ORIGINAL ARTICLE

Does maternal concern about childrens weight

affect childrens body size perception at the age
of 6.5?FA community-based study
R Mitchell
1
, M Wake
1,2,3
, L Canterford
1,2
and J Williams
1,2,4
1
Centre for Community Child Health, Royal Childrens Hospital, Melbourne, Australia;
2
Murdoch Childrens Research
Institute, Melbourne, Australia;
3
Department of Pediatrics, University of Melbourne, Melbourne, Australia and
4
Centre
for Adolescent Health, Royal Childrens Hospital, Melbourne, Australia
Objective: Though overweight is often established by school entry, not all mothers of such children report weight concerns.
Enhancing concern might assist lifestyle change, but could lead to child body dissatisfaction. We investigated (i) perceived/
desired body size and body dissatisfaction in mothers and their 6.5-year-old children, and (ii) the impact of earlier maternal
concern about overweight on childrens body mass index (BMI) status and body dissatisfaction.
Design: Prospective community study.
Setting: Melbourne, Australia.
Subjects: 317 motherchild dyads.
Main exposures: Child and maternal BMI (kgm
2
) at 4.0 and 6.5 years; maternal concern about child overweight at 4.0 years.
Outcome measures: Paired perceived and desired body size on 7-point figural rating scales self-reported by mothers and
children, and reported by mothers regarding children; dissatisfaction (desired minus perceived) score.
Results: For all three actual BMI perceived size pairings (mother self-report, mothers report on child and child self-report), BMI
correlated with perceived body size (r 0.82 (mother self-report); r 0.65 (mother reporting on child); r 0.22 (child self-
report); all Po0.001). Similarly, all three dissatisfaction scores were greater with increasing BMI status. Childrens own
dissatisfaction scores correlated with their actual BMI, but were not related to mothers own body dissatisfaction scores or with
mothers dissatisfaction with childrens body size. Maternal concern about overweight at the age of 4 years was not associated
with BMI change, or child body dissatisfaction by the age of 6.5. Most mothers of overweight and obese children (88 and 90%,
respectively) regarded their child as the middle figure (that is, 4) or thinner.
Conclusions: Despite low rates of recognition of child overweight, maternal perceptions of the childs body correlated strongly
with the childs actual BMI. Maternal concerns about child BMI did not appear to impact on child BMI change or child body
dissatisfaction.
International Journal of Obesity (2008) 32, 10011007; doi:10.1038/ijo.2008.12; published online 26 February 2008
Keywords: body mass index; body image; child; mothers; longitudinal studies
Introduction
Consistent with the global experience,
13
overweight and
obesity in Australian children have risen rapidly since 1985
and now affect approximately one-quarter of all primary
school children.
4
Although the epidemic is attracting
substantial health, media and public attention, parent
perceptions about the overweight status of their own children
often seem to be inconsistent with the magnitude of public
health concern. Assuming that addressing the childhood
obesity epidemic requires at least some individual change, it
is important to understand underlying perceptions at the
individual level. However, there is also concern that focusing
on a childs weight could foster the development of body
image problems, mental health issues and disordered eating.
5
The challenge is therefore to understand not only the problem
of lack of recognition for individual children, but also howand
when body image problems come about.
According to behavior change theories, such as the stages
of change model,
6
parents would need to recognize the issue
Received 26 July 2007; revised 5 January 2008; accepted 7 January 2008;
published online 26 February 2008
Correspondence: Dr R Mitchell, Centre for Community Child Health, Royal
Childrens Hospital, Flemington Road, Parkville, Victoria 3052, Australia.
E-mail: [email protected]
International Journal of Obesity (2008) 32, 10011007
& 2008 Nature Publishing Group All rights reserved 0307-0565/08 $30.00
www.nature.com/ijo
of overweight in themselves or their own child to motivate
the behavior change needed to achieve weight reduction.
However, multiple reports (mainly studying mothers) have
shown a striking picture of low rates of recognition; between
1798% of parents do not recognize overweight in their
overweight children.
711
The wide range may be attributed
to study differences on factors such as the obesity cut points
used, child age and the method of questioning. Mothers are
more accurate with older and heavier children,
10
and better
at estimating body size when pictorial or photographic
figures, rather than the words overweight and obese, are
used.
