 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
a Centre for Community Child Health
e Clinical Epidemiology and Biostatistics Unit, Royal Children's Hospital, Parkville, Victoria, Australia
b Murdoch Children Research Institute, Parkville, Victoria, Australia
c Department of Paediatrics, University of Melbourne, Parkville, Victoria, Australia
d Griffith Psychological Health Research Centre, Griffith University, Queensland, Australia
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONAPPENDIX: PARENTING ITEMSREFERENCES |
|---|
PARTICIPANTS AND METHODS. Participants were composed of all 4983 of the 4- to 5-year-old children in wave 1 of the nationally representative Longitudinal Study of Australian Children with complete BMI and maternal parenting data. Mothers and fathers self-reported their parenting behaviors on 3 multi-item continuous scales (warmth, control, and irritability) and were each categorized as having 1 of 4 parenting styles (authoritative, authoritarian, permissive, and disengaged) using internal warmth and control tertile cut points. Using a proportional odds model, odds ratios for children being in a higher BMI category were computed for mothers and fathers separately and together, after adjustment for factors associated with child BMI, including mothers' and fathers' BMI status.
RESULTS. The sample was composed of 2537 boys and 2446 girls with a mean age 56.9 months; 15% were overweight and 5% were obese (International Obesity Task Force criteria). Mothers' parenting behaviors and styles were not associated in any model with higher odds of children being in a heavier BMI category, with or without multiple imputation to account for missing maternal BMI data. Higher father control scores were associated with lower odds of the child being in a higher BMI category. Compared with the reference authoritative style, children of fathers with permissive and disengaged parenting styles had higher odds of being in a higher BMI category.
CONCLUSIONS. This article is the first, to our knowledge, to examine the parenting of both parents in relation to preschoolers' BMI status while also adjusting for parental BMI status. Fathers' but not mothers' parenting behaviors and styles were associated with increased risks of preschooler overweight and obesity. Longitudinal impacts of parenting on BMI gain remain to be determined.
Key Words: parenting • mothers • fathers • obesity • child • preschool
Abbreviations: LSAC—Longitudinal Study of Australian Children • SEIFA—Socio-Economic Indexes for Areas • OR—odds ratio • CI—confidence interval
The importance of childhood obesity, "modernity's scourge,"1 lies in its "ominous implications for the development of serious diseases, both during youth and later in adulthood."2 It is not yet clear how much effective solutions will require societal (eg, through legislation, community engineering, and taxation) as opposed to individual (eg, through personal and family behavior change) responsibility for reducing its prevalence. Obesity is now nearly as prevalent among preschool children3–5 as among older children, and those who are overweight or obese at school entry typically remain so during the primary school years.6 Given that the development of obesity reflects both nutritional and physical activity behaviors and that during the preschool years these behaviors occur largely within the family unit, it is timely to scrutinize the roles of parents and parenting in the preschool years.
