Diet and Nutrition Survey of Infants and Young Children in Scotland, 2011

Chapter 2 Methodology

2.1. Preliminary/development work

A series of pilot studies were undertaken to determine the best way to measure what infants and young children eat and drink in a national survey such as the Diet and Nutrition Survey of Infants and Young Children (DNSIYC) and to assess the use of utensils in the estimation of portion sizes in this age group. This pilot work was undertaken by the Human Nutrition Research Centre at Newcastle University. In addition a Dress Rehearsal of the entire protocol was conducted by the survey consortium to ensure all components of the survey were effective and viable. These preliminary studies covered:

• A consultation designed to ascertain from mothers their views on the most convenient and feasible way of measuring children's food consumption.
• Two pilot studies to test food diaries and equipment.
• A Dress Rehearsal to test the overall response to the survey and all components included in the protocol.

See Annexe B of the UK report for details of the consultation and pilots and Annexe C of the UK report for details of the Dress Rehearsal.

The following recommendations were made at the end of the pilot work:

• The mean food intake (g) and mean energy intake (kcal) recorded by the estimated intake^[1] method, supported by the use of measuring tubs and spoons, were statistically different to those of weighed intakes. The estimated intake method slightly underestimated the amount of food (on average 17g) and slightly overestimated energy intake (on average 32kcal). All estimates of food intake and 92% of estimates of energy intake lay within 50% of the weighed intake. The estimated intake method was, therefore, considered a feasible and viable method for use in the main survey.
• Parents had a preference for recording the amount of food served and left over, rather than calculating themselves the amount consumed so this was incorporated into the diary for the main survey.

Both these recommendations were adopted during the Dress Rehearsal. The recommendations made at the end of the Dress Rehearsal were numerous and are outlined in section 9, Annexe C of the UK report. These were all adopted during the mainstage survey.

Two significant outcomes were:

• It was decided not to use measuring equipment to assess quantities of foods consumed as it involved an avoidable step of converting volumes to weights. Household measures were used instead.
• Head circumference and infant length as well as weight were measured by interviewers.

2.2. Ethical approval

Following reviews by HNR's internal Research Governance Committee, an application for external ethical approval for the survey was submitted on 4 November 2009, for review by the Cambridgeshire 4 Research Ethics Committee (REC) at their meeting on 26 November 2009. A favourable provisional opinion was given on 4 December 2009, subject to further information requested from the consortium and some administrative issues. Full ethical approval for the survey was received from the Cambridgeshire 4 REC on 18 January 2010.

Further information about the ethical process can be found in Appendix A.

2.3. Sample design

A sample of infants and young children representative of the Scottish population aged 4 to 18 months was drawn using a multi-stage random probability design. Individuals were randomly selected from Child Benefit (CB) records provided by Her Majesty's Revenue and Customs (HMRC)^[2]; the Healthy Start (HS) sample^[3] was drawn from the HS recipient database provided by the Department of Health. The CB and HS samples were stratified by Government Office Region, Index of Multiple Deprivation (IMD) scores and population density to ensure representativeness (see Appendix A for details). At the time of the survey, CB was a universal credit with a high rate of take up (around 98%) making it a good sampling frame for a survey such as this. However, it does have limitations and these are discussed in Appendix A.

The Diet and Nutrition Survey of Infants and Young Children in Scotland (DNSIYCS) sample comprised three parts:

1) The children from Scotland who were part of the core sample of children selected at random from CB records covering all four countries of the UK.

2) A boost sample of children in Scotland, also drawn from CB records.

3) The children from Scotland who were part of a boost sample of children on the HS scheme selected at random from the DH's HS database. This boost sample was combined with children on the HS scheme in the core Scotland sample to form the Scottish Healthy Start sample.

These three groups comprised those reported together in DNSIYCS. Weighting factors were applied to ensure that the results were representative of the Scottish population (see section 2.3.5). More details about the sample selection process are provided in Appendix A.

2.3.1. Age

The sample drawn from the CB Register was selected in two waves. This was done to try to ensure that the sample contained sufficient numbers of children at each end of the eligible age range, i.e. from 4 to 18 months. Interviews with the parents of participants aged 17 months were prioritised to reduce the number of children who would become ineligible through being beyond the specified age range at the time of interview. Similarly, interviews with the parents of participants aged four months were prioritised to maximise participation of children of this age.

