Dietary Surveillance and Nutritional

Assessment in England:
What is measured and where are the gaps?

March 2010

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 1

NOO is delivered by Solutions for Public Health

Executive summary

Purpose and scope of the paper

This paper identifies, describes and discusses the main sources of data for national-level
dietary intake and nutritional status in adults and children in England. It is intended to
support the development and monitoring of public health nutrition policies and
services.

Descriptions are provided for the available sources of data, the methods used for their
collection, and their limitations.

The focus is on data that are systematically and regularly collected for nutritional
assessment and analyses of trends in food consumption and nutritional status of the
national population.

The paper does not systematically cover sources of regionally or locally generated data,
or nutritional assessment data that have been collected to evaluate the impact of
specific interventions.

Key sources of surveillance data and the data they provide

Key sources of surveillance data on the dietary habits of adults are:

z Health Survey for England consumption of fruit and vegetables, bread,

butter and spreads, milk, cereal, pasta, potatoes, cheese, meat, fried food,
fish, snacks, cakes and pastries, sweets and chocolates, fizzy drinks and salt.
z National Diet and Nutrition Survey detailed dietary intake and
nutritional status data.
z Low Income Diet and Nutrition Survey detailed dietary intake and
nutritional status in a sample of households categorised as being from the
15% most materially deprived in the population.
z Expenditure and Food Survey detailed data relating to expenditure
on food and drink in a sample of households over a two-week period.
z Primary Care Trust Patients Survey number of days in a week
respondents eat five portions of fruit or vegetables.

Key sources of surveillance data on the dietary habits of infants and children are:

z Health Survey for England consumption of fruit and vegetables and in

some years, consumption of other foods as per adult data collection.
z TellUs Survey - the average number of portions of fruit and vegetables
consumed each day.
z National Diet and Nutrition Survey detailed dietary intake and
nutritional status data.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 2

 z Infant Feeding Survey estimates on the incidence, prevalence and
duration of breastfeeding and weaning practices.
z School Food Trust Local Authority Survey national, regional and
local information about the take-up of school lunches in maintained
schools.

Discussion and conclusions

There are many different elements to dietary intake and nutritional assessment. The
nature of the tool or tools best used to assess these elements depends on a
combination of the nature of the information required and the context within which
the information will be collected. This makes it challenging to summarise the topic or
to offer concrete suggestions. However, a number of points still stand out as worthy of
consideration.

There are several national surveys that systematically collect different types of data
(using a variety of collection methods) to enable estimates to be made for the
nutritional status and dietary intake of the population. The most comprehensive of
these surveys in terms of nutritional assessment, is the National Diet and Nutrition
Survey (NDNS). In 2008 this became a rolling programme of annual data collection from
a sample of approximately 500 adults and 500 children. In time, this will provide useful
national trend data. The NDNS will not be suitable for sub-national analyses until
pooled year-on-year data are available for regional analyses.

The Primary Care Trust Patient Survey and the TellUs survey have sufficiently large
sample sizes to allow Primary Care Trust (PCT) and local authority level analyses, but
they only gather data relating to consumption of fruit and vegetables using questions
for which there is no evidence of validation.

The Health Survey for England (HSE) has tracked fruit and vegetable consumption in
large samples of adults and children since 2001. In some years, the HSE collected wider
dietary data using a Food Frequency Questionnaire (FFQ) which, to our knowledge, has
not been validated. The HSE does not provide sub-regional level data.

The Expenditure and Food Survey provides comprehensive data for tracking and
monitoring trends in food purchasing, and derives food and nutrient intake data that
includes an allowance for waste. Data are estimated for individuals as they are
collected at household rather than individual level.

Increased sample sizes in the NDNS and HSE would be needed to allow tracking of
trends within sub-groups, and robust sub-national analyses.

Additional testing and development of self-report data collection tools could result in
more valid and reliable data. This could potentially provide tools that could be used in
local areas to gather accurate and consistent data from their populations.

The development of a set of proxy indicators of dietary intake (beyond fruit and
vegetable intake), that are indicative of a healthy diet and overall energy intake, could
be developed and validated for use in both national and local level surveillance.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 3

There would be significant benefit from harmonising questions across national surveys
as part of a more generally strategic approach to the collection of these kinds of data.
If a standardised set of dietary intake and behaviour questions were included in all such
surveys this would simplify their analyses and allow direct comparison of data from a
range of different sources. Establishing standards such as these could improve
outcomes and increase efficiency savings.

The Integrated Household Survey (HIS) and the new Health and Social Care Survey
provide opportunities to review and refine current approaches to national-level dietary
data collection.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 4

1. Introduction to nutritional assessment

Surveillance of the diet and nutritional status of populations is an important

component of a public health approach to informing nutrition policy, promoting
healthy eating and reducing obesity. Data on dietary intake and nutritional status can
help improve the targeting of interventions, while trend data can show the impact of
policy over time.

The measurement of dietary intake and nutritional status is complex and presents
significant challenges, particularly at a population level. The method of measurement
will depend on the objectives of the surveillance and the type of data required. For
example, a study to estimate the proportion of people in a population eating five
portions of fruit and vegetables a day will require a different method from one that
assesses the proportion of the population that consumes sufficient levels of iron.

The objectives of the surveillance will determine the types of information it should
produce. For example, a policy maker may be interested in dietary intake data to
measure the behavioural response to a campaign to encourage healthy eating, or to
estimate the differences between the nutrient intakes of particular population sub-
groups. By contrast, a commissioner may require more detailed information about the
nutritional status of a population to determine how nutrition and dietetics services
should be resourced and targeted.

