The Nordic Nutrition Recommendations 2022 – prioritisation of topics for de novo systematic reviews

Background As part of the process of updating national dietary reference values (DRVs) and food-based dietary guidelines (FBDGs), the Nordic Nutrition Recommendations 2022 project (NNR2022) will select a limited number of topics for systematic reviews (SRs). Objective To develop and transparently describe the results of a procedure for prioritisation of topics that may be submitted for SRs in the NNR2022 project. Design In an open call, scientists, health professionals, national food and health authorities, food manufacturers, other stakeholders and the general population in the Nordic and Baltic countries were invited to suggest SR topics. The NNR2022 Committee developed scoping reviews (ScRs) for 51 nutrients and food groups aimed at identifying potential SR topics. These ScRs included the relevant nominations from the open call. SR topics were categorised, ranked and prioritised by the NNR2022 Committee in a modified Delphi process. Existing qualified SRs were identified to omit duplication. Results A total of 45 nominations with suggestion for more than 200 exposure–outcome pairs were received in the public call. A number of additional topics were identified in ScRs. In order to omit duplication with recently qualified SRs, we defined criteria and identified 76 qualified SRs. The NNR2022 Committee subsequently shortlisted 52 PI/ECOTSS statements, none of which overlapped with the qualified SRs. The PI/ECOTSS statements were then graded ‘High’ (n = 21), ‘Medium’ (n = 9) or ‘Low’ (n = 22) importance, and the PI/ECOTSS statements with ‘High’ were ranked in a Delphi process. The nine top prioritised PI/ECOTSS included the following exposure–outcome pairs: 1) plant protein intake in children and body growth, 2) pulses/legumes intake, and cardiovascular disease and type 2 diabetes, 3) plant protein intake in adults, and atherosclerotic/cardiovascular disease and type 2 diabetes, 4) fat quality and mental health, 5) vitamin B12 and vitamin B12 status, 6) intake of white meat (no consumption vs. high consumption and white meat replaced with red meat), and all-cause mortality, type 2 diabetes and risk factors, 7) intake of n-3 LPUFAs from supplements during pregnancy, and asthma and allergies in the offspring, 8) nuts intake and cardiovascular disease (CVD) and type 2 diabetes in adults, 9) dietary fibre intake (high vs. low) in children and bowel function. Discussion The selection of topics for de novo SRs is central in the NNR2022 project, as the results of these SRs may cause adjustment of existing DRVs and FBDGs. That is why we have developed this extensive process for the prioritisation of SR topics. For transparency, the results of the process are reported in this publication. Conclusion The principles and methodologies developed in the NNR2022 project may serve as a framework for national health authorities or organisations when developing national DRVs and FBDGs. This collaboration between the food and health authorities in Denmark, Estonia, Finland, Iceland, Latvia, Lithuania, Norway and Sweden represents an international effort for harmonisation and sharing of resources and competence when developing national DRVs and FBDGs.

S ystematic reviews (SRs) (1) are the preferred method to summarise the current evidence on the causal relationship between nutrient-or food group exposure and a health outcome. Whilst several thousand SRs have been published in the field of diet and nutrition, only a limited number of SRs have adhered to the extensive principles and methodologies required to be identified as 'qualified SRs' (2-4) (see Step 3 later) by the Nordic Nutrition Recommendations 2022 (NNR2022) project. Qualified SRs will be the main foundation when the NNR2022 project revises national dietary reference values (DRVs) and food-based dietary guidelines (FBDGs) for the Nordic and Baltic countries. Production of qualified SRs is costly, and there are few dedicated, stable and long-term funding opportunities for the production of qualified SRs by any national food or health authorities, or international food and health organisation (5).
