SMJ 2003: 48(2) 32-37
J Armstrong
Lecturer
E-mail: J.Armstrong@gcal.ac.uk
Senior Lecturer
E-mail: jjr2@clinmed.gla.ac.uk
Child Health Information Team
Hospital & Community Information Unit
E-mail: saskia.gavin@isd.csa.scot.nhs.uk
Abstract
Objective: To
assess whether anthropometric data, routinely collected as part of the Scottish
Child Health Surveillance System (CHSP- PS,
pre-school children; CHSP-S, school age children) could provide a means
of monitoring/surveillance for obesity and undernutrition at national and health
board level.
Design: A survey of 15 health boards and both
surveillance systems to identify the nature of data collected, format of data,
and extent to which data were accessible (e.g. via Information and Statistics
Division of the Common Services Agency). Measurements of weight and height
collected as part of the CHSP-PS and CHSP-S were extracted from ISD. They were
then audited and missing values or implausible values quantified, and degree of
dispersion of values used as an index of quality of measurements.
Setting: Health Board Child Health Surveillance
Systems and Information and Statistics Division, Edinburgh.
Results: Data
on height and weight are currently available for 9 health boards for pre-school
children and 4 health boards for school age children. This represents coverage of around 80 % of the pre-school
child population. Analysis of a
data extract from the 39-42 month check in 1998/99, used as an example, revealed
that 8% of weight and height data were missing, and approx. 1 % were implausible
measures. Population and health board level estimates of prevalence of obesity
and undernutrition were possible and are presented.
Data on height and weight are routinely collected in school age children
in all health boards, however only 4 health boards have growth data
electronically available via the school CHSP.
Conclusions: Growth
data routinely collected as part of child health surveillance for Scotland can
be used to estimate population prevalence of undernutrition and obesity.
These can in turn be used to monitor trends at local and national level,
to monitor achievement in relation to public health targets, identify risk
factors and high risk groups, and to follow cohorts over time.
We describe a system of surveillance for undernutrition and obesity and
identify its strengths and weaknesses.
Key words: obesity;
children; surveillance;
cadiovascular disease.
Introduction
The UK experienced an epidemic of childhood obesity
during the 1990s
5;6
, and even pre-school children were affected
7;8
. Obesity is now arguably the most common medical
problem of childhood. Paediatric obesity is associated with a number of
co-morbidities in childhood
9
, and with increased risk of adult disease,
particularly cardiovascular disease
10-12
. In
addition, there is evidence that obesity in adolescence and young adulthood is
associated with low self-esteem
13
, lifetime social and economic disadvantage
14;15
and
there may be social inequality in childhood obesity risk
6
.
Growth monitoring is used to identify and follow up
children with undernutrition. There
are short term and potentially long term health risks
16
associated with undernutrition and lack of catch up
growth
17;18
. The
short and long term developmental effects of severe undernutrition and specific
nutrient deficiencies in developing countries are well documented
19;20
. However,
the lasting effects of undernutrition in developed countries are less clear and
in any setting are likely to be modified by the severity and duration of the
insult, its timing, and other environmental factors
20
. While social class gradients in intra-uterine
growth and nutrition are well known, whether social deprivation has a more
persistent effect on nutritional status during childhood is less clear.
Childhood obesity and under-nutrition are therefore
relevant to child health, adult cardiovascular disease, and social inequality in
health, all-important concerns of the Scottish Health Service
21;22
. The
impact of childhood obesity on the nations health is highlighted in recent major
reviews. Despite increasing
national and local concern over increases in childhood obesity no data on the
population prevalence of overweight, obesity, or undernutrition were available
for the population of Scottish children until recently. Cross-sectional surveys,
such as the Scottish Health Survey have recently provided estimates of
prevalence for small cross sectional samples, providing limited analyses by age
and sex. However there is currently
no system for population surveillance of childhood obesity: no analyses of
prevalence by health board, no information on cohorts followed longitudinally,
and limited opportunity to assess prevalence in different groups, or to identify
risk factors for obesity in Scottish children.
The availability of population based data provides an important
opportunity to address these points and have been used increasingly in child
obesity surveillance in areas of England
6;8
.
