Dietary data in the Hertfordshire Cohort Study

Learn about the Hertfordshire Cohort Study (HCS) and its dietary measurements

Summary of cohort
Dietary data collection
Estimation of nutrient intake
Response
Key findings

Summary of cohort

From 1911 until the National Health Service (NHS) was formed in 1948, records of birth weight, child illness, development and infant feeding were kept in Hertfordshire and summarised in handwritten ledgers [16]. By linking these records to mortality data through the NHS Central Register, Professor David Barker and colleagues were able to link markers of early experience to later health, most notably showing that lower birth weight was associated with increased risk of death from cardiovascular disease [17]. These initial studies that had included men and women born between 1920 and 1930 were followed by the Hertfordshire Cohort Study (HCS). Participants of the HCS included men and women who (a) were born 1931-1939, (b) had early life information from the ledgers, and (c) were still alive and registered with a General Practitioner in Hertfordshire between1998-2002 (traced using the NHS central registry). Approximately 3,000 men and women agreed to a home visit with a trained research nurse and a majority of them also completed a clinic visit [16]. Sub-samples of these participants have also participated in a number of follow-up studies, principally focused on musculoskeletal outcomes.

The main objective of the HCS is to examine the interactions between genes, pre- and post-natal environments and adult diet and lifestyle behaviours in the aetiology of chronic disorders in later life [16].

Dietary data collection

Infant diet

The original Hertfordshire ledgers (from birth up to the first year of life) summarised information about infant feeding in the first year, including the type of milk feeding in infancy categorised as: breastfed only, bottle and breast-fed, or bottle-fed only. Records from individual home visits with more detailed information about type of feeding and duration were not retained [18]. The questionnaire can be found at: https://discovery.closer.ac.uk/item/uk.mrcleu-uos.hcs/17b6214b-9cbf-4a4b-a4c6-3397404e7dfe/1.

Diet in adults

Information about dietary intake was collected during the baseline home visit (1998-2002) using a nurse-administered FFQ (questionnaire can be found at: https://discovery.closer.ac.uk/item/uk.mrcleu-uos.hcs/70bd04df-78cd-43f8-b25c-78151a5b90e4/1). The FFQ was modified from the European Prospective Investigation of Cancer (EPIC) questionnaire. This FFQ has been previously validated for use in a UK population [19]. The FFQ includes 129 food groups and foods and asks the participants to record the average frequency of consumption (never, <1/month, 1-3/month, 1/week, 2-4/week, 5-6/week, 1/day, 2-3/day, 4-5/day, ≥6/day) over the 3 months preceding the interview. The frequencies of consumption of foods not listed on the FFQ were recorded if they were consumed ≥1/week. Daily amounts of milk and sugar consumed were also recorded. Prompt cards listing example foods included in each food group were used to help standardise responses to the FFQ. At the end of the FFQ, participants were asked about their use of dietary supplements during the previous three months. If they answered yes to using supplements, further details including the name and brand of the product as well as dose and frequency of use were requested. A total of 954 different dietary supplements were used by the cohort participants [20]. A shorter version of this FFQ was developed to assess diet quality in later data collections in Hertfordshire, including a sub-group of the participants in the East Hertfordshire follow-up (n=442) studies as part of a wider European project on osteoarthritis [21, 22].

24-hour food diaries were also collected at baseline but they have not been used in any publications. In 2014, a series of focus groups to explore the influences on diet were held among 92 participants aged 74-84 years whose diets had already been assessed twice; once in 1998-2001 and once in 2011 [23].

Estimation of nutrient intake

Following the general procedure outlined in Estimating nutrient intakes from DATs, standard portion sizes were allocated to each food in the FFQ based on previously published references [24]. Nutrient intakes were calculated by multiplying the frequency of consumption of a portion of each food by its nutrient content based on the UK national food composition databases (McCance and Widdowson) [25] or manufacturers’ composition data where appropriate. Nutrient intakes from dietary supplements were calculated using the frequency and dose reported by the participant, and manufacturers’ supplement composition data.

Response

Response to dietary measures in HCS

Year	Age (y)	N interviewed	Response to diet question (n(%))
1998-2002	59-73	3,225	3,217 (99%)
2011	71-80	592*	442 (75%)

Notes.
*Invited to interview.

Key findings

Infant diets

Participants of the HCS born between 1931 and 1939 unsurprisingly had different infant feeding patterns compared to what can be seen today. For instance, over half of the participants were breastfed and there was no association between socioeconomic position (SEP) at birth and type of infant feeding [18, 26]. This is useful because unlike associations from later generations, relationships between infant feeding and health outcomes are unlikely to be confounded by SEP. Breastfeeding in this cohort was associated with greater adherence to a prudent dietary pattern with authors of this study suggesting that early feeding may be linked to later food choice [10, 18].

Adult dietary patterns

Dietary data from the baseline FFQ have been used to identify dietary patterns (DP) using principal component analysis (PCA). Two main dietary patterns were identified using data from the HCS participants at baseline: 1) a diet characterised by high consumption of fruit, vegetables, oily fish and wholemeal cereals termed a “prudent” dietary pattern and 2) a diet characterised by high consumption of vegetables, processed and red meat, fish and puddings called a “traditional” dietary pattern [27]. The FFQ data in HCS have also been used to develop a shorter 24-item FFQ that has the ability to define the prudent dietary pattern in a comparable way to the full FFQ [22]. This tool has the potential to benefit future studies that are interested in capturing information on diet quality of older participants but do not have the capacity to complete a long FFQ.

Learn about the other studies covered by this guide and their dietary measurements:

Get background detail on this guide:

Learn about harmonisation in the context of dietary data:

Discussion & illustration of dietary data harmonisation

Further information:

This page is part of the CLOSER resource: ‘A guide to the dietary data in eight CLOSER studies’.