The Healthy Weight, Healthy Lives strategy1 published in January 2008 details the
governments ambition to reverse the rising tide of obesity and overweight in the
population by ensuring that all individuals are able to maintain a healthy weight. In
order to monitor progress towards this the strategy highlights the importance of
nutritional assessment data relating to:

z nutrient intake
z consumption (and/or sales) of foods high in fat, salt and sugar foods
z proportion of the adult population consuming five portions of fruit and
vegetables per day.

1.1. Nutritional status

Nutritional status can be assessed by combining measurements of dietary intake,

anthropometric indicators, biomarkers and clinical or physical indicators.2 These
markers can be used in combination or in isolation depending on the objectives of the
surveillance.3

Anthropometric indicators are measurements of the human body which, when

compared with standards that are typical of a reference population, can indicate
abnormal nutritional status.4 Biomarkers are biochemical indicators that can be
assessed in blood, bodily fluids, body tissues or excreta. They provide an indication of a
limited number and range of nutrient levels, or can be used, with varying precision,5 -
as a proxy measure for the intake of some nutrients. A number of nutritional
deficiencies can be identified from the physical appearance of the body when clinically
examined.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 5

For the NDNS, food diaries and anthropometric measures are taken and analysed
alongside non-fasting blood and urine samples. Blood samples are analysed to measure
the concentrations and functional adequacy of nutrients absorbed by the body,
including amounts held in body stores.6 Together, these provide an indicator of
nutritional status. Urine analysis is used to give indications of sodium and thus, is used
as a proxy for salt intake.

1.2. Use of dietary reference values

Nutrient intake in a population can be compared against the UK Dietary Reference

Values (DRVs) defined by the British Committee on Medical Aspects of Food (COMA) in
1991.7 These values provide a series of estimates for the amount of energy,
macronutrients, and some of the micronutrients required for good health in different
populations within the UK. Within these values are three types of estimates: lower and
upper reference nutrient intakes (LRNIs and URNIs), and estimated average
requirement (EAR).

RNIs are used for protein, vitamins and minerals. They provide an estimate of the
amount required to meet the needs of the vast majority of the target population.

LRNI values refer to the mean of the populations requirement, minus two standard
deviations. Intakes below the LRNI are almost certainly inadequate for most people,8
and are unlikely to prevent signs of deficiency.

URNI values refer to the mean of the populations requirement, plus two standard
deviations. They refer to the point at which it is unlikely that even 2.5% of the
population would not achieve adequacy.9 These are reference values as opposed to
minimum targets, and they are not intended to be used as recommendations for
intakes by individuals or particular groups, such as people with illnesses.7

For some nutrients, there are not sufficient data available to provide LRNIs or URNIs.
Therefore, safe intake levels have been described. These are considered to be adequate
for most peoples needs but are insufficient to cause undesirable effects (some
nutrients, such as Vitamin A, can be toxic above certain levels).

The Estimated Average Requirement (EAR) is used only for overall energy intake. This is
because it is not desirable to recommend energy intake above the average
requirement.

1.3. Food intake and derived nutrient intake

In the context of public health nutrition, self-report methods are commonly used to
collect food intake data because they usually use fewer resources than other methods.
This can be important for public health nutritionists who aim to influence behaviour.3

Assessment of food intake is potentially subject to many sources of both random and
systematic error.10 Studies frequently rely on the accurate reporting of habitual food
intakes by a sample of free-living individuals within a population.11 The recall ability
and psychological characteristics of individuals can influence dietary reporting. For
example, an individual may be aware that her diet is unbalanced and so may be

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 6

reluctant to provide honest answers to questions or, recollection of intake may simply
be flawed. There is also evidence that participants may report behaviour that they
perceive as socially desirable rather than accurate.12

The instrument used to collect the data (see 1.4 below) can have an under-or over-
estimating effect. Food composition tables13 are frequently used to calculate estimates
of nutrient values from food intake data. These tables make assumptions about the
formulation of specific food, but such assumptions may not be accurate.

Wherever possible, it is important to identify and quantify the sources of potential

errors when assessing food intake, particularly when deriving nutrient intake.
Errors can be minimised through careful design of the study and analyses of the data,
but they must be taken into account when analysing data or interpreting existing
analyses.

1.4. Self-report methodologies

There are several different methods or instruments for obtaining self-reported

information about food intake. In principle, all of them should be tested for reliability
and validity.14

Reliability refers to the likelihood that an instrument will measure the same thing each
time it is used - either with the same or a different respondent. Validity refers to how
accurately the instrument reflects the actual behaviour.

For example, an instrument that was developed to monitor population level nutrient
intake may not be valid and reliable for evaluating change in dietary intake for a
smaller group participating in a healthy eating intervention. Therefore, testing for
validity and reliability should be appropriate to the nature of the surveillance and
target population.

Self-report tools can be tested against more reliable and objective methods of
assessment,3 such as using doubly labelled water to measure energy intake or nitrogen
or protein intake. Biomarkers such as these are not subject to the same sources of error
and bias that a self-report instrument would be. However, they are expensive and are
usually impractical in a public health surveillance context.

1.4.1. Food frequency questionnaires

FFQs give an indication of the habitual consumption of particular foods or nutrients

over a specific period of time. Information is collected on the frequency and,
sometimes, portion size of foods eaten. Nutrient intake estimates can be derived15 from
this. FFQs differ greatly in length and characteristics, depending on the objectives of
the study.

FFQs use a closed list of foods which has the advantage of reducing the burden on the
participant. However, this assumes that the list accurately reflects the most commonly
consumed foods within the population being examined. This assumption increases the
potential for error. It may, therefore, be appropriate to rank subjects as opposed to
giving estimates of absolute intakes.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 7

FFQs should be designed and validated for the specific purpose they are intended.16
Careful consideration must be given to the suitability of pre-existing questionnaires
before they are used or adapted for other surveys.17

1.4.2. Food intake records and diaries

Food intake records require an individual to record everything they eat over a specific
period, often four to seven days. They are considered to be the gold standard of self-
reporting techniques and have been used to validate other methods.

