See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/221737412

Associations between dietary patterns at 6 and 15 months of age and

sociodemographic factors

Article  in  European journal of clinical nutrition · January 2012

DOI: 10.1038/ejcn.2011.219 · Source: PubMed

CITATIONS READS

60 122

9 authors, including:

Lisa G Smithers Laima Brazionis

University of Adelaide University of Melbourne
94 PUBLICATIONS   1,417 CITATIONS    107 PUBLICATIONS   872 CITATIONS   

SEE PROFILE SEE PROFILE

Rebecca K Golley Murthy Mittinty

Flinders University University of Adelaide
88 PUBLICATIONS   1,882 CITATIONS    73 PUBLICATIONS   546 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Centre of Research Excellence [CRE] - Diabetic Retinopathy View project

Food choices: Perceptions and Experiences of Australian Fathers View project

All content following this page was uploaded by Lisa G Smithers on 13 March 2014.

The user has requested enhancement of the downloaded file.

 European Journal of Clinical Nutrition (2012) 66, 658 - 666
& 2012 Macmillan Publishers Limited All rights reserved 0954-3007/12
www.nature.com/ejcn

ORIGINAL ARTICLE
Associations between dietary patterns at 6 and 15 months of age
and sociodemographic factors
LG Smithers1,5, L Brazionis1,5, RK Golley2, MN Mittinty1, K Northstone3, P Emmett3, SA McNaughton4, KJ Campbell4 and JW Lynch1,3

BACKGROUND/OBJECTIVES: Data from the Avon Longitudinal Study of Parents and Children (ALSPAC) have previously shown
that dietary patterns are observable by 3 years. However, it is not clear when dietary patterns emerge. We aimed to describe
dietary patterns in early life and their associations with maternal and infant sociodemographic characteristics.
SUBJECTS/METHODS: Principal component analysis was applied to diet questionnaires of ALSPAC participants at 6 months
(n ¼ 7052) and 15 months (n ¼ 5610) to extract dietary patterns. The sociodemographic factors associated with dietary patterns
were investigated using regression analyses.
RESULTS: Four dietary patterns were extracted at both 6 and 15 months. A traditional-style pattern characterized by
home-prepared meats, vegetables and desserts, a second pattern characterized by ready-prepared baby foods and a third
pattern characterized by discretionary foods such as biscuits, sweets and crisps were identified at both ages. At 6 months,
the fourth pattern was characterized predominantly by breastfeeding and at 15 months, by contemporary-style foods
including herbs, legumes, nuts, raw fruit and vegetables. Higher maternal age and education, number of siblings and lower
body mass index (BMI) was associated with higher scores on the breastfeeding or meat, vegetables and desserts patterns,
whereas higher discretionary food pattern scores were associated with younger maternal age, lower education, higher BMI
and more siblings. Associations between sociodemographic factors and the ready-prepared baby food pattern scores were
inconsistent across ages.
CONCLUSIONS: Dietary patterns emerge from infancy and are associated with sociodemographic characteristics.

European Journal of Clinical Nutrition (2012) 66, 658 -- 666; doi:10.1038/ejcn.2011.219; published online 11 January 2012
Keywords: infant; dietary patterns; longitudinal study; ALSPAC

INTRODUCTION
Principal component analysis (PCA) is a multivariate technique
that uses correlations between foods to identify underlying
patterns of foods that are often consumed together. PCA has
previously been used to characterize the diets of children enroled
in a population-based birth cohort called the Avon Longitudinal
Study of Parents and Children (ALSPAC). Three similar diet
patterns were identified in ALSPAC children at 3, 4, 7 and 9 years,
suggesting that diet patterns are established by age 3.1 These
patterns showed agreement over time, indicating that diets may
track across childhood. It is important to understand when diet
patterns emerge and the composition of such patterns, so that
we can identify when and how to intervene to ensure children
have the best nutritional start to life. However, in our systematic
review of the literature, we found only two studies that examined
dietary patterns of children o2 years.2 Two similar patterns were
reported from these studies, although they were conducted in
different countries (United Kingdom and Norway). One pattern
was associated with healthier behaviours, including adherence to
dietary recommendations, and the other was characterized by
snack foods such as biscuits and crisps.3,4
The aims of this study are twofold. The first aim is to apply PCA
to identify major dietary patterns of ALSPAC participants at 6 and
15 months of age. These ages are ideal for examining dietary
patterns because at 6 months, the diet will reflect the first types of
foods introduced and by 15 months, most children should have
completed the transition to family foods. The second aim is to
examine associations between dietary patterns and sociodemo-
graphic characteristics of children and their mothers.
SUBJECTS AND METHODS
ALSPAC is a longitudinal birth cohort study that aims to examine the
determinants of health and development.5 All pregnant women residing in
Avon, southwest England, who were expected to deliver between April
1991 and December 1992 were invited to participate. The core ALSPAC
sample consists of 14 541 pregnancies with 13 988 infants alive at 1 year.
ALSPAC participants have been followed up almost annually since birth
and children are now aged 17 - 19 years. Information about the ALSPAC
cohort is published on the internet (http://www.bristol.ac.uk/alspac/). The
study sample was broadly representative of the population, although
ethnic minorities, single parents and unmarried couples were slightly

1
Discipline of Public Health, School of Population Health and Clinical Practice, University of Adelaide, Adelaide, South Australia, Australia; 2Public Health, Sansom Institute for
Health Research, University of South Australia, Adelaide, South Australia, Australia; 3School of Social and Community Medicine, University of Bristol, Clifton, Bristol, UK and 4Centre
for Physical Activity and Nutrition Research, Deakin University, Burwood, Victoria, Australia. Correspondence: Dr L Smithers, Discipline of Public Health, Mail Drop DX 650 550,
University of Adelaide, Adelaide, South Australia 5005, Australia.
E-mail: lisa.smithers@adelaide.edu.au
Contributors: LGS, RKG and JWL conceived the project, developed the research plan and oversaw the conduct of the project; PE collected and KN prepared data; LGS and LB
analysed the data; the PCA protocol and face validity were developed by LB; all authors assisted with interpretation of the analysis; LGS prepared the first draft of the paper, all
authors contributed to further drafts of the manuscript, and all have read and approved the final version; LGS, LB, RKG, MNM and JWL have primary responsibility for the final
content.
5
These two authors contributed equally to this work.
Received 15 July 2011; revised 4 November 2011; accepted 5 December 2011; published online 11 January 2012
 Dietary patterns at 6 and 15 months of age
LG Smithers et al

