National nutrition surveys in Europe: a review on the current status in the 53 countries of the WHO European region

Objectives The objectives of this study were (1) to determine the coverage of national nutrition surveys in the 53 countries monitored by the World Health Organization (WHO) Regional Office for Europe and identify gaps in provision, (2) to describe relevant survey attributes and (3) to check whether energy and nutrients are reported with a view to providing information for evidence-based nutrition policy planning. Design Dietary survey information was gathered using three methods: (1) direct email to survey authors and other relevant contacts, (2) systematic review of literature databases and (3) general web-based searches. Survey characteristics relating to time frame, sampling and dietary methodology and nutrients reported were tabled from all relevant surveys found since 1990. Setting Fifty-three countries of the WHO Regional Office for Europe, which have need for an overview of dietary surveys across the life course. Subjects European individuals (adults and children) in national diet surveys. Results A total of 109 nationally representative dietary surveys undertaken post-1990 were found across 34 countries. Of these, 78 surveys from 33 countries were found post-2000, and of these, 48 surveys from 27 countries included children and 60 surveys from 30 countries included adults. No nationally representative surveys were found for 19 of 53 countries, mainly from Central and Eastern Europe. Multiple 24hr recall and food diaries were the most common dietary assessment methods. Only 22 countries reported energy and nutrient intakes from post-2000 surveys; macronutrients were more widely reported than micronutrients. Conclusions Less than two-thirds of WHO Europe countries have nationally representative diet surveys, mainly collected post-2000. The main availability gaps lie in Central and Eastern European countries, where nutrition policies may therefore lack an appropriate evidence base. Dietary methodological differences may limit the scope for inter-country comparisons.

burden. Other risk factors include alcohol, tobacco misuse and physical inactivity (2). In Europe, the four most common NCDs account for 77% of disease and almost 86% premature mortality (1).
NCDs and related conditions, including overweight and obesity, have significant and growing economic and social costs (1), which traditional clinical approaches are increasingly unable to address (3). Mozaffarian et al. (3) call for a shift in emphasis from such pharmacological treatments to primary prevention through addressing lifestyle risk factors such as dietary patterns in order to reduce cardiovascular risk and NCD-associated problems.
Dietary surveys thus have an important role in assessing dietary patterns in the whole population. Nutrition and health surveys formed the main source of information for dietary risk factors and physical inactivity in a systematic analysis of disease risk in 21 regions worldwide across two decades (4). Such surveys can provide a means of monitoring trends, identifying areas of concern and inequality and evaluating policy impact, thereby ultimately contributing to the promotion of best practice across the region (1). The WHO European Food and Nutrition Action Plan (1) explicitly encourages member states to 'strengthen and expand nationally representative diet and nutrition surveys'.
Many western European countries currently have established dietary surveys that assess food and nutrient intake. A global review of country-specific surveys from 1990 to 2010 only reported dietary fat and oil intake (5). A comprehensive, updated review of total nutrient and food intakes across different populations and subgroups in Europe is needed, the results of which could identify where in Europe there is a need to improve diets and whether inequalities exist. This paper makes the first step in this regard, establishing which countries have nationally representative dietary surveys and highlighting gaps in nutrition survey provision across Europe.
This review aims to identify which of the 53 countries in the WHO European region have conducted nationally representative dietary surveys of whole diets at an individual level and those that have not. It identifies key characteristics, centred on time frame, sampling and dietary methodology, of known surveys undertaken since 1990 for adults and children and aims to lay the foundations in establishing a clear picture of the current situation. Following this, future papers will examine energy and nutrient intakes in different population groups across Europe to better assess where both gaps in knowledge and dietary inadequacies lie. Information from dietary surveys can be used as a means for governments and health bodies to monitor and reduce the diet-related risk of NCDs and related conditions across Europe, thereby contributing to the goals set out in the WHO action plan.

Methods
We used three key approaches to identifying national diet surveys: (1) contacting authors of surveys, (2) systematic literature review and (3) general web-based searches.