Participants are asked to weigh food or are given instructions on how to record detail
such as portion sizes or the brand of what they have consumed. The method therefore
relies heavily on a participants accurate recording and conscientious adherence to the
methodology prescribed.

Food records can be highly burdensome for the participant and there is evidence that
recording can become increasingly inaccurate over time.18,19 Participants may alter their
normal eating behaviours in order to ease the burden of recording, or as a response to
increased awareness of their food choices through participating in the study.3 There is
also evidence that individuals tend to under-report energy intake.11

Some of these problems can be partially resolved through the use of supporting
equipment such as pictorial aids, voice recorders, computer-recording and self-
recording scales.15

1.4.3. Dietary recall

Dietary recall methods are typically administered by an interviewer and can be carried
out either face-to-face or by telephone. The respondent is usually asked to recall all
food and drink they have consumed in the past 24-hours. Interviewers can ask probing
questions to prompt memory. This method can be carried out without the respondent
having prior warning, which reduces the likelihood of diet alterations arising as a result
of the measurement process itself. This method places a much lower level of burden on
the participant, requires less completion time, and does not involve weighing food.
Dietary recall reporting is also suitable for use in populations with low levels of literacy.

The main disadvantage is that individuals do not always report their intake accurately.
In addition, this technique does not give a picture of habitual eating as it only relates
to a single days intake. In order to overcome these disadvantages, some surveys - such
as the Low Income Diet and Nutrition Survey (LIDNS), carry out a number of 24-hour
recall interviews within a longer period. This provides a measure of within-subject
variability to help to establish how accurately respondents have reported their habitual
diet.6 Visual aids for portion sizing can also be used to improve accuracy of recall.

Multiple pass methods have been used, for example in the National Health and
Nutrition Examination Survey in the US.20 This is where an initial list is given of all the
foods eaten. Additional passes are then carried out that involve probing for additions
to the foods stated. These methods provide opportunities for reviewing the initial list
and recalling further detail.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 8

1.4.4. The use of new technologies

Recent studies have incorporated the use of digital cameras, mobile phones and the
internet to improve the ease and accuracy of data collection relating to diet. They can
be particularly useful when working with children. One example is the Synchronised
Nutrition and Activity Program (SNAP). It uses web-software with a child-friendly
interface to carry out 24-hour recall questionnaires.21

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 9

2. National surveillance of dietary intake and nutritional status

2.1. Health Survey for England

The Health Survey for England (HSE) is an annual survey designed to measure health
and health-related behaviours in adults and children living in private households in
England. It has been undertaken since 1991. The survey consists of an interview for all
participants, and a nurse visit for a sub-sample of participants. In recent years, sample
sizes have been around 16,000. This is not large enough to enable robust analyses of
data below Strategic Health Authority level.

The HSE is modular but has a number of core elements which are included each year.
Since 2001 these have included a number of 24-hour dietary recall questions relating to
fruit and vegetable consumption. A FFQ covering eating habits has been included in
most surveys of adults since 1993 and in surveys of children aged 2 to 15 since 2005 by
interview. The FFQ has been amended slightly over the years to reflect changing diets.
The HSE 2007 FFQ covers consumption frequency of bread, butter and spreads, fried
foods, milk, sugar in tea and coffee, cereal, pasta, potatoes, cheese, meat, fish, snacks
(such as chocolates, crisps, nuts or biscuits), puddings (cakes, pies, puddings or pastries),
sweets and chocolate, fizzy drinks and salt.

Typically the HSE includes around 2,000 children in the core sample (depending on the
adult sample size). However, in 2006 and 2007 the number of children was boosted to
7,257 and 7,504 respectively.22

The HSE is currently undergoing review. It is important that the new Health and Social
Care Survey includes components relating to dietary intake in order to retain this as a
source of public health nutrition data. It is also important to ensure that the questions
that measure dietary intake are valid and reliable; there is no evidence that the current
recall and FFQ questions have been formally validated for dietary monitoring in
populations of adults and children.

2.2. National Diet and Nutrition Survey

The NDNS was established in 1992 by the then Ministry of Agriculture, Fisheries and
Food (MAFF) and the Department of Health (DH). It is currently funded and managed
by the Food Standards Agency (FSA), with a contribution from the DH. The NDNS was
originally set up as a series of cross-sectional surveys of diet and nutritional status of
the population covering four age groups: pre-school children in 1992 to 1993; older
adults in 1994 to 1995; school-age children in 1997; and adults in 2000 to 2001. One
survey was carried out every two to three years. Data on consumption were gathered
using a weighed intake dietary record for four to seven days.

In April 2008, the NDNS changed to a rolling programme with data collected annually
from approximately 500 adults and 500 children (older than 18 months old). The survey
sample is designed to be representative of the UK population. Sample boosts have been
carried out in Scotland, Wales and Northern Ireland. There is scope for boosts in other
population groups, or add-on studies.

The NDNS has a number of components that provide an estimate of food consumption,
nutrient intake and derived nutritional status. Fasting blood samples are taken and

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 10

analysed for a range of indicators. Height and weight measurements, and waist and hip
circumferences are also taken from all adults, and these contribute to the assessment of
nutritional status. Urine samples are collected to estimate sodium and, thus, salt intake.
Data on an individuals usual dietary behaviour are collected through face-to-face
interviews. Collected data includes whether the individual:

z has access to amenities for storage and preparation of food

z is vegetarian
z takes dietary supplements
z is currently trying to lose weight
z uses artificial sweeteners .