The family unit is the primary context for providing the nurturance, resources, and opportunities essential for healthy development.7 Key parenting skills associated with positive child outcomes in early and middle childhood include warm, affectionate interactions that are responsive to children's needs ("warmth"), firm discipline in terms of the setting of developmentally appropriate limits and expectations for children's behavior ("control"), and an absence of irritable, angry affect ("irritability").7,8 These behavioral dimensions can be combined to classify a number of "styles" of parenting. The most widely used typology is that originally developed by Baumrind,9 in which parenting is categorized into 4 main styles, each associated with a different pattern of developmental outcomes. The best developmental outcomes (including greater child self-esteem and social and cognitive skills and fewer emotional and behavioral problems) are associated with "authoritative" parenting, characterized by high levels of warmth combined with high control. "Authoritarian" parenting, characterized by high control and low warmth, is associated with a lack of social competence and self-esteem, aggressiveness, and poor academic achievement; "permissive" parenting, characterized by high warmth and low control, is associated with impulsive, aggressive behavior, and substance use problems; and "disengaged" (sometimes called "neglectful") parenting, in which both warmth and control are low, is associated with impulsivity, behavioral and emotional problems, school dropout, substance use, and delinquency.10,11
Very little is known about relationships between these aspects of parenting and childhood overweight and obesity. In 2001, Davison and Birch12 emphasized the importance of understanding how parenting styles and family characteristics contribute to the causal pathways leading to childhood obesity. However, parenting was narrowly conceptualized as those characteristics that might directly relate to children's nutrition (child feeding practices, types of food available in the home, nutritional knowledge, and parent dietary intake, food preferences, and weight status) and physical activity (parent monitoring of child television viewing, encouragement of child activity, and parents' own preferences for activity and activity patterns). More recently, Rhee et al, noting that broader parenting style may "have a greater impact on... overweight risk of children than selected parenting or feeding practices alone,"13 reported that children ascertained at 54 months old as having authoritarian, permissive, or neglectful mothers had substantially higher odds of being overweight at 85 months (odds ratios [ORs]: 4.9, 2.8, and 2.7, respectively) than children of authoritative mothers. However, this study neither adjusted for maternal BMI status (the strongest known predictor of child overweight or obesity) nor considered the parenting behaviors or styles of fathers. It is common for fathers' potential contributions to children's outcomes to be overlooked in child health-related parenting research,14 perhaps because researchers assume either that parenting is similar across mothers and fathers or that fathers' parenting does not add anything to children's outcomes, over and above the contribution of mothers.
In this article, we report on a large, nationally representative sample of Australian preschool children that, unusually, included BMI data for mothers, fathers, and children and parallel self-reported parenting data for both mothers and fathers. The first aim was to investigate whether children's BMI status was associated with any of 3 parenting indicators (warmth, control, and irritability) for mothers and fathers separately. The second aim was to similarly investigate associations between children's BMI status and the 4 classic parenting categories, hypothesizing that, compared with authoritative mothers or fathers, children with authoritarian, permissive, or disengaged mothers or fathers would have an increased odds of being in a heavier BMI category. The third aim was to explore children's BMI status more broadly in models in which mothers' and fathers' parenting behaviors and styles were examined simultaneously.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: PARENTING ITEMSREFERENCES |
|---|
1 of 10 Australian postcodes ultimately included in the study. Children born between March 1999 and February 2000 were then randomly selected using the Health Insurance Commission Medicare database, in which 98% of all 4-year-old Australian children are enrolled. Of the 10596 children included in the sampling frame, 8391 were still resident within that postcode and could be contacted, and of these, 4983 (59%) took part.
Procedures
Trained professional interviewers conducted a 2.5-hour face-to-face interview in the child's home with the study child's primary carer (usually the biological mother). The primary and secondary carer (usually the father) also each completed written questionnaires provided at the interview. Written informed consent was obtained for each participating child, and the study was approved by the Australian Institute of Family Studies ethics committee.
Outcome (Child BMI Status)
Children's weight was measured in light clothing to the nearest 50 g by using digital scales (Code 79985, Salter; Springvale, Victoria, Australia) and height to the nearest 0.1 cm by using a portable rigid stadiometer (Code IPO955, Invicta; Oadby, Leicester, United Kingdom).16 The averages of 2 height measurements were used in analyses; where the 2 differed by >0.5 cm, a third measurement was taken, and the average of the 2 closest was used. Children were classified as nonoverweight, overweight, or obese according to the International Obesity Task Force age- and gender-specific criteria17 for BMI (kilograms per meter squared).
