appetite 60 - qut eprintseprints.qut.edu.au/54016/4/54016.pdf · 2 abstract the purpose of this...
TRANSCRIPT
This is the author’s version of a work that was submitted/accepted for pub-lication in the following source:
McMeekin, Sascha, Jansen, Elena, Mallan, Kimberley M., Nicholson, Jan,Magarey, Anthea, & Daniels, Lynne(2013)Associations between infant temperament and early feeding practices. Across-sectional study of Australian mother-infant dyads from the NOURISHrandomised controlled trial.Appetite, 60(1), pp. 239-245.
This file was downloaded from: https://eprints.qut.edu.au/54016/
c© Copyright 2012 Elsevier Ltd.
This is the author’s version of a work that was accepted for publication in Appetite.Changes resulting from the publishing process, such as peer review, editing, corrections,structural formatting, and other quality control mechanisms may not be reflected in thisdocument. Changes may have been made to this work since it was submitted for publica-tion. A definitive version was subsequently published in Appetite, Volume 60, 1 January2013. DOI: 10.1016/j.appet.2012.10.005
License: Creative Commons: Attribution-Noncommercial-No DerivativeWorks 4.0
Notice: Changes introduced as a result of publishing processes such ascopy-editing and formatting may not be reflected in this document. For adefinitive version of this work, please refer to the published source:
https://doi.org/10.1016/j.appet.2012.10.005
1
Associations between infant temperament and early feeding practices: a cross-sectional
study of Australian mother-infant dyads from the NOURISH randomised controlled
trial
Authors: Sascha McMeekin a, Elena Jansen a, Kimberley Mallan a, Jan Nicholson b,c,
Anthea Magarey a,d, Lynne Daniels a,d.
a Institute of Health and Biomedical Innovation, School of Exercise and Nutrition
Sciences, Queensland University of Technology, 60 Musk Avenue, Kelvin Grove,
Queensland 4059, Australia.
b Parenting Research Centre, 232 Victoria Parade, East Melbourne, Victoria 3002,
Australia.
c Centre for Learning Innovation, Queensland University of Technology, Brisbane 4059,
Australia.
d Nutrition and Dietetics, Flinders University, Adelaide, South Australia 5001,
Australia.
2
Abstract
The purpose of this study was to investigate the association between temperament in
Australian infants aged 2–7 months and feeding practices of their first-time mothers
(n=698). Associations between feeding practices and beliefs (Infant Feeding
Questionnaire) and infant temperament (easy-difficult continuous scale from the Short
Temperament Scale for Infants) were tested using linear and binary logistic regression
models adjusted for a comprehensive range of covariates. Mothers of infants with a
more difficult temperament reported a lower awareness of infant cues, were more likely
to use food to calm and reported high concern about overweight and underweight. The
covariate maternal depression score largely mirrored these associations. Infant
temperament may be an important variable to consider in future research on the
prevention of childhood obesity. In practice, mothers of temperamentally difficult
infants may need targeted feeding advice to minimise the adoption of undesirable
feeding practices.
Key words: feeding practices, feeding beliefs, infant temperament, obesity, feeding
relationship.
3
Introduction
In 2010, it was estimated that nearly 12 % of children from birth to five years in
developed countries were overweight or obese (de Onis, Blössner, & Borghi, 2010).
These children are at a greater risk of becoming overweight adults (Margarey, Daniels,
Boulton, & Cockington, 2003; Power & Parsons, 2000), with associated increased
morbidity and mortality risk (Reilly et al., 2003). Direct costs of obesity are estimated
to be around 9 % of the total health care costs in the USA, between 1 % and 5 % in
Europe and 2 % in Australia (Minister for Health and Ageing, 2007).
Early maternal feeding practices and beliefs, hereafter referred to as ‘feeding practices’,
have been associated with child weight status (Ventura & Birch, 2008). The feeding
relationship is implicitly bi-directional in that both mother and infant contribute to the
exchange (DiSantis, Hodges, Johnson, & Fisher, 2011). Infants are required to
contribute clear hunger and satiety cues, and the caregiver must accurately interpret and
respond to those cues in order for responsive feeding to occur. If feeding is discordant
to the child’s hunger and satiety the child’s innate self-regulatory capacity may be
hindered (Birch & Deysher, 1986; Kral et al., 2007), and over-feeding, accelerated
weight gain and increased risk of future overweight may result (DiSantis, et al., 2011).
The infant’s role in the feeding relationship, that is clearly expressing hunger and
satiety, may be influenced by their temperament.
Infant temperament is a set of traits that characterises individuals’ behaviour (Goldsmith
et al., 1987; Maccoby & Martin, 1983). These traits have a genetic component that is
4
influenced by a complex interplay with environmental factors (Li Fein, & Grummer-
Strawn, 2008; Worobey, Islas Lopez & Hoffman, 2009). Temperament has been linked
with weight gain during infancy (Carey, 1985; Darlington & Wright, 2006; Niegel,
Ystrom, & Vollrath, 2007) and later weight status (Agras, Hammer, McNicholas, &
Kraemer, 2004; Pulkki-Raback, Elovainio, Kivimaki, Raitakari, & Keltikangas-
Jarvinen, 2005). The mechanisms by which temperament influences weight are yet to
be clearly elucidated. One possible pathway is via the influence infant temperament
may have on the development of feeding practices adopted by mothers, however very
few studies have explored this relationship.
Evidence supports the notion that negative, or ‘difficult’, temperament dimensions are
associated with both undesirable feeding practices, and feeding difficulties during
infancy. Of note, all studies reviewed here used maternal-report measures of infant
temperament and feeding practices. In a cross-sectional study of six-month-old infants
(n=99) from the United Kingdom (Farrow & Blissett, 2006a), unadaptable and fussy-
difficult infant temperament were associated with maternal reports of feeding
difficulties. A longitudinal follow-up of the above study (n = 62) concluded that infant
difficultness reported by mothers at six months predicted restrictive feeding practices at
two years of age (Blissett & Farrow, 2007). In a US study of older children (n = 55, six
to 12 years) a relationship between infant temperament and maternal feeding practices
was also found, via the association found between the ‘difficult’ temperament
dimension low adaptability, (Horn, Galloway, Webb, & Gagnon, 2011) and the feeding
practice using pressure to eat. These findings, though limited, do appear to support that
5
infant temperament may be a contributing factor in the development of maternal feeding
practices.