659
underrepresented compared with the 1991 National Census of women the latent structures.9,10 The number of components extracted was based
residing in Avon. Ethical approval for the study was obtained from on the second break in the Scree plot and interpretability, which is
the ALSPAC Law and Ethics Committee and Local Research Ethics common for identifying major dietary patterns.11
Committees. Food items were entered into PCA as frequency of consumption (times/
week). Some caregivers reported their child consuming a food ‘all the
Dietary data time’. These responses were re-coded as the median intake of participants
who consumed that food item. Food items were not grouped for PCA as
Food and beverage data were collected by questionnaires that were
associations between patterns and outcomes may differ depending on
posted to the primary caregiver when the study child was 6 and 15 months
how foods are grouped.12 Pattern scores were calculated for each
old. Postal reminders and phone prompts were given if questionnaires
participant by summing the product of a standardized frequency of
were not returned within 1 month. The questionnaire listed 43 food
consumption of each item by factor loadings.13,14 Patterns were
and beverage items at 6 months, increasing to 70 items at 15 months in
approximately normally distributed, had a mean of 0 and s.d. of 1.
accordance with the wider exposure to foods with age. The questionnaires
Patterns were named based on foods with loadings X|0.3| to aid
were designed by an experienced dietician (PE) who selected items based
discussion.
on foods commonly fed to infants. Questionnaires included information
on breastfeeding, formula feeding, as well as feeding behaviours and
difficulties.6 Specific questions asked whether babies were fed commer- Statistical analysis
cially made baby foods (from jar, tin or packet) such as savoury meat, fish,
Participants with complete information on all food items at 6 and 15
vegetables, fruit and milk puddings, and whether their baby was fed
months were compared with nonrespondents using independent t-tests
home-cooked foods such as potatoes, meat, egg and puddings. Caregivers
for continuous variables and w2 tests for categorical variables. Face validity
were asked to complete questionnaires documenting whether their child
of dietary patterns was examined by comparing frequency of intakes
had ever had each food/beverage item and if so, the age at which their
across quartiles of pattern scores. Multivariable linear regression was used
child first had the item. Participants who had never consumed a food item
to evaluate whether maternal or infant characteristics were independently
were assigned an intake of zero. The age at introduction of a food/
associated with dietary pattern scores. The number of components
beverage was used to exclude items introduced after 6 or 15 months of
extracted, food loadings and association between patterns and socio-
age. Current intake was obtained by asking caregivers how many times per
demographic variables were consistent between analyses based on
week they fed their child an item ‘nowadays’ and these data were used
complete case data and using all available data. Analyses using all
in the analysis described in this paper. Portion size and exclusivity of
available data are reported. Regression coefficients and 95% confidence
breastfeeding was not obtained. The questionnaire has not been directly
intervals were used to evaluate the strength and precision of associations
validated; however, food records in a subsample (n ¼ 852) at 4 months
between sociodemographic factors and dietary patterns. Statistical
confirms the consumption of these foods.7 Only questionnaires with
analyses were conducted using STATA (Intercooled 11.0, StataCorp,
complete information on all dietary questions were entered into PCA
College Station, TX, USA).
analysis. Questionnaires used in the present study are available at
http://www.bristol.ac.uk/alspac/documents/ques-cb02-my-daughter.pdf
(6-month questionnaire) and http://www.bristol.ac.uk/alspac/documents/
ques-cb03-my-infant-daughter.pdf (15-month questionnaire). RESULTS
At 6 months, 11 485 questionnaires and at 15 months, 11 073
Maternal sociodemographic and health behaviours questionnaires were returned, with the majority of questionnaires
received within 2 months of the target age at 6 (n ¼ 10 728; 93%)
Information on maternal age, education, social class, pre-pregnancy weight
and 15 months (n ¼ 10 520; 95%). The flow of participants through
and height, marital status, current or previous tobacco smoking and
the study is shown in Figure 1. Mothers who responded with
number of children (o16 years) living in the family home was collected
complete diet information at 6 months (n ¼ 7052; 61% of
by postal questionnaires between 8 and 32 weeks of gestation. Maternal
respondents) and 15 months (n ¼ 5610; 51% of respondents)
education was reported as the highest completed level of five categories
differed from nonrespondents. The proportion of mothers in each
from Certificate of Secondary Education (CSE), Vocational training, O
education and social class category differed, and respondents
(ordinary)-level (taken by the top 25% of CSE at 16 years), A (advanced)-
level (involving 2 years of study beyond O-level) and Xdegree. The CSE,
O-levels and A-levels are completed at secondary school. Social class was
categorized by maternal occupation during pregnancy rather than at the
diet assessments as mothers may have left the workforce to care for their Mothers recruited during
child.8 Body mass index (BMI, kg/m2) was calculated from maternal report pregnancy
of pre-pregnancy weight and height. n=14,541

Infant information Live births

n=14,062
Infant sex, birth weight and singleton/multiple birth information was
collected by ALSPAC staff at birth, from medical records or birth
notification. Ethnicity was obtained by questionnaire sent to the mother Follow up of infants at 6 Complete dietary data Complete data on all
during pregnancy. months for 6-month-old PCAa covariables at 6 monthsb
n=11485 n=7052 n=5129

Dietary pattern analysis

Follow up of toddlers at 15 Complete dietary data Complete data on all
PCA was conducted using PASW, version 17.0 (SPSS Inc., Chicago, IL, USA) months for 15-month-old PCAa covariables at 15 monthsb
to explore latent diet patterns on a continuous scale. PCA is a multivariable n=11073 n=5610 n=3998

data-reduction technique that uses correlations between foods to identify

latent components or ‘patterns’, such that foods with wide frequencies of
consumption have a stronger influence on the pattern. Rotation of the
correlation structure maximizes the variance within components, and
hence loadings on each component are more distinguishable. An oblimin
rotation was applied in this study because it is more flexible for examining
Figure 1. Participants at enrolment, birth, data collection and
analysis. PCA, principal component analysis. aDietary pattern
analysis was conducted on participants with complete data on all
food variables at 6 or 15 months of age. bAnalysis of the
sociodemographic predictors of dietary patterns were conducted on
participants for which all covariable and dietary data were available.

& 2012 Macmillan Publishers Limited European Journal of Clinical Nutrition (2012) 658 - 666
 Dietary patterns at 6 and 15 months of age
LG Smithers et al

660
Table 1. Characteristics of surviving ALSPAC infants in the core cohort and those included in complete case PCA analysis at 6 and 15 months

6 months 15 months Core cohort Comparison at Comparison at

(n ¼ 7052) (n ¼ 5610) (n ¼ 13 978) 6 monthsa 15 monthsa

Maternal characteristics
Mother’s age at child’s birth (years)b 28.6 (4.7) 28.5 (4.7) 28.0 (5.0) 1.2 (0.99, 1.3) 0.8 (0.7, 1.0)
Not recorded/missingc 0 (0) 0 (0) 0 (0) [o0.001] [o0.001]

Highest education achievedc

CSE 1055 (15.0) 788 (14.0) 2504 (17.9)
Vocational 616 (8.7) 494 (8.8) 1224 (8.8)
O-levels 2429 (34.4) 2071 (36.9) 4296 (30.7) [o0.001] [o0.001]
A-levels 1700 (24.1) 1336 (23.8) 2794 (20.0)
Degree 1020 (14.5) 737 (13.1) 1600 (11.4)
Not recorded/missing 232 (3.3) 184 (3.3) 1560 (11.1)

