Interdisciplinary safety evaluation is needed for learning and decision-making, for understanding the evolving product safety profile, and for ensuring effective risk management strategies.
We need a framework for safety evaluation throughout the product development life-cycle, a process of learning and decision-making by a team of safety and data scientists working with evolving data sources, which engages critical team members, supports the iterative nature of the process, and recognizes different data sources used across the life-cycle of the product.
This approach will produce quantitative assessments that are product-specific and decisions that are driven by medical judgment, facilitating consistent and authoritative communication of the safety story in scientific evaluations and public disclosures to healthcare providers and patients.
For a deeper understanding, you could read the book Willful Ignorance by Herbert Weisberg.
“Practitioners are focused to a large extent on ambiguity. They are in the business of resolving ambiguity in productive ways and are often uncomfortable with statistical analyses. Researchers, on the contrary, generally regard their empirical data and mathematical models as imperfect but objective representations of reality. They are in the business of generating inferences in the form of probability statements and are often uncomfortable with qualitative insight. Each of these alternative perspectives is of limited scope. The path toward some kind of fruitful reconciliation must be based on a conception of uncertainty that is more integrative.”