Guidance on protocol development for EFSA generic scientific assessments

Abstract

EFSA Strategy 2027 outlines the need for fit-for-purpose protocols for EFSA generic scientific assessments to aid in delivering trustworthy scientific advice. This EFSA Scientific Committee guidance document helps address this need by providing a harmonised and flexible framework for developing protocols for EFSA generic assessments. The guidance replaces the 'Draft framework for protocol development for EFSA's scientific assessments' published in 2020. The two main steps in protocol development are described. The first is problem formulation, which illustrates the objectives of the assessment. Here a new approach to translating the mandated Terms of Reference into scientifically answerable assessment questions and sub-questions is proposed: the 'APRIO' paradigm (Agent, Pathway, Receptor, Intervention and Output). Owing to its cross-cutting nature, this paradigm is considered adaptable and broadly applicable within and across the various EFSA domains and, if applied using the definitions given in this guidance, is expected to help harmonise the problem formulation process and outputs and foster consistency in protocol development. APRIO may also overcome the difficulty of implementing some existing frameworks across the multiple EFSA disciplines, e.g. the PICO/PECO approach (Population, Intervention/Exposure, Comparator, Outcome). Therefore, although not mandatory, APRIO is recommended. The second step in protocol development is the specification of the evidence needs and the methods that will be applied for answering the assessment questions and sub-questions, including uncertainty analysis. Five possible approaches to answering individual (sub-)questions are outlined: using evidence from scientific literature and study reports; using data from databases other than bibliographic; using expert judgement informally collected or elicited via semi-formal or formal expert knowledge elicitation processes; using mathematical/statistical models; and - not covered in this guidance - generating empirical evidence ex novo. The guidance is complemented by a standalone 'template' for EFSA protocols that guides the users step by step through the process of planning an EFSA scientific assessment.

Keywords: APRIO; PECO; PICO; generic scientific assessments; guidance; methods; planning; problem formulation; protocol development.

Figure 1

Steps in problem formulation

1. Note: Steps 2–4 are optional. Step 2 is often needed in EFSA mandates. ToR: Term of Reference; AQ: assessment question; SQ: sub‐question; PF: problem formulation; APRIO: Agent, Pathway, Receptor, Intervention, Output.

Figure 2

Chlorinated paraffins in food (hypothetical mandate)

1. In this example, there is one assessment question corresponding to multiple Tier 1 and Tier 2 sub‐questions. More granular sub‐questions are shown by the arrow from top to bottom. The examples of sub‐questions are illustrative and not necessarily exhaustive for the mandate. The assessment question and sub‐questions are reported in the form of statements and are not yet ‘formulated’ (Section 6.2). ADME: Absorption, Distribution, Metabolism and Excretion; MoA: Mode of Action; ToR: Term of Reference.

Figure 3

High‐pressure processing technique for food (hypothetical mandate)

1. In this example, there are two Terms of Reference (ToRs) and six sub‐ToRs. Each sub‐term of reference is translated into one assessment question. The assessment questions are broken down into Tier 1 sub‐questions (i.e. assessment questions 1 and 4) or not split (i.e. assessment questions 2, 3, 5 and 6). More granular sub‐questions are shown by the arrow from top to bottom. The assessment questions and sub‐questions are reported in the form of statements and are not yet ‘formulated’ (Section 6.2 and Appendix A.4). HPP: high‐pressure processing; P: pressure; t: time; UHT: ultra‐high temperature.

Figure 4

Illustrative example of a conceptual model displaying all A‐P‐R‐I‐O elements for the various pathways linking the agent to the output

1. This is a simplified example of a quantitative pest risk assessment that addresses one of the risk assessment steps, i.e. the entry.

Figure 5

Illustrative example of a conceptual model developed using A‐P‐R‐O. In this case, the intervention (I) is not applicable

Figure 6

Example of APRIO as a tool to harmonise question and sub‐question formulation across different domains

1. The three mandates are hypothetical. Further details are given in Appendix A.2. More granular sub‐questions are shown by the arrow from top to bottom. RA: Risk Assessment.

Figure 7

Superchilled fresh fishery products (SFFP) (illustrative, simplified example) where the aim is to evaluate if the superchilled fresh fishery products stored/transported in boxes without ice are at least as safe (from a microbiological food safety perspective) as the currently authorised practice in boxes with ice (conventional fresh fishery products, CFFP)

1. To answer the assessment question, the sub‐questions are answered in a logical sequence. More granular sub‐questions are shown by the arrow from top to bottom. ToR: Terms of Reference.

Figure 8

Example of a conceptual model based on the pathway to harm approach for the cultivation of genetically modified (GM) maize in EU areas infested with teosinte, which interlinks the assessment questions, sub‐questions and relevant events of the pathway (simplified from EFSA GMO Panel, (EFSA, 2022))

1. Bt‐maize: genetically engineered maize that expresses insecticidal proteins from the bacterium Bacillus thuringiensis (Bt); NTO: non‐target organism; TO: target organism.