Mother’s social classc,d

I 387 (5.5) 283 (5.0) 591 (4.2)
II 1896 (26.9) 1517 (27.0) 3169 (22.7) [o0.001] [o0.001]
III (non-manual) 2495 (35.4) 2032 (36.2) 4304 (30.8)
III (manual) 384 (5.4) 319 (5.7) 789 (5.6)
IV 503 (7.1) 391 (7.0) 992 (7.1)
V 95 (1.3) 95 (1.7) 218 (1.6)
Not recorded/missing 1292 (18.3) 973 (17.3) 3915 (28.0)

Parityc
Primiparous 3088 (43.8) 2510 (44.7) 5772 (41.3)
Multiparous 3708 (52.6) 2926 (52.2) 7157 (51.5) [0.06] [0.003]
Not recorded/missing 256 (3.6) 174 (3.1) 1049 (7.5)

Smokingc
Never 3473 (49.2) 2714 (48.4) 5804 (41.5)
Quit 2009 (28.4) 1582 (28.2) 3457 (24.7) [o0.001] [0.001]
In pregnancy 1133 (16.1) 912 (16.3) 2173 (15.5)
Not recorded/missing 437 (6.2) 402 (7.2) 2544 (18.2)

Infants characteristics
Sexc
Male 3615 (51.3) 2865 (51.1) 7220 (51.7) [0.34] [0.25]
Female 3437 (48.7) 2745 (48.9) 6756 (48.3)
Not recorded/missing 0 (0) 0 (0) 2 (0)

Multiplicityc
Singleton 6885 (97.6) 5477 (97.6) 13 617 (97.4) [0.07] [0.20]
Twin 167 (2.4) 133 (2.4) 361 (2.6)

Birth weight (kg)c 3.41 (0.55) 3.41 (0.55) 3.39 (0.56) 0.04 (0.02, 0.06) 0.03 (0.05, 0.01)
Not recorded/missing 84 (1.2) 74 (1.3) 180 (1.3) [o0.001] [0.001]
Abbreviations: ALSPAC, Avon Longitudinal Study of Parents and Children; CSE, Certificate of Secondary Education; PCA, principal component analysis.
a
Statistical tests compared characteristics of participants with PCA analysis against those with no PCA analysis at 6 and 15 months. Continuous variables
were compared by using independent t-test and categorical variables were compared by using w2 test. Values show the mean difference (95% CI) and
[P-value]. bValues are mean (s.d.). cValues are number (%). dSocial class was categorized according to maternal occupation using UK standard classifications
of occupation.8

were slightly older, nonsmokers, primiparous and had fewer
missing data compared with nonrespondents (Table 1).
Food intake and dietary patterns at 6 months
The 10 foods consumed most frequently by participants at 6
months of age were (mean (median), nonconsumers assigned 0
times/week): infant formula 13.4 (21), breastfeeds 7.9 (0), baby rice
cereal 4.6 (7), baby fruit pudding 3.7 (3), home-prepared
vegetables 3.7 (3), baby vegetables 3.3 (3), baby meat 3.11 (3),
home-prepared potatoes 3.1 (3), fruit drink 2.8 (1) and plain rusks
2.0 (1). The addition of sugar or cereal to infant bottles occurred in
8.6% and 2.7% of infants, respectively.
Four dietary patterns were extracted at 6 months of age. Table 2
shows loadings for food items on each pattern. Foods loading
strongly on the first pattern included (25th -- 75th percentiles in
times/week); home-prepared vegetables (1 -- 7), potato (0 -- 5), meat
(0 -- 3), fish (0 -- 1), fruit puddings (0 -- 3), milk puddings (0-0) and egg
(0 -- 0), which are associated with a traditional English-style diet,
and therefore the first pattern was named ‘Meat, Vegetables and
Desserts’. Biscuits (0 -- 0), chocolate (0 -- 0), tea (0 -- 0), sweets (0 -- 0),
bread (0 -- 2), cola (0 -- 0), other fizzy drinks (0 -- 0) and adding sugar
to infant bottles loaded on the second pattern, named ‘Biscuits,
Sweets and Crisps’. The third pattern, labelled ‘Ready-prepared baby
foods’, was characterized by commercially manufactured foods for
infants including tinned/jar meat (0 -- 5), fruit pudding (1 -- 7),
vegetables (1 -- 5), milk pudding (0 -- 5) and fish (0 -- 1). The fourth
pattern was termed ‘Breastfeeding’ as it had a strong positive
loading on breastfeeding (0 -- 14), strong negative loading on
infant formula (0 -- 21) and moderate loadings on raw fruit (0 -- 2)
and vegetables (0 -- 0). A complete description of the mean
frequency of food intakes across quartiles of pattern scores is

European Journal of Clinical Nutrition (2012) 658 -- 666 & 2012 Macmillan Publishers Limited
 Dietary patterns at 6 and 15 months of age
LG Smithers et al

661
Table 2. Oblimin-rotated structure matrix loadings of foods on four dietary patterns extracted by principal component analysis at 6 months of
age (n ¼ 7052)

n (%) respondents Dietary patternsa

who consumed food
Meat, vegetables Biscuits, sweets Ready-prepared Breastfeeding
and desserts and crisps baby foods

Baby foods
Breast milk 2018 (28.6) 0.005 0.15 0.16 0.78
Infant formula 4853 (68.8) 0.07 0.02 0.10 0.76
Follow-on formula 491 (7.0) 0.02 0.04 0.08 0.09
Hypoallergenic formula 10 (0.1) 0.01 0.02 0.02 0.05
Soy milk 117 (1.7) 0.01 0.00 0.03 0.07
Goat milk 7 (0.1) 0.01 0.07 0.03 0.06
Baby milk pudding 3514 (49.8) 0.01 0.05 0.62 0.11
Baby meat 5118 (72.6) 0.19 0.01 0.71 0.12
Baby fish 1869 (26.5) 0.07 0.04 0.48 0.05
Baby vegetables 5569 (79.0) 0.19 0.07 0.69 0.07
Baby fruit pudding 5458 (77.4) 0.15 0.01 0.69 0.05
Rice cereal 2169 (30.8) 0.03 0.09 0.11 0.19
Flavoured rice cereal 1322 (18.7) 0.04 0.06 0.22 0.19
Plain rusks 3803 (53.9) 0.05 0.18 0.19 0.02
Sweetened rusks 1366 (19.4) 0.01 0.25 0.16 0.02
Gripe waterb 1633 (23.2) 0.03 0.18 0.18 0.08
Sugar added to bottles 604 (8.6) 0.12 0.37 0.01 0.08
Cereal added to bottles 193 (2.7) 0.02 0.15 0.02 0.10