Identifying authors of national diet surveys
We identified authors of national surveys within the WHO Europe remit using listed contact names and other information from two main reports of national dietary surveys (5,6). If no response was obtained from those authors, Internet searches of nutrition organisations by country and the survey titles listed in the review of 1990-2010 surveys (5) and the European Food Consumption Survey (6) were carried out to find other potentially useful contact details. For countries where this approach did not yield usable contact details, Internet searches using various search terms were performed on organisations specialising in nutrition, including known government and public health agencies. WHO also provided contact details for some of those countries for which they had relevant associates. Contacts identified were asked to complete a questionnaire (Appendix 1) to provide information on nationally representative dietary surveys conducted at an individual level since 1990, including links or references to relevant reports.

Systematic database search
For countries where no contact could be identified, systematic searches were undertaken across Web of Science, Medline and Scopus for nationally representative dietary surveys that collected data at an individual level from 1990 to June 2016. The following query terms were run without language restrictions: (survey* OR research* [TS]) AND (nutrition* OR diet* OR food* [TS]) AND (list of countries).
The title of each paper generated by the database searches was screened for relevance according to the criteria in Table 1; those that are not relevant were excluded. The remaining papers were screened by title and abstract, and full article where available, and their appropriateness for inclusion was checked by a second reviewer. Further surveys, related papers and nutrition expert contact names were gathered by general Internet searching to capture any recently released information, targeting known government and public health agencies using various search term combinations in order to maximise returns. Although there were no language restrictions in the initial search, the WHO Regional Office for Europe, Division of Noncommunicable Diseases and Promoting Health through the Life-Course, conducted an additional database search of papers in the Russian language as an extra check to maximise returns in the 12 Central and Eastern European countries where Russian is an official or widely spoken language (Armenia, Azerbaijan, Belarus, Georgia, Kazakhstan, Kyrgyzstan, Moldova, Russia, Tajikistan, Turkmenistan, Ukraine and Uzbekistan). However, no papers or reports that met the inclusion and exclusion criteria were found. The databases searched were PubMed, Web of Science and Google Scholar, using the search terms mentioned above, translated into Russian. Further searches with these terms were undertaken in three specific Russian language databases: Kazakh Academy of Nutrition; 1st Moscow Medical Academy named after Sechenov and Electronic Scientific Library in Russian.

Database extraction
Where long-running surveys had multiple collection waves, for example, the French INCA 1 and INCA 2 or UK NDNS 2000-1 and NDNS 2008-12, each collection wave was counted as a separate survey (see Table 2). Survey characteristics were extracted and tabled from the relevant publications, which were accessed in various forms, including summary reports, academic articles and completed questionnaires (see Table 2). The survey characteristics included the following: country name, survey name, year of survey (data collection), information source, sample size and age range, dietary methodology, nutrient composition database and reference. The availability of energy and selected nutrients from the latest surveys collected after 2000 are listed in Appendices 1 and 2.

Data extracted
A total of 109 nationally representative surveys that obtained data on whole diets (rather than focusing only on certain foods) at an individual level since 1990 were found for 34 out of the 53 countries in the WHO office region. Table 2 shows the characteristics of these surveys and that the majority of countries with national dietary surveys (NDS) had conducted multiple surveys. Of the 34 countries with NDS, almost half (n = 16) had long-running surveys with waves conducted over various years; 10 of these also had stand-alone surveys (  Table 2 for the characteristics of all dietary surveys conducted since 1990.  Information regarding nutrient composition databases has been added for those surveys for which energy and nutrient intakes were reported and graphed. ‡ Y = energy intakes were taken from the latest survey for which they were reported; N = energy and nutrient intakes were either not reported or were not extracted because intakes for that country were available in a later survey. ‡ ‡ The Slovakian surveys were not truly nationally representative, but were country-wide and designed to 'recruit a diverse sample of subjects of different age categories and socio-economic background' (121).
NB -The EFSA guidance for the standardised collection of national food consumption data was released in 2009.
No nationally representative surveys were found by any method that collected dietary intake of whole diets at individual level for 19 European countries (see Table 3 and Fig. 2). Although one survey of children was found for Croatia, it was not nationally representative (7  Non-nationally representative dietary surveys were found for eight countries (Croatia, Czech Republic, Germany, Greece, Iceland, Luxembourg, Russia and Switzerland), but because of our exclusion criteria they were not included in the list of nationally representative surveys in Table 2. Additionally, 16 studies conducted in Central and Eastern European countries were returned from the systematic literature search in English and 49 from the WHO Russian language database search and were not included in any tables; common reasons for rejection were no or partial dietary intake collected, data not collected at individual level, duplicate and sample size too small (<200). Eight countries completed the WHO STEPwise approach to noncommunicable disease risk factor surveillance (STEPS) adult survey (8)(9)(10)(11)(12)(13)(14)(15). However, although these were nationally representative population-based surveys with large sample sizes, they were not included in this review because they only covered specific food groups, not whole diets, and as such did not meet our inclusion criteria.