Dietary intake is recorded using a diary of estimated (un-weighed) food over four
consecutive days.

The new rolling NDNS allows for more timely results and better tracking of trends over
time. More regular surveillance gives increased flexibility to enable modules to be
added at short notice and to be more responsive to policy needs.23 Headline results will
be published annually. The results of the first of the new surveys are to be published in
early 2010. Subsequent reporting will be in April each year. A more detailed report
covering four years of data collection from 2008 to 2012 will be published in spring
2013.

Prior to 2008, the NDNS reported on the types and quantities of food consumed. This
included:

z nutrient intake compared with Dietary Reference Values (DRVs) 7

z the main food sources of key nutrients
z intakes of salt
z intakes of alcohol
z overall nutritional status by age and sex.

The NDNS routinely reports on energy and macronutrients (protein, total carbohydrate,
non-milk extrinsic sugar, total fat and saturated fat) and vitamins and minerals (vitamin
A, thiamin, riboflavin, niacin, vitamin B6, vitamin B 12, folate, vitamin C, vitamin D,
iron, calcium, magnesium, potassium, zinc, iodine and copper). The data are held in the
public archive.

As stated above, the new methodology for the NDNS includes a sample of children.
Information is collected from them using a questionnaire on eating habits and a food
diary. For children aged 18 months - 10 years, parents are asked to answer questions
and complete a food diary on behalf of their child, whilst the child is present. Children
aged 11-15 years are asked to answer the questions for themselves. Children aged eight
years and above may also be given a written questionnaire to fill in themselves.
Anthropometric and other relevant measurements (such as blood pressure, biomarker
measurements from urine and blood samples) are taken to give a picture of nutritional
status. Nutrient intake and nutritional status are recorded for both children and adults.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 11

2.3. Low Income Diet and Nutrition Survey

The (LIDNS) was commissioned by the FSA as a one-off addition to the NDNS. It
provides nationally representative evidence on food and nutrient intakes, sources of
nutrients and nutritional status of people on low income. In addition, it examined the
relationship between dietary intake and factors associated with food choice in low-
income populations. The survey was carried out by a consortium of three organisations
led by the Health Research Group at the National Centre for Social Research (NatCen).6

A total of 3,728 people from 2,477 households from the most deprived populations in
the UK were included in the survey. It took place between November 2003 and January
2005. The survey included adults (aged 19 and above) and children (aged 2- 18). Data
collection methods included four 24-hour dietary recalls. These were undertaken on
random days (including at least one weekend day) within a ten-day period. Physical
measurements of height, weight and blood pressure were also collected. Blood samples
were taken from those aged eight and over to give an indication of nutritional status.
Additional information on environmental, economic and social factors was collected in
a face-to-face interview and via a self-completed questionnaire.

The survey report gives detailed information regarding the types of foods consumed by
the study sample gained from 24-hour dietary recall. These data were then linked to a
nutrient databank to calculate intake of macronutrients, micronutrients and alcohol.6

2.4. Infant Feeding Survey

The Infant Feeding Survey has been conducted every five years since 1975. The most
recent survey was conducted in 2005. The main aim is to provide national figures on the
incidence, prevalence, and duration of breastfeeding and other feeding practices
adopted by mothers in the first eight to ten months of the childs life. It also allows for
wider examination of trends in infant feeding and weaning practices over recent years.
These include investigation of variations among different socio-demographic groups,
and estimates for the proportion of mothers who smoke and drink alcohol during
pregnancy.

The 2005 survey provides separate estimates for all four countries in the UK as well as
for the UK as a whole. It is based on an initial representative sample of mothers who
were selected from all births registered in the UK during August and September 2005.
The mothers were asked to complete questionnaires at three stages. Stage 1 was when
babies were around four to ten weeks old. Stage 2 was around four to six months, and
Stage 3 was around eight to ten months. A total of 9,416 mothers, a response rate of
47%, completed and returned all three questionnaires.24

2.5. Expenditure and Food Survey and Family Food Reports

The Expenditure and Food Survey (EFS) has been run on an ongoing basis since April
2001 by the Social Surveys Division of the Office of National Statistics and The
Department for Environment, Food and Rural Affairs (DEFRA). The data are based on
food purchased. This provides proxies for consumption of foods, energy and nutrient
intake. Data are collected from a sample of private UK households using self-report
diaries of all purchases of food and drink (including food eaten out) over a two-week

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 12

period.25 These purchases are then assigned to the individuals in the household
according to a standard formula.

Family Food26 annual reports based on the analysis of the survey are published by
DEFRA. The most recent Family Food report derived from the EFS, is for 2007 and
includes data from 14,647 individuals in 6,141 households.26

The EFS analysis for the Family Food reports includes estimates for energy and nutrient
intake broken down by macro and micronutrients as an average intake per person and
as a proportion of the RNI. The report also provides comparisons between: countries in
the UK; regions in England; age and sex by consumption of food type; energy; and
nutrient intake.

Energy and nutrient intakes are calculated using standard nutrient composition data
which are supplemented by values from manufacturers and retailers. Nutrient
composition is updated on a rolling basis to keep up-to-date with new products and
the reformulation of processed goods.

From January 2008, the EFS was joined with other major national surveys on topics such
as smoking, housing and employment to form the Integrated Household Survey (IHS),
with the intention of increasing the range of the statistical outputs. The first reporting
from this new combined survey is due in 2010.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 13

3. Surveillance of fruit and vegetable intake: national surveys
providing local-level data

Local authorities and primary care trusts may carry out routine postal or telephone
surveys in their local populations for a number of purposes. These include managing
performance; assessing need; consulting on services or policies; and targeting and
evaluating interventions or campaigns. Due to the practical constraints of budgets,
resources, time and expertise, it is common for one survey to cover several topics, and
for the questionnaire to be kept simple for ease of completion and analysis.