Family foods
Cow’s milk 1228 (17.4) 0.24 0.25 0.03 0.05
Home-prepared milk puddings 1329 (18.8) 0.46 0.17 0.03 0.004
Home-prepared meat 3165 (44.9) 0.74 0.03 0.07 0.04
Home-prepared fish 1814 (25.7) 0.60 0.07 0.03 0.04
Home-prepared egg 924 (13.1) 0.35 0.31 0.01 0.04
Home-prepared potatoes 5239 (74.3) 0.82 0.03 0.13 0.05
Home-prepared vegetables 5306 (75.2) 0.82 0.03 0.20 0.15
Raw vegetables 1066 (15.1) 0.25 0.07 0.05 0.36
Raw fruit 2755 (39.1) 0.37 0.04 0.08 0.41
Home-prepared fruit pudding 2973 (42.2) 0.59 0.08 0.17 0.25
Apple juice 1392 (19.7) 0.18 0.020 0.11 0.16
Blackcurrant and rosehip drink 2322 (32.9) 0.13 0.25 0.26 0.13
Other fruit drink 3606 (51.1) 0.21 0.07 0.19 0.11
Bread or toast 2667 (37.8) 0.35 0.42 0.08 0.24
Other cereal 5432 (77.0) 0.26 0.04 0.20 0.07
Biscuits 1594 (22.6) 0.22 0.58 0.17 0.04
Chocolate 1314 (18.6) 0.06 0.57 0.14 0.04
Sweets 80 (1.1) 0.01 0.44 0.03 0.08
Crisps 122 (1.7) 0.08 0.32 0.02 0.01
Herbal drink 2564 (36.4) 0.03 0.08 0.16 0.09
Cola 27 (0.4) 0.00 0.42 0.04 0.05
Other fizzy drinks 39 (0.6) 0.04 0.41 0.02 0.05
Tea 447 (6.3) 0.08 0.50 0.03 0.01
Coffee 36 (0.5) 0.01 0.38 0.05 0.04
Alcohol 44 (0.6) 0.03 0.31 0.04 0.10
a
Loadings X|0.3| are shown in bold. bGripe water is a solution used to relieve infants from colic and gastrointestinal discomfort.

shown in Supplementary Table S1. Together, Table 2 and
Supplementary Table S1 show that foods with strong positive
loadings have a gradient increasing across quartiles of pattern
scores and foods with negative loadings decrease across quartiles.
Foods consumed infrequently (for example, goat milk) or very
frequently by many participants (other cereals, cow’s milk) do not
load on any pattern as there is little variation in intake and no
gradient across quartiles.
Food intake and dietary patterns at 15 months
At 15 months, the foods consumed with a mean and median
intake X4 times/week were (mean (median), nonconsumers
assigned 0 times/week): cow’s milk 11.5 (14), breakfast cereal 5.7
(7), bread or toast 6.6 (7), biscuit 5.6 (5), meat products 4.4 (4),
potatoes 5.4 (5), other vegetables 6.3 (7), yoghurt 4.4 (4) and raw
fruit 4.3 (4).
Four patterns were extracted at 15 months of age (Table 3).
Foods ’loading on the first pattern included herbs (25th -- 75th
quartiles in times/week; 1 -- 4), legumes (1 -- 2), raw vegetables (1 -- 4)
and fruit (3 -- 7), spices (1 -- 3) and cheese (2 -- 4), and was named
‘Herbs, Raw Fruit and Vegetables’. Fizzy drinks (0 -- 2), sweets (1 -- 3),
crisps (1 -- 3), biscuits (4 -- 7) and ketchup (1 -- 2) loaded on the
second pattern and because of similarity with the second pattern
identified at 6 months, it was named ‘Biscuits, Sweets and Crisps’.
Consistent with 6 months, commercially made baby foods loaded
strongly on the third pattern at 15 months, which was named
‘Ready-prepared baby foods’ (includes rice cereal (0 -- 0), other baby
cereal (0 -- 0) rusks (0 -- 2), baby meat (0 -- 1), baby vegetables (0 -- 1),
baby milk pudding (0 -- 0) and baby fruit pudding (0 -- 2)). The

& 2012 Macmillan Publishers Limited European Journal of Clinical Nutrition (2012) 658 - 666
 Dietary patterns at 6 and 15 months of age
LG Smithers et al

662
Table 3. Oblimin-rotated structure matrix loadings of foods on four dietary patterns extracted by principal component analysis at 15 months of age
(n ¼ 5610)

n (%) respondents Dietary patternsa

consumed food
Herbs, raw fruit Discretionary Ready-prepared Reverse meat,
and vegetables foods baby foods vegetables and desserts

Types of milk
Breast milk 240 (4.3) 0.17 0.01 0.02 0.20
Infant formula 340 (6.1) 0.01 0.04 0.23 0.06
Follow-on formula 271 (4.8) 0.05 0.03 0.18 0.01
Hypoallergenic formula 3 (0.1) 0.01 0.02 0.01 0.01
Soya formula 91 (1.6) 0.06 0.03 0.06 0.09

Commercially made baby foods, toddler foods and junior foods (from jar, tin or packet)
Baby milk pudding 787 (14.0) 0.07 0.00 0.70 0.06
Baby meat 1176 (21.0) 0.16 0.00 0.75 0.00
Baby fish 530 (9.4) 0.04 0.00 0.61 0.06
Baby vegetables 1071 (19.1) 0.05 0.01 0.78 0.04
Baby fruit pudding 1213 (21.6) 0.06 0.02 0.75 0.01
Rice cereal 155 (2.8) 0.05 0.03 0.36 0.01
Other baby cereal 457 (8.1) 0.01 0.01 0.43 0.04
Rusks 1262 (22.5) 0.06 0.04 0.40 0.01
b
Gripe water 157 (2.8) 0.00 0.03 0.17 0.01