7
Postulated reasons for this low recognition include
the normalization of overweight, denial, a genuine lack of
concern or a belief that childhood puppy fat is healthy;
however, the evidence about the reasons remains scant.
Concluding that the laypersons perception of average
weight and the clinical definition of overweight are now in
conflict, Jeffery reported 40% of overweight mothers and
45% of overweight fathers judged their own weight about
right and that a large proportion were unconcerned about
their weight.
It is possible that parents are more concerned about their
childs future than current weight status. In Campbells
recent study, while only 5% of mothers of overweight and
obese children had concerns about current overweight in
their 4-year-old children, three times as many (16%) were
concerned that their child would later become overweight.
12
Parents are more likely to be concerned about future
overweight if their children are already overweight or they
themselves are overweight.
9
However, it is not yet clear
whether parental concern about either current or future
overweight actually influences weight change over time or
subsequent body perception in young children.
Paradoxically, despite this under-recognition, negative
body image and its associated problems are prevalent.
Several studies indicate that children, from as young as 5,
want body sizes thinner than they have.
1315
Truby and
Paxton, using a sophisticated figural scale developed to
represent actual body mass index (BMI), showed 48% of girls
and 36% of boys of primary school age wanted to have a
smaller body figure.
14
One study found that 5-year-old girls
with higher weight status had lower body self-esteem,
especially when their parents were overtly concerned about
their weight.
16
As children mature, overweight and obesity
become risk factors for disordered eating behavior. The 2002
US population-based Project EAT showed that 76% of
overweight teenage girls and 55% of overweight teenage
boys reported unhealthy weight control behavior, such as
skipping meals, smoking, fast food substitutes and eating
very little food. More extreme behaviors, such as binge
eating, vomiting, laxatives, diuretics and weight control pills
were reported by one in five overweight girls.
5
In this paper, we draw on data from an established
community-based longitudinal cohort to report on (1)
cross-sectional relationships between actual, perceived
and desired body mass, as reported by mothers and their
6.5-year-old children; (2) longitudinal relationships between
earlier maternal concern (at the age of 4 years) about child
overweight and the childs own subsequent BMI gain,
perceived and desired body mass and body satisfaction and
(3) whether these relationships are influenced by factors
such as the mothers own weight status, childs weight status,
socioeconomic status and sex of the child.
Methods
Design and participants
Participants were all 6.5-year-old children in the PEAS
(Parent Education and Support) Kids Growth Study, an
established longitudinal community-based study.
17
Between
June 1998 and December 1999, all community-based well-
child nurses in three local government areas (inner urban,
suburban and semirural) in Melbourne, Australia, sequen-
tially approached parents of first-born newborns, with 493
of the parents approached (70% response) ultimately recruited.
The 402 still-contactable families were invited to participate
in semiannual follow-ups between 4 and 6.5 years, of which
317 provided data for this paper. When children were
between ages 4 and 6.5 years, parents completed written
questionnaires and children were weighed and measured; at
the age 6.5 years, the children themselves completed a
questionnaire by interview and mothers were also weighed
and measured. The study was approved by the Royal
Childrens Hospital Ethics in Human Research Committee,
and parents provided written informed consent at both
recruitment and 4 years.
Measures
Concern about child overweight (age 4 years). Mothers
responded on identical five-point scales to two statements
(I am worried that my child will become overweight and
I am worried that my child is overweight right now), which
were then dichotomized into not concerned (disagree a
lot, disagree a little or no strong feelings) vs concerned
(agree a lot or agree a little). Prevalence of responses by
child sex, BMI and demographic categories have been
reported previously elsewhere.
12
Body perceptions (age 6.5 years). To assess perceptions of
body mass and satisfaction, we used Stunkards
18
and
Collins
13
sex-specific figural rating scales, a series of seven
line-drawings, numbered 1 (thin) to 7 (obese), depicting an
adult (Stunkard) or child (Collins) with increasing levels of
adiposity. Excellent reliability has been reported for the adult
scale.
