PANCAKE – Pilot study for the Assessment of Nutrient intake and food Consumption Among Kids

Summary

Project developed on the procurement project CFP/EFSA/DATEX/2009/02

Aim

The objective of the Pilot study for Assessment of Nutrient intake and food Consumption Among Kids in Europe (PANCAKE) project was to develop, test, and evaluate tools and procedures for a future harmonised pan-European food consumption survey (EU Menu) among infants, toddlers, children (up to ten years), and breastfeeding women. With regard to the dietary assessment tool and related administration protocols it was aimed to develop and test two versions: Version 1) a three-day food diary checked with the parent/caretaker and data entry afterwards using EPIC-Soft ®; and Version 2) two non-consecutive one-day food diaries followed by an EPIC-Soft completion interview with the parent/caretaker, one by telephone and one face-to-face at the participant’s home.

Tools and protocols

Developed tools and databases for the collection of food consumption data among the required age groups included:

Developed protocols and guidelines included:

Validity of developed picture series

Picture books for infants, toddlers and children were developed and validated. Identical validation sessions were conducted in three European countries. In each country 45 foods were evaluated using the 38 common picture series of six pictures each. For each food 36 pre-weighed portion sizes (each picture was evaluated six times by means of two portions lower, two portions equal and two portions higher than the depicted portion) were evaluated by convenience samples of parents having children (three months - ten years). Percentages of participants choosing the correct picture, and the adjacent or distant picture to the correct picture were calculated and performance of individual pictures within the series was assessed. For 20 food items the picture series performed acceptably. Additional 12 food items were rated acceptable after adjustment for density differences, while, some other series (e.g. on baby food) became acceptable after in-depth analyses at country level. However, performance of the series of salads and cakes remained problematic. It was concluded that all of the picture series were acceptable for inclusion in the PANCAKE picture book. However, picture series on baby food, (mixed) salads and cakes either can only be used for very similar foods as depicted on the pictures or need to be substituted by another quantification tool.

Recruitment and sampling in the pilot studies to test the developed tools and protocols

Two pilot studies were conducted using the two alternative dietary assessment methods; each with 94-96 participating children in Belgium and Czech Republic. In both countries, infants (from three months onwards), toddlers, and children were sampled from the national registers. In order to deal with non-response three age-sex matched replacement individuals were connected to each selected individual (called a cluster), and reserve clusters of individuals were drawn additionally. Recruitment was country-specific. In Czech Republic, recruitment was by post, with door-step or telephone recruitment in case of no reply to the invitation letter. In Belgium, recruitment was directly through the door-step or telephone after sending an explanation letter. Response varied between 47-52% across countries and pilot studies. The number of breastfeeding women attached to the infants in Belgium was very low (two and eight for both protocols as compared to 20/21 in the protocols in Czech Republic). Overall, sampling and recruitment took a lot of time, but it was feasible. For the future, when the study covers the whole year, repeated selection of participants is needed in short periods to cover the youngest respondents. Direct oversampling rather than working with matched replacement individuals is recommended to limit time restrictions. With regard to breastfeeding mothers it is encouraged to look for alternative sampling frames.

Administration protocol with the three-day food diary

For the pilot study with the three-day food diary, field work was performed in two home visits, followed by data entry in EPIC-Soft. In each country this was done by seven to eight interviewers that were trained dietitians. The first home visit consisted of: a) Instruction for completing the three-day food diary, including the food recording booklet for eating out of home; b) Face-to-face administration of the general questionnaire on paper; c) Handing over the food propensity questionnaire and the evaluation questionnaire to be filled out before the second home visit. This home visit lasted on average 35 minutes. Then the caretaker kept the food diary for three-days. Afterwards, the second home visit consisted of: a) Measurement of weight and height of the subject; b) Checking of the three-day food diary on completeness, required details and consumed quantities; c) Measurement of household measures indicated in the diary; d) Collection and quick check of the food propensity questionnaire; e) Completion of the evaluation questionnaire by the participant; f) Handing over of an incentive to the subject (allowed in Czech Republic only). This visit took about 50 minutes on average. Afterwards, the interviewer entered the three-day diary information in the computer using an adapted version of EPIC-Soft (i.e. the EPIC-Soft data entry application). This data entry was done by the same person who checked the diary, and preferably as soon as possible after the second home visit, but at latest within two days. On average, this took 33 minutes in Belgium and 36 minutes in Czech Republic. If important information appeared to be missing, the interviewer was allowed to contact the subject (or day-care for out of home consumption) by telephone if the subject had given permission.

