AI Models Bounded Rationality in Drug Shortage Decisions

A recent study published on arXiv explores the decision-making processes of hospital pharmacists when faced with uncertainty and time constraints during drug shortages. Through interviews, researchers found that pharmacists concentrate on a limited range of medications, directing their cognitive resources toward the most pressing situations. They established a decision-making framework based on bounded rationality and attention, which categorizes drugs into two groups: one for high-cost analysis and another for low-cost oversight. The study introduced two agents: an Expert Agent that utilizes attention weights derived from pharmacist interviews and a Learner Agent that adjusts its attention distribution based on experience. Simulations demonstrated that attention-guided strategies facilitate consistent decision-making without requiring complete state analysis, indicating that the key decision lies in how to prioritize cognitive effort.