18
For grades 13 children (mean age 8.0 years),
Collins
13
reported an overall correlation of 0.37 between
the figural rating scale score and BMI, although it was
stronger for older children. Similarly, using the Collins
drawings arranged nonlinearly, Williamson and Delin
14
have
reported accurate identification of body size with Australian
Maternal and child body image
R Mitchell et al
1002
International Journal of Obesity
children as young as five. Regarding the adult female
pictures, mothers were asked their perceived body size
(Which picture looks most like you?) and their desired
body size (Which picture looks most like you want to
look?). Regarding the child sex-specific pictures, mothers
were asked Which picture looks most like your child? and
Which picture looks most like how you want your child to
look? Children were asked Which picture looks most like
you? and Which picture looks most like you want to look?
by a trained interviewer who recorded the figure pointed
to by the child.
Body dissatisfaction scores were calculated for three
paired sets of ratings (mother self-report, mother report on
child, child self-report) by subtracting the numerical rating
of the perceived figure from that desired. Possible dissatisfac-
tion scores were 6 to 6, with a positive score indicating
a desire for a heavier body and a negative score a desire for
a lighter body.
Anthropometry (ages 4 and 6.5 years). Height and weight
were recorded at both time points by trained research
assistants according to the standardized protocol.
19
Height
was measured to the nearest 0.1 cm using an Invicta portable
stadiometer and weight to the nearest 0.1kg on digital scales,
without shoes and wearing light clothing. BMI (kg m
2
) was
calculated at each time point. Child BMI was transformed to
a z-score according to the UK 1990 Growth Reference
20
and
children were categorized as nonoverweight, overweight or
obese using sex- and age-specific International Obesity Task
Force (IOTF) cut points.
21,22
Maternal BMI was calculated
from self-reported height and weight
23
at the 4-year time
point and direct measurement at the 6.5-year time point;
BMI data for pregnant women were excluded. Mothers were
categorized as nonoverweight (o25 kg m
2
), overweight (25
to o30kg m
2
) and obese (X30kg m
2
).
Demographic data. Socio-Economic Indexes for Areas (SEIFA)
disadvantage index scores were assigned according to the
family postcode of residence data at the age of 6.5 years and
were analyzed by population-based quintiles. SEIFA values
are standardized scores derived from the 2001 national
population census data, which can be linked to geographic
areas and summarize the social and economic conditions
(national mean 1000, s.d. 100; higher values represent
greater advantage).
24
Statistical analyses
w
2
tests were used to compare children in the original birth
cohort who were retained at the age of 6.5 years (n317) to
those who were not retained (n176) on child sex and
maternal age and education status at the birth of the child.
Children in the cohort at 4 years (n341) who were retained
at 6.5 years (n317) were further compared to those who
were not retained (n24) on child and maternal BMI status
when the child was 4 years of age.
Box and whisker plots and Pearsons correlation coeffi-
cients were generated to summarize the relationship of
actual BMI with each of perceived and desired body image
for mother self-report, mother report on child and child self-
report. For each of the three sets of dissatisfaction scores,
Cuzicks nonparametric Wilcoxon-like test for trend was
carried out to test for linear trend across the categories of the
relevant BMI status variable.
Three multivariable linear regression models were fitted to
estimate the strength of longitudinal associations between
maternal concern about childrens current and future over-
weight status at the age of 4 years and (a) change in child
BMI z-score from the age of 4 to 6.5 years, (b) maternal body
dissatisfaction score regarding the child at 6.5 years and (c)
child dissatisfaction score at the age of 6.5 years. The
following variables were added to each of the three models
as covariates: maternal BMI status at the age of 4 years, child
sex and SEIFA disadvantage index quintile. Child BMI z-score
at the age of 4 years was added as a covariate for models (b)
and (c), but not for model (a) because of the potential
introduction of bias.
25
Two multivariable linear regression models were fitted to
estimate the strength of the association between child
dissatisfaction score at the age of 6.5 years with (i) maternal
dissatisfaction score regarding herself and (ii) maternal
dissatisfaction score regarding the child. Both crude and
adjusted (for maternal BMI status at the age of 6.5 years,
child sex, SEIFA disadvantage index quintile and child BMI z-
score at the age of 6.5 years) coefficients were calculated.
Linear regression analysis assumptions were checked for all
models and were found to be nonviolated. Ordinal and
continuous covariates were tested for departure from
linearity and are presented as linear effects where the test
result was nonsignificant at the 5% level. All analyses were
carried out using Stata version 9.0 software (Statacorp 2005,
College Station, TX, USA).