Dietary methodologies
The most common dietary assessment methodologies used across the 109 nationally representative surveys were the 24hr recall and food diary. Of these surveys, 45 used 24hr recall, 35 of which were surveys conducted since 2000 (Table 2). Of the 45 surveys using 24hr recall, the range of daily recalls was 1-4; 29 surveys used multiple 24hr recalls, 26 of which were conducted post-2000. Table 2 illustrates that where countries used both 24hr recall and food diaries, this was a combination of methodological changes in waves of long-running surveys, different surveys using different methodologies or both methods being employed within the same survey for different population groups, for example, adults and children. A 2×24hr recall is the method recommended by the European Food Safety Authority (EFSA) for adults' NDS (16). Countries with surveys conducted post-2000 using multiple 24hr recall were Austria, Belgium, Bulgaria, Czech Republic, Estonia, Finland, France, Greece, Iceland, Kazakhstan, Latvia, the Netherlands, Norway, Portugal, Slovenia, Spain, Sweden, the former Yugoslav Republic of Macedonia and the United Kingdom. Spain calculated usual nutrient intake from 24hr recall and a 3-day dietary diary.
Food diaries were used as a primary method by 47 surveys, 33 of which were conducted post-2000. The range of diary days per survey was 1-7. Thirty-eight surveys used multiple day diaries as the primary method, and 26 of these were conducted post-2000 from the following countries: Austria, Cyprus, Denmark, France, Greece, Hungary, Ireland, Italy, the Netherlands, Norway, Sweden and the United Kingdom. The majority of these were performed over consecutive days. Weighed diaries were used as the sole method by some surveys in France, Ireland, Italy and the United Kingdom, but also as a primary method by one survey in Germany.
Food frequency questionnaires (FFQs) were used by 12 surveys, 5 of which were conducted post-2000 (Estonia, Ireland, Norway, Romania and Slovenia). FFQs were used by Ireland, Norway and Slovenia in pre-2000 surveys and as a supplementary, rather than primary, dietary assessment tool by other countries (Andorra, Belgium, Greece, Hungary, Iceland, Latvia, Lithuania, the Netherlands, Poland, Slovakia, Spain and Turkey).
Of the 28 surveys that reported energy and nutrient intakes (see Table 2 for older NDS approaches where available), 10 used interviews -these were primarily (n = 8) face-to-face rather than telephone-based, and 3 of these were electronic, for example, computer or tablet-based.

National nutrition surveys in Europe
Respondents self-completed in 11 surveys, which were all food diaries. Electronic resources were utilised in five surveys, just two of which were web-based. Five surveys used multiple approaches -these were mainly a combination of face-to-face and telephonic interviews with the exception of Spain, which used both interview forms plus a tablet and camera-photos.  Table 2 identifies these 28 surveys and illustrates their differing methodological approaches. All 28 surveys included energy and also carbohydrate, fibre, fat and protein intakes. Most surveys (n = 25) included intake data on saturated fat (Germany and the Irish child and teen surveys did not): MUFAs (n = 25) (Germany, Irish child and teen surveys did not) and PUFAs (n = 24) (Germany, Irish child and teen surveys, and the Dutch DNFCS young children did not). See Appendix 2 and Fig. 3 for tabular and graphical summaries of the macronutrients included by each survey. The majority of surveys (n = 21) included intake levels of sugars in some form, either as total sugars or as added sugars or sucrose. However, Cyprus, Germany, the Irish child and teen surveys, Latvia, the Spanish ENIDE survey and Turkey included neither. Given current concerns about sugar consumption, this is an important gap. Few surveys (n = 6) included data on starch intakes and less than half (n = 9) included trans-fatty acid (TFA) intakes (see Appendix 2).