In the field of cancer, the World Cancer Research Fund International (WCRF) regularly produces qualified SRs on diet, obesity and physical activity and their causal relationship with the 17 most common cancers (6). The 'Dietary Guidelines for Americans' project (7), which is updated every 5 years, and the joint US-Canadian 'Dietary Reference Intakes' project (8) organised by The National Academy of Sciences, Engineering and Medicine also produce qualified SRs for the selected exposurehealth outcome pairs. Some additional national food and health authorities or international food and health organisations also produce a limited number of qualified SRs. These are precious and authoritative sources for national health authorities developing DRVs and FBDGs.
In the NNR2022 project, we have considered multiple health outcomes from 51 nutrient or food group exposures, representing in total several hundred possible exposure-health outcome pairs. Thus, the available qualified SRs from national food and health authorities and international food and health organisations cover only a subset of all possible nutrient/food group relationships with the main outcomes considered when setting DRVs and FBDGs in the NNR2022 project. The NNR2022 project plans to use the available budget to develop a limited set (i.e. 9) of de novo SRs, which adhere to the extensive principles and methodologies for qualified SRs.
National authorities have most often used an ad hoc procedure when prioritising topics for SRs. Recently, a more systematic and transparent approach has been set out (5,(9)(10)(11). The NNR2022 project has developed an open and transparent process for selecting topics for de novo SRs, which builds on and further extends these procedures.
The process of selection of topics for SRs is important since these topics are selected in areas where it is possible or conceivable that the DRVs and FBDGs will be adjusted compared to the previous edition of NNR. Whilst this process never can be totally objective, the NNR2022 Committee has strived to select topics with the best intentions and based on the best of our knowledge, without ideological, commercial, political, or other types of subjective biases.
This paper describes the results of the six-step procedure to identify topics that will be prioritised for de novo SRs by the NNR2022 project (Fig. 1).

3
(page number not for citation purpose)

The Nordic Nutrition Recommendations 2022
September 2019. Deadline for the submission of topics was December 31, 2020. The submitted nominations consisted of three components: 1) a cover letter with a rationale and a description of why a review on a specific topic was warranted and how it related to health issues in Nordic and Baltic populations; 2) a list of references for scientific papers; and 3) a simple 'PI/ECOTSS' statement covering the elements 'population', 'intervention/exposure', 'outcome', 'timing', 'setting' and 'study design'.
A total of 45 nominations with suggestion for more than 200 exposure-outcome pairs were received. Two nominations were excluded because they were incomplete; they were more like comments (see the complete list at the NNR2022 project website (12)). Forty-three of the nominations fulfilled all elements described earlier. The complete list of nominations, with their rationale and arguments, is available on the NNR2022 project website (12) and as an Electronic Supplementary Table 1. All submissions were considered by the NNR2022 Committee. Several of the nominations were overlapping, and some nominations needed to be interpreted and translated to a scientific question by the NNR2022 Committee. The NNR Committee developed a summary table of the nominations, where overlapping nominations were combined, that represents 43 exposure-outcome pairs (Table 1).
Whilst only a limited number of topics made it through to the final list of SR prioritisations due to limited resources, all public nominations will be evaluated carefully by the NNR2022 Committee and various chapter experts when the DRVs and FBDGs are developed and formulated.
Step 2. Scoping reviews on all nutrients and food groups considered in NNR2022 To develop candidate topics for prioritisation of de novo SRs, members of the NNR2022 Committee performed 51 scoping reviews (ScRs), one for each of the nutrients and food group chapters that will be part of the final NNR2022 report. An ScR is a relatively new approach to explore existing evidence (13). It differs from SRs both in its purpose and methodology. The purpose of an ScR is to provide an overview of available research without producing a synthesis and grading of total strength of evidence for a specific research question. An ScR should follow the procedures of the PRISMA Extension for Scoping Reviews (PRISMA-ScR) defined by the Equator Network (13). The methodology is much simpler than the extensive and more costly methodology for qualified SRs.
