Using decision modeling to guide drug allocation during a shortage

Background: Drug shortages require clinical teams to decide how to allocate drugs in limited supply among their patients. Ethical frameworks are invaluable for promoting rational approaches to drug distribution, but gaps remain between ethical theory and clinical application. The goal of this work was to explore how decision modeling could supplement ethical frameworks to inform drug distribution from the perspective of a clinical team.

Procedure: We created a hypothetical pediatric oncology clinic with a limited supply of 50,000 mg of methotrexate (MTX) and 21 patients due for treatment on one of six regimens. We constructed a simple decision analytic model to compare the effectiveness of MTX in milligrams per life year saved for each regimen. The robustness of the model was tested under various conditions including alternative drug effectiveness and time horizons. Effects on outcomes and distribution by substituting alternative dosing were explored for each regimen.

Results: Prescribed therapy for this group of patients required 108,791 mg MTX. Two regimens for three patients required ≥20,000 mg/m² . If distributed in order of arrival, only seven patients could receive full treatment. If distributed in order of efficiency, 19 patients could receive treatment. If less effective regimens were substituted, 20 patients could receive treatment. The primary driver of efficiency was dose per square meter.

Conclusions: In this hypothetical drug shortage, no allocation scenario exists that does not result in a worse outcome for some patients. Evidence of drug efficacy affected the decisions to substitute alternative treatments. First-come-first-served allocation resulted in fewer patients receiving treatment than allocation based on efficiency.