As already described, dietary surveillance is complex, and it is hard to measure diet and
nutrition accurately. Local surveys tend to ask simple questions to measure specific
aspects of dietary intake such as numbers of portions of fruit and vegetables consumed.

Whilst the measuring of fruit and vegetable intake is useful in the context of national
campaigns and policy, the relationship between such intake and overall quality of diet
is unclear.

It is often unclear if the questions have been tested for validity and reliability and,
therefore, how accurately they measure intake of fruit and vegetables. For example,
the contribution of vegetables in a composite meal to the overall intake of fruit and
vegetables in a day can easily be missed.

The surveys described below are local surveys that are nationally prescribed in terms of
how they are carried out and what they measure.

3.1. Primary Care Trust Patient Survey

The National Patient Survey programme is administered by the Healthcare Commission

and has been operating since 2002. Participation in the survey is mandatory for PCTs.
The most recent survey was carried out in Spring 2008, with each PCT sending
questionnaires to 1200 patients, following a strict protocol. An average response rate
of 40% was achieved.

While the main focus of the survey is on patients experiences of their primary care
services, the 2008 survey also asked a number of health and lifestyle-related questions.
One of these was: How regularly do you eat the recommended five portions of fruit or
vegetables a day?. Respondents could answer every day, most days, 1 - 3 times a
week, or less often than 1 day per week.27

This provides a local-level indication of fruit and vegetable intake in adults that can be
tracked year on year. However, the sample is only taken from individuals who are
registered with a GP practice so the study is liable to sampling bias. For example, there
are sub-groups of the population, such as some black and minority ethnic groups, and
young adults who may be studying or moving house frequently that are less likely than
the general population to be registered with a GP practice. There is no evidence that
the survey question has been tested for validity or reliability, and it may be influenced
by social desirability bias, i.e. the respondent may provide what they perceive as the
desired answer.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 14

3.2. TellUs

TellUs is a series of annual online surveys to gather quantitative information on the

views and experiences of children and young people. It is commissioned by the Office
for Standards in Education, Childrens Services and Skills (Ofsted) and the Department
of Children, Schools and Families, and carried out by Ipsos MORI. The surveys have been
running since 2006. They provide data for the National Indicators for local authorities
and local authority partnerships. They are intended to help local authorities to judge
the impact of their services on perceived quality of life for children and young people.

TellUs asks children and young people questions relating to the five outcomes in Every
Child Matters28 and these include healthy eating. Children are asked to say how many
portions of fruit and vegetables they normally eat in a day. The response options are
none, 1-2, 3-4, and 5 or more. There is also a separate attitude question on
childrens perceptions of the information and advice they get on eating healthy food.

TellUs is completed by a sample of children in all local authority areas across England.
Pupils complete questionnaires online via a dedicated website. The sample size is
calculated with a view to obtaining enough responses to allow robust analysis at local
area level. Sampling methods take account of different types and sizes of schools,
together with socio-economic factors. The sample includes maintained schools, pupil
referral units, academies, and city technology colleges. Data are collected at school
level, and aggregated up to local authority level.

The questions were developed through cognitive testing with children to ensure that
they understood the questions and answered them in a meaningful way. However,
published reports on the survey do not provide any detail as to whether the survey
questions have been formally tested for reliability or validity.29 The most recent survey
report is from TellUs 3 which took place in spring 2008.30

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 15

4. Other relevant data sources

Large amounts of data have been collected to evaluate national interventions relating
to fruit and vegetable intake. Although not strictly surveillance data, they can be very
useful. Two such sources of data with large sample sizes are briefly described below.

4.1. Evaluation of the National School Fruit and Vegetable Scheme

The National School Fruit and Vegetable Scheme involves 3,703 children aged four to
six years. They took part in a non-randomised controlled trial in March, June and
November 2004.31 The Child and Diet Evaluation Tool (CADET)32 was used to estimate
the usual dietary intake (including portions of fruit and vegetables), energy, and
nutrients. Comparison studies have been carried out against a one-day semi-weighed
food diary. These have shown CADET to typically overestimate fruit intake, and
underestimate vegetable intake, though it was able to rank individuals into levels of
intake. Repeatability tests were carried out against a food diary for two days. The
second round of CADET was seen to correlate better to the food diary than the first
round.

4.2. Evaluation of the 5-A-DAY Programme

The 5-A-DAY local community initiative encourages people in 66 programme areas to

eat at least five portions of fruit or vegetables a day. In addition, it aims to improve
awareness of, knowledge about, and attitudes to healthy eating.

The evaluation aimed to measure change in intake of fruit and vegetables and changes
to attitudes, knowledge and awareness. This was in representative samples of the adult
household population targeted for intervention in each of the participating 66
programme areas, both pre-intervention in 2003 and post-intervention in 2005. Intake
was also measured pre- and post-intervention in control areas.

Data were collected using a self-completed postal questionnaire titled the Five-A-Day
Consumption Evaluation Tool (FACET).33 The questionnaire was completed by 13,151
adults from the intervention areas, and by 2,640 from the control areas.34 FACET was
found to overestimate fruit and vegetable consumption by 1.5 portions when
compared with a food diary. Consequently, the results of the evaluation were
expressed both as adjusted and unadjusted figures.