A number of other infant and maternal characteristics may also play a role in the
feeding relationship. Infant gender has been shown to modulate associations between
maternal controlling feeding practices and children’s food intake, which was
significantly associated in female children and not male children (Fisher & Birch,
1999). Child’s weight is another important covariate: variations in maternal feeding
practices (in particular controlling feeding and pressure to eat) have been associated
with child weight and weight gain during infancy (Farrow & Blissett, 2006b; Webber,
Hill, Saxton, Van Jaarsveld, & Wardle, 2009). Mothers’ own weight status has been
implicated in the initiation and duration of breastfeeding, and may therefore influence
feeding practices as well (Donath & Amir, 2000; Li, Jewell, & Grummer-Strawn, 2003).
In addition, mothers with a university education, who breastfeed and are older have
been shown to use more ‘positive child feeding behaviours’ (Hendricks, Briefel, Novak,
& Ziegler, 2006). Maternal mood may also play a role in the feeding practices reported
by mothers, as it has been shown that mothers with depression reported less responsive
feeding practices (Hurley, Black, Papas, & Caufield, 2008). It is therefore appropriate
to consider (i.e., “statistically adjust for”) these infant and maternal characteristics as
covariates in the study of feeding practices.
Although there is emerging evidence to support the proposal that infant temperament
could be associated with maternal feeding practices, the limited number of studies
investigating this relationship typically have varied methods to define both temperament
6
and feeding practices, small samples, and control for few if any covariates. The aim of
this study was to investigate whether infant temperament is associated with maternal
feeding practices and beliefs in a large sample of Australian first-time mothers of
infants aged 2-7 months, adjusting for a comprehensive range of potential covariates.
Methods
Participants
This cross-sectional study reports a secondary analysis of baseline data from the
NOURISH randomised controlled trial (Australian and New Zealand Clinical Trials
Registry Number [ACTRN] 12608000056392), a primary prevention intervention for
childhood obesity, commencing when infants were aged 2–7 months. The trial protocol
and recruitment have been described in detail elsewhere (Daniels et al., 2009). A total
of 698 first-time mothers and their infants, from Brisbane and Adelaide in Australia,
consented to participate in the NOURISH trial. First-time mothers were included in the
NOURISH study as it was expected they would have less experience in infant feeding
and as a result the intervention would develop rather than modify their feeding
practices. Eligible participants for the study were ≥ 18 years old who delivered a
healthy term infant (birth weight > 2500 g, and ≥ 37 weeks gestation) and were willing
and able to attend intervention sessions. Mothers were excluded if they did not have
facility with English or had a documented history of domestic violence, intravenous
substance abuse, mental health problems or self-reported eating disorders.
7
Procedures
A two-stage recruitment process was conducted. Stage 1 involved approaching a
consecutive sample of eligible (as above) first-time mothers in hospital within 72 hours
of delivery to obtain consent to be contacted again. Participants agreeing to later
contact at Stage 1 provided basic demographic data, as did approximately half those
who declined to be re-contacted. Three to four months later, at Stage 2, mothers were
re-contacted by mail to request consent for study participation. Prior to allocation when
infants were 2–7 months old, consenting mothers were mailed a questionnaire which
was returned during an appointment to gather basic anthropometric data at a local child
health clinic. Ethical approval was obtained from relevant university and hospital
institutions (Queensland University of Technology HREC 00171 Protocol 0700000752)
(Daniels, et al., 2009).
Measures
Infant temperament. Infant temperament was assessed using a maternal self-report
measure entitled the “Short Temperament Scale for Infants” (STSI) (Sanson, Prior,
Garino, Oberklaid, & Sewell, 1987). The STSI is based on the temperament research
conducted in the New York Longitudinal Study (Thomas, Chess, Birch, Hertzig, &
Korn, 1963) and measurement tools developed by Carey and colleagues (1978). The
STSI has been validated for use in Australian samples of infants aged 4–8 months (n =
2443) (Sanson, et al., 1987). The questionnaire has 12 items assessing three factors; (i)
approach, the infant's level of comfort with new situations or people (4 items, e.g. “My
baby’s first reaction (at home) to approach from strangers is acceptance”); (ii)
cooperation, the adaptability of the infant to new activities (4 items, e.g. “My baby
8
stays still during procedures like hair brushing and nail cutting”); and (iii) irritability,
the infant's agitation and difficulty to soothe (4 items, e.g. “My baby continues to cry in
spite of several minutes of soothing”). All items were measured on a 6–point scale
from “almost never” (1) to “almost always” (6) and high scores related to greater
approach, cooperation and irritability. The 12-item STSI was administered to mothers
when their infants were aged 2–7 months. An overall continuous easy-difficult
temperament scale ranging from 1–6 was calculated from the three factors by first
recoding individual items such that high scores were always indicative of
temperamental difficulty (i.e. low cooperation, low approach and high irritability), and
generating a mean score across all items (Sanson, et al., 1987). The internal reliability
estimate for the easy-difficult temperament scale (higher scores indicate more difficult
temperament) was acceptable: Cronbach’s α = 0.74.
Feeding practices and beliefs. The “Infant Feeding Questionnaire” (IFQ) (Baughcum et
al., 2001) was self-administered by consenting mothers to assess current self-reported
maternal feeding practices. Original validation of the 20–item questionnaire as a
retrospective measure of feeding practices during the first year of life revealed a seven
factor solution (Baughcum, et al., 2001). The questionnaire required alterations to make
it a suitable measure of feeding practices in an Australian sample with anticipated high
rates of breastfeeding. The tense was changed from past to present, Australian wording
adapted (e.g. “being unsettled” instead of “fussiness”), and a “not applicable” response
category was added for three items that assumed formula feeding was occurring (e.g.
adding cereal to the bottle). These three items, and hence two of the original seven
scales, were subsequently excluded from the questionnaire as the majority of the sample
9
responded with “not applicable”. The remaining five factors and the reliability
estimates in our sample were: awareness of infant hunger and satiety cues (4 items, e.g.