Family foods
Cow’s milk 5104 (91.0) 0.04 0.08 0.06 0.24
Goat’s milk 20 (0.4) 0.03 0.02 0.00 0.06
Soya milk 44 (0.8) 0.17 0.02 0.07 0.06
Other milk 42 (0.7) 0.07 0.01 0.04 0.04
Cheesec 4878 (87.0) 0.47 0.08 0.08 0.16
Yoghurt 5159 (92.0) 0.15 0.14 0.11 0.35
Milk puddingc 4145 (73.9) 0.12 0.09 0.08 0.47
Meat or meat productsc 5227 (93.2) 0.03 0.19 0.10 0.60
Liver/pate 794 (14.2) 0.13 0.07 0.04 0.06
Kidney 246 (4.4) 0.08 0.04 0.00 0.05
Smoked/cured food 1899 (33.9) 0.29 0.18 0.04 0.07
Fish or fish productsc 4849 (86.4) 0.30 0.01 0.04 0.31
Shell fish 305 (5.4) 0.25 0.20 0.05 0.10
Barbecue food 309 (5.5) 0.13 0.18 0.04 0.06
Eggc 3904 (69.6) 0.28 0.16 0.07 0.06
Nuts/nut products 328 (5.8) 0.39 0.11 0.02 0.16
Baked beans 4891 (87.1) 0.07 0.32 0.05 0.17
Green peas 5118 (91.2) 0.06 0.23 0.01 0.39
Other legume 1198 (21.4) 0.58 0.16 0.01 0.12
Potatoesc 5509 (98.2) 0.05 0.20 0.05 0.68
Other vegetablesc 5478 (97.6) 0.39 0.02 0.07 0.59
Raw carrot 2024 (36.1) 0.37 0.01 0.03 0.03
Other raw vegetables 707 (12.6) 0.54 0.08 0.01 0.06
Raw apple 4436 (79.1) 0.43 0.06 0.00 0.10
Other raw fruit 5220 (93.0) 0.50 0.09 0.02 0.15
Fig/fig products 161 (2.9) 0.20 0.03 0.03 0.02
Fruit puddingc 4397 (78.4) 0.29 0.11 0.10 0.45
Apple juice 1867 (33.3) 0.35 0.17 0.01 0.03
Blackcurrant and rosehip drink 2691 (48.0) 0.04 0.16 0.06 0.16
Other fruit juice 3491 (62.2) 0.22 0.02 0.07 0.09
Any other fruit drink 3638 (64.8) 0.13 0.29 0.12 0.15
Bread or toast 5496 (98.0) 0.29 0.02 0.05 0.23
Breakfast cereal 5012 (89.3) 0.05 0.07 0.08 0.21
Biscuits 5381 (95.9) 0.02 0.34 0.07 0.27
Chocolate 4492 (80.1) 0.07 0.55 0.07 0.09
Sweets 1651 (29.4) 0.01 0.52 0.02 0.04
Mints 203 (3.6) 0.06 0.27 0.02 0.08
Crisps 3418 (60.9) 0.03 0.49 0.03 0.03
Other savoury snacks 3447(61.4) 0.02 0.31 0.08 0.16
Packet microwave meal 254 (4.5) 0.09 0.14 0.04 0.02
Canned soup 1915 (34.1) 0.08 0.29 0.06 0.03
Packet soup 337 (6.0) 0.01 0.18 0.01 0.00
Added salt 1141 (20.3) 0.04 0.29 0.06 0.01
Gravy or soy sauce 4382 (78.1) 0.21 0.32 0.01 0.40
Herbs 2015 (35.9) 0.60 0.06 0.10 0.02
Spices 845 (15.1) 0.48 0.08 0.04 0.11
Tomato ketchup 1421 (25.3) 0.17 0.44 0.04 0.05

European Journal of Clinical Nutrition (2012) 658 -- 666 & 2012 Macmillan Publishers Limited

Table 3 (Continued )
n (%) respondents
consumed food
Herbs, raw fruit
and vegetables
Added sauce (other) 513 (9.1)
Added sugar 1482 (26.4)
Herbal drink 485 (8.6)
Sports drink 12 (0.2)
Cola 724 (12.9)
Other fizzy drink 790 (14.1)
Tea 1939 (34.6)
Coffee 320 (5.7)
Alcohol 189 (3.4)
a
Loadings X|0.3| are shown in bold. bGripe water is a solution used to relieve infants from colic and gastrointestinal discomfort. cFoods identified in the
15-month questionnaire as not being ready-prepared baby or toddler varieties.
fourth pattern extracted at 15 months was negatively associated
with home-prepared, traditional British foods and was named
‘Reverse Meat, Vegetables and Desserts’. The reverse direction of the
association indicates that lower scores are associated with higher
consumption of home-prepared potatoes (4 -- 7), meat products
(3 -- 7), vegetables (5 -- 7) and milk puddings (1 -- 4).
Similar to 6 months, a gradient of food intake across quartiles of
dietary pattern scores at 15 months is evident for foods with high
loadings (Supplementary Table S2). Foods consumed in high
quantities by the majority of infants (for example, cow’s milk,
breakfast cereal) do not load on any pattern and have a
consistently high intake across all quartiles of pattern scores.
Dietary patterns and sociodemographic characteristics
After adjustment for covariables, maternal age, education, marital
status, BMI and infants’ number of siblings were independently
associated with pattern scores at 6 months (Table 4). For example,
Xdegree education of mothers was associated with a 0.31 (95%
confidence interval 0.19, 0.43) higher Meat, Vegetables and Desserts
pattern score than the lowest education level. Higher scores on the
Meat, Vegetables and Desserts pattern were associated with married
mothers and infants of white ethnicity, whereas lower scores were
associated with smoking during pregnancy, maternal BMI 430 kg/m2
and infants with more siblings. Associations between the Biscuits,
Sweets and Crisps pattern and sociodemographic variables were
generally in the opposite direction, with higher scores associ-
ated with younger maternal age, lower education, smoking in
pregnancy, higher BMI and more siblings. Higher scores on the
Ready-prepared baby foods pattern were associated with younger
maternal age, lower education and smoking in pregnancy.
Breastfeeding pattern scores were positively associated with
maternal age, education and more siblings, and negatively with
smoking during pregnancy and maternal BMI X25 kg/m2.
At 15 months, the associations between sociodemographic
characteristics and scores on the Biscuits, Sweets and Crisps pattern
were generally consistent with that at 6 months (Table 5). For
example, maternal BMI X30 kg/m2 was associated with a 0.43
(0.31, 0.55) higher score on the Biscuits, Sweets and Crisps pattern
at 15 months. Lower scores on the Reverse Meat, Vegetables and
Desserts pattern were associated with being married 0.25 (0.33,
0.16) and the reverse direction indicates that a higher mean
intake of traditional-style foods such as cooked meat and
vegetables is associated with lower scores. Older mothers, not
being married and not having other children predicted higher
scores on the Ready-prepared baby foods pattern. Higher maternal
age and education were positively associated, whereas being
married and white ethnicity were negatively associated with
Herbs, Raw Fruit and Vegetables pattern scores.
& 2012 Macmillan Publishers Limited
Dietary patterns at 6 and 15 months of age
LG Smithers et al
663
Dietary patternsa
Discretionary Ready-prepared Reverse meat,
foods baby foods vegetables and desserts
0.23 0.04 0.00 0.04
0.07 0.30 0.00 0.03
0.01 0.05 0.13 0.06
0.01 0.09 0.02 0.06
0.01 0.54 0.01 0.15
0.03 0.53 0.00 0.14
0.16 0.36 0.00 0.02
0.05 0.24 0.01 0.03
0.14 0.19 0.02 0.10
DISCUSSION
The present study involving a large population-based cohort
provides evidence that dietary patterns are identifiable at 6
months of age and social factors are associated with these
patterns. A strength of this analysis is that each pattern is clearly
distinguishable by the type of foods loading on it. Of the four
patterns extracted, three were similar at 6 and 15 months and
these shared similarity to patterns identified in the Southampton
Women’s Study (SWS) at 6 and 12 months (n ¼ 1432) and the
Norwegian Mother and Child Cohort (MoBa) at 18 months
(n ¼ 27 763).3,4 For example, the Meat, Vegetables and Desserts
pattern is similar to the SWS Infant guidelines and MoBa
Wholesome patterns, whereas the Biscuits, Sweets and Crisps
pattern is consistent with the SWS Adult Foods and MoBa
Unhealthy patterns. The Ready-prepared baby foods pattern is
similar to the SWS Baby jar foods at 6 months. Loadings of key
foods such as breastfeeding and formula was strong at 6 months
(rB0.8) declining to levels similar to those observed in the SWS at
15 months (rB0.1 -- 0.2). Foods loading on the Meat, Vegetables
and Desserts, the Biscuits, Sweets and Crisps and Herbs, Raw Fruit
and Vegetables patterns were similar to the Traditional, Processed
and Healthy patterns identified in the ALSPAC cohort at 3 -- 9
years,1,15 suggesting that patterns present in childhood may
emerge during infancy. However, not all early patterns are evident
at later ages (for example, Breastfeeding and Ready-prepared baby
foods), which is related to providing foods only appropriate for
infants. Data collected in the United States at a similar period to
ALSPAC show similarities in proportions of children being fed
breast milk (B5%) and cow’s milk (B90%), but there are large
differences in other foods.16 For example, 0% of 15-to-18-month-
old US children were fed meat-based baby foods, 54% were fed
non-infant cereals and 40 -- 50% were fed bread and biscuits,
compared with 21%, 89% and 98% of ALSPAC children,
respectively.16 Therefore, further work is necessary to determine
whether the findings are generalizable outside of England and
Europe.
The mean frequency of food intakes across quartiles of pattern
scores demonstrates the face validity of the patterns (Supple-
mentary Tables S1 and S2). For example, the Meat, Vegetables and
Desserts pattern is associated with higher intakes of home-cooked
vegetables. However, foods consumed by many participants do
not load on any pattern because the food is not unique to that
style diet, and thus someone who scores highly on the Biscuits,
Sweets and Crisps patterns may regularly consume cereals.
We extracted four dietary patterns at both ages, whereas the
MoBa and SWS studies reported two or three patterns. It is
possible that grouping foods may have resulted in such
differences, although the number of patterns extracted may also
European Journal of Clinical Nutrition (2012) 658 - 666
 Dietary patterns at 6 and 15 months of age
LG Smithers et al