In Scotland, information on childrens health is
gathered through a number of systems at a local level for pre-school and school
age children. This forms the Scottish component of the National (UK) Child
Health Surveillance Programme (CHSP), reviewed regularly by the Royal College of
Paediatrics and Scottish Child Health working group
23
. In Scotland there are two branches of this
programme: pre-school (CHSP-PS) and school (CHSP-S). The aims and objectives for the CHSP-PS are set by the Child
Health Surveillance Project Board and National User Groups and these include
standardised techniques for taking measurements
24
. The Health Services in Schools report lays out
recommendations for the CHSP school health service, including height and weight
measurements
25
. Today health information on a child gathered in
different systems e.g. CHSP, SIRS (immunisation record) can be linked through
the CHI system providing a fuller health profile for each child.
In Scotland, population-based data which can provide
estimates of childhood undernutrition and obesity, are collected at routine
assessments made by health visitors and school nurses for the purposes of growth
monitoring the entire child population. These
growth data are then made available at health board level, but the
accessibility, quality, nature and use of the data for epidemiological purposes
were unknown prior to the present study and were known to differ markedly
between health boards, partly depending on their participation in the national
CHSP. The aims of the present study were therefore to:
(a) Describe availability of existing data at health board and national
level (b) Provide an example
of how these data may be used to monitor childhood undernutrition and obesity
prevalence in the child population (c)
Identify some weaknesses in the system and make recommendations for a
surveillance system for under and overnutrition in Scottish children. d) Provide
baseline epidemiology on obesity and under-nutrition in Scottish children.
Health
Boards : growth monitoring
During the period November 2000 - February 2001 we surveyed the 15 Scottish
Health Boards to establish which height and weight measures are taken as part of
the routine health reviews for children. Contact
was made by telephone or by letter to the child pre-school health services and
school health service through Community Child Health in each health board.
We asked relevant personnel, usually lead health visitor (for pre-school)
or lead school nurse (for school) or community paediatrician at what ages height
and weight measurements were taken from children as a part of their routine
health review. In addition to
assess the accessibility of data collected we asked when the measurements were
taken, where the data was recorded, collated and if it was stored in paper or
electronic format (in some boards this questions was also directed to the health
information section).
Child
indices of obesity and undernutrition: Body Mass Index
The best simple index of population prevalence of
undernutrition, overweight, and obesity in children is provided by the (BMI)
body mass index weight (kg)/height2(m2)
26
. Since
BMI changes with age, in children this value must be interpreted by comparison
with UK 1990 population reference data
27
, using cut-off points in the distribution of BMI (centiles
or standard deviation scores). We have defined obesity as BMI >95th
centile, severe obesity as BMI >98th centile, low body mass index
as BMI < 5th centile and undernutrition as BMI
The definitions for obesity have high specificity (low false positive rate) and
moderately high sensitivity in identifying the fattest children
within the population, and so are informative for individual children as well as
populations. The definition of
obesity was that which is widely used and recommended
28-30
and is
not arbitrary. Children with BMI
>95th centile are those most likely to experience co-morbidity,
obesity defined in this way is likely to persist, and is associated with the
presence and clustering of cardiovascular risk factors
12;28
.
The definitions for undernutrition are based on
methods used to identify children who are failing to thrive
31
.
32
Children who have a BMI < 2nd Centile
are recommended for clinical referral
33
and the majority of active cases of failure to
thrive have a nutritional problem
34
and a low BMI, usually < 10th centile
32
. The most appropriate BMI cut off point for
population prevalence of undernutrition has not been universally agreed
and based on current evidence we have used a cut off of BMI < 2nd
centile to define undernutrition.
Sample
: Pre-school children aged 3-4 years
When we commenced the project in August 2000, 9/15
health boards (Table 1) were using the CHSP PS information system,
representing 80% of the Scottish child population. As an example or model of utilising current growth data we
extracted records of 74500 children who had received their 39-42 month health
check in 1998/99 for the following fields: age, gender, weight, height, post
code and health board area. We
linked to their earlier record (where available) for birth weight, type of
feeding (recorded at 6-8 weeks old), weight and height/length at age 8 months
and 21 months. In Scotland the size of the birth cohort in 1999 was 52,987
children, this giving an indication of the relative size of our sample to a
population birth cohort. Guidelines for taking routine weight and height
measurements are detailed elsewhere
24
.