Energy and nutrient coverage
All surveys with the exception of the Spanish ANIBES study included some micronutrients of interest (see Appendix 3 and Fig. 4). However, none of the micronutrients investigated was reported by every survey. Vitamin A, riboflavin, thiamine, vitamin B6, vitamin B12, vitamin C, vitamin D, calcium, magnesium and iron were reported by 26 or more surveys. Copper (13), iodine (13), selenium (11) and fluoride (1 -not tabled) were reported by fewer than half the surveys.

Data collection
This report details the initial findings of a review into dietary surveys across the 53 countries within the WHO Europe remit (17). Nationally representative surveys which collected data on whole diets at individual level   since 1990 were found for only 64% of countries, the main gaps clearly lying in 17 countries in the Central and Eastern European region of the WHO Europe remit. Although eight countries without NDS had recently completed a comprehensive WHO STEPS survey, including questions on fruit and vegetable intake, salt consumption and use of fats and oils in cooking and eating, the survey does not address whole diets and only included adults; therefore, this represents a knowledge gap. However, non-nationally representative surveys were found in two countries that had no other NDS, which demonstrates that although some countries have no nationally representative surveys, other initiatives are in place and the expertise and fieldwork experience needed to conduct NDS may be present. All Western European countries had published survey information after 2000. Of countries with NDS, 16 conducted long-running surveys with multiple collection waves, which could generate important information for trends analysis. Fewer surveys were available that measured diet in children than adults; again gaps were primarily in Central and Eastern European countries. This implies that nutrition policies in this region are based on limited data, which is of concern, as overweight and obesity have tripled in some of these countries since 1980 and NCD prevalence rates are reaching those of Western Europe (1).
Emailing nutrition experts and general Internet searches were the most successful data gathering methods. A major source for contacts and survey information was a global survey review from 1990 to 2010 (5). Few academic papers met the pre-set inclusion criteria in the systematic database search performed for countries -particularly Central and Eastern countries -with no surveys or contacts mentioned in previous reviews, which also minimises the risk of bias. A possible explanation is that survey results and characteristics may be published as government or other official reports rather than academic papers. However, we also undertook wider web-based searches, targeting known government and public health agencies using various search terms to account for this. Another reason is that dietary assessment in large-scale studies like national diet surveys is costly, due to the labour-intensive nature of study preparation and data collection, and therefore may not be undertaken by some countries (18). This could explain the disproportionate concentration of gaps in survey provision in Central and Eastern European countries, which tend to have lower national incomes (19). This highlights a need to clarify major barriers and work with countries to establish mechanisms to overcome these and subsequently to devise and implement NDS.

