On the evidentiary standards for nutrition advice
Introduction
During the last decades, the tenets of evidence-based medicine (EBM) have been adopted in several fields outside clinical medicine, for example, in public health, social policy, and library science (e.g., Parkhurst & Abeysinghe, 2016). This paper critically evaluates the application of evidentiary standards originating from EBM in the fields of nutrition science and advice. Hierarchies of evidence and the understanding that randomized controlled trials (RCTs) provide the highest level of evidence for making practical recommendations have become tools for assessing claims on nutrition and diets (e.g., Blumberg et al., 2010; Ioannidis, 2013). Concurrently, referring to the lack of evidence from RCTs has become a common argumentation strategy for criticizing nutrition recommendations. In what follows, I will show that this is problematic. The argument is two-fold. First, I argue that, due to practical, ethical, and methodological issues, it is difficult to conduct rigorous RCTs for acquiring evidence that is relevant for achieving the goals of population-level nutrition recommendations. Second, I will show that given the non-epistemic goals of the recommendations, the evidence assessment has to take into consideration the values of the target population and risks that follow from acting on the basis of the evidence. Consequently, the criteria of acceptable evidence should be adapted to the goals of the practice and the practical, ethical, and methodological constraints of the situation. Epistemic robustness and social robustness (Carrier & Krohn, 2016; Carrier, 2017) can serve as criteria for evaluating advice that has to be given under uncertainty often faced while science-based practical guidelines are given.
Evaluating the evidentiary standards in nutrition science and advice is a topic that is relevant to several current philosophy of science debates. After the recent “epistemic turn” (Stegenga, Kennedy, Tekin, Jukola, & Bluhm, 2017) in philosophy of medicine, the use of different forms of evidence in making clinical recommendations has been under intense scrutiny. The developing field of philosophy of epidemiology has brought light to the philosophical underpinnings of the population-level observational studies (Broadbent, 2013). Lately philosophers of science (e.g., Carrier & Krohn, 2016; Douglas, 2009; Wilholt, 2013) have become increasingly interested in examining the role of values in science in a context in which its results are applied to make policy recommendations. The molding of evidence into nutrition recommendations is an area in which science, political concerns and cultural habits intersect. This gives rise to philosophically interesting questions. What is the role of non-epistemic values in nutrition advice? How can different types of evidence be amalgamated into practical recommendations? Despite the social relevance of the topic and its potential for raising philosophically interesting questions, nutrition research and population-level nutrition guidelines are understudied topics in philosophy of science.1 This article aims at starting to fill this gap.
Before continuing a few caveats are in order. This article deals with nutrition research which is used for informing health policy that aims at maintaining health and preventing chronic diseases at the population-level. I shall not focus on how the criteria of evidence should be set for guidelines used for treating diseases such as diabetes. The epistemic challenges of evidence production and application are different in these cases. The standards of evidence that are typical of EBM can in theory be applied to evaluating guidelines on how to manage disease states, while conducting RCTs on how nutrition affects the development of chronic diseases is not without problems for reasons that shall be discussed further below (Blumberg et al., 2010, p. 476). Moreover, I shall be by-passing several methodological challenges of nutrition science. For example, I will not discuss the issues related to self-reported dietary data (e.g., Subar et al., 2015). This is because my aim is not to defend nutrition science against all methodological criticism. Instead, my goal is to first show that RCTs are not capable of delivering evidence for answering many important questions that have to be solved for official nutrition guidelines and then to give an argument why this should not prevent the publication of nutrition recommendations. Lastly, it is not my intention to take a stand for or against what the current guidelines recommend as healthy diet but to examine the controversy surrounding the evidentiary standards in nutrition advice from the meta-level.