The objective of the 51 ScRs was to contribute to the shortlisting of topics. The major outcome of the ScRs was the formulation of shortlisted SR topics, formulated as PI/ECOTSS. Forty-nine topics were shortlisted based on the literature search. The literature search for the ScRs is presented in Electronic Supplementary Table 2.
When developing the search strategy for the ScRs, the aim was to identify possible topics that might be chosen for de novo SRs. We assumed that any topic with a significant amount of new data since the last edition of NNR Fig. 1. Multi-step process for prioritisation of topics for systematic reviews.
Shortlisting of potential SR topics (formulated as PI/ECOTSSs) Grading of SR topics into high, medium and low importance Prioritization of SR topics of high importance (two rounds modified Delphi process)

Final prioritisation
An open web-based nomination process for SR topics ScRs for all nutrients/food groups considered in NNR2022 Identification of qSRs to omit duplications Step 1 Step 2 Step 3 Step 4 Step 5 Step  The Nordic Nutrition Recommendations 2022 would likely have been covered in a recent review article. We selected to set the bar at the level of 'reviews', rather than 'systematic reviews'. By selecting reviews as the bar, we assume that we would pick up research activities that had not yet been dealt with in an SR. Thus, by choosing 'reviews', we have had a more open search with lower threshold than if we had selected 'systematic reviews'. In the NNR chapters, however, the initial ScR search string will be carefully adjusted and modified (e.g. by including 'systematic reviews', 'meta-analysis', 'Mendelian randomisation studies' and other types of relevant literature) when appropriate.
An evaluation of the results of the open public call (Electronic Supplementary Table 1) was included in each relevant ScR. Each ScR was considered by the NNR2022 Committee. The final version of the ScRs was formulated in a consensus process after several rounds of consultations in the NNR2022 Committee.
The criteria for shortlisting and prioritisation included evidence of significant new and relevant research since the previous edition of NNR (NNR2012) (14) and relevance to current public health concerns in the Nordic or Baltic countries (Box 1).

Step 3. Identification of qualified SRs to omit duplications
In order to omit duplication of recent qualified SRs, we established a process to identify relevant qualified SRs.
The definition of a qualified SR was based on the inclusion and exclusion criteria (Box 2) pre-specified by the NNR2022 project (2)(3)(4).
The search for qualified SRs was based on searches in PubMed/Medline and inspection of the websites of national and international food and health authorities as described by the Food and Agriculture Organization (FAO) of the United Nations (15). We also contacted the following major national food and health authorities and organisations directly for information on previous or planned SRs: All identified qualified SRs that fulfilled the inclusion and exclusion criteria are listed in Table 2.
• Relevance: The topic is within the scope of NNR2022. • Within scope (examples): Healthy populations/individuals; prevention purposes (e.g. population health topics, clinically oriented topics are not included and people with increased genetic risk for chronic diseases are included); covers different age groups, and pregnant and lactating women; increased requirements during short-term mild infections; etc. • Outside scope (examples): Long-term infections; malabsorption; various metabolic disturbances; treatment of persons with a sub-optimal nutritional status; clinical guidelines on dietary supplementation. • Importance: The topic has new, relevant and significant data or is an emerging topic in an area of substantial public health interest or concern. • Substantial public health concern (examples): Overweight, obesity and adiposity-related illness; metabolic syndrome and diabetes mellitus type 2; atherosclerotic cardiovascular diseases; cancer; osteoporosis; neurodegenerative diseases; mental health; oral health; multi-morbidity and mortality; reproductive health; optimal growth. • Relevant and significant: Refers to the overall scientific quality of the evidence, the number of studies, consistency of results and whether new study results appear to expand the DRV-and FBDGs-related information available in the previous edition of NNR. • Potential national impact: The SR may potentially inform national food and health policies and programs. An SR with the specific topic may result in a new or an adjustment of previous DRVs or FBDGs. • If the research question is within the scope of NNR2022 and covers an outcome of substantial health concern to the Nordic and Baltic countries, then it has potential national impact. In other words, it may inform DRVs, FBDGs and national food and health policies and programs. Step 4: Formulation and shortlisting of PI/ECOTSS statements All shortlisted topics from the ScRs and the public call were formulated by the NNR2022 Committee as initial PI/ ECOTSS statements (2)(3)(4). The shortlisted PI/ECOTSS statements were then compared with topics covered in the list of qSRs (Table 2), and overlapping PI/ECOTSS statements, which had not been removed in a previous stage, were excluded from the shortlisting. The initial formulation of PI/ECOTSS statements was adjusted by the NNR2022 Committee during several steps of this process to improve the precision of the scientific question. Consultation with topic experts, the members of the NNR SR Centre and the Scientific Advisory Group was helpful in formulating the final PI/ECOTSS statements. Elimination of PI/ECOTSS statements due to overlap with qSRs was continuously updated in accordance with the ongoing adjustments in PI/ECOTSS statements. Table 3 presents the 52 PI/ECOTSS statements that were shortlisted.
