PDF
Citation
Print
INTRODUCTION
Childhood obesity is a major health problem faced by low-, middle-, and high-income sectors [1]. The prevalence of obesity among school-aged children and adolescents aged 5-19 has risen more than 10-fold over the last 40 years [2]. In developed countries, children with low socioeconomic status are at a higher risk of being overweight than those with middle/high socioeconomic status [3]. In South Korea, the prevalence of overweight/obesity was more than twice that of overweight/obesity (26.7%) of the general child group (11.4%) [4].
Childhood obesity has multiple physiological, behavioral, and socio-environmental determinants [5]. In the absence of resolution of these causes, it lasts into adulthood and is linked to the development of chronic diseases [6]. In this context, identifying antecedent determinants of childhood obesity is critical, which may lead to the prevention and management of obesity in childhood, subsequently leading to the prevention of chronic diseases in adulthood [2].
Childhood obesity is mainly affected by parents, who are major agents who execute their parenting behaviors [7]. Promoting parenting behaviors may be an essential behavioral strategy for controlling a child’s obesity status. Parenting behavior such as stimulating children’s healthy food intake and taking responsibility for them was associated with lower levels of body mass index (BMI) among children, while parenting behaviors such as parents’ concern and restriction about their child's food intake were related to higher levels of BMI among children [8].
Childhood obesity may be indirectly influenced by parental behavioral determinants, such as parenting self-efficacy and parenting style through parenting behavior [9]. Reportedly, parents of obese children are less likely to have confidence in managing children's lifestyle behaviors and more likely to execute permissive and authoritarian parenting styles than parents of healthy-weight children [10]. Parents with low socioeconomic status have little opportunity and time to learn healthy parenting practices because of their busy work schedules [11], so they are more likely to have poor levels of parenting behavior in preventing or managing their children’s obesity status [11]. In this regard, parental behavioral determinants as exogenous factors of obesity-specific parenting behavior should be identified in socioeconomically vulnerable families so that children’s obesity status can be effectively prevented and managed based on their cultural background.
Among parenting behavioral determinants, parenting self-efficacy may promote obesity-specific parenting behaviors. Self-efficacy is a well-known function of behavioral change based on Bandura's social cognitive theory [12]. Parenting self-efficacy is defined as the belief that parents can overcome or solve specific parenting problems [13], which has been reported to be associated with successful general parenting behavior. Meanwhile, parenting styles characterized as general parenting behavior may empirically fulfill the path to specific parenting behavior, that is, obesity-specific parenting behavior [14]. Basically, parenting styles were identified by interaction frequencies manifested between parent and child and specifically categorized by two characteristics of responsive (i.e., nurturing and warmth) and demanding (i.e., control such as establishing and enforcing boundaries) by Baumrind [15]. The authoritative style was characterized by responsive and demanding; the authoritarian style by unresponsive and demanding; and the permissive style by responsive and undemanding [15]. Reportedly, the authoritative parenting style is related to parenting behaviors with a high level of monitoring and responsibility for children’s dietary intake [16], while authoritarian and permissive parenting styles are related to parenting behaviors with a low level of monitoring and a high level of restriction of children's eating [16]. However, little is known about whether each parenting style is associated with obesity-specific parenting behaviors in socioeconomically vulnerable families.
In the above context, no previous studies have yet reported an integrated conceptual pathway of “parenting self-efficacy,” “parenting styles,” “obesity-specific parenting behavior,” and “child obesity status.” This study assumed that “parenting self-efficacy” and “parenting styles” might be parental behavioral determinants for obesity-specific parenting behavior. It tested hypotheses of 1) whether the two determinants would be significantly associated with obesity-specific parenting behaviors and 2) whether the obesity-specific parenting behavior would be significantly associated with a child's obesity status among socioeconomically vulnerable children.