Dietary methodologies of post-2000 surveys
The most common methods of collecting dietary intake used in the 78 post-2000 surveys were the 24hr recall and food diary, the majority of which were collected over multiple days. Although 24hr recalls are known for under-reporting (20), their increased use could reflect their advantage in being less onerous for respondents and potentially providing more consistent results across all age and sex groups compared with other methods (21). Retrospective dietary recalls can provide detailed information on eating patterns and exert less influence on food choice than food diaries (22), thereby generating a more accurate and realistic report on population nutrient intake. However, such short-term dietary assessment methods are associated with within-person errors and wider variation of intakes within the population, particularly when intakes of only 1 or 2 days are collected, the latter as recommended by EFSA (16). Although FFQs provide long-term assessment, they nevertheless can present inflated energy and nutrient intakes (21), which could explain why few post-2000 surveys used FFQs as the primary dietary assessment method.
Prospective weighed and non-weighed food diaries allow very detailed information to be gathered on multiple days (22) and are sometimes used to validate other methods using a small sub-sample, but have a high respondent burden and like the 24hr recall, are susceptible to under-reporting (23). Food dairies with weighed intake are particularly burdensome and prone to response bias and respondent fatigue (24) -most likely the reason why fewer studies used it as a primary assessment method and the United Kingdom moved from weighed intake to estimated intake.
Many studies used multiple tools to collect food intake. Of the 22 countries for which energy and nutrient intakes were reported, all surveys that collected dietary intake using more than one tool generated energy and nutrient intake data from a primary method and used the other method(s) as a means of validation and calibration. The exception was Spain, which was the only country that used a truly mixed methods approach. Food diaries and 24hr recalls do not provide insight into usual intakes, whereas FFQs are less accurate in estimating individuals' absolute intakes; combining methods could help rectify these shortcomings (24). Spain, Belgium and the Netherlands estimated 'usual' intakes using the Statistical Program to Assess Dietary Exposure (SPADE), although the Dutch intakes presented by age group in this report reflect the average of actual intakes reported by individuals. Of the other countries employing FFQ as a supplementary method, Greece and Iceland also explicitly stated that this was used to estimate usual intake. This approach is designed to overcome within-person errors and wider intake variations when only 2 days of intake have been collected, although methodological limitations cannot be fully negated.
Of the 23 surveys that sampled children only, over half (n = 15) used some form of food diary. This could be because children are expected to remember less retrospectively, so prospective methods of capturing intake, although subject to under-reporting and the limitations mentioned above, are deemed preferable and more accurate. This also fits with EFSA guidance on the collection of national food consumption data, which recommends countries '…use the dietary record method for infants and children and the 24-hour recall method for adults' (16). EFSA further recommend data be collected on two nonconsecutive days and that they be supplemented with a food propensity questionnaire (16). It remains to be seen whether more countries will move towards non-consecutive diaries in future surveys; at present, the majority of multiple food diaries are conducted on consecutive days. More detailed methodological recommendations for NDS of children are available via the Pilot study for the Assessment of Nutrient intake and food Consumption Among Kids in Europe (PANCAKE) project (25).
Of the 28 surveys that reported energy and nutrient intakes, Austria, Estonia, Iceland and Norway moved to 2×24hr recall in the latest NDS, perhaps to comply with the latest EFSA guidance (16). The United Kingdom switched from a 7-day weighed to a 4-day estimated food diary, which is more likely a move to reduce respondent burden. Although methodological changes make comparisons problematic across survey waves, the move towards a common approach will ease comparisons between countries in the long term and should be actively encouraged in line with EFSA recommendations. Although this could be logistically and financially challenging, it would assist in making inter-country comparisons and identifying vulnerable groups, thereby enabling the effective targeting of policy resources.

Technology in national dietary surveys
Care is needed in any dietary assessment method to minimise measurement error. Many dietary assessment methods require highly skilled interviewers, which increases survey costs and presents a potential barrier to conducting NDS (24). Technology like computer-administered interviews and image-capture could help overcome this obstacle and also promote standardised practices. The European Prospective Investigation into Cancer and Nutrition (EPIC)-Soft software package developed by the EPIC Study provided uniform templates for various aspects of NDS including conducting 24hr recall, which has since been modified by the European Food Consumption Validation (EFCOVAL) Study and renamed 'Globodiet'. It aimed for Europe-wide use, but is limited by the need for professionals to be trained in its use (26).
At present, none of the surveys identified used mobile technologies to collect dietary information; although Belgian, German and Portuguese surveys employed electronic interviews, the Spanish ANIBES used tablets and the Norwegian Ungkost3 and Swedish Riksmaten used a web-based food diary. This current lack of use may be due to the lack of validation or differential usability across population groups. However, web-based dietary assessments with self-administered record or recall methodologies have the potential to reduce data entry expense and allow data collection for large numbers on multiple days over different time periods (27). They could therefore be more cost-effective and encourage countries for which cost has been a significant barrier to undertake surveys. For example, myfood24 is an online 24-hour dietary assessment tool that can be used for either of the EFSA-approved (16) 24hr recall or a food diary methods (27). It employs country-specific food databases and is currently in operation in Denmark, Germany and the United Kingdom. Technologies like this could reduce the onus on researchers by automatically coding food records (27). These benefits could encourage countries that historically lack national diet survey provision to undertake surveys and enable countries that already undertake surveys to implement these at more regular intervals. This would serve to increase the amount of dietary and nutrient intake data available in the WHO Europe remit, directly contributing to the WHO objective of strengthening and expanding nationally representative diet and nutrition surveys WHO (1).