2.3.2. Inclusion/exclusion criteria

As stated in the introduction, exclusive breastfeeding is recommended for the first six months of an infant's life. However, complementary feeding often commences earlier than six months. Therefore, the sample included infants and children from four months to capture those infants who were receiving complementary foods earlier than six months. The full age range of the survey included children aged 4 to 18 months. Those who had turned 18 months by the time of the first interview were screened out.

Children with a birth weight under 2kg and/or those fed through a gastric tube at or after one week of age were excluded at the main interview stage. The latter criterion was chosen to exclude those individuals with congenital abnormalities likely to affect feeding practices^[4].

Participants who no longer lived at the selected address when the sample was drawn were still considered eligible to participate if they had moved locally and a new address could be obtained. The parent of the participant who had the most involvement in the feeding of the child was selected as the interviewee. Only parents or legal guardians could be selected to complete the interview and to provide consent.

All children for whom at least three days of the food diary had been completed were classed as fully productive participants.

2.3.3. Selecting postcode sectors

Postcode sectors were selected by NatCen using CB claimant records provided by HMRC and information about HS recipients from DH. See Appendix A for more information on selecting postcode sectors and participants.

2.3.4. Seasonality

Ideally, it would have been desirable to assess seasonal variation in diet across the sample, however, it was not possible for fieldwork to span an entire year due to limited access to CB records. Consequently, it was not possible to detect seasonal variation in the diet in DNSIYC or DNSIYCS and it is not reported.

2.3.5. Weighting factors

DNSIYCS required a set of weighting factors to adjust the sample for differences in sample selection and response. The weighting factors adjust for differential selection probabilities of boost sample members, non-response to the interviewer stage.

An interview weighting factor was required for participants who responded to the individual interview and completed three or four food diary days. This weighting factor was generated using a combination of logistic regression modelling and calibration. The aim was to reduce bias resulting from sampling error and differential non-response.

Further information about weighting the sample can be found in Appendix B.

2.4. Overview of methodology

This section provides an overview of the methodology employed on DNSIYCS. The survey aimed to collect data from a representative sample of 500 infants and young children aged 4 to 18 months living in private households in Scotland. This sample size was chosen to enable robust analyses of the different age groups in Scotland. The key components were:

• Face-to-face interviews, conducted using Computer Assisted Personal Interviewing (CAPI) with the parent most involved in feeding the sampled child.
• Dietary data collection, using a four-day estimated food diary.
• Anthropometric measurements (maternal height and weight; infant length, weight and head circumference).

Copies of all survey documents are included in Appendices C to H. The following diagram sets out the movement of participants through the survey.

Figure 2.A. Summary of stages included in DNSIYCS

Two weeks in advance of starting fieldwork, a letter describing the purpose of the survey was sent to each sampled child's parent or legal guardian (referred to as 'the parent'). This letter provided the parent or legal guardian of the sampled child with the opportunity to 'opt-out' of the survey prior to the beginning of fieldwork.

Parents of participants who took part in the CAPI interview and completed a food diary for at least three days were classified as 'fully productive'.

2.5. Overview of survey components and fieldwork procedures

Further details of all survey components and fieldwork procedures can be found in Appendix A.

Fieldwork assignments were issued to interviewers in two waves:

Wave 1: 6 January - 1 March 2011

Wave 2: 28 March - 23 May 2011

The interview visit consisted of:

• Detailed face-to-face interview using CAPI to collect background information on family dietary habits, socio-demographic status and health information, as well as information about the child such as feeding practices, eating patterns, developmental stages, sunlight exposure and gastrointestinal symptoms (see Appendix C).
• Dietary data collection (estimated food diary, completed for four consecutive days) to provide a quantitative estimate of food consumption and nutrient intakes (see Appendices D and E).
• Physical measurements (height and weight of mother; length, weight and head circumference of child) in order to assess growth (see Appendix A).

Details of documents used by interviewers are provided in Appendices F to H. The full methods, and benefits and limitations of various methods are discussed in Appendix A.

2.6. Feedback to participants and General Practitioners (GPs)

Various types of feedback were provided to participants' parents and General Practitioners (GPs) with parental agreement. These are described in Appendix A and examples can be found in Appendix I.

2.7. Fieldwork quality control

Quality control measures for fieldwork were put in place to monitor the quality of completion of the food diaries and also to monitor interviewer measurements. Refer to Appendices A of this report and Appendix L of the UK report for further details.

2.8. Tokens of appreciation

At the final interviewer visit, the interviewer gave a token of appreciation (£30 in high street vouchers) to all who completed at least three food diary recording days.