4.3. Attitudes and knowledge data

Eating behaviours are linked to a number of factors such as knowledge about, and
attitudes towards healthy eating. In 2007, the HSE included questions relating to these
for a sample of children aged 11 to 15, and adults aged 16 and over.35

There are other surveys in the public domain, as well as a number of commercially
available datasets covering these areas; they are beyond the scope of this report.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 16

5. Discussion: possible gaps in the data

5.1. Surveillance of energy and nutrient intake data at local level and in sub-
groups

There is a shortage of data available for surveillance of nutrition and diet at local level
in adults and children. The Primary Care Trust Patient Surveys provide data on adult
consumption of fruit and vegetables at PCT level, and the TellUs survey provides similar
data for children. However, there is no information in the reports from either of these
surveys relating to the validity or reliability of the questions asked.

As stated earlier, fruit and vegetable intake is only one element of dietary surveillance,
and it does not give an accurate picture of overall energy intake or nutritional status.

There is a shortage of validated nutritional assessment tools that are practical and
suitable for use at local population level. Locally developed lifestyle surveys are prone
to using fruit and vegetable intake as a proxy for a healthy diet.

The NDNS has not, to date, had a large enough sample to enable analysis at local level
or within some sub-groups (excluding the LIDNS which targets a specific sub-group).
The rolling programme will allow for boosts in particular areas, or within sub-groups.
However, it is unlikely that the core sample will allow for local level or sub-group
analyses until several years of pooled data are available.

Similarly, the HSE allows analysis of adult data at government office region and
Strategic Health Authority level, but not at a more local level. Analysis by some sub-
groups has been made possible by booster samples in some years. In addition, a
collaboration between the Association of Public Health Observatories (APHO) and
NatCen has provided synthetically modelled estimates at super output area level, using
pooled HSE data from between 2003 to 2005.36 These are for adult obesity and
consumption of fruit and vegetables. They are not suitable for target setting or
performance monitoring but can be useful for planning services locally or
benchmarking across PCTs.

5.2. Availability of trend data

The NDNS (pre-2008) has provided a picture of population diet and nutritional status
compared with DRVs. However, it has been difficult to track changes and trends over
time due to large gaps between data points. The new methods for the NDNS will
enable better tracking of trends and more frequent reporting. One drawback is the
small sample sizes which mean that tracking trends at lower geographical levels or in
demographic sub-groups will be possible only when pooled year-on-year data become
available.

The HSE has collected consistent dietary recall questions relating to fruit and vegetable
intake from a sample of adults and children each year since 2001. This allows for the
tracking of trends nationally. The HSE FFQ has been used with a sample of adults in
1994, 1997, 2003, 2006 and 2007 (and among ethnic minority groups in 2004), and in
children in 1997, 1998, and since 2005.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 17

It would be helpful if all questions relating to diet could be included in the core
questions each year, and for the same questions to be asked in the NDNS. This would
provide a valuable check of validity.

The EFS provides useful population level trends in average purchases, expenditure and
derived nutrient intakes. However, the data collection is conducted at household level,
and figures are estimated for at individuals. As a result, individual consumption or sex
differences cannot be identified.

5.3. Validity of data collection methods

The HSE provides consistent data on dietary intake in adults and children, particularly
for fruit and vegetable consumption. However the FFQ has not been tested for validity
or reliability as a population level surveillance tool. The FFQ is based on a primary care
dietary brief assessment tool that was developed and validated in 1994.37 Whilst the
tool itself has been updated and amended for use in the survey, it has not undergone
more recent, appropriate testing.

Similarly, published reports on the Primary Care Trust Patient Surveys and the TellUs
surveys provide little detail on the validity of the methods used. The Family Food
reports of the EFS do not give any detail about the extent to which reported average
figures for foods purchased have been validated against more robust methods for
measuring food consumed.

All research tools should be validated against more robust measures. The nature of the
measure will depend on the aim of the tool. For example, a tool estimating dietary
intake could be validated in a smaller sample or pilot against a diary that records
weighed food. A tool aiming to estimate overall energy intake could be validated
against total energy expenditure from doubly labelled water.

There are a limited number of biomarkers accepted as measurements or proxy

indicators of nutritional status. This means that robust validation of some tools is a
challenge. In addition, the practicalities of carrying out more clinical exercises for the
purposes of validation are expensive, resource intensive and potentially invasive.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 18

6. Conclusions

Following the recent revision of the NDNS, data from a sample representative of the UK
will be available annually. These data will relate to: consumption of types and
quantities of foods; nutrient intake compared with DRVs; and nutritional status. The
relatively small sample size means that analysis of the data at lower geographical levels,
or for tracking trends in sub-groups of the population, will produce results with large
confidence intervals. This will remain the case until several years of data can be pooled,
or when boosted samples are available.

The HSE is another way in which dietary intake data are collected. It provides
consistent, annual data for fruit and vegetable consumption in adults and children
dating back to 2001. In some years, an FFQ was used to provide estimates of the
consumption of other food groups.

Whilst the HSE sample is larger than the NDNS, it still does not provide robust figures at
local level (sub-regional) or within some sub-groups without boosted samples. The FFQ
used has also not been appropriately validated and may be liable to systematic and
random error. In addition, it does not allow for nutrient intake estimations to be
derived.

It is important that the tools used in the HSE are valid and reliable. A validation study
should be undertaken comparing the methods for data collection in the HSE with more
robust, objective methods such as those used in the NDNS. Improved sub-group and
local level analysis would require more regular boosts to the HSE to increase the sample
size.

The EFS has provided comprehensive data for tracking and monitoring trends in food
purchasing, and derived dietary and nutrient intake. A similar method may be used in
the IHS. If this is the case, a study to investigate the correlation between purchased and
consumed food should be completed to support more robust conclusions about food
intake.