Results
Characteristics of participants retained and lost through the
study are shown in Table 1, highlighting that fewer of
the retained than lost mothers were overweight or obese. For
the 317 motherchild dyads contributing to these analyses,
97.5% of the mothers and 99.7% of the children completed
the figural rating scales.
Table 2 shows characteristics of the sample at the age of
4 and 6.5 years; it can be seen that the two most
disadvantaged SEIFA quintiles were relatively underrepre-
sented. Overall, there was little change between 4 and 6.5
years in the percentages of overweight and obese children. At
the age of 6.5 years, 90% of mothers and 82% of children saw
the child as being the middle weight or thinner (p4), even
though 19.5% of children were classified as overweight or
obese.
Maternal and child body image
R Mitchell et al
1003
International Journal of Obesity
Relationship between body mass, body image and body
dissatisfaction
Figure 1 shows the three pairs of perceived and desired body
image ratings in relation to BMI: (a) mother self-report in
relation to maternal BMI, (b) mother report on child
in relation to the childs BMI z-score and (c) child self-report
in relation to the childs BMI z-score. Maternal BMI
correlated strongly with maternal perceived figural rating
(r 0.82, Po0.001) and moderately with maternal desired
figural rating (r 0.51, Po0.001); in other words, heavier
mothers reported perceiving themselves as heavier, but also
desiring heavier bodies, than thinner mothers. Nonetheless,
a substantial proportion (42%) of the overweight and some
(5%) of the obese women still regarded themselves as the
middle figure (that is, 4) or less. Trends were similar but
weaker when mothers reported on their children, with
correlations between the childs BMI z-score and the
perceived and desired figural ratings of 0.65 (Po0.001) and
0.37 (Po0.001), respectively. For the child self-report, the
childs actual BMI z-score bore only weak relationships to
either the childs perceived (r 0.22, Po0.001) or desired
(r 0.10, P0.08) figural rating. Virtually, all mothers of
overweight and obese children (88 and 90%, respectively)
perceived the child as best represented by the middle figure
or thinner, in other words seeming to view them as
considerably thinner than they actually were. Similar to
their mothers, almost all overweight (83%) and obese (83%)
children perceived that they were best represented by the
middle figure or thinner. Overall, the desire to be heavier
than the middle figure was expressed by very few mothers or
children for themselves and by very few mothers for their
children.
Table 3 shows the body dissatisfaction results according to
the BMI status. For all three results (mother self-report,
mothers report on child and child self-report), there was
strong evidence (all Po0.001) of a linear relationship
between increasingly negative dissatisfaction scores (that is,
desire to be thinner) and increasing BMI status, with the
strongest gradient in the mother self-report and smallest in
the child self-report. These child self-report findings for BMI
status contrast with the minimal correlation demonstrated
in the figure between childrens BMI z-score and their own
perceived or desired body image, suggesting that the desire
to be thinner was focused among those at the upper end
Table 1 Comparison of characteristics between those retained at the age of 6.5 years and (i) the total lost to follow up by the age of 6.5 years and (ii) the subset of
those retained at the age of 4.5 years but lost by 6.5 years
Characteristics Retained (n317) Lost to follow up (n176)
Total lost by 6.5
years (n176)
Retained at 4.5 but lost
by 6.5 years (n23)
Males (%) 48.9 55.7 F
Maternal age at birth of child (years) 29.9 29.0 F
Maternal university degree (%) at birth of child 40.4 32.9 F
Child BMI status at 4.5 year time point
Not overweight 80.3 F 82.6
Overweight 16.5 F 17.4
Obese 3.2 F 0.0
Maternal BMI status at 4.5 year time point
Not overweight 64.7 F 52.6
Overweight 23.7 F 36.8
Obese 11.7 F 10.5
Abbreviation: BMI, body mass index.