The structure of the article is as follows. In section 2, I will briefly introduce the main tenets of evidence-based medicine and the criticism philosophers of medicine have raised against them. In section 3, I will give an outline of what nutrition research as well as nutrition advice are and provide an overview of the public criticism of nutrition advice, focusing on the objection that the lack of RCTs threatens the trustworthiness of population-level nutrition guidelines. In section 4, I will show that conducting randomized controlled trials on how nutrition can either prevent or cause chronic diseases faces several problems. Here I shall demonstrate that carrying out sound RCTs that could be used for guiding dietary advice is not always possible. In section 5, I will argue that nutrition guidelines as a form of scientific expertise should be evaluated according to the criteria of epistemic and social robustness. Taking into account the practical goals of translating evidence into health advice entails that the criteria of adequate evidence can differ from context to context. Finally, some conclusions are drawn.
Section snippets
RCT as the Gold Standard – on the ideals of evidence-based medicine
A notable part of recent literature in philosophy of medicine has focused on evaluating claims that are made in EBM concerning the required evidence base for clinical practice, especially the epistemic status of RCTs (Stegenga et al., 2017). The aim of EBM has been to improve patient care by reducing the impact of subjective judgments on clinical decision-making: Instead of consulting senior practitioners or referring to their own anecdotal experiences, clinicians should rely on research
On nutrition science and advice
Does red meat cause cancer? Is butter a better choice than margarine? How to combine meals and what foodstuffs should you avoid if you want to stay healthy? These are some of the questions nutrition research aims at answering. Lagua and Claudio (2012) give the following definition of the discipline: “science of food, the nutrients and other substances therein, their action, interaction and balance in relation to health and disease, and the processes by which the organism ingests, absorbs,
RCTs and research on nutrition
To recap: According to one central line of criticism of nutrition advice, depicted, for instance, in the article by Erickson et al. (2016), the lack of evidence from RCTs renders nutrition advice untrustworthy and confuses the laymen. However, as I will show next, in many situations nutritional RCTs cannot deliver evidence that could serve the aims of nutrition guidelines due to practical, ethical, and methodological reasons. Some of the issues that RCTs face in the context of nutrition science
Nutrition guidelines as scientific expertise
Above I have shown that carrying out rigorous RCTs on the effects of diet is complicated and that using RCTs to investigate certain important questions concerning food and health is not even possible. What does this imply with respect to giving dietary advice and nutrition recommendations? If evidence needed for nutrition guidelines cannot satisfy the criteria originating from EBM, can we justify releasing guidelines?
Nutrition advice is not the only area where experts have to take a stand
Conclusion
The tenets of EBM have become prominent also outside clinical medicine. The official nutrition guidelines are often criticized for not satisfying the evidential criteria that originate from this context. According to the critics, the fact that dietary recommendations are not based on RCTs threatens their trustworthiness. However, above I have shown that RCTs are no panacea for nutrition science and advice. For several practical, ethical and methodological reasons, RCTs often cannot be conducted
Acknowledgements
I thank Anke Bueter, Stefano Canali, Martin Carrier, Torsten Wilholt, and two anonymous referees for their helpful comments. The research towards this paper was supported by funding from Deutsche Forschungsgemeinschaft project 2073.
References (67)
The role of epidemiology in developing nutritional recommendations: Past, present, and future
Journal of the American College of Nutrition
(1999)Disease
Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences
(2002)Defining ‘health’and ‘disease’
Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences
(2009)- et al.
Newspaper debates on milk fats and vegetable oils in Finland, 1978–2013: An analysis of conflicts over risks, expertise, evidence and pleasure
Appetite
(2016) - et al.
Understanding nutritional epidemiology and its role in policy
Advances in Nutrition: An International Review Journal
(2015) - et al.
Addressing current criticism regarding the value of self-report dietary data
Journal of Nutrition
(2015) When are observational studies as credible as randomised trials?
The Lancet
(2004)The ethics and science of placebo-controlled trials: Assay sensitivity and the Duhem–Quine thesis
Journal of Medicine and Philosophy
(2006)- et al.
In vitro bioaccessibility of carotenoids, flavonoids, and vitamin C from differently processed oranges and orange juices [Citrus sinensis (L.) Osbeck]
Journal of Agricultural and Food Chemistry
(2015) - et al.
The one chart you need to understand any health study
(2015 January 17)