Step 5. The grading of SR topics into high, medium and low importance Subsequently, the NNR2022 Committee members graded individually the PI/ECOTSS into 'High' (n = 21), 'Medium' (n = 9) or 'Low' (n = 22) importance (Table 3), based on the criteria described (Box 1). The final grading was then decided in a consensus process. This process took more than 6 months and included careful evaluation of all the 51 ScRs as well as secondary literature-and citation searches.
Step 6. The ranking of SR topics of high importance The ranking of PI/ECOTSS statements with high importance was performed in a modified Delphi process amongst the NNR2022 Committee members. The Delphi process is a general, structured, interactive technique involving a panel of experts. It can also include face-to-face meetings. Delphi is based on the principle that decisions from a structured group of individuals are more accurate than those from unstructured groups. The experts answer questionnaires in two or more rounds. After each round, a facilitator provides an anonymised summary of the experts' voting from the previous round as well as the reasons they provided for their judgments. Thus, experts are encouraged to revise their earlier answers in light of the replies of other members of their panel. It is assumed that during this process, the range of the answers will decrease, and the group will converge towards a consensus (16).
The NNR2022 Committee individually prioritised the 21 PI/ECOTSS statements graded 'High importance' by giving each PI/ECOTSS statement a priority between 1 and 21.
An anonymised summary table, including arguments for prioritisation, was presented for the whole Committee by the NNR2022 project secretary. The Committee members were encouraged to revise their initial prioritisations in light of the discussion in the Committee meetings. A new anonymised summary table was then presented to the whole Committee in the next meeting. This procedure was repeated three times before a consensus was reached. The ranked list of the SR topics,

Inclusion criteria for SRs:
• Commissioned by national food or health authorities, or international food and health organisation • Authored by a group of multidisciplinary experts • Consist of an original systematic review of the evidence for a nutrient/diet-health relationship • Includes at least one nutrient/food topic and its relationship to at least one outcome related to a chronic disease or condition that is of public health interest in Nordic of Baltic countries; includes a clear description of the systematic review methodology, which should be similar to the methodology used NNR2022 (2, 3) • Includes an assessment of the quality of primary studies • Provides an evidence grade for the overall quality of the evidence • English language • Recent: Refer to SRs that have been published after the previous edition of the NNR Exclusion criteria for SRs: • Commissioned or sponsored by industry or an organisation with a business or ideological interest • Authors with strong ties to industry or ideological organisations • Later updated in another qualified SR on the same topic • Focused on an outcome outside the scope of the NNR (e.g. disease management or food safety) The Nordic Nutrition Recommendations 2022 Table 2. Qualified systematic reviews were identified based on the inclusion and exclusion criteria described in Box 1

Topic
Year Authors/organisation (country)          and the main arguments for ranking, is presented in Table 4. The formulation of the PI/ECOTSS was adjusted during the prioritisation process; thus, the formulation of the PI/ECOTSS in Table 4 is more specific compared with Table 3. The first five top prioritised topics, as well as all relevant background documentation, was submitted to the NNR SR Centre for their comments. In a dialog between the NNR SR Centre and the NNR2022 Committee, the final PI/ECOTSS statements for the five prioritised topics were formulated and agreed on by January 13, 2021 ( Table 4). The four remaining PI/ECOTSS statements was agreed on in June 2021. Results from step 1 to 6 in the procedure are summarised in Fig. 2. (page number not for citation purpose) The Nordic Nutrition Recommendations 2022       The topic is currently addressed through other qSRs   A protocol (17)(18)(19)(20)(21) will be developed for all de novo SRs by the SR Centre and published in PROSPERO (https:// www.crd.york.ac.uk/prospero/). The NNR2022 Committee and the topic experts (i.e. the scientists recruited to author the respective nutrient or food group chapters in NNR2022) will be consulted when finalising the protocols.

