Mechanisms and the evidence hierarchy
Exploratory AwardProject Team:
Principal Investigator:
Professor Jon Williamson, University of Kent
Co-Investigators:
Dr Brendan Clarke (UCL)
Professor Donald Gillies (UCL)
Dr Phyllis Illari (Hertfordshire & UCL)
Dr Federica Russo (Brussels & Kent)
Award Information:
This Exploratory Award focused on Evidence-Based Medicine. One of the main strands of the project explored the consequences of recent work in the philosophy of causality and history of medicine for the key question of whether mechanisms should have a place in the evidence hierarchy.
Since the advent of Evidence-Based Medicine (EBM), hierarchies of evidence have been widely used to guide the evaluation of evidence for causal claims. These hierarchies focus on statistical evidence for causal claims, and tend to champion the use of randomised controlled trials (RCTs).
This project explored the question of whether evidence of mechanisms should have a place in evidence hierarchies, complementing statistical evidence such as that obtained from RCTs.
The project concluded that mechanistic evidence should indeed play a more prominent role in explicit hierarchies of evidence, for two reasons. First, recent work in the philosophy of causality and the history of medicine suggests that in order to establish a causal claim one needs to establish both a statistical connection between the putative cause and the putative effect and a mechanistic connection that can explain the statistical connection. Second, discussion with medical scientists and public health policy makers revealed that such practitioners often do take mechanistic evidence into account, but mostly in a tacit or implicit way, in contravention to the explicit recommendations of the hierarchies of evidence.
Given the importance of these hierarchies in medicine, public health policy and the social sciences, the need for their reform is urgent.
Further information:
A Case Study (PDF) of this AHRC Science in Culture Theme Exploratory Award is available to download here.