Administration protocol with the two one-day food diaries

The field work of this pilot study was performed in four steps, including data entry with EPIC-Soft. In each country, seven or eight trained dietitians conducted the four steps. The first home visit consisted of: a) Instruction for completing the food diary including the food recording booklet for eating out of home; b) For infants, measurement of weight and length of the subject; c) Measurement of generally used household measures; d) Face-to-face administration of the general questionnaire on paper; e) Confirmation of the appointment for the telephone interview on the day after completing the food diary, or two days later if this should be a Sunday f) Confirmation of the appointments for recording and interview of the second record day g) Handing over of the food propensity questionnaire and the evaluation questionnaire to be filled out before the second home visit. This visit on average took 38 minutes in Belgium and 42 minutes in Czech Republic. Then the caretaker kept the food diary on the fixed day. The day after or with one day in between, the first EPIC-Soft completion interview using the EPIC-Soft data entry application (with activated ‘probing’), in combination with the food diary was executed by telephone. Thus, the subject could use the diary to answer the questions of the recall and the interviewer did not see this diary. The telephone call lasted about 35 minutes on average. About two weeks later the interviewer called the caretaker to remind her/him to record the consumption of food and drink in the day-two diary. After this, a second home visit was conducted. This home visit consisted of: a) Measurement of weight and height of the subject (second measurement in case of infants); b) The second EPIC-Soft completion interview in combination with the food diary; c) Collection and quick check of the food propensity questionnaire; d) Completion of the evaluation questionnaire; e) Handing over the incentive to the subject (only allowed in Czech Republic) and taking along day-one and day-two diaries. The visit on average lasted 42 minutes in Belgium, and 51 minutes in Czech Republic. In Czech Republic participants were given the choice for an alternative approach: to conduct the completion interview by telephone call and have a subsequent home visit for the other measurements, but this option was not often used.

Evaluation of both administration and interview protocols with related tools

Both administration protocols were evaluated as feasible and some specific recommendations were made for further improvement. The most important are:

a) the food propensity questionnaire should include an answer category ‘never’ and should preferably be shortened;

b) the equipment for measuring the height and weight should be as light as possible; height and weight can better be measured during the first home visit;

c) one specific bug in the data entry tool should be fixed, the tool could be extended with forms used during recruitment and an overview function;

d) in the food diaries the recording of cooking fat should be stressed; the out-of-home booklet is not needed in connection to the one-day food diaries;

e) EPIC-Soft should be best installed by IT staff, and should include automatic backup procedures.

f) all tools related to EPIC-Soft (e.g. Interview Manager) as part of the dietary e-Standardised Methodologies Platform (e-SMP) should be finalised and sufficiently tested before implementation in the EU Menu project;

g) for the administration of the food diaries, instructions on the distribution of days of the week and maximum number of days between the steps of the protocol should be further specified

h) a form to report the quality indicators of the study could be added.

Evaluation of the most appropriate dietary assessment method

It was concluded that the two tested dietary assessment methods and related protocols were both feasible and each had their own advantages and disadvantages. The advantages of the protocol with two one-day food diaries were 1) that the diet recording is a smaller burden to the participants as compared to the three-day food diary, 2) the computer assisted completion interview is easier for the interviewers than the visual check of the three-day diary, 3) the collected consumption data seemed a little better in quality (less missing/unspecified values) than those collected through the three-day food diary, 4) the non-consecutive days of recording are probably more independent than consecutive days (requirement for usual intake modelling), and 5) the protocol with two one-day food diaries is more similar to the recommended protocol for adults. On the other hand, the protocol with the three-day food diary was 1) connected to shorter home visits as compared to the protocol with two one-day food diaries, 2) the logistics of the protocol with the three-day food diaries was easier, 3) costs seemed slightly less, and 4) for foods that are rarely consumed, more consumers can be found during three instead of two days of diet recording. The results do not allow to conclude whether one method is superior with regard to under- or over-reporting. Overall, it was concluded that the two non-consecutive one-day food diaries followed by an EPIC-Soft completion interview with the parent/caretaker offered more advantages for the future pan-European survey in children aged zero to ten years than the three-day food diaries.

Overall conclusion

The PANCAKE project provided tools and protocols for a future harmonised pan-European food consumption survey in infants, toddlers and children, with specific recommendations for further improvement and on the preferred dietary assessment method. The tools and protocols that enable the collection of standardized and detailed food consumption survey data were perceived feasible by participants and personnel involved in the survey. Study preparation and data collection are however labour intensive. Applicability of the conclusions to other member states should be considered, as well as aspects related to performing food consumption surveys on a larger scale. To ensure sufficient quality and efficiency, it is strongly recommended that inexperienced organisations will get step-to-step guidance and training in all aspects of the study methodologies, particularly those related to EPIC-Soft.