OBJECTIVES: To describe the methods adopted by economic models of CML interventions, assess their strengths and limitations, and develop best practice recommendations.  METHODS: Examples of different economic modeling approaches used to assess the cost effectiveness of CML interventions were identified in MEDLINE and EMBASE. The studies were reviewed to map the method employed, how and why these approaches were selected, and lessons learned by the authors. RESULTS: 22 unique CML models were reviewed. The large majority of these models were published in the last 10 years, with almost half being published in 2011. All but 1 of the models adopted a Markov structure, based around the following health states: chronic phase; accelerated phase; blast phase; and death. In line with best practice recommendations, over 75% of studies modeled progression and survival based on response to treatment. Extrapolation of trial data used a wide range of statistical models. Contrary to best practice recommendations, the fit of these models to the trial data and the validity of the extrapolation were not always tested.  A variety of approaches were employed to estimate the health related quality of life associate with health states, including the direct valuation of health states, the use of standard health instruments (such as EQ5D), and mapping methods. CONCLUSIONS: Several approaches to the economic modeling of CML interventions were identified in the literature. A number of examples of good practice were identified, including the use of disease response outcomes when modeling progression and survival, and the systematic testing of the fit of survival distributions to trial data.  Key challenges facing CML modeling are the validity of extrapolations of trial data given the long time periods over which these extrapolations are required, and the lack of data against which to validate them.