Sample
: School children aged 5-14
When we commenced the school part of the project in
January 2001 3/15 Health Boards were participating in the national CHSP-S
information system: Lanarkshire, Borders and Lothian (West Lothian only).
As an example or model of utilising current growth data from the CHSP we
extracted the health records of 55647 School age children taken in Primary 1, 3,
7 and Secondary 3 spanning the years 1997-1999.
Other
Data
Since any surveillance system for under and
overnutrition should permit assessments of differences in prevalence (e.g.
between health boards; between deprivation category), or changes over time, for
our sample we determined availability of data at health board level, quantified
missing data on deprivation category, and availability of data over time.
Deprivation was defined using Carstairs Deprivation Categories (Depcat). There
are seven separate categories, ranging from very low deprivation (depcat 1) to
very high (depcat 7)
35
.
Analysis
of data sets
The BMI for each child was calculated and converted
to an SD score relative to the UK reference standard using a growth foundation
software package
33
. The
prevalence of obesity and undernutrition was compared to the UK reference using
a chi-squared goodness of fit test. In
the pre-school sample we were able to test the association between obesity and
undernutrition and potential explanatory variables (deprivation category,
breastfeeding, birth weight) using binary logistic regression analysis. In the
sample of school children the prevalence of obesity at different ages was tested
using a chi-squared test for trend. Statistical
analyses were carried out using SPSS for Windows version 10.1.
Health
Board : growth monitoring
In May 2001 there were 9/15 health boards and 4/15
health boards participating in the CHSP-PS (pre-school) and CHSP-S (school)
respectively (summarised in Table 1). As part of the CHSP height and weight
measurements taken at routine health reviews by health visitor (CSP-PS) and
school nurses (CHSP-S) are collated and stored in electronic format and
available for analysis at ISD. The number of health board in which growth data
is available electronically is increasing as more health board participate fully
in the CHSP. In Health Boards who
do not participate in the national CHSP, the information from routine child
health reviews is collected and stored either in paper format within the
childs medical (for pre-school) or in their school health record at a local
level. Table 1 indicates at
which reviews routine weight and height measures were taken in the health boards
at the time of the study.
The following are the stages recommended by CHSP for
taking routine height/length and weight as part of child health reviews.
6-8 weeks
Primary 1
8-9 months
*Primary 3 or 4
22-24 months
*Primary 7
39-42 months
*Secondary
*Pre-school 48 months
* variable at the local level and may be limited to follow up, self referral or
may include height only
The WHO
expert committee recommend using the observed standard deviation of height/weight or BMI SDS
distribution to assess the
spread of height and weight survey
Pre-school
Children aged 3-4 years
a)
Prevalence of obesity and undernutrition
As an example of what is possible when using CHSP-Pre-school
and CHSP-School for surveillance, prevalence of undernutrition (BMI <2nd
centile), obesity (BMI >95th centile) and severe obesity
(BMI>98th centile) are given for Scottish children aged 3-4 years
in 1998/99.
The prevalence of Scottish children aged 3-4 years
with obesity (Girls 8.1%, Boys 9.0%) and severe obesity (Girls 4.1% and Boys
4.4% ) (Table 3 )was
significantly higher than the 1990 UK reference standard of 5% and 2%
respectively (p <0.0001). The
prevalence of children with low BMI (Girls 6.3% and Boys 5.9%) and
under-nutrition (Girls 3.3% , boys 3.2%) (Table 3) was significantly
higher than the 1990 UK reference standard of 5% and 2% respectively (p
<0.0001).
b)
Pre-school children: Health Board Area
The variation in prevalence of children with
undernutrition ( BMI < 2nd Centile) and severe obesity (BMI >
98th centile) at age 3-4 by Health Board Area is shown in Figure
1. The varied prevalence between health board areas and is likely to reflect
in part differences in distribution of social deprivation between health boards.
c)
Pre-school children: other risk factors for under-nutrition and obesity
Using the same pre-school data as an example we went
on to explore other factors and their relationship with undernutrition and
obesity.
To assess the relationship between birthweight and
BMI at age 3-4 years we categorised birthweight as follows : <2.5Kg,
2.5-3.0Kg, 3.0-3.5kg, 3.5-4.0Kg and >4.0Kg. The
prevalence of undernutrition and obesity at age 3-4 years by birth weight
category is shown in figure 2.