664
Table 4. Associations between dietary patterns scores at 6 months and sociodemographic characteristics (n ¼ 5129)a

Dietary patterns

n Meat, vegetables Biscuits, sweets Ready-prepared Breastfeeding

and desserts and crisps baby foods

ba 95% CI b 95% CI b 95% CI b 95% CI

Maternal and family characteristics

Maternal age
p20 years 175 Referent
21 - 25 years 628 0.109 (0.059, 0.276) 0.276 (0.429, 0.122) 0.131 (0.295, 0.034) 0.023 (0.140, 0.187)
26 - 30 years 2923 0.111 (0.047, 0.270) 0.582 (0.728, 0.436) 0.262 (0.418, 0.106) 0.024 (0.131, 0.179)
X31 years 1403 0.110 (0.056, 0.276) 0.671 (0.824, 0.519) 0.258 (0.421, 0.095) 0.205 (0.043, 0.367)
Maternal education
CSE 589 Referent
Vocational 424 0.112 (0.011, 0.235) 0.044 (0.157, 0.069) 0.166 (0.045, 0.287) 0.048 (0.073, 0.168)
O-level 1873 0.146 (0.052, 0.239) 0.125 (0.211, 0.039) 0.086 (0.006, 0.178) 0.168 (0.077, 0.260)
A-level 1380 0.329 (0.226, 0.431) 0.209 (0.303, 0.115) 0.024 (0.077, 0.125) 0.369 (0.269, 0.470)
XDegree 863 0.311 (0.190, 0.431) 0.292 (0.403, 0.181) 0.125 (0.243, 0.006) 0.757 (0.639, 0.875)
b
Maternal social class
I 357 Referent
II 1704 0.051 (0.171, 0.069) 0.009 (0.102, 0.119) 0.066 (0.184, 0.052) 0.074 (0.191, 0.044)
III (non-manual) 2228 0.171 (0.298, 0.043) 0.052 (0.065, 0.169) 0.089 (0.215, 0.036) 0.239 (0.364, 0.115)
III (manual) 323 0.000 (0.163, 0.163) 0.124 (0.026, 0.275) 0.059 (0.220, 0.102) 0.059 (0.219, 0.101)
IV 437 0.025 (0.130, 0.180) 0.070 (0.072, 0.213) 0.068 (0.220, 0.084) 0.130 (0.281, 0.022)
V 80 0.100 (0.351, 0.152) 0.301 (0.070, 0.532) 0.157 (0.404, 0.091) 0.138 (0.384, 0.108)
Smoking
Never 2798 Referent (0.056, 0.066) 0.022 (0.035, 0.078) 0.033 (0.027, 0.093) 0.002 (0.058, 0.062)
Quit 1577 0.005
In pregnancy 754 0.083 (0.167, 0.001) 0.169 (0.092, 0.246) 0.110 (0.027, 0.192) 0.131 (0.213, 0.049)
Marital status
Not married 933 Referent
Married 4196 0.074 (0.003, 0.150) 0.049 (0.120, 0.021) 0.042 (0.033, 0.117) 0.067 (0.141, 0.008)
Mother’s BMI (kg/m2)
o20 950 0.058 (0.129, 0.013) 0.003 (0.062, 0.068) 0.060 (0.130, 0.009) 0.049 (0.020, 0.118)
20 to o25 3213 Referent
25 to o30 722 0.017 (0.096, 0.062) 0.074 (0.001, 0.146) 0.090 (0.012, 0.168) 0.089 (0.166, 0.012)
X30 244 0.127 (0.255, 0.001) 0.196 (0.078, 0.313) 0.057 (0.068, 0.183) 0.117 (0.241, 0.008)
Number of older siblings
None 2423 Referent
One 1905 0.021 (0.082, 0.039) 0.078 (0.022, 0.133) 0.052 (0.007, 0.112) 0.054 (0.005, 0.113)
Two or more 801 0.099 (0.181, 0.016) 0.222 (0.146, 0.298) 0.045 (0.037, 0.126) 0.160 (0.079, 0.241)

