OBJECTIVES: To conduct a cost-effectiveness analysis of romiplostim as first-line ITP treatment in adult splenectomised patients who are refractory to other treatments and as second-line treatment in adult non-splenectomised patients for whom surgery is contra-indicated vs. eltrombopag. METHODS: A Markov model with embedded decision tree containing three health status (platelets≥50×10/L; platelets<50×10/L; and dead) was developed from a Colombian Health Ministry perspective and evaluated at 4-week cycles over a lifetime horizon. Efficacy was characterized by initial response; mean time to response, and duration of response and was estimated from literature review. Used resources and treatment patterns were obtained by a modified Delphi panel from a group of four hematologist. Costs include drug administration, visits, laboratory tests, rescue therapy, intracranial, GI and gynecological bleeding. Social Security costs (ISS+30) are used for procedures, visits and laboratory tests; and SISMED-2014 prices for drugs. Clinical benefits and costs are discounted 5% per annum. RESULTS: Total expected treatment cost for romiplostim arm was $1,276,302,002 (romiplostim cost $408,991,91; subsequent treatment lines $4,612,365; rescue therapy (IVIg and IV steroids) $859,929,341; and bleedings $2,768,379) vs. $1,315,173,138 for eltrombopag arm (eltrombopag cost $191,795,316; subsequent treatment lines $5,836,389; rescue therapy $1,113,981,314; and bleedings $3,560,119). Use of romiplostim, compared with eltrombopag, increased 4.46 years response duration, prevented 4.5 bleeding episodes and 1.5 admissions over a lifetime horizon. Romiplostim proves to be the dominant approach compared with eltrombopag. CONCLUSIONS: Use of romiplostim in the ITP treatment pathway, compared with eltrombopag, improves clinical outcomes, by increasing and maintaining platelet count, reducing bleeding events and rescue therapy need. These benefits generate cost savings and positioning romiplostim as a dominant approach.