Energy and nutrient intakes
Energy and nutrient intake provision was documented from the latest survey collected after 2000 for each country for which we could locate intake data. For some countries, more recent surveys had been conducted (see Table 2), but intake data were not yet available in all cases. An additional limitation on data availability was the range of nutrients each survey covered. Of the countries that specified nutrient intakes, Germany and Belgium were the most likely to have gaps in reported intakes of macro-and micronutrients, respectively, and the Spanish ANIBES survey (28)(29)(30) only reported macronutrient data (see Appendix 3). This suggests that the reporting of nutrient intakes is inconsistent, making it harder to assess nutrient coverage and make intercountry comparisons.
Inconsistent age groupings across countries also make inter-country comparisons potentially problematic. In Andorra, the youngest age group spanned adults and children, meaning that although children were sampled, intake levels would not be valid in any comparisons. Future investigation could be undertaken using raw data and consistent age groups to obtain more reliable conclusions.
Differences in dietary methodologies may be a limiting factor when making inter-country comparisons. The relatively low levels seen in Turkish adult and child energy intakes compared to other countries could potentially be explained by methodological differences. The Turkish survey used a single 24hr recall, whereas the Belgian, Danish, German, Hungarian, Dutch, Norwegian and Spanish surveys, whilst using different methodologies (see Table 2), all collected data on multiple days. Collection on a single day is more likely to result in error due to less control over day-to-day variation (31).
Lack of alignment and completeness of national food composition databases and classification systems is a further limitation. For example, some food composition databases may not be updated to account for reformulated products, which could introduce differences and potential error in the energy and nutrient content of foods and therefore population intakes as reported in NDS. Common approaches to food composition databases are set out in more detail in the EFCOVAL study (144). Energy and nutrient intake values will be reported and discussed in more detail in future publications (145).

Strengths and limitations
The strength of the current review is that it presents a unique, current overview of dietary survey characteristics in all WHO Europe countries since 1990. The existence of newer studies such as Bel-Serrat et al. (146) illustrates the fluidity of the situation and the need for updated, comprehensive reviews. This review includes surveys covering both adults and children; therefore, it provides a full picture of the current state of dietary survey provision across the life course. It also discusses methodologies, enabling insights into common methods and paving the way for future exploration of best practice and policy recommendations.
However, the surveys employed different methodologies, both between surveys and within long-running surveys with multiple collection waves, potentially making the task of comparing countries problematic. Despite this, we feel that there is still a need to use the available information to make inter-country comparisons where possible. Another limitation of the review was that we were unable to establish contact with nutrition experts or government officials who may be working in nutrition in some of the 19 countries where no surveys were found, which were mainly Central and Eastern European countries. Therefore, we cannot ascertain that these countries do not have any relevant dietary surveys. We also cannot assure that there are no other nationally representative surveys in countries where we obtained survey information from contacts. However, it is likely that these contacts would be aware of other surveys in their countries; in the distributed questionnaire, contacts were asked for details of all surveys in their country.

Conclusion
This review found that less than two-thirds of the 53 countries in the WHO European region conducted national diet and nutrition surveys since 1990, with only 22 countries reporting nutrient intake data since 2000. The main survey gaps for both adults and children lie in the Central and Eastern European countries, where nutrition policies may lack an appropriate evidence base. Differing dietary assessment methodologies may have impact on the ability to make inter-country comparisons; existing efforts to harmonise NDS across all ages, particularly guidelines set by EFSA (16), should be encouraged, including beyond Western Europe. It would therefore be beneficial to target future efforts at standardising methodologies and filling knowledge gaps for the countries that have no surveys post-2000 in order to increase the information available for evidence-based policy planning. By establishing which countries have NDS, this review lays the foundation for a future review and stratified analyses of actual nutrient intakes across population groups in Europe.

Appendix 1. Questionnaire concerning nationally representative diet and nutrition surveys and their methodologies
Please complete one questionnaire per diet and nutrition survey (DNS) for questions 1-3; if necessary make multiple copies. If there any questions in sections 1-3 that you cannot answer, please provide contact details of a person(s) who may be able to answer the outstanding questions.
Please email the completed questionnaire(s) to Holly Rippin fshr@leeds.ac.uk at the University of Leeds, who will be collating this information for the European Office of the World Health Organization. please fill in the below information: Please note that any survey to be included should meet the following criteria: • The survey should collect dietary intakes across all food groups which are then converted into nutrient values. • The survey uses national population-based samples or representative regional samples. • The survey should not be restricted to a specific part of the population (e.g. children, occupational groups or patients). • Preferably there should be plans to repeat the survey later, unless it already has been repeated. You can also record standalone surveys.