Infant characteristics
Sex
Male 2663 Referent
Female 2496 0.059 (0.113, 0.006) 0.021 (0.070, 0.078) 0.001 (0.052, 0.054) 0.070 (0.015, 0.119)
Ethnicity
Non-white 176 Referent Referent Referent
White 4953 0.172 (0.025, 0.320) 0.024 (0.112, 0.159) 0.024 (0.121, 0.169) 0.055 (0.199, 0.090)
Multiple birth
Singleton 5020 Referent
Twin 109 0.077 (0.110, 0.264) 0.045 (0.217, 0.127) 0.085 (0.100, 0.268) 0.150 (0.333, 0.033)
a
Abbreviations: BMI, body mass index; CSE, certificate of secondary education. Results were obtained from linear regression models with diet pattern
score as the independent variable and all covariables as independent predictors. Data shown are regression model b coefficients and their 95% confidence
intervals (95% CIs). bSocial class was categorized according to maternal occupation using standard classifications of occupation.8

be the result of broader extraction criteria. Our four-component
solution was robust to different rotation methods (none, Varimax,
Oblimin) and truncating intakes to be less right skewed (data not
shown). The use of a food frequency questionnaire rather than
other measures of diet is not a limitation of this work as food
frequency questionnaires are considered appropriate for large
population-based studies;17 however, portion-size data were not
included and the food frequency questionnaire used here has not
been validated against other diet assessments. It is important to
consider that attrition, missing data or that socioeconomic factors
collected during pregnancy and not when diet was assessed may have
influenced associations between diet and socioeconomic factors.
Diet patterns were associated with maternal age, education,
smoking in pregnancy and the number of siblings, with associ-
ations consistent at both ages and with patterns at age 3.18
Indeed, the SWS reported that higher maternal education, age and
lower BMI were associated with healthier patterns at 6 and 12
months, and reverse associations were found with unhealthier
patterns,3 suggesting agreement across studies and time. Never-
theless, the association between the number of siblings and
dietary patterns seemed contradictory. A greater number of
siblings was associated with higher scores on Biscuits, Sweets and
Crisps pattern, but was also associated with the Breastfeeding
pattern. The association between siblings and Breastfeeding scores

European Journal of Clinical Nutrition (2012) 658 -- 666 & 2012 Macmillan Publishers Limited
 Dietary patterns at 6 and 15 months of age
LG Smithers et al

665
Table 5. Associations between dietary patterns scores at 15 months and sociodemographic characteristics (n ¼ 3998)

Dietary patterns

n Herbs, raw fruit Biscuits, sweets Ready-prepared Reverse meat,

and vegetables and crisps baby foods vegetables and desserts

ba 95% CI b 95% CI b 95% CI b 95% CI

Maternal and family characteristics

Maternal age
p20 years 118 Referent
21 - 25 years 493 0.186 (0.000, 0.371) 0.379 (0.554, 0.203) 0.044 (0.148, 0.236) 0.045 (0.148, 0.238)
26 - 30 years 2318 0.301 (0.123, 0.478) 0.636 (0.803, 0.468) 0.151 (0.032, 0.334) 0.019 (0.165, 0.204)
X31 years 1069 0.483 (0.298, 0.668) 0.772 (0.947, 0.597) 0.292 (0.101, 0.483) 0.016 (0.176, 0.209)
Maternal education
CSE 425 Referent
Vocational 337 0.075 (0.056, 0.206) 0.065 (0.189, 0.058) 0.026 (0.161, 0.109) 0.014 (0.122, 0.151)
O-level 1526 0.214 (0.114, 0.315) 0.202 (0.297, 0.107) 0.094 (0.198, 0.010) 0.166 (0.271, 0.061)
A-level 1092 0.556 (0.445, 0.666) 0.274 (0.378, 0.170) 0.093 (0.207, 0.021) 0.236 (0.351, 0.121)
XDegree 618 0.932 (0.800, 1.063) 0.274 (0.652, 0.403) 0.093 (0.261, 0.010) 0.085 (0.222, 0.051)
b
Maternal social class
I 258 Referent
II 1355 0.045 (0.088, 0.178) 0.040 (0.166, 0.085) 0.004 (0.134, 0.141) 0.074 (0.213, 0.064)
III (non-manual) 1752 0.146 (0.287, 0.006) 0.066 (0.068, 0.199) 0.010 (0.156, 0.135) 0.035 (0.182, 0.111)
III (manual) 239 0.050 (0.230, 0.129) 0.132 (0.038, 0.302) 0.060 (0.125, 0.246) 0.033 (0.154, 0.220)
IV 322 0.051 (0.221, 0.119) 0.115 (0.046, 0.275) 0.099 (0.077, 0.274) 0.038 (0.216, 0.139)
V 72 0.282 (0.538, 0.026) 0.302 (0.061, 0.544) 0.236 (0.028, 0.500) 0.034 (0.232, 0.300)
Smoking
Never smoked 2172 Referent
Quit 1225 0.135 (0.070, 0.199) 0.007 (0.054, 0.068) 0.003 (0.063, 0.070) 0.016 (0.051, 0.083)
In pregnancy 601 0.021 (0.108, 0.066) 0.124 (0.042, 0.206) 0.013 (0.077, 0.103) 0.023 (0.067, 0.114)
Marital status
Not married 717 Referent
Married 3281 0.116 (0.197, 0.036) 0.043 (0.033, 0.119) 0.117 (0.200, 0.034) 0.245 (0.329, 0.162)
Mother’s BMI (kg/m2)
o20 710 0.018 (0.058, 0.094) 0.083 (0.154, 0.011) 0.050 (0.123, 0.036) 0.051 (0.131, 0.028)
20 to o25 2487 Referent
25 to o30 596 0.004 (0.078, 0.085) 0.201 (0.124, 0.278) 0.042 (0.042, 0.126) 0.118 (0.202, 0.033)
X30 205 0.061 (0.069, 0.191) 0.429 (0.306, 0.552) 0.078 (0.212, 0.056) 0.009 (0.144, 0.127)
Number of older siblings
None 1927 Referent
One 1446 0.041 (0.023, 0.105) 0.367 (0.306, 0.428) 0.186 (0.252, 0.120) 0.044 (0.023, 0.111)
Two or more 625 0.072 (0.160, 0.016) 0.354 (0.271, 0.437) 0.242 (0.333, 0.152) 0.104 (0.013, 0.196)

