The EU Health Technology Assessment Regulation, which has applied to new cancer medicines and advanced therapy medicinal products since January 2025, was designed so that a health technology developer submits clinical evidence once at Union level and Member States stop repeating the same clinical assessment. The Regulation also requires the assessment scope of a joint clinical assessment to be inclusive and to reflect Member States’ needs, and the scopes agreed so far have contained several distinct population, intervention, comparator and outcome (PICO) questions. Debate about this PICO multiplicity has concentrated on procedural feasibility, on how many questions simulation studies predict, and on whether indirect comparison methods can answer them. Much less attention has been paid to the economic question that follows, because every additional question consumes scarce assessor and developer capacity and that capacity has an opportunity cost. This article reframes PICO generation as a resource allocation problem. We define four quantities, namely question generation cost, analytical cost, feasibility loss and expected decision value, and use them to develop three ideas. These are a PICO feasibility frontier beyond which the submitted evidence cannot support further questions, analytical deadweight loss, and a cost incidence account of who requests a question, who answers it and who gains from the answer. We propose four tests for prioritising PICOs and argue that the efficiency of joint assessment should be judged by decision value generated per unit of assessment resource rather than by the number of questions addressed.