OBJECTIVES Network Meta Analyses (NMAs) are often used to parameterise efficacy in decision models for economic evaluation. A common source of heterogeneity in NMA arises from the fact that treatments may be given at different doses. This variation may manifest as unexplained heterogeneity in standard NMA models and propagates through to the decision analysis. We aim to explore how dose-response NMA can be used to inform cost-effectiveness analysis using a cost-utility analysis of treatments for acute migraine. METHODS We conducted four NMAs with different assumptions around dose-response to inform an economic evaluation in acute migraine. Separate 1-level NMAs were conducted where interventions were ‘lumped’ at the ‘dose’, ‘treatment’ and ‘class’ levels and a multi-level NMA was conducted, assuming monotonic dose-response. All NMAs were used to inform effect sizes in an economic model; the model structure, costing methods and utility inputs from the NICE Headaches guideline were adopted. The NMA models were compared in terms of heterogeneity and Deviance Information Criteria (DIC).  We report the results of the economic analyses using cost-effectiveness acceptability frontiers (CEAFs). RESULTS Dose-response parameterisation lead to NMA models with lower heterogeneity and better fit. Different dose-response parameterisations substantially changed the resource allocation decision, particularly at lower willingness to pay thresholds. For the more complex NMA model, it is unclear from a decision making perspective which effect size estimates should be selected as inputs to the decision model and we show that careful consideration should be given to the relevance of individual doses and confounding bias. CONCLUSIONS Dose-response NMA provides a useful and arguably more appropriate method for conducting NMA for decision analysis. Careful consideration should be given to how to treat doses of the same intervention in NMA since different parameterisations may lead to biased effect sizes and sub-optimal conclusions around cost-effectiveness.