The following points suggest how dietary surveillance and national assessment can be
improved:

z Increased sample sizes to enable improved tracking of trends within sub-

groups, and robust analyses at lower geographical levels in relation to
energy, dietary and nutrient intake.
z Testing of widely used self-reporting tools such as the FFQ in the HSE to
ensure their validity and reliability.
z Continuing work to develop more robust validated population level dietary
assessment tools. This includes investigating the use of new technologies
such as text-messaging and web-based data collection.
z Reaching consensus on the appropriate tools to use for different purposes.
Local and national level surveillance systems could then be harmonised to
improve comparability of datasets and evaluation of interventions.
z Identifying and developing biomarkers of intake (as opposed to nutritional
status).

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 19

 z Identifying which elements of dietary intake provide the most robust proxy
indicators of a healthy diet. Simple, self-reported data could then be
collected in local surveys (such as the TellUs survey) and could be used more
confidently as a proxy for overall diet patterns in local populations. This
would enable local areas to gain valuable baseline data to include in their
local area agreements.

The emergence of the new IHS, and the adaption of the NDNS to a more flexible and
regular rolling programme, are positive moves towards improved availability of dietary
data. The plans to expand the HSE into a broader health and social survey from 2011
may provide a further opportunity for cross-government working that will improve the
existing mechanisms for the collection of diet and nutrition data.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 20

References

1. Department of Health. Healthy weight, healthy lives: a cross government strategy

for England [Online] London: COI; 20008. Available from:
http://www.dh.gov.uk/prod_consum_dh/groups/dh_digitalassets/documents/digital
asset/dh_084024.pdf [Accessed 16th February 2010].

2. Rutishauser, IHE and Black, AE. Measuring food Intake. In: Gibney, M.J, Vorster,
HH and Kok, F.J. Introduction to human nutrition. Oxford: Blackwell Science Ltd;
2002.

3. Pattersen, RE and Pietinen, P. Assessment of nutritional status in individuals and

populations. In: Gibney, M.J, Margetts, B.M, Keanrney, J.M and Arab, L. Public
health nutrition. Oxford: Blackwell Scienc Ltd; 2004. pp. 66-82.

4. Barasi, M. Human Nutrition a health perspective. 2nd ed. London: Hodder

Arnold; 2003.

5. Nelson, M, Beresford, AA and Kearney, JM. Nutritional epidemiology. In: Gibney,

M.J, Margetts, BM, Keanrney, JM and Arab, L. Public health nutrition. Oxford:
Blackwell Science Ltd; 2004. pp. 26-55.

6. Nelson, M, Erens, B, Bates, B., Church, S, and Boshier, T. Low income diet and
nutrition survey summary of key findings. Norwich: TSO; 2007.

7. Department of Health. Dietary reference values for food energy and nutrients for
the United Kingdom. London: The Stationery Office; 1991.

8. British Nutrition Foundation. Nutrient requirements. [Online]. Available from:

http://www.nutrition.org.uk/nutritionscience/nutrients/nutrient-requirements
[Accessed 16th February 2010].

9. Younger, KM. Dietary reference standards. In: Gibney, MJ, Vorster, HH and Kok, F.
Introduction to human nutrition. Oxford: Blackwell Science Ltd; 2002. pp. 116-124.

10. Gibson, R. Principles of nutritional assessment. New York: Oxford University Press;
2005.

11. Livingstone, MBE, Prentice, AM, Strain, JJ, Coward, WA., Black, AE, Barker, ME,
McKenna, PG and Whitehead, RG. Accuracy of weight dietary records in studies of
diet and health. British Medical Journal 1990; 300: 708- 712.

12. Herbert JR, Clemow I, Pbert L et al. Social desirability bias in dietary self report
may compromise the validity of dietary intake measures. International Journal
Epidemiology. 1995; 24: 389-398.

13. West, CE and Schonfeldt, HC. Food composition. In: Gibney, MJ, Vorster, HH and
Kok, F.J. Introduction to human nutrition. Oxford: Blackwell Science Ltd; 2002. pp.
249-262.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 21

14. Roberts K, Cavill N, Rutter H. Standard evaluation framework for weight
management interventions. [Online]. Oxford: National Obesity Observatory; 2009.
Available from:
http://www.noo.org.uk/uploads/doc/vid_3534_NOOSEFreportJuly09.pdf
[Accessed 16th February 2010].

15. Thompson, FE and Subur, AF. Dietary assessment methodology. In: Coulston, AM
and Bouchey, CJ. (eds). Nutrition in the prevention of treatment and disease.
London: Elsevier; 2008. pp. 3-22.

16. Cade JE, Burley VJ, Warm DL, Thomson RL, Margetts BM. Food frequency
questionnaires: a Review of their Design, Validation and Utilisation. Nutrition
Research Reviews 2004; 17: 5-22.

17. Cade J, Thompson R, Burley V, Warm D. Development, validation and utilisations

of food frequency questionnaires a review. Public Health Nutrition 2001; 5(4):
567-587.

18. Gerovitz, M, Madden, J, Smiciklas-Wright, H. Validities of the 24-hour dietary

recall and seven-day record for group comparisons. Journal of the American
Dietetics Association 1978; 73:48-55.

19. Johnson, K. Dietary intake How do we measure what people are really eating?
Obesity Research. 2002; 10 (1): 63-68.

20. US Department of Health and Human Services. NHANES measuring guides for the
dietary recall interview. Hyattsville: Centers for disease control and prevention;
2009. Available from:
http://www.cdc.gov/nchs/nhanes/measuring_guides_dri/measuringguides.htm
[Accessed 16th February 2010].

21. Moore, HJ, Ells, LJ, McLure, SA., Crooks, S, Cumbor, S, Summerbell, CD and
Batterham AM. The development and evaluation of a novel computer program to
assess previous-day dietary and physical activity behaviours in school children: The
Synchronised Nutrition and Activity Program (SNAP) British Journal of Nutrition.
2008: 99, 6; 1266-74.