METHODS
1. Study design
A secondary analysis was conducted based on a cross-sectional correlation study design using baseline data from a parent study. The parent study examined the effects of the Healthy Children, Healthy Families, and Healthy Communities Program for obesity prevention, tailoring socioeconomically vulnerable children to improve their healthy lifestyle behaviors and obesity status [17].
2. Hypotheses
The hypotheses of this study are as follows. First, parenting self-efficacy would be significantly associated with obesity-specific parenting behavior (hypothesis 1). Second, each parenting style would be significantly associated with obesity-specific parenting behavior (hypothesis 2), authoritative parenting style would be significantly associated with obesity-specific parenting behavior (hypothesis 2-1), authoritarian parenting style would be significantly associated with obesity-specific parenting behavior (hypothesis 2-2), and permissive parenting style would be significantly associated with obesity-specific parenting behavior (hypothesis 2-3). Third, obesity-specific parenting behavior would be significantly associated with children’s obesity status (hypothesis 3).
3. Participants & setting
Parent study participants were was 61 parents and 107 children, of which 61 parents and 61 paired children were included for the secondary analysis of the present study. Parent study participants were recruited from eight community child centers in Seongbuk-gu Municipal County, Seoul, South Korea. The inclusion criteria for the present study were: 1) children in elementary school grades (3rd-6th grades), 2) children who were enrolled in the public welfare system of community child centers in Seongbuk-gu, and 3) children who were living with parents or legal representatives. The exclusion criteria for the present study were: 1) children with mental and physical disabilities and 2) children who had either parents or legal representatives who could not understand the Korean language. Among the children recruited for the parent study, 61 were eligible for the present study. The minimum sample size in this study was 50 participants, considering the five independent variables. According to Kline [18], the appropriate sample size should always be 10 times the number of parameters in path analysis. The 61 participants in this study met the minimum sample size based on this evidence.
4. Measurements
The data were collected by trained researchers using questionnaire surveys and physical measurements from June 12-28, 2017.
1) General characteristics of children and their parents
Children’s demographic characteristics (i.e., age and sex) were self-reported. Parents’ age, sex, household income, educational status, and employment status were self-reported. The average monthly household income was classified as less than 2 million won (approximately $1,630) versus more than 2 million won [19]. Education status was classified as less than college-educated versus greater college-educated. Parents' employment status was classified as currently employed versus unemployed. Parental BMI (kg/m2) was calculated using self-reported height (cm) and weight (kg).
2) Child’s obesity status
The child’s obesity status was assessed using the value of BMI z-scores, indicating standardized BMI scores adjusted for the child’s age and sex at a population level calculated using the World Health Organization AnthroPlus [20]. Children’s body weight (kg) and height (cm) were measured with no shoes or wear using an electronic weight/body fat scale (HBF-212; Omron, Kyoto, Japan) and a standing height scale (Seca 213; Seca GmbH & Co. KG., Hamburg, Germany). Before anthropometric measurements, children fasted for eight hours and emptied their bladders [17].
3) Parenting self-efficacy
Parenting self-efficacy was measured by a five-point Likert scale developed by Dumka et al. [21] Three Korean nursing scholars translated the English version of parenting self-efficacy, and the three Korean versions were discussed, confirmed, and consolidated into a single Korean version. The Korean version was back-translated by a native English speaker. The back-translated English version was again confirmed by nursing scholars who translated it into Korean. The scale consists of five items that evaluate confidence levels in successfully performing a parental role [21] and reports the mean score of the five-item responses. The higher the mean parenting self-efficacy score, the higher the level of parenting self-efficacy. Cronbach's alpha coefficient was 0.80 in a previous study [21] and 0.87 in this study.