Increasing birth-weight was significantly associated with the prevalence
of obesity and negatively associated with undernutrition after adjusting for
gender, deprivation and feeding at 6-8 weeks (p<0.001). A repeat of the
analysis after exclusion of children born <37 weeks gestation gave a similar
result. In addition low birth
weight babies were more likely to be <2nd centile for BMI at age
3-4 years.
School
children aged 5-14 years
The school sample of 55647 children there was BMI SDS
available for 47185 (84.8%) and data for the others was missing or implausible.
In the school year 1999/2000 for the Primary 1 children eligible for a
review 71%, 81% and 87% had received a review in Lanarkshire, Borders and West
Lothian respectively. Figure 3
shows the prevalence of obesity in children in primary 1, primary 3, primary 7
and secondary 3 in the three health board areas for the years 1997-1999.
The
prevalence of obesity (BMI >95th centile) at each stage was higher
than the UK reference of 5% (P<0.0001). There was an increasing trend in obesity prevalence with age
in girls (chi squared for linear trend = 29.7, P<0.0001) and boys
(chi squared for linear trend = 54.1, p<0.0001). Figure
4 shows the prevalence of obesity by deprivation category, with the highest
prevalence in the most deprived categories 6 and 7 (chi squared for linear trend
= 8.9, P<0.005)
Routine
Growth Monitoring
Analysis of routinely collected health data can be used to inform and direct
health policy and facilitate planning of health services. Analysis of population
data can provide community diagnoses by assessing the health of the
population and the need for preventive strategies.
Routinely collected child health data can also be used to monitor trends,
and, potentially, to follow cohorts of children longitudinally. Population data
can also be used in a more exploratory fashion, to identify or test risk factors
for a disease for example. The value of routinely collected growth data to
estimate obesity has been shown in a number of recent studies in England
6;8
. We
have used the data in this way in a separate communication, in order to test the
hypothesis that breastfeeding is protective against childhood obesity
37
and that deprivation is associated with both
undernutrition and obesity at age 3-4 years
38
. Surveillance
of childhood obesity and under-nutrition at local and national level within
Scotland contributes to achieving the applications outlined above, but this
depends on preserving and developing the systems for monitoring growth. The
primary aim of the present study was to explore the existing systems, CHSP-PS
and CHSP-S, in order to assess their potential as a surveillance system for
undernutrition and obesity.
The quality of the data is high in terms of coverage
and population distribution and confirms findings from preliminary English data
39
. Not all health boards participate in the system but
based on our analysis of pre-school data for 1998/99 the proportion of missing
height and weight parameters, implausible measures and the data distribution,
the data quality was suitable for the epidemiological purposes illustrated here.
A significant proportion of missing data was observed (8.4% of height and/or
weight) and a small number were implausible anthropometrics measurements (0.9%).
This must be regarded as inevitable when using routinely collected data, though
the % of missing data at 39 months in this study was smaller than in surveys
which used routinely collected child health data in England2,4.
In addition, further efforts on data available from the 39 month
check would have recovered some of the missing
data (from families where the check was carried out later at a re-scheduled
visit). There was a small but significant bias of more missing data from the
most deprived group. However a
detailed audit of the quality of measures and procedures was beyond the scope of
this project.
Pre-school
children: obesity and undernutrition
The CHSP PS covers around 80% of the child
population up to school age. It can be used to estimate the population
prevalence of obesity and undernutrition at local and national level relative to
national reference data. The prevalence of obesity, 8.6% in 3-4 year olds in
Scotland in 1998/99, was significantly higher than UK 1990 prevalence (5%), and
consistent with reports of an epidemic of childhood obesity. A significant
increase in childhood and adolescent obesity has occurred during the 1990s
and the problem in school age children is more marked in Scotland than England15.
The development of obesity can begin early in childhood and the longer it
persists and the more severe the obesity the more likely an obese child will
become an obese adult. The
identification of risk factors for obesity and key stages of development is
central to tackling the problem at the public health level.