Discussion
Given the extent of scientific publications in the field of nutrition and health, and the limited resources available to summarise present research status rigorously and transparently, we have developed a procedure for prioritisation of topics that may be selected for SRs. The selection of topics for de novo SRs is central in the NNR2022 project, as the results of these SRs may cause adjustment of existing DRVs and FBDGs. That is why we have developed this extensive process for prioritisation of SR topics. The current paper describes the results of this procedure used to prioritise topics for de novo SRs in the NNR2022 project. The nine prioritised PI/ECOTSS statements include the following exposure-outcome pairs: 1) plant protein intake in children and growth, 2) pulses/legumes, and cardiovascular disease and type 2 diabetes, 3) plant proteins, and atherosclerotic/cardiovascular disease and type 2 diabetes, 4) fat quality and mental health and 5) vitamin B 12 and vitamin B 12 status, 6) intake of white meat (no consumption vs. high consumption and white meat replaced with red meat), and all-cause mortality, type 2 diabetes and risk factors, 7) intake of n-3 LPUFAs from supplements during pregnancy and asthma and allergies in the offspring, 8) nuts intake, and CVD and type 2 diabetes in adults, 9) dietary fibre intake (high vs. low) in children and bowel function (Table 4). Small adjustments of the PI/ECOTSS may occur during the development of the protocols. The final wording will be available in the published protocols.
The nine top SR topics are given high priority since significant new evidence within these topics might change the current recommendations. Additionally, increased adherence and more focus on plant-based diets and an environmentally sustainable diet were also important arguments for several of the SR priorities. Health effects of such changes must be considered and evaluated before potentially adjusting DRVs and FBDGs. The topic on vitamin B 12 status is also partly due to the aging population and related health consequences. The rational for the prioritisations is given in Table 4.
A delicate balance must be considered when PI/ ECOTSS statements are formulated. They may be too narrow to be generalisable. Additionally, it is always tempting to broaden the scope, for example, the exposure, the population or the outcome, but this may massively influence the resources needed for performing the SR. Too broad PI/ECOTSS statements may also be more imprecise and mask specific questions. In this process, we have tried, openly and explicitly, to identify the most relevant PI/ ECOTSS for adjusting DRVs and FBDGs in the Nordic and Baltic countries, but, at the same time, use the limited resources available in the most cost-effective manner.
Traditionally, the working group responsible for developing national DRVs and FBDGs select SR topics based on their own scientific knowledge and after consultation with appointed scientists in the field of interest. In the NNR2022 project, we have involved numerous scientists, health professionals, national food and health authorities, food manufacturers, other stakeholders and the general population to generate a large and representative pool of potential SR topics. This pool of topics was valuable when the NNR2022 Committee performed the prioritisation process in the modified Delphi process. Selection of SR topics can never be a fully objective exercise. Some stakeholders may be more proactive than others. The NNR2022 Committee tried to use all available information, independent of subjective engagement by stakeholders. In the end, selection of SR topics was the decision of the NNR2022 Committee.