4) Parenting styles
The parenting styles were measured by a five-point Likert scale, that is, the Parenting Styles and Dimension Questionnaireshort version developed by Robinson et al. [22] This scale was also developed using translation and back-translation processes, as previously described in detail. Parenting styles consist of 32 items and three typologies based on Baumrind’s conceptualization: authoritative, authoritarian, and permissive parenting styles [22]. The 32 items can be classified into three types and seven dimensions of parenting [22]. Parents should be informed of how often they use the specific behavior described for each item. The authoritative parenting is divided into 15 items into three dimensions: support/affection, regulation, and autonomy [22]. The authoritarian style consists of 12 items and consists of three dimensions: physical coercion, verbal hostility, and punishment [22]. The permissive style consist of one dimension such as indulgence [22]. Therefore, the score for each parenting style ranges from 1 to 5, with higher scores indicating more frequent use of each parenting style. Cronbach's alpha coefficients were 0.64-0.86 in a previous study [22] and 0.52-0.90 in this study.
5) Obesity-specific parenting behavior
The obesity-specific parenting behavior was defined in this study as encouraging a child’s healthy eating habits. It was measured using the Child Feeding Questionnaire (CFQ), a five-point Likert scale developed by Birch et al. [23]. This scale was also developed using translation and back-translation processes, as previously described in detail. This scale consists of 21 items with five subscales comprising perceived responsibility (three items), monitoring (three items), concern about the child's weight (three items), restriction of children's eating (eight items), and pressure to eat (four items) [23]. The CFQ scores report the total score after reversing the subscale scores of concerns about children's weight, pressure to eat, and restriction of children's eating. A higher total CFQ score indicates a higher level of obesity-specific parenting behaviors. Cronbach's alpha coefficients in the previous study were in the ranges of 0.70-0.9323) and the ranges of 0.61-0.96 in this study.
5. Data analysis
All statistical analyses were performed using SPSS 26.0 (IBM, Chicago, IL, USA) and AMOS 27.0. (IBM, Chicago, IL, USA) Statistical significance was set at P<0.05. The participants’ sociodemographic characteristics were analyzed using frequency, percentage, mean, and standard deviation. Pearson correlation coefficients were used to examine the correlations of the main study variables (i.e., parenting self-efficacy, parenting styles, obesity-specific parenting behavior, and children’s obesity status). We performed path analysis using maximum likelihood estimation and bootstrapping to test our hypothetical path model. Based on the hypothetical path model, we calculated standardized estimates (beta) of the direct, indirect, and total effects of exogenous variables (i.e., parenting self-efficacy and parenting styles) on endogenous variables (i.e., obesity-specific parenting behavior and the child’s obesity status). Finally, the validity and fit of the path model in the present study were evaluated using statistical values from chi-square, the goodness-of-fit index (GFI), adjusted goodness-of-fit index (AGFI), and root mean square error of approximation (RMSEA) and reported by squared multiple correlations (SMC), indicating an R2 value for the explanatory power of exogenous variables.
6. Ethical considerations
All participants provided written informed consent after receiving an explanation of the study’s purpose. Children and parents provided consent to participate, and parents provided consent for their own participation or that of their children. This study was approved by the Institutional Review Board of Korea University (No. 1040548-KU-IRB-17-82-A-2).
RESULTS
The average age of the children was 10.0 years (range, 8.0-12.0 years), with a mean BMI z-score of 0.9, with 13.1% being obese and 16.4% being overweight. Of the children, 52.5% were female (Table 1).
The average age of the parents was 45.8 years, ranging from 27.0 to 62.0 years (Table 1). Of the parents, 86.9% were females, 34.4%’s monthly household income was less than 2 million won (approximately $1,630), 59.0% were not college-educated, and 85.2% were employed. The parents had a mean BMI of 22.6 kg/m2. The parenting self-efficacy had a mean of 3.7 scores (range, 1.4-5.0). The authoritative, authoritarian, and permissive parenting styles had means of 3.6 (range, 2.0-5.0), 1.9 (range, 1.1-3.8), and 2.5 (range, 1.4-4.2) scores, respectively. The obesity-specific parenting behavior had a mean score of 16.1 (range, 7.9-20.9).