Table 2 Characteristics of the sample (n 317)
Characteristics 4.0 years 6.5 years
Child
Males (n (%)) 155 (48.9)
Age in years (mean (s.d.)) 4.2 (0.2) 6.6 (0.1)
BMI
a
Median (interquartile range) 16.4 (15.617.2) 16.2 (15.317.2)
z-score (mean (s.d.)) 0.53 (0.90) 0.46 (0.94)
BMI status
a
Not overweight (n (%)) 253 (80.3) 255 (80.4)
Overweight (n (%)) 52 (16.5) 53 (16.7)
Obese (n (%)) 10 (3.2) 9 (2.8)
Mother
Age in years (mean (s.d.)) 34.1 (4.2) 36.5 (4.2)
BMI
b
(median (interquartile range)) 23.6 (21.426.5) 24.7 (22.227.7)
BMI status
b
Not overweight (n (%)) 194 (64.7) 150 (52.3)
Overweight (n (%)) 71 (23.7) 96 (33.5)
Obese (n (%)) 35 (11.7) 41 (14.3)
Family SEIFA quintile
c
(n (%))
First (most disadvantaged) F 10 (3.2)
Second F 41 (12.9)
Third F 86 (27.1)
Fourth F 68 (21.5)
Fifth (least disadvantaged) F 112 (35.3)
Abbreviation: BMI, body mass index. All percentages quoted are valid
percentages.
a
n 315 at 4 years.
b
n 300 at 4 years and 287 at 6.5 years.
c
Based on childs most recent postcode (i.e. at 6.5 years).
Maternal and child body image
R Mitchell et al
1004
International Journal of Obesity
of the BMI spectrum. Neither maternal dissatisfaction with
maternal body size nor maternal dissatisfaction with the
childs body size predicted the childs self-reported dissatis-
faction score. The only predictor of child self-dissatisfaction
was increasing child BMI z-score (b coefficient 0.41, 95% CI
0.57 to 0.26, Po0.001).
Effect of earlier maternal concern on BMI change and body
dissatisfaction
Table 4 shows the results of the three linear regression
analyses investigating associations between early maternal
concern (at 4 years of age) about current and/or future
overweight, change in BMI z-score and body dissatisfaction
outcomes at the age of 6.5 years. Change in child BMI z-score
rose with high maternal BMI status (0.10 increase for each
step from not overweight to overweight to obese) and being
a female child (0.15), whereas fell by 0.07 for each quintile
of increasing social advantage; it was not associated with
either current or future concern about overweight at the age
of 4 years. Mothers who were concerned about the childs
current, but not future, overweight at 4 years were the most
dissatisfied (that is, desired their child to be thinner than
perceived) at the age of 6.5 years, but there was no evidence
for an association between these concerns and children
themselves having large dissatisfaction scores at the age
of 6.5 years. The childs BMI z-score at the age of 4.0 years
was the strongest predictor of both child and mother
dissatisfaction with the childs body at the age of 6.5 years,
but mothers own BMI status was not predictive. Both child
and maternal ratings of dissatisfaction with the childs body
were slightly higher for female children, but this reached
statistical significance only for the maternal ratings
of dissatisfaction with the childs body.
Discussion
Although mothers were poor at recognizing overweight and
obesity in their children at the age of 6.5 years (as at the age
of 4 years in this same cohort
12
), on the whole, their ratings
correctly ranked children along the BMI spectrum on a
figural rating scale. Earlier maternal concern about the child
being overweight was linked to subsequent maternal dis-
satisfaction with the childs body size. However, it appeared
to have neither positive (reduction in BMI gain) nor negative
20
30
40
50
M
a
t
e
r
n
a
l

B
M
I
1 2 3 4 5 6 7
20
30
40
50
M
a
t
e
r
n
a
l

B
M
I
1 2 3 4 5 6 7
-2
0
2
4
C
h
i
l
d

B
M
I

z
-
s
c
o
r
e
1 2 3 4 5 6 7
-2
0
2
4
C
h
i
l
d

B
M
I

z
-
s
c
o
r
e
1 2 3 4 5 6 7
-2
0
2
4
C
h
i
l
d

B
M
I

z
-
s
c
o
r
e
1 2 3 4 5 6 7
-2
0
2
4
C
h
i
l
d

B
M
I

z
-
s
c
o
r
e
1 2 3 4 5 6 7
r = 0.82 r = 0.51
r = 0.65 r = 0.37
r = 0.22 r = -0.10
d e r i s e D ) i i ( d e v i e c r e P ) i (
Mother
self-report
Child self-
report
Mother
report on
child
Figure 1 Body image and body mass index (BMI). Box-and-whisker plots depicting relationships between measured BMI and scores on (i) perceived and (ii)
desired body size. Graphs depict (a) mother self-report in relation to maternal BMI, (b) mother report on child in relation to the childs BMI z-score and (c) child self-
report in relation to the childs BMI z-score.