22. The Information Centre. Health Survey for England 2007 Volume 2. [Online].
Leeds: The Information Centre; 2008. Available from:
http://www.ic.nhs.uk/webfiles/publications/HSE07/HSE%2007-Volume%202.pdf
[Accessed 16th January 2010].

23. Swan, G, Bush, M, Farron-Wilson, M, Strather, M, Tedstone, A, Bates, B, Teucher

GB Stephen A. The UK National Diet and Nutrition Survey (NDNS) Challenges in
moving to a rolling programme. Food Standards Agency, National Centre for
Social Research, MRC Human Nutrition Research. July 2009.

24. Boiling, K, Grant, C, Hamlyn, B, Thorton, A. Infant Feeding Survey 2007. Leeds:
The Information Centre; 2007.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 22

25. The Information Centre. Statistics on obesity, physical activity and diet: England,
February 2009. [Online]. Leeds: The Information Centre; 2009. Available from:
http://www.ic.nhs.uk/webfiles/publications/opan09/OPAD%20Feb%202009%20fin
al.pdf [Accessed 16th February 2010].

26. National Statistics. Family food 2007. London: TSO; 2008. [Online]. Available from:
https://statistics.defra.gov.uk/esg/publications/efs/default.asp. [Accessed 16th
February 2010].

27. Martin, J. The key findings report of the 2008 national survey of the local health
services. Oxford: Picker Institute Europe; 2008. [Online]. Available from:
www.nhssurveys.org. [Accessed 16th February 2010].

28. Department for Children, Schools and Families. Every Child Matters. Norwich: TSO;
2003. Available from:
http://www.dcsf.gov.uk/everychildmatters/about/aims/outcomes/outcomescyp/
[Accessed 16th February 2010].

29. Roberts, K, Flaherty, SJ, Jenner, D, and Ells, L. A review and needs assessment of
local Level Diet and Nutrition Data. Mansfield: EMPHO; 2008. [Online] Available
from: http://www.empho.org.uk/pages/viewResource.aspx?id=11024

30 . Ofsted. Tellus3 National Report. [Online]. Available from:

http://www.ofsted.gov.uk/Ofsted-home/Publications-and-research/Browse-all-
by/Documents-by-type/Statistics/TellUs3-National-Report [Accessed 16th February
2010].

31. Ransley, J, Greenwood, DC, Cade, JE, Blenkinsop, S, Schagen, I, Teeman, D, Scott E,
White, G, Schagen, S. Does the School Fruit and Vegetable Scheme improve
childrens diet? A non randomised controlled trial. Journal of Epidemiology &
Community Health 2007; 61: 699-703.

32. Cade, JE, Frear, L and Greenwood. DC. Assessment of diet in young children with
an emphasis on fruit and vegetable intake: using CADET Child and Diet
Evaluation Tool. Public Health Nutrition 2006; 9:5018.

33. Ashfield-Watt, PAL, Welch, AA, Godward, S and Bingham, SA. Effect of a pilot
community intervention on fruit and vegetable intakes: use of FACET (Five-a-day
Community Evaluation Tool). Public Health Nutrition 2007; 10(7); 671-680.

34. Bremner, P, Dalziel, D, Evans, L. Five A Day Programme Final Report. London: Big
Lottery Fund; 2006. Available from:
http://www.biglotteryfund.org.uk/er_eval_5aday_report_evaluation.pdf [Accessed
16th February 2010].

35. The Information Centre and NatCen. Healthy Lifestyles: knowledge, attitudes and
behaviour: summary of key findings. [Online]. London: NatCen; 2007. Available
from:
http://www.ic.nhs.uk/webfiles/publications/HSE07/HSE%2007-Volume%201.pdf
[Accessed 16th February 2010].

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 23

36. The Information Centre. Neighbourhood statistics: LA model based estimates of
healthy lifestyle behaviours 2003-2005. [Online]. London: NatCen; 2007. Available
from: http://www.ic.nhs.uk/webfiles/Popgeog/Stage%201%20report.pdf [Accessed
16th February 2010].

37. Roe, L., Strong, C, Whiteside, C., Neil, A, and Mant, D. Dietary intervention in
primary care: validity of the DINE method for diet assessment. Family Practice
and International Journal 1994; 11(4): 375-381.

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 24

Reader Information

Title Dietary Surveillance and Nutritional Assessment in

England: what is measured and where are the gaps?
Author(s) Kath Roberts

Reviewer(s) Dr Susan Jebb, Human Nutrition Unit, Medical Research

Council
Dr Janet Cade, University of Leeds
Alison Lennox, Human Nutrition Unit, Medical Research
Council
Sheela Reddy, Department of Health
Dave Jenner, East Midlands Public Health Observatory
Sarah Flaherty, East Midlands Public Health Observatory
Publication date March 2010

Target audience Obesity/ physical activity/ nutrition professionals in: central

government; regional government; local authorities;
Primary Care Trusts; Strategic Health Authorities.
Public Health Observatories.
Academics and researchers working in the field of
nutrition and obesity.
Description This paper identifies, describes and discusses the main
sources of data for national-level dietary intake and
nutritional status in adults and children in England. It is
intended to support the development and monitoring of
public health nutrition policies and services.
How to cite Roberts K. Dietary Surveillance and Nutritional Assessment
in England: what is measured and where are the gaps?
Oxford: National Obesity Observatory, 2010

Contact National Obesity Observatory

www.noo.org.uk
[email protected]

Electronic location http://www.noo.org.uk/NOO_pub/briefing_papers

Copyright National Obesity Observatory

National Obesity Observatory

References

NOO | DIETARY SURVEILLANCE AND NUTRITIONAL ASSESSMENT IN ENGLAND 25