Primary Exposures (Parenting Measures)
Parents self-rated their parenting behaviors along 3 continuous parenting dimensions administered at the face-to-face interview for the primary carer and the written questionnaire for the secondary carer, as follows. Six items from the Child Rearing Questionnaire18 addressed the frequency with which they displayed warm affectionate behaviors toward their child (warmth, eg, "How often do you have warm, close times together with this child?"); 5 items from the National Longitudinal Survey of Children and Youth19 addressed the frequency with which they set and enforced clear expectations and limits for their children's behavior (control, eg, "When you discipline this child, how often does he/she ignore the punishment?" reverse scored); and 4 items from the National Longitudinal Survey of Children and Youth19 addressed the frequency with which their interactions with the child entailed behaviors such as disapproval, lack of praise, and anger (irritability, eg, "How often are you angry when you punish this child?"). Responses were on 5-point Likert scales and ranged from 1 ("never/almost never") to 5 ("all the time"). Items were averaged to obtain 3 summary scores (Cronbach's 
= .83 for warmth, .61 for irritability, and .73 for control), which have demonstrated construct validity when modeled against a wide range of concurrent family and parental characteristics.20
Warmth and control scores were dichotomized and combined to approximate the 4 categorical parenting styles.11 Because there are no standard cut points for these scales, and scores were skewed toward the positive, the summary scores were dichotomized at the least positive tertile separately for mother and father data. The combination of high warmth and high control was classified as authoritative; low warmth and high control as authoritarian; high warmth and low control as permissive; and low warmth and low control as disengaged.
Covariates were those identified in previous analyses of this sample to independently contribute to child BMI status.16 Parent-reported child variables were gender (male or female), number of siblings in the household (0, 1, 2, or 
3), and language other than English spoken at home by the child (yes or no). Maternal/paternal variables were highest completed educational level (<12 years, 12 years [ie, school completion], or tertiary) and BMI status (nonoverweight [<25 kg/m2], overweight 25 to <30 kg/m2], or obese [30 kg/m2] calculated from mothers' and fathers' self-reported height and weight21). The single neighborhood or geographical variable was the Socio-Economic Indexes for Areas (SEIFA22) disadvantage index at the postcode of residence level, categorized using whole-of-Australia population quintiles. SEIFA values are standardized scores by geographic area compiled from 2001 Australian census data to numerically summarize the social and economic conditions of Australia (national mean: 1000; SD: 100; higher values represent greater advantage). Because parenting behaviors vary with family structure,20 another family covariate of family type (1- or 2-parent home) was also included. For analyses involving parenting style, irritability was also included, because this dimension may vary independent of parenting style.
Statistical Analysis
Separate analyses were conducted for mothers, for fathers, and for mothers and fathers combined. For each of these analyses, the primary parenting exposures were included in 2 ways: the 3 parenting dimensions and the 4 parenting styles. All of the analyses were conducted by using Stata 9.2 (Stata Corp, College Station, TX).23
Weights were applied to take account of the design and response patterns, as specified in the LSAC data set, and 95% confidence intervals (CIs) and SEs were estimated using first-order Taylor linearization to account for the weights and the clustered design. We used multiple imputation with the method of chained equations to account for missing maternal data for children with a mother in the household.24 In addition to the mother's BMI status (missing for 1085 [22%] of the children), imputation was conducted for 4 maternal covariates with few (<1%) missing cases (education [n = 13], warmth [n = 47], control [n = 49], and irritability [n = 48]). Paternal data were not imputed, because >30% of data were missing for each of paternal weight status, warmth, control, and irritability. Estimates obtained from the multiple imputations were pooled to obtain a single set of results.25 Analyses were also conducted on the completed case data sets and compared.
Univariate and multivariable ordinal logistic regression (using the proportional odds model) were used to assess associations between the parenting variables and the odds of higher BMI status in the child (nonoverweight, overweight, or obese). P values and 95% CIs were obtained by using Wald tests adjusted for the survey design.26 Interactions of parenting variables with the BMI status of the corresponding parent, for all but the combined parent analyses, were investigated and only included in the final model if statistical significance was <5%.