Table 3 Body dissatisfaction (mean (s.d.) by BMI status
Groups Body dissatisfaction score (mean (s.d.)) z
a
P-value
a
Not overweight Overweight Obese
Mothers self
b
0.54 (0.64) 1.20 (0.60) 1.95 (0.78) 9.99 o0.001
MotherChild
c
0.26 (0.53) 0.31 (0.47) 0.63 (0.52) 7.62 o0.001
Childs self
d
0.35 (1.23) 0.79 (1.57) 1.56 (1.33) 3.35 0.001
Abbreviation: BMI, body mass index.
a
Cuzicks nonparametric Wilcoxon-like
test for trend across categories.
b
Not overweight: n 147; overweight: n93;
Obese: n 40.
c
Not overweight: n250; overweight: n 51; Obese: n 8.
d
Not overweight: n 254; overweight: n53; Obese: n9.
Maternal and child body image
R Mitchell et al
1005
International Journal of Obesity
(a greater desire by the child to be thinner) impacts,
suggesting that mothers may not act on or impart their
concerns to children of this age. It is also possible that this
outcome may reflect a lack of knowledge about how to assist
their children or a fear that they may damage their childs
self-esteem by addressing the problem of overweight. The
only predictor of childrens self-reported body dissatisfaction
was their own BMI.
Strengths of this study include the use of a longitudinal
community cohort of generally healthy children with high
interwave cohort retention. Rates of maternal overweight/
obesity and child overweight reflected current Australian
population rates using the conservative IOTF cut points,
although obese children were underrepresented (3%, rather
than the expected 56% at this age). This study is unusual in
that we were able to compare and analyze body image data
from mothers and young children simultaneously, and
to explore longitudinal developments and influences on
weight and body image.
One limitation is the mixed use of maternal self-report and
measured BMI; the known tendency to underreport weight
and overreport height may have inflated the estimate of the
increase in maternal BMI demonstrated in Table 2. However,
this would not have affected the cross-sectional analyses
at the age of 6 years (Figure 1 and Table 3) for which
measured BMI was used. For the longitudinal analyses, this
may have slightly attenuated the small effect for maternal
BMI status noted in Table 4, but we do not believe would
have altered the other (negative) longitudinal findings. Our
findings are also limited by the age of the children in our
study, as young children are less accurate than older children
on linear figural rating scales.
13
Clearly, the relationship
between body dissatisfaction in children and their mothers
needs to be studied in older age groups. If different results
were found, this could imply either greater understanding of
the task at hand or changing predictors of body image with
age. As children mature, their parents may be increasingly
concerned about,
11
and focused on, overweight, and this
may lead to children developing views that reflect their
parents beliefs about their body size and influence body
image.
Consistent with the current literature,
711
mothers tended
to regard their heavier children as having smaller body sizes
than the measured BMI would objectively suggest, even
though there were strong correlations between BMI and
maternal body image of self and child. Overall, larger women
and mothers of larger children desired and perceived
themselves or their children to be heavier compared with
smaller women or mothers of smaller children. The finding
that some heavier mothers or mothers of heavier children
actually desired for themselves or their children to be
comparatively heavier was concerning. Desire for a heavier
body size in this situation may reflect a misguided judgment
about what body mass is healthiest, or perhaps an accep-
tance of overweight as the normal state.
Body dissatisfaction was found in all three reporting
scenariosFmother self-report, the mothers report on the
child and child self-report. Interestingly, although children
were poor at recognizing body size, they were increasingly
dissatisfied with their bodies with increasing body mass. The
lack of statistical evidence that child dissatisfaction was
linked to maternal self-dissatisfaction, maternal dissatisfac-
tion with the childs body size or earlier maternal concern
about the childs weight status contrasts with Krahnstoever
Davisons conclusion that parental concern about weight
status is associated with negative self-evaluation in 5-year-
old girls.
16
However, in KrahnstoeverDavisons study, low
self-esteem was only found in girls whose fathers expressed
concern, and concern was measured through a series of
questions that included comments on concern about the girl
eating too much or needing to diet, in addition to concern
about overweight. Relationships between fathers views and
how they relate to their young childrens self-perceptions
require further research, as do community and societal
perceptions of overweight in young children.