1. Correlations among the study variables
Parenting self-efficacy significantly correlated with obesity-specific parenting behavior (r=0.49, P<0.001) but not with the child’s obesity status (Table 2). Among the parenting styles, the authoritative (r=0.38, P=0.003) and authoritarian styles (r=-0.42, P=0.001) significantly correlated with the obesity-specific parenting behavior but not with the child’s obesity status; however, the permissive style was not significantly correlated with obesity-specific parenting behavior and the child’s obesity status. Furthermore, obesity-specific parenting behavior significantly correlated with the child’s obesity status (r=-0.37, P=0.003).
2. Evaluation of the path model
Our hypothetical path model was consolidated based on 1) the associations of parenting self-efficacy and parenting styles with obesity-specific parenting behavior and 2) the association of obesity-specific parenting behavior with the child’s obesity status (Figure 1). The model contains four exogenous variables (i.e., parenting self-efficacy and three parenting styles) and two endogenous variables (i.e., obesity-specific behavior and childhood obesity). Regarding testing model validity and fitness, our hypothetical path model manifested a saturated model, which indicates a perfect fit to the data based on the results of the chi-square test and the (χ2=1.22, GFI=0.993), and satisfying other goodness-of-fit indices: root mean square residual (RMR)=0.016, normed fit index (NFI)=0.988, and comparative fit index (CFI)=1.000.
3. Hypothesis testing
According to the path model indicated in Figure 1, parenting self-efficacy was significantly and positively associated with obesity-specific parenting behavior (β=0.40, P=0.007). Authoritarian parenting, but not authoritative or permissive parenting, was significantly and inversely associated with obesity-specific parenting behavior (β=-0.39, P=0.032). Moreover, obesity-specific parenting behavior was significantly and inversely associated with the child’s obesity status (β=-0.37, P=0.001). The fitness of the model (with authoritarian style, not with either authoritative or permissive style) indicated a perfect fit to the data based on the results (χ²=1.22, degree of freedom=4, RMR=0.016, RMSEA<0.001, GFI=0.993, AGFI=0.965, NFI=0.988, CFI=1.00) (Figure 1). Hypothesis 1 was supported for the path of parenting selfefficacy→ obesity-specific parenting behaviors (β=0.40, P=0.007). (Table 3, Figure 1). Thus, hypothesis 2 is partially supported. Hypothesis 2-1 was not supported for the path of authoritative parenting→obesity-specific parenting behaviors (β=0.03, P=0.844). Hypothesis 2-2 was met for the path of authoritarian parenting→obesity-specific parenting behaviors (β=-0.385, P=0.032). Hypothesis 2-3 was not met (β=0.18, P=0.202), as expected from the results obtained in the correlation analysis (Table 1). Hypothesis 3 was supported for the path of obesity-specific parenting→child obesity (β=-0.37, P=0.001). Finally, based on SMC (R2) values, parenting self-efficacy and authoritarian parenting style explained 33% of the total variance in obesity-specific parenting behavior. Parenting self-efficacy, authoritarian parenting style, and obesity-specific parenting behavior explained 14% of the total variance in childhood obesity.
DISCUSSION
This is the first study to integrate conceptual relationships into a path model: lower levels of parenting self-efficacy and higher levels of authoritarian parenting were significantly associated with a lower level of obesity-specific parenting behavior defined as encouraging a child’s healthy lifestyle, which was subsequently and significantly associated with a lower child’s obesity status among socioeconomically vulnerable families. However, the paths from authoritative or permissive parenting styles to obesity-specific parenting behavior were insignificant.
First, we found that a higher level of obesity-specific parenting behavior was significantly associated with lower BMI z-scores. Shelton et al. [24] found that children's BMI was improved by providing parents of overweight children with a 4-week parent-focused group education program as a parental strategy for dealing with behavioral problems associated with children’s food intake and physical activity [24]. These findings support the importance of the parental role in promoting healthy eating and activity behaviors in preventing and managing childhood obesity. For future research, a child obesity prevention and management program should be parent-focused by designing interventional strategies to increase obesity-specific parenting behaviors targeting children’s healthy lifestyles. It must be recognized that such a parent-focused intervention may effectively reduce childhood obesity among socioeconomically vulnerable families.
