Response to the Opinion paper ‘Why do we disagree?’

In the Opinion piece by Woodward et al. (2020),1 we appreciate their explanations and proposals to resolve disagreement among epidemiologists. However, one glaring omission in their framework relates to the funding source as a major reason for uncompromising disagreement in the profession. The funding source, broadly interpreted, includes not only money, but also professional ambition tied to professional security. The scope and impact of this problem are deep and, as shown by Kramer and Soskolne (2017) in Figure 1, have grown steeply since about 2000.2 The funding source impacts not only the choice of research topic, but also the design, analysis and interpretation of studies over which disagreement often results.3–5 The funding source also results in influences that foment uncertainty and cast doubt,6–9 resulting in disagreement. With the increasing dependence of academic institutions on corporate funding, as well as the influence of politics on federal research priorities and funding, the problem is further exacerbated. Funding results in conflicted interests with entrenched stakes at play, so the question becomes one of, in the context of conflict, whether an honest conversation between epidemiologists about the design, analysis and interpretation of a study could take place. In the context of conflicting interests, would bringing social and behavioural scientists into a panel discussion change anyone’s position if it threatens their funding support and/or career? Based on the body of evidence, it would not. Indeed, it would be magical thinking to anticipate anything different. An adversarial role is taken by otherwise well-meaning epidemiologists in the presence of vested interests, especially when money is involved. Long overdue for broad consensus, and a topic that we speculate would help to minimize the likelihood of disagreement of the type proposed by Woodward et al. (2020),1 is the need for a more formal requirement that familiarity with ethics guidelines10 be required in epidemiology training programmes, as well as in mentoring programmes. Ethics guidelines have been put forward for the profession since the 1990s. However, their uptake, determined from ad hoc surveys at professional meetings, as well the small number of times they are cited in publications, has been poor. The International Society for Environmental Epidemiology),10,11 as well as the American College of Epidemiology,12 were leaders in approving their ethics guidelines. The guidelines, formally adopted by these prominent professional epidemiology organizations, provide normative standards for epidemiological practice and behaviour based on expressed core values and ethical principles. In so doing, a framework is provided for constructive argument regarding actions that need to be taken for advancing knowledge. The guidelines also provide detailed mechanisms for community engagement, including oversight bodies that determine the relevance of the very question being addressed epidemiologically, to its interpretation and dissemination, impacting on its effective translation into policy. We are left wondering whether the concerns expressed by Woodward et al.1 would exist had adequate attention over the past 20 years been given to integrating ethics training into the curriculum, and into mentorship programmes. Shoring up, in any way we can, the integrity of our discipline is needed if the public interest is to be protected.13 The purpose behind ethics-guideline development is to be proactive in providing a transparent framework for assessing normative ethical options anchored in transparent professional values. Guidelines facilitate discourse, including constructive disagreement, among well-intentioned people. However, when conflicted interests arise through financial interests, we submit that agreement is nigh on impossible among parties with entrenched positions. None declared.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.