“I know when my baby is hungry”, Cronbach’s α = 0.72), using food to calm infant’s
fussiness (2 items, e.g. “When your baby gets upset, is feeding him/her to first thing you
do?”, Cronbach’s α = 0.66), concern about infant over-eating and becoming overweight
(3 items, e.g. “Do you worry that your baby is feeding too much?”, Cronbach’s α =
0.66), concern about infant under-eating or becoming underweight (4 items, e.g. “I am
worried that my baby will become underweight”, Cronbach’s α = 0.76), and feeding on
demand vs. schedule (2 items, e.g. “Do you let your baby feed whenever s/he wants
to?”, Cronbach’s α = 0.83) 1. Items were measured on a 5-point Likert scale from
“never” (0) to “always” (4). Internal consistency on all five scales in our sample was
higher than reported in the original validation study (Baughcum, et al., 2001).
Covariates. Data collected at first contact with participants in hospital included infant
gender, infant birthweight (from hospital records), maternal age, and maternal education
level (dichotomised into university education vs. no university education). Scores on
the Edinburgh Post Natal Depression Scale (EPDS) were calculated as the sum of
responses to the 10 items (Cox, Holden, & Sagovsky, 1987). Each item was scored
from 0–3 giving a maximum total score of 30. For descriptive purposes only, mothers
scoring between 10 and 13 were classified as “borderline”, and those scoring above 13
were labelled “depressive” based on approximate thresholds advised in the validation
study (Cox, et al., 1987). Breastfeeding status was categorised into three levels: 1)
exclusive breastfeeding, breastmilk only, 2) any breastfeeding, including combination
feeding with formula and fully breastfeeding with or without some complementary
10
foods, 3) no breastfeeding, formula feeding (no breastmilk) with or without some
complementary foods. These data were dummy coded into two variables for the
analyses. Trained study staff collected anthropometric measures of both mother and
child using standard protocols to the nearest 0.01 kg and 0.01 m for weights and heights
respectively. Three infant nude weight and recumbent length measures were averaged.
Infant weight gain (g/wk) was calculated by subtracting the infant’s birthweight from
their weight at 2–7 months and dividing by the infant’s exact age in weeks. Z-scores for
weight for age and birthweight were calculated using the software program WHO
Anthro version 3.0.1 and macros (World Health Organisation, 2006). Raw change in
weight for age Z-score from birth to baseline was calculated, for descriptive purposes,
by subtracting birthweight Z-score from weight-for-age Z-score at 2–7 months.
Maternal weight and height were measured once without shoes. Maternal body mass
index (BMI) was calculated from these height and weight measures (kg/m2) (World
Health Organisation, 1995). Maternal perception of infant weight status, measured with
the item “Do you think your baby is?”, was considered for inclusion as a covariate;
however, as the response category “normal weight” was overrepresented and just small
percentages of mothers classed their infants as “underweight” or “overweight” this
variable was only included to describe the sample (see Table 1).
Analytic strategy
Statistical analysis was conducted using SPSS 18.0.3. Tests for normality were first
conducted for the easy-difficult temperament scale and the potential covariates,
revealing that just the EPDS score had a non-normal distribution. The square root
transformation of the EPDS data was therefore used in all further analyses to improve
11
the sensitivity of the measure. Using food to calm and awareness of cues were found to
be normally distributed, whereas concern about over- and underweight were positively
skewed and feeding on demand was negatively skewed. Non-parametric bivariate
analyses were therefore used to explore the association between potential covariates
with each of the feeding practices. Spearman’s correlations were used for continuous
variables, and Independent samples Mann-Whitney U tests for categorical variables
(Field, 2009). Statistical significance was set at P < 0.05. The covariates found to have
a significant association with at least one of the feeding practices (outcome variables)
were included in the regression analyses that followed.
Square root, log and reciprocal transformations were performed on the three non-
normally distributed IFQ factors, which failed to improve the distribution in any
instance. These three IFQ factors were then split into roughly equal tertiles, from which
the ‘outlier’ tertile (upper tertile for concern about over- and underweight, lower tertile
for feeding on demand vs. schedule) was separated from the other two tertiles to create
binary variables that were entered as the outcome variable into hierarchical binary
logistic regression models. This method is in accordance with that used by Burdette et
al. (2006). Using food to calm and awareness of cues were left as continuous scores
and entered into hierarchical linear regression models. Step 1 of each of the 2-step
linear and binary logistic regression models included infant gender, maternal age at
delivery, maternal education, maternal depression score (EPDS), breastfeeding status,
infant weight gain (g/wk) and maternal BMI. The easy-difficult temperament scale
(independent variable) was incorporated into each model in Step 2. Two multivariate
outliers were detected based on Mahalanobis distance values > 26.12, P < 0.001.
12
Exclusion of these two participants did not affect the results and therefore the full
dataset was used in the analyses. Calculation of Cross’s distance values showed there
were no influential data points in the sample. Post-hoc univariate analysis of variance
were run to test the moderating effects of infant gender and feeding mode combined
with infant temperament on maternal feeding practices with the covariates included.
Results
The characteristics of the sample are shown in Table 1. Forty-four percent of the
sample consented to study participation at Stage 2 of recruitment, excluding those who
could not be contacted or became ineligible. Differences between consenters and non-
consenters at Stage 2 were established from the data collected at Stage 1. Compared
with mothers who either declined consent or could not be recontacted (at Stage 2),
mothers who consented to participate were older (M = 30.1 ± 5.3 vs. M = 27.4 ± 5.6; P
< 0.001), more likely to have completed a university degree (58 % vs. 33 %; OR = 2. 9;
CI95% = 2.4 to 3.5; P < 0.001), and were more likely to report that they intended to
breastfeed their baby exclusively (88 % vs. 75 %; OR = 1.8, CI95% = 1.3 to 2.5; P <
0.001).