School
children : obesity
The CHSP - School can estimate prevalence of obesity
in those areas participating in this information system. The prevalence of
obesity in school age children increased from 9.0% in primary 1 to 15.1% in
secondary year 3, in three health board areas, and was significantly higher at
each age than the UK 1990 value (5%). In
children the prevalence of obesity increases with age
7
and adolescent obesity is strongly predictive of
obesity in adult life
40
. The national study of health
and growth
41
showed in a cross sectional survey of Scottish children (age 4-11 years)
little change in the prevalence of obesity between 1974 and 1984 but the
prevalence doubled between 1984 and 1994. In addition the prevalence increased
with age and was consistently higher in Scottish children compared to English
children. In
the short term obese children are at risk of suffering a number of
co-morbidities
10
13
and are at greater risk of Type II Diabetes during
adolescence
11
. Obese
school children are more likely to be physically inactive which exacerbates the
risk to their health. The human
costs of adult obesity are well-described
42
.
A major limitation is that to date only 4 of 15
health boards are participating in the CHSP (allowing data accessible in
electronic format at ISD), although it is planned that by 2003 all health boards
will participate in the CHSP. Nevertheless
in all health boards height and weight measurements are administered and
recorded in school health records at one or more points throughout the childs
school years.
There are a number of weaknesses in the existing
system for surveillance. First at present, not all health boards participate in
the pre-school and school systems (table 1), and so complete national coverage
is not possible until all health boards are fully in the system. Second, data
collection in the current form began only recently for most health boards so
that the system can provide information on weight and height for those years in
the system only (limits retrospective analysis). A further complication in the assessment of population
undernutrition, overweight, and obesity is that height and weight data must be
converted to BMI, and then expressed as a centile or SD score.
This is relatively straightforward given expertise in handling large data
sets and the availability of inexpensive software (Child Growth Foundation,
London) for analysis of BMI data.
Growth
monitoring : obesity and undernutrition
The current CHSP- PS and S could provide a population
profile (trends and cohorts) of child growth, in identifying children at risk
and having the power to identify risk factors. In addition it could provide
baseline information to support community and school based population approaches
to improving health and tackling childhood malnutrition.
Traditionally child health surveillance covers the
whole child population. However the
value of this approach has been questioned due to limited evidence of
effectiveness and that surveillance targeted at children with identified high
risk may be more appropriate
2;43
. Review of the child health surveillance system is
likely to recommend a reduction in measurement and recording of height and
weight in future
1
. A
targeted approach would limit epidemiological analysis at a population level, as
demonstrated in this study and would reduce the opportunity to identify
overweight children at risk of developing obesity or with persisting
undernutrition. Thus Scotlands has an important asset of a national
co-ordinated system of growth monitoring through the National Users Group and
Child Health Information Team ISD. This system of central collection and
management of data and single point access can provide data on overnutrition and
undernutrition for most of the Scottish child population. The value of child
growth data in identifying risk factors and predicting future health of the
population is demonstrated by recent epidemiological studies in a number of
countries
6;8;16;44
.
Acknowledgements
This study was funded by the Chief Scientists Office.
We are grateful for the assistance of Dr Jennie Jackson and Ahmed Dorosty.
Arch.Dis.Child 2000;82:197-201
3.
Voss
LD,.Mulligan J. Child obesity and body-mass index. Lancet
1999;353:2070.
4.
Voss
LD,.Mulligan J. Too short or too fat: should we be monitoring weight? Lancet
1999;353:413-4.
5.
Reilly
JJ, Dorosty AR, Emmett PM. Prevalence of overweight and obesity in British
children: cohort study. BMJ
1999;319:1039.
6.
Kinra
S, Nelder RP, Lewendon GJ. Deprivation and childhood obesity: a cross sectional
study of 20,973 children in Plymouth, United Kingdom. J.Epidemiol.Community
Health 2000;54:456-60.
7.
Reilly
JJ,.Dorosty AR. Epidemic of obesity in UK children. Lancet
1999;354:1874-5.
8.
Bundred
P, Kitchiner D, Buchan I. Prevalence of overweight and obese children between
1989 and 1998: population based series of cross sectional studies. BMJ
2001;322 :326-8.
9.
Dietz
W. Health consequences of obesity in youth: childhood predictors of adult
disease. Pediatrics 1998;101(S):518-25.
10.
Freedman DS,
Dietz WH, Srinivasan SR, Berenson GS. The relation of overweight to
cardiovascular risk factors among children and adolescents: the Bogalusa Heart
Study. Pediatrics 1999;103:1175-82.
11.
