To the Editor Dr Figueroa and colleagues1 recognized the importance of social determinants in shaping health outcomes. Echoing a recommendation to study marginalized populations,2 they recommended multisectoral data linkage as a promising means of examining these complex relationships. We agree with this recommendation but urge caution in pursuing development of a polysocial risk score for 4 reasons. First, unlike genes, social determinants vary in intensity and duration of exposure, exerting different influences at different ages and in different social, cultural, and economic circumstances. To adequately capture this heterogeneity in a single risk score would be difficult. Second, the optimal risk score could differ across health outcomes and, as Figueroa and colleagues1 noted, may vary over time, such that a single polysocial risk score for poor health outcomes writ large may have limited utility. Third, the mechanistic conceptualization of social determinants “causing” poor health outcomes, just as genetic factors “cause” disease, may be reductionist. The effect of psychosocial exposures on health is mediated by numerous factors including resilience, itself a complex and multifaceted construct that remains poorly understood.3 Fourth, unlike in the field of genetics, the “treatment” for social determinants of health is often a policy shift. Demonstrating an association between a polysocial risk score and health outcomes may reaffirm the importance of social determinants but is insufficient to inform decision makers, who are often only able to pull one policy lever at a time. Figueroa and colleagues1 argued that stratification on polysocial risk score could “…facilitate testing of social interventions that may be effective in these populations,” but this begs the question: Which interventions are indicated for individuals with a high score? The argument for a polysocial risk score evokes comparison with the literature on adverse childhood experiences (ACEs). A global meta-analysis revealed that individuals exposed to 4 or more ACEs were at increased risk of poor health outcomes in adulthood.4 Although compelling, it does not follow that a poly-ACE risk score would facilitate targeted policy responses. Similarly, although the proposed polysocial risk score may predict poor health outcomes, it may have limited utility for individuals tasked with developing tangible responses.5 Multisectoral data linkage is an important tool for examining social determinants of health and can provide data necessary for generating polysocial risk scores. However, although such scores may be able to predict poor health outcomes, they may do little to help prevent or improve those outcomes.