Infant characteristics
Sex
Male 2047 Referent
Female 1951 0.043 (0.013, 0.100) 0.020 (0.033, 0.065) 0.012 (0.046, 0.071) 0.069 (0.010, 0.127)
Ethnicity
Non-white 119 Referent
White 3879 0.365 (0.533, 0.198) 0.089 (0.247, 0.070) 0.166 (0.339, 0.007) 0.424 (0.599, 0.250)
Multiple birth
Singleton 3913 Referent
Twin 85 0.187 (0.385, 0.012) 0.220 (0.407, 0.032) 0.145 (0.060, 0.350) 0.626 (0.833, 0.419)
a
Abbreviations: BMI, body mass index; CSE, certificate of secondary education. Results were obtained from linear regression models with diet pattern score
as the independent variable and all covariables as independent predictors. Data shown are regression model b coefficients and their 95% confidence
intervals (95% CIs). Food loadings on the Reverse Meat, Vegetables and Desserts pattern were negative and therefore lower scores reflect higher intake of foods
that load strongly this pattern (such as meat, home-cooked vegetables and desserts). bSocial class was categorized according to maternal occupation
using standard classifications of occupation.8

may be because of previously breastfeeding an infant,19 and the
impact of more siblings on poorer diet quality has been
documented in ALSPAC and elsewhere.15,20 There is little empirical
evidence on how diets change as the number of siblings increase
and this may be an important area for further research.
In summary, dietary patterns can be identified at 6 months
and these patterns show face validity for detecting variation in
food intake across pattern scores. Whether these patterns are
associated with later health and development is the subject of
future research. Maternal age, education, BMI, smoking and
number of siblings were the characteristics most strongly
associated with dietary patterns. These data illustrate the natural
history of diet and its social patterning.
CONFLICT OF INTEREST
Both PE and KN have received support from commercial infant food manufacturers
and have undertaken invited lectures. The other authors declare no conflict of interest.

& 2012 Macmillan Publishers Limited European Journal of Clinical Nutrition (2012) 658 - 666
 Dietary patterns at 6 and 15 months of age
LG Smithers et al
666
ACKNOWLEDGEMENTS
We are extremely grateful to all the families who took part in this study, the midwives
for their help in recruiting them, and the whole ALSPAC team, which includes
interviewers, computer and laboratory technicians, clerical workers, research
scientists, volunteers, managers, receptionists and nurses. The UK Medical Research
Council, the Wellcome Trust and the University of Bristol provide funding for the
ALSPAC study. JWL is supported by an Australia Fellowship and RKG with a
Postdoctoral Training Fellowship from the National Health and Medical Research
Council of Australia. LGS, LB and MNM are supported by funds from the Australia
Fellowship awarded to JWL. KN and PE are partly supported by the UK Arthritic
Association and partly by funding from the European Community’s Seventh
Framework Programme (FP7/2007-2013 Grant number 245012).
REFERENCES
1 Northstone K, Emmett PM. Are dietary patterns stable throughout early and
mid-childhood? A birth cohort study. Br J Nutr 2008; 100, 1069 - 1076.
2 Smithers LG, Golley RK, Brazionis L, Lynch JW. Characterizing whole diets of young
children from developed countries and the association between diet and health:
a systematic review. Nutr Rev 2011; 69, 449 - 467.
3 Robinson S, Marriott L, Poole J, Crozier S, Borland S, Lawrence W et al. Dietary
patterns in infancy: the importance of maternal and family influences on feeding
practice. Br J Nutr 2007; 98, 1029 - 1037.
4 Ystrom E, Niegel S, Vollrath ME. The impact of maternal negative affectivity on
dietary patterns of 18-month-old children in the Norwegian Mother and Child
Cohort Study. Matern Child Nutr 2009; 5, 234 - 242.
5 Golding J, the ALSPAC Study Team. Children of the nineties: a resource for
assessing the magnitude of long-term effects of prenatal, perinatal and
subsequent events. Contemp Rev Obstet Gynaecol 1996; 8, 89 - 92.
6 Emmett P. Dietary assessment in the Avon Longitudinal Study of Parents and
Children. Eur J Clin Nutr 2009; 63 (Suppl 1), S38 - S44.
7 Noble S, Emmett P. Differences in weaning practice, food and nutrient intake
between breast- and formula-fed 4-month-old infants in England. J Hum Nutr Diet
2006; 19, 303 - 313.
8 Office of Population Censuses & Surveys. Standard Occupational Classification. Her
Majesty’s Stationery Office: London, 1991.
Supplementary Information accompanies the paper on European Journal of Clinical Nutrition website (http://www.nature.com/ejcn)
European Journal of Clinical Nutrition (2012) 658 -- 666
View publication stats
9 Field A. Chapter 17: Exploratory factor analysis. In: Discovering Statistics Using
SPSS, 3rd edn. SAGE Publications Ltd: London, 2006, pp 627 - 685.
10 Costello AB, Osborne JW. Best practices in exploratory factor analysis: four
recommendations for getting the most from your analysis. Practical Assess Res
Eval 2005; 10. Available online at: http://pareonline.net/getvn.asp?v=10&n=7.
11 Gorusch RL. Factor Analysis. Lawrence Erlbaum: Hillsdale, NJ, 1983, pp 142 - 174.
12 McCann SE, Marshall JR, Brasure JR, Graham S, Freudenheim JL. Analysis of
patterns of food intake in nutritional epidemiology: food classification and in
principal components analysis and the subsequent impact on estimates for
endometrial cancer. Public Health Nutr 2001; 4, 989 - 997.
13 DiStefano C, Zhu M, Mindrila D. Understanding and using factor scores:
considerations for the applied researcher. Practical Assess Res Eval 2009; 14.
Available online at: http://pareonline.net/getvn.asp?v=14&n=20.
14 Tabachnick BG, Fidell LS. Principal components and factor analysis. In: Hartman S
(ed) Using Multivariate Statistics, 5th edn. Pearson Education Inc.: Boston, MA,
2007, pp 607 - 675.
15 Northstone K, Emmett P, ALSPAC Study Team. Multivariate analysis of diet in
children at four and seven years of age and associations with socio-demographic
characteristics. Eur J Clin Nutr 2005; 59, 751 - 760.
16 Briefel RR, Reidy K, Karwe V, Jankowski L, Hendricks K. Toddlers’ transition to
table foods: impact on nutrient intakes and food patterns. J Am Diet Assoc 2004;
104 (1 Suppl 1), s38 - s44.
17 Hu FB, Rimm E, Smith-Warner SA, Feskanich D, Stampfer MJ, Ascherio A et al.
Reproducibility and validity of dietary patterns assessed with a food-frequency
questionnaire. Am J Clin Nutr 1999; 69, 243 - 249.
18 North K, Emmet P, Avon Longitudinal Study of Pregnancy and Childhood
(ALSPAC) Study Team. Multivariate analysis of diet among three-year-old children
and associations with socio-demographic characteristics. Eur J Clin Nutr 2000; 54,
73 - 80.
19 Kramer MS, Fombonne E, Igumnov S, Vanilovich I, Matush L, Mironova E
et al. Effects of prolonged and exclusive breastfeeding on child behavior and
maternal adjustment: evidence from a large, randomized trial. Pediatrics 2008;
121, e435 - e440.
20 Golley RK, Hendrie GA, McNaughton SA. Scores on the Dietary
Guideline Index for children and adolescents are associated with nutrient
intake and socio-economic position but not adiposity. J Nutr 2011; 141,
1340 - 1347.
& 2012 Macmillan Publishers Limited