The hierarchical linear and binary logistic regression analyses, shown in Table 2 and 3,
identified that the easy-difficult temperament scale added significantly to the variance in
four of the five feeding practices with the full models explaining 12-19 % of the total
variance in these feeding practices (Adjusted R2 for linear regression, Nagelkerke R2 for
binary logistic regression). Inclusion of the easy-difficult temperament variable elicited
13
a change in variance of in the awareness of cues model (R2 = 0.09, P < 0.001),
indicating mothers of infants with a more difficult temperament reported less awareness
of infant cues. Mothers of infants with a more difficult temperament were also more
likely to use food to calm (R2 = 0.02, P = 0.001). High concern regarding both
overweight (OR = 1.75, CI95% = 1.23 to 2.49, P = 0.002) and underweight (OR = 2.00,
CI95% = 1.41 to 2.84, P < 0.001) was more likely in mothers of infants with a more
difficult temperament. Feeding on demand vs. schedule was not associated with
temperament. The covariate maternal depression was independently associated with all
feeding practices. No significant interaction effects were found with infant gender (P =
0.430–0.815) and breastfeeding status (P = 0.197–0.615) when coupled with infant
temperament.
Discussion
To our knowledge the current study is one of the largest to investigate the association
between infant temperament and maternal feeding practices and beliefs. The results
showed that infant temperament was associated with maternal feeding practices and as
such is likely to play a role in the early feeding relationship. Mothers of infants scoring
high on “difficult” temperament reported less awareness of infant hunger cues, more use
of food to calm, and high concern about over- and underweight.
The association of ‘difficult’ temperament (low approach, low cooperation, high
irritability) (Sanson, et al., 1987) and lower maternal awareness of infant hunger and
satiety cues is potentially a manifestation of both maternal and infant characteristics and
14
behaviours. By definition, infants with a difficult temperament react negatively to new
situations, unfamiliar people and novel activities. Mothers of temperamentally difficult
infants may therefore have trouble distinguishing cues that specifically signal hunger
and satiety from the distress cues given in response to the broad range of stimuli
mentioned above. As a consequence of the lack of clarity around an infant's signals of
hunger and satiety and mother's recognition and interpretation of these cues, discordant
feeding is more likely in infants with a difficult temperament and may contribute to the
erosion of the child’s innate ability to regulate intake (DiSantis, et al., 2011).
Alternatively, maternal characteristics may influence the way the mother perceives her
child’s behaviour as well as her capacity to appropriately interpret and respond to cues.
For example, a mother with depression may both perceive her infant as more difficult
and be less responsive to infant cues resulting in nonresponsive feeding styles (Ganley,
1989). Our finding that maternal depression independently influences awareness of
cues lends support to this notion.
Higher temperamental difficulty was associated with greater use of food to calm, a
feeding practice that, at least in older children, has been postulated to compromise self
regulation of intake by providing incidental learning for the child to eat for reasons
other than hunger (DiSantis, et al., 2011). Two studies report links between
temperament and specific feeding practices that the authors argue are a practical
manifestation of using food to calm. The first study, conducted in the USA with “low-
income black mothers” and their three-month-old infants, examined associations
between infant temperament and complementary feeding before four months of age.
This study found that early introduction of solids was associated with high “distress-to-
15
limitations” in infants (e.g. “When placed on his/her back, how often did the infant fuss
or protest?”), supporting the authors’ hypothesis that “a fussy infant temperament may
lead parents to use food as a soothing technique” (Wasser, et al., 2011). In a study of
older infants mothers of children with “distress-prone temperaments” (n = 40266, 18-
month-olds), both externalising (e.g. “the child gets upset easily”) and internalising (e.g.
“too fearful and anxious”), were more likely to feed sweet foods and drinks. The
authors of this study proposed a similar theory to Wasser and colleagues (2011) to
explain their findings. They suggested that mothers respond to “difficult” temperament
traits by attempting to calm and soothe with sweets foods or drinks, as sweets foods
“can have a calming effect as a result of...neurological mechanisms” (Vollrath, Tonstad,
Rothbart, & Hampson, 2010). Our results provide empirical support for the theories
developed to explain these feeding responses to difficult infant temperament.
Overall, the majority of mothers in the NOURISH sample perceived their infant as
having a normal weight status, consistent with the mean raw weight-for-age Z-score
change from birth to baseline of –0.43, which represents healthy weight gain as infants
are closely tracking growth percentiles (Ong, et al., 2006). Also consistent with these
data is the relatively low concern regarding intake and weight status, on the two relevant
IFQ scales (Baughcum, et al., 2001), as indicated by means of 1 from a 5-point Likert
scale where 5 indicates high concern. However, follow up data (not shown) indicates
that when infants reached 14 months of age the mean factor scores for the concern about
over- and underweight had risen to 1.5 and 2.0 respectively. Perceptions and concern
are likely to be important drivers of feeding practices (Farrow & Blissett, 2006b; Teng,
Bartle, Sadeh, & Mindell, 2012). For example, mothers may be more likely to
16
encourage their child to feed more frequently, finish the bottle or eat more (i.e. pressure
their child to eat) if they think their child is, or might become, underweight. Such
discordant feeding practices are linked theoretically with reduced self regulation of
intake and increased weight status in older children (Birch, 2006; DiSantis, et al., 2011).
In our study high concern about the child's weight (both over- and underweight) was
associated with higher temperamental difficulty in infants. This concern may arise from
the feeding difficulties commonly reported by mothers of infants high in temperamental
difficulty. As mentioned previously, Farrow and Blissett (2006a) found that feeding
difficulties were associated with inadaptable and ‘difficult’ infant temperament.
Similarly, Galler et al. (2004) showed that Barbadian mothers (n = 226) who rated their
infant as having a difficult temperament were also more likely to rate them as difficult
to feed. It is not surprising that infants that are identified as “difficult” generally are also
perceived as “difficult” feeders. The association between difficult infant temperament
and greater maternal concern for their child’s intake and weight may be mediated by
perceived feeding difficulty, though temporal relationships between these variables
cannot be established from the data available.
Feeding on demand vs. schedule was not associated with infant temperament. Feeding
on a schedule or on demand is likely to be influenced by factors such as mother’s
working commitments, socio-cultural norms (Dettwyler & Fishman, 1992) and advice
of health professionals and significant others. Mothers who return to work soon after
the birth of their child would have a reduced capacity to breastfeed, and more
specifically would have difficulty maintaining a demand feeding regime (Gjerdingen &
Debra, 1991; Van Esterik, 2002).
