Why his framework matters
Goldacre’s Bad Science framework identifies specific recurrent patterns of unreliable medical claim: retrospective pattern-matching without pre-registration; absence of a control group; hypothesis-first rather than hypothesis-testing reasoning; absence of peer review; heavy reliance on narrative over data. Each of these is a warning sign for medical-evidence unreliability.
Applied to the Crown’s expert evidence at the Letby trial, each warning sign is present. The Crown’s causation expert (Dr Evans) conducted retrospective pattern-matching against selected events; there was no control group of similar-acuity preterm collapses without Letby present; the hypothesis came first and the evidence was read to fit; none of the methodology was published in peer-reviewed form; and the Crown’s closing speech was heavily narrative.
Professional background
- Professor of Evidence-Based Medicine, University of Oxford.
- Director of the Bennett Institute for Applied Data Science.
- Author of Bad Science (2008), Bad Pharma (2012), and the long-running Guardian column of the same name.
- Past winner of the Royal Statistical Society President’s Medal for services to statistical communication.
- Co-author of the OpenPrescribing platform and of the Goldacre Review of NHS data (2022).
The Bad Science framework applied to the trial evidence
- Retrospective pattern-matching. The Letby shift-rota chart and medical-evidence framework started from the pattern, not from the data.
- Absence of control. No comparison group of similar-acuity preterm deterioration events without Letby present was adduced.
- Hypothesis-first reasoning. The Crown’s causation expert worked from the suspect hypothesis toward confirming findings.
- Non-peer-reviewed methodology. Dr Evans’s methodology was not published in peer-reviewed form. The Panel’s methodology is.
- Narrative heaviness. The Crown’s closing speech framed the case as pattern and narrative, not as quantitative evidence.
Why the framework is the right tool
Bad Science is a popular-reference framework, but it encodes specific methodological criteria that professional EBM research applies formally. The same criteria professional epidemiologists use to evaluate clinical-claim reliability apply to expert evidence in criminal trials. Goldacre’s framework is the accessible version; Heneghan’s EBM framework is the formal version. Both reach the same conclusion on the Letby evidence.