Commentary on the position paper of the German Nutrition Society (DGE) 5 December 2024 By Dr Gregor Zwirn, Managing Director of G. Z. Research & Consulting KG and Research Associate at the University of Cambridge
The idea of basing policy decisions on solid evidence is a principle few would argue against. However, there is a troubling trend where evidence-based policy seems to be morphing into policy-driven evidence which implies using selective data to push ideological agendas.
Take, for instance, the Canadian Centre on Substance Use and Addiction (CCSA) that recently proposed new Canadian drinking guidelines recommending abstinence and defining ‘low risk’ as consuming no more than two Canadian standard drinks per week. Even if the modeled evidence supporting these recommendations were accurate, ‘low risk’ refers to such marginal risk increases (maybe a modelled loss of around 3 days on average over a lifetime) that the renowned risk expert Professor Sir David Spiegelhalter described the approach as ‘tackling a non-problem’.
Moreover, the evidence supporting the CCSA's recommendations carries a number of weaknesses. For instance, the evidence lacks replicability: the structure, input data sources, assumptions, and model codes are not transparently disclosed. Or, the evidence relies on selective use of studies and inaccurate claims: some studies cited as ‘new’ were published before the previous drinking guidelines were released in 2011, which remain the official standard.
It appears that the CCSA aims to persuade Canadian consumers through fear, shock, or alarm to adopt the ideology of temperance movements – namely, complete abstinence from alcohol – rather than providing balanced information about the risks associated with different drinking patterns.
A similar Canadian research group published a meta-analysis in 2023 (Zhao et al.), which has also been cited as ‘new’ evidence to support complete alcohol abstention. For instance, the German Nutrition Society (DGE) published a position paper, which draws heavily on CCSA and Zhao et al (2023). However, Zhao et al. found that increased all-cause mortality risk is associated only with alcohol consumption exceeding 45g/day (equivalent to more than four standard drinks). No increased all-cause mortality risk was observed for consumption below this threshold.
In fact, numerous studies over the past two decades have concluded that light-to-moderate alcohol consumption is associated with the lowest all-cause mortality risk in the general population. For instance, Wood et al. (2018) found that consuming up to 100g/week (approximately 10 standard drinks) is associated with the lowest all-cause mortality risk. This study, like Zhao et al, also accounted for the so-called ‘sick-quitter’ and ‘abstainer bias’.
Why This Matters
The shift from evidence-based policy to policy-driven evidence is more than a technical issue – it is a threat to public trust. When data is selectively interpreted to align with ideology, it undermines confidence in health recommendations. For policies to genuinely improve public health, they must be transparent, rely on a comprehensive review of all relevant evidence, and honestly communicate risks without exaggeration. Clear, balanced information allows individuals to make informed choices. Let’s ensure that health guidelines remain a tool for empowerment and education, rather than confusion or fear.