17
Although not a primary aim of the study, our data also provide information on the
independent associations between a range of covariates and feeding practices in a large
sample. Of most interest, mothers with higher scores on the EPDS (Cox, et al., 1987)
were more likely to use food to calm, were less aware of infant cues, more concerned
for their child’s intake and weight and more likely to feed on demand, largely mirroring
the associations between infant temperament and feeding practices. Similar to the
associations of difficult infant temperament, maternal depression has been found to be
associated with reported feeding difficulty and potentially undesirable feeding practices
in other studies. Farrow & Blissett (2006a) found that prenatal and postnatal maternal
cognitions (e.g. higher levels of prenatal emotional deprivation entitlement core beliefs,
and postnatal lower maternal self-esteem) were implicated in the development of
feeding difficulty in six-month-old infants. In 18-month-olds it was found that mother’s
scoring high in negative affectivity (anxiety and depression) during pregnancy were
more likely to feed their child caloric drinks at night (Vollrath, et al., 2010). Our
findings that, in the same sample, maternal depression and difficult infant temperament
show a similar pattern of association with maternal feeding practices is inherently
plausible. However, an alternative factor may be that mothers with depression are more
likely to perceive their infant’s temperament as more difficult. In a sample of 139
women from the USA, depressed mothers reported higher difficult temperament scores
when infants were both two and six months-of-age than their non-depressed
counterparts (McGrath, Records, & Rice, 2008). The previously mentioned study
conducted in Barbados found maternal despair and anxiety at seven weeks and six
months were both associated with reported ‘difficult’ temperament traits in infants
18
(decreased adaptability, reduced approach, negative mood and increased sensory
threshold) at six months of age (Galler, et al., 2004). Mothers who rated their infant as
having a difficult temperament were also more likely to rate them as difficult to feed
(Galler, et al., 2004). Unlike our study, this study did not examine the relationship
between maternal mood and feeding practices. Further studies are needed with analyses
designed to specially evaluate the complex, multi-directional relationships between
maternal depression (both pre- and post partum), infant temperament ratings and
maternal feeding practices to better understand the relationship between these variables.
Strengths of this study include a large sample of very young infants which has enabled
contemporaneous assessment of early developmental stages in feeding and related
feeding practices without the limitations of recall bias. Temperament was assessed
using a tool that was validated in a representative Australian sample and the concurrent
use of the modified IFQ performed well in this sample of first-time mothers, compared
with its reliability as a retrospective measure. Also, the analyses adjusted for a
comprehensive range of maternal and infant covariates.
There are also limitations that should be considered. The generalisability of these
results is affected by the sample of primiparous, mostly well educated Caucasian
mothers and a high prevalence of exclusive breastfeeding. Maternal self-report is
subject to potential social desirability bias in responses which may limit the accuracy of
data (Harden, Klein, & Durbin, 2005). However, questionnaires were self-administered,
labelled with study ID only and completed at baseline prior to any personal contact with
19
study staff. Furthermore, the alternative of direct observation has limitations in terms of
assessing usual behaviour (Haycraft & Blissett, 2008) and was not feasible in such a
large sample. Finally, these analyses are cross-sectional and as such causal
relationships cannot be determined and require longitudinal exploration.
Conclusions
Infant temperament was found to be associated with the feeding practices of first-time
mothers. Overall, the results suggest that mothers with infants of more difficult
temperament tend to adopt feeding practices that have been associated with indicators of
later obesity risk. As such, feeding practices are a potential mediator for the previously
demonstrated associations between infant temperament and weight outcomes and
further analysis of this relationship is recommended. Mothers of temperamentally
difficult infants may benefit from additional guidance and support with feeding to
minimise the adoption of undesirable feeding practices. The findings that maternal
depression has an independent influence on feeding practises suggests that the provision
of support for infant feeding concurrent with treatment of postnatal depression may also
assist the feeding relationship. Longitudinal prospective studies in a range of
population groups are required to extend our understanding of the relationships between
temperament and infant feeding practices. It is recommended that both temperament
and maternal depression be included as covariates in future studies of feeding practices,
child eating, and obesity risk.
Conflict of interest
20
The authors have indicated that they have no personal financial relationships relevant to
this article to disclose.
Acknowledgements
NOURISH was funded 2008-2011 by the Australian National Health and Medical
Research Council (grant 426704). Dr Kimberley Mallan has occupied the Heinz
Postdoctoral Fellowship funded by HJ Heinz. Additional funding was provided by
Meat & Livestock Australia (MLA), Department Health South Australia, Food
Standards Australia New Zealand (FSANZ) and Queensland University of Technology,
and NHMRC Career development Award to JMN (grant 390136). We acknowledge the
NOURISH investigators: Professors Diana Battistutta, Ann Farrell, Geoffrey Cleghorn
and Geoffrey Davidson. We express appreciation for Professor Karen Thorpe’s
contribution to the manuscript. We sincerely thank all our participants, recruiting staff
and study staff including Dr Carla Rogers, Jo Meedeniya, Gizelle Wilson, Chelsea
Mauch.
21
References
Agras, W. S., Hammer, L. D., McNicholas, F., & Kraemer, H. C. (2004). Risk factors
for childhood overweight: A prospective study from birth to 9.5 years. The
Journal of Pediatrics, 145(1), 20-25.
Baughcum, A. E., Powers, S. W., Johnson, S. B., Chamberlin, L. A., Deeks, C. M., Jain,
A., & Whitaker, R. C. (2001). Maternal feeding practices and beliefs and their
relationships to overweight in early childhood. Journal of Developmental &
Behavioral Pediatrics, 22(6), 391-408.
Birch, L., & Deysher, M. (1986). Caloric compensation and sensory specific satiety -
evidence for self regulation of food-intake by young-children. [Article]. Appetite,
7(4), 323-331.
Birch, L. L. (2006). Child feeding practices and the etiology of obesity. Obesity, 14(3),
343-344.