The Brant test27 was used to test the proportional odds constraint that the regression coefficients for the comparison of categories (nonoverweight versus overweight and obese and nonoverweight and overweight versus obese) were similar. If this test was statistically significant at the 5% level for any individual predictor in a model, a partial proportional odds model was fitted, allowing the constraint to be relaxed for the predictor in question.28
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: PARENTING ITEMSREFERENCES |
|---|
|
.69). Similarly, after adjustment for the covariates, there was no evidence for an association between maternal parenting style and child BMI status (P = .85). These findings were very similar when the nonimputed data set was analyzed (results not shown).
|
.79). After adjustment for the covariates, strong evidence was also found for an association between paternal parenting style and child BMI category (P = .002). Compared with those with fathers in the reference (authoritative) category, the odds of a child being in a heavier BMI category increased by 59% (95% CI: 25%–103%) for those with permissive and by 35% (95% CI: 2%–80%) for those with disengaged fathers. Results for the combined model were consistent with those for mothers and fathers separately (Table 3). After adjustment for the covariates and for all of the maternal and paternal parenting dimensions, higher paternal control score was strongly associated with decreased odds of the child being in a heavier BMI category (OR: 0.75; 95% CI: 0.65–0.86; P < .001). After adjustment for the covariates and for maternal and paternal parenting style, the odds of a child being in a heavier BMI category increased by 59% (95% CI: 24%–114%) for those with permissive and by 35% (95% CI: 1%–80%) for those with disengaged fathers, compared with those with fathers in the reference (authoritative) category (overall P = .002).
|
.05). For these predictors, the odds of a child being of normal weight (compared with overweight or obese) differed from the odds of a child being of normal weight or overweight (compared with obese). Conclusions did not alter when the 2 paternal models and the 2 combined models were refitted, relaxing the proportional odds constraint for the relevant predictors, and provided, if anything, evidence of even stronger relationships (results available on request). No evidence was found for interactions between maternal and paternal parenting variables and the BMI status of the respective parent for any model (all P values >.05), and these were not included in the final models. |
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: PARENTING ITEMSREFERENCES |
|---|
Although unexpected, these findings are consistent with the intervention study by Stein et al29 in which fathers' but not mothers' parenting (warmth and support) predicted better maintenance of weight loss after a behavioral parenting intervention for pediatric obesity. The authors speculated that fathers' parenting may be an important determinant of the extent to which the family environment as a whole is supportive of children's attempts to lose weight. Paternal dietary and physical activity behaviors have also been associated with children's weight, diet, and physical activity, either independently or in combination with mothers' behaviors.30,31 More generally, studies with older children and adolescents have found that fathers' parenting influences children's health risk behaviors and psychological well-being.32,33
This study directly contrasts with the finding of Rhee et al13 of a strong association between child BMI status and the maternal authoritarian parenting style (characterized by high control, as well as low warmth and sensitivity). We do not believe that design differences (the design of Rhee et al13 was longitudinal, whereas the current study was cross-sectional) account for this, because children's BMI status is already very stable by school entry, suggesting that reanalysis using wave 2 LSAC longitudinal BMI status as the outcome will most likely yield congruent results. The 2 studies used different measures as proxies for the parenting dimensions of warmth and sensitivity, both with arbitrary cut points to dichotomize the parenting dimensions, so that they could have measured different underlying constructs. It is possible (although it seems unlikely) that associations between maternal parenting style and child BMI status may vary between Australian and American families. Perhaps most importantly, the conclusions of Rhee et al13 were based on only 97 overweight children, making chance associations more likely than with the 1000 overweight and obese children in the current study.
The current study has some notable strengths in comparison with the existing literature.13,34,35 First, it fills a gap in the literature by examining the association between parenting style and child BMI status for fathers, in addition to mothers. Second, it adjusts for both maternal and paternal BMI status, which are important potential confounders of any association between parenting style and child BMI status.16 Other strengths include its large scale, nationally representative design, and the recency of the data, enabling firm conclusions to be drawn about parenting and obesity as they relate to today's preschoolers and parents throughout Australia. Children's height and weight were directly measured using standardized measurement techniques, and the prevalence of overweight and obesity mirrored that in other recent surveys in Australia36 and elsewhere. The results were highly internally consistent across conceptualizations of parenting (dimensional and categorical), across univariate and multivariate analyses, and across analyses with and without imputation for missing data. Because the LSAC is an "omnibus" study, the overt focus at data collection was not on parenting as it relates to obesity, reducing the likelihood of social desirability and other biases affecting the relationships observed.
