Ecological risk assessments typically are organized using the processes of planning (a discussion among managers, stakeholders, and analysts to clarify ecosystem management goals and assessment scope) and problem formulation (evaluation of existing information to generate hypotheses about adverse ecological effects, select assessment endpoints, and develop an analysis plan). These processes require modification to be applicable for integrated assessments that evaluate ecosystem management alternatives in terms of their ecological, economic, and social consequences. We present 8 questions that define the steps of a new process we term integrated problem formulation (IPF), and we illustrate the use of IPF through a retrospective case study comparing 2 recent phases of development of the Fire Program Analysis (FPA) system, a planning and budgeting system for the management of wildland fire throughout publicly managed lands in the United States. IPF extends traditional planning and problem formulation by including the explicit comparison of management alternatives, the valuation of ecological, economic and social endpoints, and the combination or integration of those endpoints. The phase 1, limited‐prototype FPA system used a set of assessment endpoints of common form (i.e., probabilities of given flame heights over acres of selected land‐resource types), which were specified and assigned relative weights at the local level in relation to a uniform national standard. This approach was chosen to permit system‐wide optimization of fire management budget allocations according to a cost‐effectiveness criterion. Before full development, however, the agencies abandoned this approach in favor of a phase 2 system that examined locally specified (rather than system‐optimized) allocation alternatives and was more permissive as to endpoint form. We demonstrate how the IPF process illuminates the nature, rationale, and consequences of these differences, and argue that its early use for the FPA system may have enabled a smoother development path.