Blissett, J., & Farrow, C. (2007). Predictors of maternal control of feeding at 1 and 2
years of age. International Journal of Obesity, 31(10), 1520-1526. doi:
10.1038/sj.ijo.0803661
Burdette, H. L., Whitaker, R. C., Hall, W. C., & Daniels, S. R. (2006). Maternal infant-
feeding style and children's adiposity at 5 years of age. Arch Pediatr Adolesc
Med, 160(5), 513-520. doi: 10.1001/archpedi.160.5.513
Carey, W. B. (1985). Temperament and increased weight gain in infants. Journal of
Developmental & Behavioral Pediatrics, 6(3), 128-131.
Carey, W. B., & McDevitt, S. C. (1978). Revision of the Infant Temperament
Questionnaire. Pediatrics, 61(5), 735-739.
22
Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression:
Development of the 10-item Edinburgh Postnatal Depression Scale. British
Journal of Psychiatry, 150, 782-786.
Daniels, L. A., Magarey, A., Battistutta, D., Nicholson, J. M., Farrell, A., Davidson, G.,
& Cleghorn, G. (2009). The NOURISH randomised control trial: Positive feeding
practices and food preferences in early childhood - a primary prevention program
for childhood obesity. Bmc Public Health, 9. doi: 38710.1186/1471-2458-9-387
Darlington, A. S., & Wright, C. M. (2006). The influence of temperament on weight
gain in early Infancy. Journal of Developmental & Behavioral Pediatrics, 27(4),
329-335.
de Onis, M., Blössner, M., & Borghi, E. (2010). Global prevalence and trends of
overweight and obesity among preschool children. The American Journal of
Clinical Nutrition, 92(5), 1257-1264. doi: 10.3945/ajcn.2010.29786
Dettwyler, K., & Fishman, C. (1992). Infant feeding practices and growth. Annual
Review of Anthropology, 21, 171-204.
DiSantis, K. I., Hodges, E. A., Johnson, S. L., & Fisher, J. O. (2011). The role of
responsive feeding in overweight during infancy and toddlerhood: a systematic
review. Int J Obes, 35(4), 480-492.
Donath, S., & Amir, L. (2000). Does maternal obesity adversely affect breastfeeding
initiation and duration? Journal of Paediatrics and Child Health, 36(5), 482-486.
doi: 10.1046/j.1440-1754.2000.00562.x
Farrow, C., & Blissett, J. (2006a). Maternal cognitions, psychopathologic symptoms,
and infant temperament as predictors of early infant feeding problems: A
23
longitudinal study. International Journal of Eating Disorders, 39(2), 128-134. doi:
10.1002/eat.20220
Farrow, C., & Blissett, J. (2006b). Does Maternal Control During Feeding Moderate
Early Infant Weight Gain? Pediatrics, 118(2), e293-e298. doi: 10.1542/peds.2005-
2919
Field, A. (2009). Discovering statistics using SPSS (Third ed.). Dubai: Oriental Press.
Fisher, J. O., & Birch, L. L. (1999). Restricting Access to Foods and Children's Eating.
Appetite, 32(3), 405-419.
Galler, J. R., Harrison, R. H., Ramsey, F., Butler, S., & Forde, V. (2004). Postpartum
maternal mood, feeding practices, and infant temperament in Barbados. Infant
Behavior & Development, 27(3), 267-287. doi: 10.1016/j.infbeh.2003.11.002
Ganley, R. M. (1989). Emotion and eating in obesity: A review of the literature.
International Journal of Eating Disorders, 8(3), 343-361.
Gjerdingen, D. K., & Debra, F. (1991). Predictors of health in new mothers. Social
Science & Medicine, 33(12), 1399-1407. doi: 10.1016/0277-9536(91)90285-
k
Goldsmith, H. H., Buss, A. H., Plomin, R., Rothbart, M. K., Thomas, A., Chess, S., . . .
McCall, R. B. (1987). Roundtable: What is temperament? Four approaches. Child
Development, 58(2), 505-529.
Haycraft, E. L., & Blissett, J. M. (2008). Maternal and paternal controlling feeding
practices: reliability and relationships with BMI. Obesity, 16(7), 1552-1558.
Hayden, E. P., Klein, D. N., & Durbin, C. E. (2005). Parent reports and laboratory
assessments of child temperament: A comparison of their associations with risk
24
for depression and externalizing disorders. Journal of Psychopathology and
Behavioral Assessment, 27(2), 89-100. doi: 10.1007/s10862-005-5383-z
Hendricks, K., Briefel, R., Novak, T., & Ziegler, P. (2006). Maternal and Child
Characteristics Associated with Infant and Toddler Feeding Practices. Journal of
the American Dietetic Association, 106(1, Supplement), 135-148. doi:
10.1016/j.jada.2005.09.035
Horn, M. G., Galloway, A. T., Webb, R. M., & Gagnon, S. G. (2011). The role of child
temperament in parental child feeding practices and attitudes using a sibling
design. Appetite, 57(2), 510-516. doi: 10.1016/j.appet.2011.06.015
Hurley, K. M., Black, M. M., Papas, M. A., & Caufield, L. E. (2008). Maternal
Symptoms of Stress, Depression, and Anxiety Are Related to Nonresponsive
Feeding Styles in a Statewide Sample of WIC Participants. The Journal of
Nutrition, 138(4), 799-805.
Kral, T. V. E., Stunkard, A. J., Berkowitz, R. I., Stallings, V. A., Brown, D. D., & Faith,
M. S. (2007). Daily food intake in relation to dietary energy density in the free-
living environment: a prospective analysis of children born at different risk of
obesity. The American Journal of Clinical Nutrition, 86(1), 41-47.
Li, R., Fein, S. B., & Grummer-Strawn, L. M. (2008). Association of Breastfeeding
Intensity and Bottle-Emptying Behaviors at Early Infancy With Infants' Risk for
Excess Weight at Late Infancy. Pediatrics, 122(Supplement 2), S77-S84. doi:
10.1542/peds.2008-1315j
Li, R., Jewell, S., & Grummer-Strawn, L. (2003). Maternal obesity and breast-feeding
practices. The American Journal of Clinical Nutrition, 77(4), 931-936.