Links between obesity, concern about overweight and
body image could provide information about factors that
may or may not motivate change when weight threatens
Table 4 Change in BMI z-score and dissatisfaction outcomes at the age of 6.5 years: results of the three linear regression models
Predictors at the age of 4 years D Child BMI z-score (n298) Dissatisfaction score for childs body at the age of 6.5 years
Rated by child (n297) Rated by mother (n293)
b-coefficient (95% CI) P-value b-coefficient (95% CI) P-value b-coefficient (95% CI) P-value
Concern about child overweight
a
Current 0.00 (0.31, 0.31) 0.998 0.03 (0.85, 0.78) 0.94 0.34 (0.64, 0.04) 0.03
Future 0.05 (0.23, 0.13) 0.61 0.02 (0.47, 0.50) 0.95 0.07 (0.24, 0.11) 0.45
Increasing maternal BMI status
b
0.10 (0.01, 0.18) 0.03 0.07 (0.29, 0.15) 0.52 0.01 (0.09, 0.07) 0.89
Child BMI z-score F F 0.34 (0.53, 0.16) o0.001 0.29 (0.36, 0.23) o0.001
Female sex
c
0.15 (0.04, 0.27) 0.01 0.20 (0.51, 0.11) 0.20 0.28 (0.39, 0.16) o0.001
Increasing SEIFA quintile
d
0.07 (0.12, 0.02) 0.005 0.07 (0.06, 0.20) 0.26 0.03 (0.02, 0.07) 0.25
Abbreviations: BMI, body mass index, CI, confidence interval. Each predictor listed is adjusted for all others in each column, i.e. for each of the three separate models.
However, child BMI z-score is not included as a predictor for the model in which D child BMI z-score is the outcome.
a
Reference category is not concerned.
24
b
Reference category is not overweight.
c
Reference category is male sex.
d
Reference category is first quintile (most disadvantaged).
Maternal and child body image
R Mitchell et al
1006
International Journal of Obesity
health, and perhaps about the likelihood of creating mental
health problems in a generation of increasingly heavy
children. This is a complex issue involving multiple factors.
Effectively moving overweight and obese children toward
a healthier body size should ideally occur without increasing
their body dissatisfaction or damaging their mental health.
In suggesting that raising rates of maternal recognition of
their young childrens overweight and obesity would not
necessarily damage their childrens body image, this study
offers some cause for optimism that this can be achieved
in this age group. However, whether increasing such
recognition would also progress readiness to change, and
whether such progression would in fact alter BMI outcomes,
remains to be tested. It will be important to ensure that the
way in which this concern is expressed to children is
carefully considered, and that parents with this concern
have adequate community resources to assist them. Finally,
as concern (even when appropriately held) was not linked to
altered BMI trajectories, it is imperative that ongoing
research focuses on developing ways of effectively translat-
ing maternal concern into improved BMI outcomes.
Acknowledgements
This paper draws on data from a larger longitudinal study,
the PEAS Kids Growth Study. RM worked on and MW
supervised all phases of the work leading to this paper; MW
founded the PEAS Study, and has been a Chief Investigator in
all phases; LC conducted all statistical analyses and prepared
the tables; JW was the Principal Investigator for the PEAS
Kids Growth Study (4.0- to 6.5-year-old waves) and advised
on the paper. RM, MW and LC co-wrote and JW edited the
paper. We thank all the parents and children who took part
in the study, and the research assistants who assisted with
data collection. We also acknowledge the major role of Dr
Michele Campbell (PhD student) in the 4.0- to -6.5-year-old
wave of the PEAS Kids Growth Study.
Financial Disclosures:
The PEAS Kids Growth Study 4.0- to -6.5-year-old waves were
funded by the Australian National Health and Medical
Research Council (NHMRC Project Grant 284509) and the
Murdoch Childrens Research Institute. Dr Wakes salary is
part-funded by NHMRC Population Health Career Develop-
ment Award no. 284556. There are no competing interests.
References
1 Rudolf MC, Sahota P, Barth JH, Walker J. Increasing prevalence of
obesity in primary school children: cohort study. BMJ 2001; 322:
10941095.