Fagot-Campagna
A, Narayan KM, Imperatore G. Type 2 diabetes in children. BMJ
2001;322:377-8.
12.
Morrison JA,
Barton BA, Biro FM, Daniels SR, Sprecher DL. Overweight, fat patterning, and
cardiovascular disease risk factors in black and white boys. J.Pediatr.
1999;135:451-7.
13.
Strauss R.S.
Childhood obesity and Self-esteem. Pediatrics
2000;105:e15.
14.
Gortmaker
SL, Must A, Perrin JM, Sobol AM, Dietz WH. Social and economic consequences of
overweight in adolescence and young adulthood. N.Engl.J.Med.
1993;329:1008-12.
15.
Sargent
JD,.Blanchflower DG. Obesity and stature in adolescence and earnings in young
adulthood. Analysis of a British birth cohort. Arch.Pediatr.Adolesc.Med.
1994;148:681-7.
16.
Eriksson JG,
Forsen T, Tuomilehto J, Osmond C, Barker DJ. Early growth and coronary heart
disease in later life: longitudinal study. BMJ
2001;322:949-53.
17.
Eriksson JG,
Forsen T, Tuomilehto J, Winter PD, Osmond C, Barker DJ. Catch-up growth in
childhood and death from coronary heart disease: longitudinal study. BMJ
1999;318:427-31.
18.
Yajnik C.
Interaction of perturbations in interuterine growth in growth during childhood
on risk of adult onset disease. Proc.Nutr.Soc.
2000;59:257-65.
19.
Tomkins A.
Malnutrition, morbidity and mortality in children and their mothers. Proc.Nutr.Soc.
2000;59:135-46.
20.
Grantham-McGregor
SM, Walker SP, Chang S. Nutritional deficiences and later behavioural
development. Proc.Nutr.Soc. 2000;59:257-65.
21.
Scottish
Office Department of Health. Towards a Healthier Scotland: A White Paper on
Health. 2-70. 1999. Edinburgh, The
Stationary Office.
24.
Pre-School
Child Health Surveillance National User Group. Child Health Surveillance
Programme Pre-School: Clinical Guidelines for Practitioners.
2001.
27.
Cole TJ,
Freeman JV, Preece MA. Body mass index reference curves for the UK, 1990. Arch.Dis.Child
1995;73:25-9.
28.
Barlow
SE,.Dietz WH. Obesity evaluation and treatment: Expert Committee
recommendations. The Maternal and Child Health Bureau, Health Resources and
Services Administration and the Department of Health and Human Services. Pediatrics
1998;102:E29.
29.
Royal
College of Paediatrics and Child Health. Growth Reference Charts for use in the
United Kingdom. Royal College of Pediatrics and Child Health, London . 2001.
32.
Wright CM.
Identification and management of failure to thrive: a community perspective. Arch.Dis.Child 2000;82:5-9.
33.
Child Growth
Foundation. UK Cross-sectional reference data: 1990/1.
1996. London, Child Growth Foundation.
35.
Carstairs V,
Morris R. Deprivation and Health in Scotland. Aberdeen University Press, 1991.
36.
World Health
Organisation. Physical status: the use and interpretation of anthropometry.
Reprot of a WHO Expert Committee. 854. 1995. World Health Organ. Tech. Rep. Ser.
38.
Armstrong
J., Reilly JJ, Child Health Information Team. Estimating the prevalence of
underweight and overweight in Scottish pre-school children using data from Child
Health Surveillance. Ann Nutr Metab
2001;45.
39.
Wright CM.
Who comes to be weighed: an exception to the inverse care law. Lancet
1997;350:642.
40.
Power C,
Lake JK, Cole TJ. Measurement and long-term helath risks of child and adolescent
fatness. International Journal of Obesity 1997;21:507-26.
41.
Chinn
S,.Rona RJ. Prevalence and trends in overweight and obesity in three cross
sectional studies of British Children, 1974-94. BMJ
2001;322:24-6.
42.
National
Audit Office. Tackling Obesity in England. National Audit Office.
1-66. 2-8-2001. London, The Stationery Office.
44.
von Kries R,
Koletzko B, Sauerwald T, von Mutius E, Barnert D, Grunert V
et al. Breast feeding and obesity: cross sectional study
BMJ 1999;319:147-50.