25
Maccoby, E., & Martin, J. (1983). Socialization in the context of the family: Parent-
child interaction. In P. H. Mussen (Ed.), Handbook of Child Psychology (5 ed.).
USA: John Wiley and Sons.
Magarey, A. M., Daniels, L. A., Boulton, T. J., & Cockington, R. A. (2003). Predicting
obesity in early adulthood from childhood and parental obesity. Int J Obes Relat
Metab Disord, 27(4), 505-513.
McGrath, J. M., Records, K., & Rice, M. (2008). Maternal depression and infant
temperament characteristics. Infant Behavior and Development, 31(1), 71-80.
Minister for Health and Ageing. (2007). Tackling Childhood Obesity in Australia
Summit. Sydney: Queensland Government.
Niegel, S., Ystrom, E., & Vollrath, M. E. (2007). Is difficult temperament related to
overweight and rapid early weight gain in infants? A prospective cohort study.
Journal of Developmental and Behavioral Pediatrics, 28(6), 462-466.
Ong, K. K., Emmett, P. M., Noble, S., Ness, A., Dunger, D. B., & and the ALSPAC
Study Team. (2006). Dietary Energy Intake at the Age of 4 Months Predicts
Postnatal Weight Gain and Childhood Body Mass Index. Pediatrics, 117(3), e503-
508. doi: 10.1542/peds.2005-1668
Power, C., & Parsons, T. (2000). Nutritional and other influences in childhood as
predictors of adult obesity. Proceedings of the Nutrition Society, 59(02), 267-272.
doi: doi:10.1017/S002966510000029X
Pulkki-Raback, L., Elovainio, M., Kivimaki, M., Raitakari, O. T., & Keltikangas-
Jarvinen, L. (2005). Temperament in childhood predicts body mass in adulthood:
the cardiovascular risk in Young Finns Study. Health Psychology, 24(3), 307-315.
26
Reilly, J. J., Methven, E., McDowell, Z. C., Hacking, B., Alexander, D., Stewart, L., &
Kelnar, C. J. H. (2003). Health consequences of obesity. Archives of Disease in
Childhood, 88(9), 748-752. doi: 10.1136/adc.88.9.748
Sanson, A., Prior, M., Garino, E., Oberklaid, F., & Sewell, J. (1987). The structure of
infant temperament: Factor analysis of the revised infant temperament
questionnaire. [DOI: 10.1016/0163-6383(87)90009-9]. Infant Behavior and
Development, 10(1), 97-104.
Teng, A., Bartle, A., Sadeh, A., & Mindell, J. (2012). Infant and toddler sleep in
Australia and New Zealand. Journal of Paediatrics and Child Health, 48(3), 268-
273.
Thomas, A., Chess, S., Birch, H., Hertzig, M., & Korn, S. (1963). Behavioral
Individuality in Early Childhood. New York: New York University Press.
Van Esterik, P. (2002). Contemporary trends in infant feeding research. Annual Review
of Anthropology, 31, 257-278.
Ventura, A., & Birch, L. (2008). Does parenting affect children's eating and weight
status? International Journal of Behavioral Nutrition and Physical Activity, 5(1),
15.
Vollrath, M. E., Tonstad, S., Rothbart, M. K., & Hampson, S. E. (2010). Infant
temperament is associated with potentially obesogenic diet at 18 months.
International Journal of Pediatric Obesity, 1-7. doi:
doi:10.3109/17477166.2010.518240
Wasser, H., Bentley, M., Borja, J., Goldman, B. D., Thompson, A., Slining, M., &
Adair, L. (2011). Infants perceived as "fussy" are more likely to receive
27
complementary foods before 4 months. Pediatrics, 127(2), 229-237. doi:
10.1542/peds.2010-0166
Webber, L., Hill, C., Saxton, J., Van Jaarsveld, C. H. M., & Wardle, J. (2009). Eating
behaviour and weight in children. International Journal of Obesity, 33, 21+.
World Health Organisation. (1995). Physical status: the use and interpretation of
anthropometry. Geneva: World Health Organization.
World Health Organisation. (2006). WHO child growth standards: Length/height-for-
age, weight-for-age, weight-for-length, weight-for height and body mass index-
for-age: Methods and development. Geneva.
Worobey, J., Islas Lopez, M., & Hoffman, D. J. (2009). Maternal Behavior and Infant
Weight Gain in the First Year. Journal of Nutrition Education and Behavior,
41(3), 169-175.
Footnote
28
1 The five factors from the IFQ (Baughcum, et al., 2001) have been shortened to:
feeding on demand, using food to calm, awareness of infant cues, concern about
overweight and concern about underweight.
29
a High scores indicate greater maternal depressive symptoms (0-30), Edinburgh
Postnatal Depression Scale (Cox, et al., 1987). These data were square root transformed
for the analyses.
b Exclusive breastfeeding, breastmilk only; Any breastfeeding, including combination
feeding with formula and fully breastfeeding with or without some complementary
foods; No breastfeeding, formula feeding (no breastmilk) with or without some
complementary foods.
