Sources of variability on the estimate of treatment effect in the PROWESS trial: Implications for the design and conduct of future studies in severe sepsis*

A learning curve appeared to be present within the PROWESS trial such that the ability to demonstrate efficacy improved with increasing site experience. This potential learning curve may have implications for design of future trials. Investigational sites may need to require a minimum level of protocol-specific experience to appropriately implement a given trial. This experience should be an important consideration in designing trials and analysis plans. Diligence by coordinating centers, site investigators, study coordinators, and sponsors is necessary to ensure that the protocol is executed as designed such that a treatment benefit, if present, will be evident.