The precautionary principle brings a special challenge to the practice of evidence-based public health decision-making, suggesting changes in the interpretative methods of public health used to identify causes of disease. In this paper, precautionary changes to these methods are examined: including discounting contrary evidence, reducing the number of causal criteria, weakening the rules of evidence assigned to the criteria, and altering thresholds for statistical significance. All such changes reflect the precautionary goal of earlier primary preventive intervention, i.e. acting on insufficient (...) evidence, the least amount, or minimum level, of evidence for causation. Evaluating the impact of these changes will be difficult without a careful study of how well the current methods of causal inference work, their theoretical foundations, and the ethical implications of their applications. That research program will be most productive if it is jointly developed by public health professionals trained in the ethics and philosophy and by bioethicists and philosophers trained in the theories, methods, and practice of public health. (shrink)