2 Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal
KM. Prevalence of overweight and obesity in the United States,
19992004. JAMA 2006; 292: 15491555.
3 Popkin BM. Global nutrition dynamics: the world is shifting
rapidly toward a diet linked with noncommunicable diseases. Am
J Clin Nutr 2006; 84: 289298.
4 Booth ML, Chey T, Wake M, Norton K, Hesketh K, Dollman J.
Change in the prevalence of overweight and obesity among
young Australians, 19691997. Am J Clin Nutr 2003; 77: 2936.
5 Neumark-Szatainer D. Can we simultaneously work toward the
prevention of obesity and eating disorders in children and
adolescents? Int J Eat Disord 2005; 38: 220227.
6 Taylor RJ, Smith BH, van Teijlingen ER. Theories of Health
Promotion. Health and Illness in the Community. Oxford University
Press: New York, 2003; 123126.
7 Jeffery AN, Voss LD, Metcalf BS, Alba S, Wilkin TJ. Parents
awareness of overweight in themselves and their children: cross
sectional study within a cohort (EarlyBird 21). BMJ 2006; 330:
2324.
8 Eckstein KC, Mikhail LM, Ariza AJ, Thomson JS, Millard SC, Binns
HJ. Parents perceptions of their childs weight and health.
Pediatrics 2006; 117: 681690.
9 Baughum AE, Chamberlin LA, Deeks CM, Powers SW, Whitaker
RC. Maternal perceptions of overweight in preschool children.
Pediatrics 2000; 106: 13801386.
10 Carnell S, Edwards C, Crocker H, Boniface D, Wardle J. Parental
perceptions of overweight in 35 year olds. Int J Obes 2005; 29:
353355.
11 Maynard LM, Galuska DA, Blanck HM, Serdula MK. Maternal
perceptions of weight status of children. Pediatrics 2003; 111:
12261237.
12 Campbell MW, Williams J, Hampton A, Wake M. Maternal
concern and perceptions of overweight in Australian pre-
school-aged children. Med J Aust 2006; 184: 274277.
13 Collins ME. Body figure perceptions and preferences among
preadolescent children. Int J Eat Disord 1991; 10: 199208.
14 Williamson S, Delin C. Young childrens figural selections:
accuracy of reporting and body size dissatisfaction. Int J Eat
Disord 2001; 29: 8084.
15 Truby H, Paxton SJ. Development of the childrens body image
scale. Br J Clin Psychol 2002; 41: 185203.
16 Krahnstoever-Davison KK, Lipps Birch LL. Weight status, parent
reaction and self-concept in five-year-old girls. Pediatrics 2001;
107: 4653.
17 The Parent Education and Support (PEAS) Program: final report.
Centre for community child health, royal childrens hospital.
Available at http://hnb.dhs.vic.gov.au/commcare/ccdnav.nsf/fid/
Accessed 2003.
18 Stunkard AJ, Sorensen T, Schulsinger F. Use of the Danish
adoption register for the study of obesity and thinness. In: Kety
SS, Rowland LP, Sidman RL, Matthysse SW (eds). Genetics
of Neurological and Psychiatric Disorders. Raven Press: New York,
1983, pp 115120.
19 National Health Data Committee 2003. National Health Data
Dictionary Version 12. Australian Institute of Health and Welfare:
Canberra, 2003. (AIHW Cat. No HWI 43).
20 Cole TJ. Growth monitoring with the British 1990 growth
reference. Arch Dis Child 1997; 76: 4749.
21 Vidmar S, Carlin J, Hesketh K, Cole T. Standardizing anthropo-
metric measures in children and adolescents with new functions
for egen [computer program]. Stata J 2004; 4: 5055. Available at:
http://www.stata-journal.com/abstracts/dm0004.pdf.
22 Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing
a standard definition for child overweight and obesity worldwide:
international survey. BMJ 2000; 320: 12401243.
23 Rowland ML. Self reported weight and height. Am J Clin Nutr
1990; 52: 11251133.
24 Australian Bureau of Statistics, Commonwealth of Australia.
Socio-Economic Indexes for Areas (SEIFA) Volume 1.1. Australian
Bureau of Statistics: Canberra, Australia, 2001.
25 Lord FM. A paradox in the interpretation of group comparisons.
Psychol Bull 1967; 68: 304305.
Maternal and child body image
R Mitchell et al
1007
International Journal of Obesity