Table 1. Sample characteristics
Variable N % (n) or mean ± SD Maternal characteristics
Maternal age at delivery (years) 698 30.1 ± 5.3 Maternal education 698 University education 58 (406) Maternal depression a 669 4.7 ± 3.9 Borderline/depressive 12 (79) Breastfeeding status b 631 Exclusive breastfeeding 45 (314) Any breastfeeding 23.7 (150) No breastfeeding 26.5 (167) Maternal BMI (kg/m2) 692 26 ± 5 Feeding practices and beliefs Awareness of hunger/satiety cues c 675 3.2 ± 0.5 Using food to calm fussiness c 671 1.8 ± 0.8 Concern over-eating/overweight d 670 High concern 219 (33) Concern under-eating/underweight d 675 High concern 238 (35) Feeding on demand vs. schedule e 664 Feeding on schedule 246 (37)
Infant characteristics Age (months) 696 4.3 ± 1.0 Baby gender 698 Female 51 (354) Infant weight gain (g/wk) 697 197 ± 46 Infant raw change in weight for age Z-score from birth 693 –0.43 ± 1.00 Maternal perception of child weight status 660 Underweight 5 (32) Normal weight 87 (573) Overweight 8 (55) Temperament d 666 2.5 ± 0.6
30
c Measured on a 5-point Likert scale from 0= “Never” to 4= “Always”, Infant Feeding
Questionnaire (Baughcum, et al., 2001)
d Outlier tertile separated from other two tertiles to create a binary variable, Infant
Feeding Questionnaire (Baughcum, et al., 2001)
e High scores indicate a more difficult temperament, Short Temperament Scale for
Infants (Sanson, et al., 1987)
31
* p < 0.05
a Measured on a 5-point Likert scale from 0 = “Never” to 4 = “Always”, Infant Feeding
Questionnaire (Baughcum, et al., 2001)
b Values taken from final model
c High scores indicate greater maternal depressive symptoms (data has been square root
transformed), Edinburgh Postnatal Depression Scale (Cox, et al., 1987)
d Exclusive breastfeeding, breastmilk only; Any breastfeeding, including combination
feeding with formula and fully breastfeeding with or without some complementary
foods; No breastfeeding, formula feeding (no breastmilk) with or without some
complementary foods.
e Infant weight gain per week from birth to 2–7-months-of-age
Table 2. Hierarchical linear regression analyses for feeding practices and beliefs and infant temperament. Feeding practices and beliefs a Awareness of cues Using food to calm N 598 596 β 95% CI β 95% CI Step 1 b R2
Adj = 0.05* R2Adj = 0.10*
Infant gender (male) 0.00 –0.08-0.09 0.09* 0.02-0.27 Maternal age at delivery (years) –0.05 –0.01-0.00 –0.01 –0.01-0.01 Maternal education (no university) –0.04 –0.12-0.05 0.04 –0.06-0.20 Maternal depression c –0.13* –0.14- –0.03 0.09* 0.01-0.17 Breastfeeding (BF) status d Exclusive BF vs any BF/no BF 0.08 –0.02- 0.18 –0.15* –0.38- –0.09 No BF vs any BF/exclusive BF –0.02 –0.14-0.08 0.17* 0.13- 0.47 Infant wt gain (g/wk) e 0.03 0.00-0.00 0.03 0.00-0.00 Maternal BMI (kg/m2) –0.02 –0.01-0.01 0.00 –0.01-0.01 F (df) 5.16 (8, 589)* 9.55 (8, 587)* Step 2 R2
Adj = 0.15* R2Adj = 0.12*
R2 = 0.09* R2 = 0.02* Temperament f –0.32* –0.38- –0.23 0.14* 0.09-0.31 F(df) 64.47 (1, 588)* 11.97 (1, 586)* F (df) 12.25 (9, 588)* 9.98 (9, 586)*
32
f High scores indicate a more difficult temperament, Short Temperament Scale for
Infants (Sanson, et al., 1987)
33
Abbreviations: B, unstandardised regression coefficient; Wχ2,Wald χ2-test; OR, Odds ratio; 95% CI, 95% CI for Odds ratio
*p < 0.05
a Dichotomised as per Table 1, Infant Feeding Questionnaire (Baughcum, et al., 2001)
b Values taken from final model
c High scores indicate greater maternal depressive symptoms (data has been square root transformed), Edinburgh Postnatal Depression
Scale (Cox, et al., 1987)
Table 3. Hierarchical binary logistic regression for feeding practices and infant temperament Feeding practices and beliefs a Concern about overweight Concern about underweight Feeding on demand vs. schedule N 596 598 588 Variables B Wχ2 OR 95% CI B Wχ2 OR 95% CI B Wχ2 OR 95% CI Step 1b Nagelkerke R2 = 0.15 Nagelkerke R2 = 0.15 Nagelkerke R2 = 0.13 Infant gender (male) –0.11 0.32 0.89 0.60-1.33 0.13 0.44 1.14 0.77-1.68 0.35 3.15 1.41 0.96-2.07 Maternal age at delivery (years) 0.03 2.15 1.03 0.99-1.07 0.03 2.73 1.03 0.99-1.07 –0.02 0.69 0.99 0.95-1.02 Maternal education (no uni) 0.10 0.22 1.10 0.73-1.66 0.62 8.91* 1.85 1.24-2.77 –0.14 0.47 0.87 0.59-1.29 Maternal depression c 0.48 14.50* 1.61 1.26-2.06 0.35 8.19* 1.41 1.12-1.80 –0.25 4.22* 0.78 0.62-0.99 Breastfeeding (BF) status d Exclusive BF vs any BF/no BF 0.00 0.00 1.00 0.63-1.60 0.17 0.56 1.19 0.76-1.87 –0.71 9.67* 0.49 0.32-0.77 No BF vs any BF/exclusive BF
0.26 0.92 1.30 0.76-2.21 –0.25 0.90 0.78 0.46-1.31 0.70 7.88* 2.01 1.23-3.27
Infant weight gain (g/wk) e 0.01 25.95* 1.01 1.01-1.02 –0.01 31.41* 0.99 0.98-0.99 0.00 1.24 1.00 1.00-1.01 Maternal BMI (kg/m2) 0.06 10.54* 1.06 1.02-1.10 –0.03 2.57 0.97 0.94-1.01 0.07 14.75* 1.08 1.04-1.12 Step 2 Nagelkerke R2 = 0.17 Nagelkerke R2 = 0.19 Nagelkerke R2 = 0.13 Goodness of fit χ2(df) =78.89(9)* Goodness of fit χ2(df) = 86.28(9)* Goodness of fit χ2(df) = 60.21(9)* Temperament f 0.56 9.56* 1.75 1.23-2.49 0.69 15.06* 2.00 1.41-2.84 –0.01 0.00 0.99 0.71-1.39
34
d Exclusive breastfeeding, breastmilk only; Any breastfeeding, including combination feeding with formula and fully breastfeeding with or
without some complementary foods; No breastfeeding, formula feeding (no breastmilk) with or without some complementary foods.
e Infant weight gain per week from birth to 2–7-months-of-age
f High scores indicate a more difficult temperament, Short Temperament Scale for Infants (Sanson, et al., 1987)