Pharmacoeconomic data may be obtained within the context of randomised clinical trials (RCTs) and from effectiveness studies in the 'real world'. The differences between the 2 types of study design have implications for the types of data that can be obtained and the interpretation of the resulting findings. Because RCTs are designed to assess the safety and efficacy of pharmaceuticals, and because the study design of RCTs emphasises internal validity over generalisability, the pharmacoeconomic data collected from them are limited. The data may not be applicable to the more heterogeneous patients encountered in actual clinical practice, and cost estimates may be inaccurate because of protocol requirements. Effectiveness studies, in which treatments are studied under real-world conditions, remedy some of these limitations. Generalisability to actual users is generally enhanced in effectiveness designs, but data may be biased in other ways. This brief review compares the 2 study designs as they relate to pharmacoeconomic evaluations in terms of the research questions they address, design differences and their implications for study bias, data collection and data analysis and the generalisability of their results.