EFSA’s Role in Scientific Assessment

A. The Functional Separation of Risk Assessment and

Risk Management

The roles of risk management and risk assessment are described as functionally separated by the founding of EFSA by the GFL in EU food policies and deci­sions related to food, food safety, and the food chain.^[123] This strict separation

should ensure that the scientific advice is completely independent and is purely an objective, transparent assessment of scientific evidence related to the ques­tion that needs to be answered.

B. The Legal Status of Scientific Opinions

As EFSA is a scientific authority and does not function as risk manager, no legally binding decisions are adopted or published by EFSA, nor does the

Scientific Evidence in Food Assessments 39 Authority have any controlling responsibilities.^[130] However, as previously indi­cated, in cases where the risk manager does not follow the risk assessor’s opin­ion, this needs to be justified by the decision maker. In practice, very little divergence seems to exist between the issued scientific opinion and the subse­quent authorization or other legal decision made by the risk manager.^[131] As put forward in the previous sections, the current potential entanglement between the risk assessor and risk manager therefore requires careful analysis.

Within the different documents that are issued by EFSA, it is also empha­sized that these cannot be considered to have any legal status. This is exempli­fied by the disclaimer found in the annex of a scientific opinion addressing the assessment of Vitamin E—related health claims by the NDA Panel: “The present opinion does not constitute, and cannot be construed as, an authorization of the marketing of the food/food constituent”.^[132] In EFSA’s general scientific guid­ance on health claim applications, the section on the legal framework reem­phasizes this by detailing that “decisions regarding the authorisation of health claims... are taken by risk managers”.^[133] The GFL in itself does not seem to introduce any possibility for legal review of any opinions issued by EFSA^[134]. The opinions of EFSA are considered an intermediate step in a procedure, while only the final decisions can be brought to court. This implies that merely the usage of the scientific assessment by the risk manager, and not the scientific assessment itself, can be contested.^[135] EFSA has, however, been brought to the European Court ofJustice related to the use of specific documents, regarding the confidentiality of data submitted to the Agency, among others.

C. Scientific Assessment Versus Risk Assessment

The work of EFSA is based on the mandates given by the different risk manag­ers and can also originate from their own initiative (often referred to as “self­mandate”). Even though throughout the GFL, the concept of “risk assessment” and the role of “risk assessor” are used to describe the Authority’s activities, one of the elements that falls within the mission of EFSA according to Article 22(5) of the GFL is the provision of advice and support on human nutrition-related issues within the European Union. This provides the legal basis to also provide

scientific advice related to health benefits (or efficacy) when mandated to do so within European food laws or upon request of one of the risk management bodies. The interesting difference in terminology of “scientific assessment,” as it is referred to by EFSA itself^[136] versus the use of “risk assessment” as main term in the GFL can therefore not be explained by the mandate itself, which entails all aspects of scientific assessment related to food.^[137]

As risk assessor, EFSA is not tasked with conducting scientific experiments for specific authorization procedures. As laid down in various specific food regulations that require foods,^[138] food improvement agents,^[139] or health effects of foods^[140] to be based on scientific evidence, these safety or efficacy considerations need to be proven by the food business operator who is interested in placing the product on the market or using the claim on a food label.^[141] Even though EFSA is involved in research projects, the main role of EFSA and the scientific experts involved in the different panels is therefore based on scientific evidence gathered by such food industrials. This has resulted in the work of EFSA being referred to as “peer review” rather than risk assessment: reviewing whether the scientific evidence is of the highest standard seems a limited interpretation of conducting risk assessment.^[142] Whether this role of EFSA as rather young scien­tific agency will change in the future can be questioned.

D.

In the past, nutrition science and nutritional research were mainly focused on studying the safety of foods and analyzing how undernourishment should be prevented by providing sufficient nutrients in foods.^[143] Because of both tech­nological and economic developments, during the last decades, the effects of nutrition in maintaining and improving health gained increasing interest.^[144] Risk assessment was already employed to deal with safety risks of, for example,

Scientific Evidence in Food Assessments 41 microbiological contamination, while it is also increasingly used to analyze the two-sided risks (of under- and overconsumption) of essential nutri­ents.^[145] The approaches focus on determining a threshold or maximum level above which the dose of an ingredient or compound would present a risk or a minimum level below which the intake of a nutrient is inadequate.^[146] Assessing benefits of nutrients is concerned with the range beyond which there is a risk, that is, within the range between inadequate and excessive intake.^[147]

The benefit assessment of nutrition under EU food law as employed within the Nutrition and Health Claim Regulation has, however, taken a slightly dif­ferent approach: subsequent to the identification and characterization of the ingredient, the effect needs to be characterized, and the benefit of this effect for human health will be analyzed. The third step in this analysis is to identify the cause-and-effect relationship between consuming the ingredient (identi­fied in step 1) and the beneficial health effect (characterized in step 2).^[148] This difference illustrates that, even though in risk assessment of single nutrients for, among others, dietary reference values, a “window of benefit” is identified, and a dose-effect relationship is essential,^[149] while for the analysis of benefits of nutrients under the Nutrition and Health Claim Regulation, a cut-off point is identified above which the health effect is considered sufficiently charac­terized.^[150] While the scientific basis of this distinction should be investigated, and methods to assess benefits and safety of foods need further clarification, the combination of these different assessment approaches into the single term “scientific assessment” also requires reflection.

There are two main considerations in this reflection. First, the type of data used in assessing whether a component could affect the human population (either positively or negatively) should be considered. The data requirements for both safety and efficacy assessments differ from each other: whereas safety assessments are mainly conducted with in vitro and animal studies, for efficacy assessments, it is required to study the effects of compounds in human trials.^[151] As summarized by Nauta et al., already within the type of assessments, there are differences when it comes to scientific evidence requirements: for health

claims, a positive physiological effect should be clearly established with suf­ficient well-established data, while dietary guidelines are based on probable associations between intake and potential benefits.^[152]

Second, the methodology used to conduct the assessment itself is of impor­tance. Is it necessary to establish one single methodology for all three types of assessments (risk assessment, benefit/efficacy assessments, and risk-benefit assessments)? Currently, risk-benefit assessments are mainly conducted for those substances that are known as essential food components to maintain health, while overconsumption could result in toxic health effects. The 2010 Guidance on human health risk-benefit assessment of foods by EFSA advises mirroring the risk assessment procedures that are already used [(1) hazard identification, (2) hazard characterization, (3) exposure assessment, (4) risk characterization] with the four following steps: (1) positive health effect iden­tification, (2) positive health effect characterization (dose-response assessment), (3) exposure assessment, and (4) benefit characterization.^[153] The initial assess­ment in this combined procedure should then identify whether the health risks far outweigh the health benefits or vice versa, which should be followed by a more refined assessment for quantitative estimates of the risk and ben­efits and finally result in a single “net health impact value,” using metrics such as disability-adjusted life years (DALYs) an quality-adjusted life years.^[154] Even though in both risks and benefits, exposure to the compounds will be key, it is questionable whether this mirroring methodology is the most accurate way of identifying an optimum or optimal range of intake of substances.

It is questionable whether all three types of assessment should be conducted in the exact same way, and it may be very reasonable for them to differ in the type of data requested for assessing risks or benefits of foods and food ingredi­ents. However, a shared overall methodology as approach to these assessments is expected to benefit all involved parties: one basic method in which the submit­ted data is systematically gathered, critically reviewed by the applicant, and sub­sequently assessed by experts who follow a transparent, step-by-step procedure will further benefit the quality and efficiency of the procedure.^[156]

Scientific Evidence in Food Assessments 43

VI.