Although consensus was reached in the NNR2022 Committee, it does not necessarily mean that we have concluded with the 'correct' selection. Several other topics might have been considered and prioritised. The question about what is most important in nutritional sciences is large and open. In the present project, we have, however, focused on topics with substantial recent data and public health concern, which is most relevant for setting DRVs and FBDGs in the Nordic and Baltic countries.
A limitation of our study is the literature search (Supplementary Table 2) used to develop the 51 ScRs. We decided initially to limit the search to reviews published in 2011 and later with the filter 'Humans'. If the search resulted in ≥500 items, we limited the search to papers with the nutrients or food groups in the title. If still ≥500 items, we included the additional requirements: 'Diet' OR 'Dietary' OR 'FOOD' OR 'Nutrition' OR 'Nutritional'. If still ≥500 items, we limited the search to only include 'Systematic reviews'. The reason why we initially selected to search for reviews published after 2010 is that it is likely that a topic with significant new and relevant data would have been discussed in a review paper published after the search date in the previous edition of NNR. In this type of strategy, we omit all original publications. However, DRVs or FBDGs are seldom, or never, revised based on one or a few original publications. In the present literature search process to identify SR topics, only original study results found important enough to be cited and discussed in review papers are candidate for SR topics.
Additionally, if a large number of reviews were identified for a single nutrient or food group (i.e. ≥500 papers), we added sequentially additional relevant limitations, High priority due to focus on sustainability of diets and might affect DRVs.
In the context of a more plantbases diet, it is important to know how B12 status is impacted in the most vulnerable groups.
This SR would identify data that facilitates setting DRVs for vulnerable groups   simply to reduce the burden of the authors of the 51 ScRs. In total, 13,992 reviews were identified and scrutinised by the ScR authors. Although we do not believe that other topics would have been prioritised with an even more comprehensive search strategy, we cannot rule out the possibility that some important topics have been missed. It is important to note that the present literature search was only used to select topics for de novo SRs. In each of the 51 nutrient and food group chapters that will be part of the final NNR2022 report, a separate literature search will be performed and described.
The organisation, the principles and the methodologies developed in the NNR2022 project build on processes similar to other national authorities or international health organisations. The procedure described in this paper, together with the three previous principle and methodology papers from the NNR2022 project (2)(3)(4), may serve as a framework that other national health authorities or organisations can adapt when developing national DRVs and FBDGs.
A large amount of resources and extensive interdisciplinary front-edge competence is needed to develop national DRVs and FBDGs. No or few single nations have these qualifications alone. Thus, international collaboration and global harmonisation of methodological approaches are highly needed. The NNR2022 project, which is a collaboration between the food and health authorities in Denmark, Estonia, Finland, Iceland, Latvia, Lithuania, Norway and Sweden, represents such an international effort for harmonisation and sharing of resources and competence.

Summary and conclusions
SRs are the preferred method to summarise the causal relationship between nutrient or food group exposure and a health outcome. They are the main fundament for developing DRVs and FBDGs. In this paper, we describe the results of an open, transparent six-step procedure to identify and prioritise topics most appropriate for de novo SRs in the NNR2022 project. The nine prioritised PI/ECOTSS include the following exposure-outcome pairs: 1) plant protein intake in children and body growth, 2) pulses/ legumes intake, and cardiovascular disease and type 2 diabetes, 3) plant protein intake in adults, and atherosclerotic/cardiovascular disease and type 2 diabetes, 4) fat quality and mental health, 5) vitamin B 12 and vitamin B 12 status, 6) intake of white meat (no consumption vs. high consumption and white meat replaced with red meat), and all-cause mortality, type 2 diabetes and risk factors, 7) intake of n-3 LPUFAs from supplements during pregnancy and asthma and allergies in the offspring, 8) nuts intake, and CVD and type 2 diabetes in adults, 9) dietary fibre intake (high vs. low) in children and bowel function.