disability.