OBJECTIVES: Treatment crossover refers to the situation in randomised controlled trials (RCTs) where patients randomised to the control group switch onto the experimental treatment. This leads to biased estimates of treatment effects if not appropriately controlled for. Several crossover adjustment methods are available, but previous research has shown that the optimal adjustment method depends upon the characteristics of the trial. This study applies crossover adjustment methods to an RCT comparing trametinib to chemotherapy in patients with BRAF V600E/K mutation-positive advanced or metastatic melanoma (NCT01245062), and investigates which adjustment method best fits this case study. METHODS: The crossover adjustment methods applied include the Rank Preserving Structural Failure Time Model (RPSFTM), Iterative Parameter Estimation (IPE) algorithm, Inverse Probability of Censoring Weights (IPCW) and a two-stage accelerated failure time model estimation procedure. Suitability of each method is compared by assessing the plausibility of the underlying assumptions of the models in this case study and analysing output and performance indicators associated with each method. RESULTS: In the primary efficacy population (patients without history of brain metastases) 67.4% of chemotherapy patients switched onto trametinib. The intention to treat (ITT) hazard ratio (HR) for overall survival (OS) was 0.72 (95% CI 0.52-1.01). Point-estimates of the adjusted HRs produced by the most plausible applications of the RPSFTM, IPE, IPCW and two-stage methods ranged between 0.43 and 0.49, consistently favouring trametinib. Results were sensitive to the technique used to apply each method. Key issues included recensoring, the active nature of the comparator, and the choice of covariates included in the analyses. CONCLUSIONS: Each of the crossover adjustment methods result in a lower HR than the ITT analysis. However, results are uncertain and sensitive to key assumptions. It is important to carefully analyse trial characteristics and model output when identifying which applications of adjustment methods are most plausible.