Weaknesses include the response rate of 59%, possibly reducing the generalizability of our findings to children of lower socioeconomic status postcodes, non–English-speaking backgrounds, and lone-parent families (all known to be underrepresented in the LSAC). Because of the substantial missing data, the paternal findings must also be accepted cautiously. Also, because parents reported their own height and weight, some overweight and obese parents may have been misclassified into a lower BMI category, but it seems unlikely that this would have greatly altered results. Using tertile cut points meant that the proportions of parents allocated to each parenting style were essentially arbitrary, but this does not negate the strong internal relationships, which were the prime concern of this article.
After seminal adoption studies >15 years ago, Stunkard et al37 concluded that "genetic influences on body-mass index are substantial, whereas the childhood environment has little or no influence." We broadly concur with the conclusion of Stunkard et al37: the strongest predictor of child overweight and obesity in our cohort was parental BMI status, and this relationship was not attenuated by inclusion of parenting dimensions or styles in the multivariable models. Nonetheless, times have changed; today's preschoolers live in an obesogenic environment that was unimaginable as recently as 1990, and nongenetic factors may now be relatively more important. It seems from our research that warm, firm paternal parenting, broadly conceptualized, may at least partly protect against preschool overweight and obesity in this new environment. Other "societal" problems also associated recently with children's BMI status (such as minority status,38 social disadvantage,39,40 and high birth weight41) may be even more important.
It is possible that early parenting behaviors and style have greater "downstream" impacts for older children. As children mature and take control of their own nutrition and physical activity, the approaches that parents use to shape these behaviors may become more important, and additional longitudinal research is clearly needed. At present, however, the literature remains conflicting regarding the role of parenting in the genesis of childhood obesity. We conclude first that fathers are important and should be included in all such research.42,43 Secondly, warm, firm, and authoritative parenting is already known to be associated with the best child outcomes. These findings provide another reason to support community initiatives (universal and/or targeted44) aiming to enhance these parenting attributes45; if this also helps the obesity epidemic, then that is all to the good.
|
APPENDIX: PARENTING ITEMS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: PARENTING ITEMSREFERENCES |
|---|
Warmth
How often do you express affection by hugging, kissing, and holding this child?
How often do you hug or hold this child for no particular reason?
How often do you tell this child how happy he/she makes you?
How often do you have warm, close times together with this child?
How often do you enjoy doing things with this child?
How often do you feel close to this child both when he/she is happy and when he/she is upset?
Control
When you give this child an instruction or make a request to do something, how often do you make sure that he/she does it?
If you tell this child he/she will get punished if he/she doesn't stop doing something, but he/she keeps doing it, how often will you punish him/her?
How often does this child get away with things that you feel should have been punished? (R)
How often is this child able to get out of punishment when he/she really sets his/her mind to it? (R)
When you discipline this child, how often does he/she ignore the punishment? (R)
Irritability
Of all the times that you talk to this child about his/her behavior, how often is this praise? (R)
Of all the times that you talk to this child about his/her behavior, how often is this disapproval?
How often are you angry when you punish this child?
How often do you feel you are having problems managing this child in general?
|
ACKNOWLEDGMENTS |
|---|
We thank all of the parents and children who took part in wave 1 of the LSAC.
|
FOOTNOTES |
|---|
Address correspondence to Melissa Wake, MD, Centre for Community Child Health, Royal Children's Hospital, Flemington Road, Parkville, Victoria 3052, Australia. E-mail: melissa.wake{at}rch.org.au
For this article we used confidentialized unit record files from the Longitudinal Study of Australian Children survey. The findings and views reported, however, are those of the authors and should not be attributed to either the Australian Department of Family and Community Services and Indigenous Affairs or the Australian Institute of Family Studies.
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: PARENTING ITEMSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. Bush Update in Pediatric Lung Disease 2007 Am. J. Respir. Crit. Care Med., April 1, 2008; 177(7): 686 - 695. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2007 American Academy of Pediatrics. All rights reserved. |
||
|
||
