Understanding Cuomo’s Paradox in Oncology Research and its Implications for Pharmacologic Studies

Abstract

Cuomo’s Paradox describes a recurrent pattern in which exposures that influence the probability of developing a disease show the opposite or neutral association with survival once the disease is present. Pharmacy and pharmacology sit at the fulcrum of this tension because pharmacists are asked to translate evidence across prevention, active treatment, and survivorship, often in the same clinical encounter. The abstract summarizes the rationale for a formal framework that separates prevention and survivorship questions, outlines a protocol for a scoping review and evidence map, and previews a typology that classifies exposures according to whether incidence and survival associations align or diverge. The article then develops practical guidance for counseling, dosing, and supportive care without presenting new estimates. Throughout, it emphasizes the role of causal structure, including collider bias, stage migration, and reverse causation, in generating apparent inversions between prevention and survival evidence. Illustrations are drawn from adiposity and immune checkpoint therapy, serum cholesterol in oncology cohorts, antioxidant supplements before and during cancer therapy, and alcohol across prevention and survivorship contexts. The central message is that the same exposure can be beneficial when the goal is to prevent disease and not beneficial, or even harmful, when the goal is to extend life or improve treatment tolerance after diagnosis. Pharmacists can reduce contradictory advice by asking which question an estimate answers, by preferring survivorship specific endpoints once disease is present, and by counseling with treatment mechanisms in mind. The article closes with recommendations for research and guideline development that require explicit declaration of whether an analysis targets prevention or survivorship and that encourage prespecified survival focused estimands in pharmacotherapy studies.