We also found that higher levels of parenting self-efficacy were significantly associated with a higher level of obesityspecific parenting behavior. Dumka et al. [21] reported consistent findings among 90 middle-income white mothers and 94 low-income Mexican immigrant mothers. Furthermore, Shumow and Lomax [25] reported that a higher level of parenting self-efficacy significantly predicted higher parental involvement, parental monitoring, and parent-child communication among 929 North American parents. Based on these findings, we suggest conducting clinical trials to determine whether an approach to enhance parenting self-efficacy would increase obesity-specific parenting behavior among socioeconomically vulnerable families.
We found that a higher level of authoritarian parenting style was significantly associated with a lower level of obesity-specific parenting behavior. A systematic review conducted by Collins et al. [26] elucidated that an authoritarian parenting style was significantly related to poor obesity-specific parenting behavior, such as pressuring a child to eat or having restrictive feeding behaviors. Children of authoritarian parents from low-income families are more likely to manifest higher emotional overeating [27]. Meanwhile, our study showed that the paths between authoritative/permissive parenting styles and obesity-specific parenting behavior were not significant. Bornstein and Bradley [28] and Hoff and Laursen [29] have reported that parents with low socioeconomic status were more likely to use an “authoritarian” parenting style than those with high socioeconomic status. Although the authoritarian parenting style in the present study scored as not higher than that of other parenting styles, its impact on parenting behaviors might be larger than those of other parenting styles. Therefore, these findings may address the importance of understanding discrepancies in the levels of parenting styles in different sociocultural and economic contexts.
Our findings showed that lower levels of parenting self-efficacy and higher levels of authoritarian parenting style were significantly associated with a lower level of obesity-specific parenting behavior, which was subsequently and significantly associated with increased children's obesity status. West et al. [30] conducted a parent-centered intervention to improve parenting skills and confidence in parents to improve their child's healthy behaviors for 12 weeks [30]. Therefore, parents increased their confidence levels in managing their children's weight-related behaviors, decreased the levels of inconsistent or authoritarian parenting behaviors, and reduced their BMI z-scores [30]. Therefore, we suggest that focusing on both conceptual constructs of parenting self-efficacy and authoritarian parenting style is necessary when community health nurses design obesity-specific parent-focused programs for socioeconomically vulnerable families in a community setting.
This study has several strengths and limitations. This study may provide a theoretical framework for designing a family-focused intervention for reducing childhood obesity and emphasize a parent-focused behavioral approach by modifying parenting determinants and behaviors. This study also has some limitations. First, caution should be exercised when interpreting the results of this study in light of the potential causality issue. Future research using a randomized controlled trial design is needed to guarantee causality among the study variables. Second, the outcome variable of this study was child obesity status. Of 61 children of the participants, 70.5% fell into the category of normal weight; this finding may limit the interpretation of the study results. Third, the obesity-specific behavior may be influenced by various family-related variables (e.g., mother vs. father or child’s gender) and parental anthropometric data (e.g., BMI) that were not able to be included in the path model. Parental BMI, not but other variables, was significantly associated with obesity-specific behavior in the present study (β=-0.24, P=0.010) (data not shown). In this regard, next studies need to identify parental variables associated with parenting behaviors thoroughly, and adjust them in the other statistical models that might be able to be included. Moreover, a small sample size (n=61) may give biased interpretation of the present findings, possibly attributable to the low scale reliability of the permissive parenting style (Cronbach's alpha=0.52). However, the present study may have some advantage to reveal a potential conceptual link of parenting behavior and its associates. Finally, the results of this sample could apply to Korean socioeconomically vulnerable families with school-aged children and may not be generalizable to other ethnic and high-income population groups. Thus, a study with larger and more diverse sample size is needed for future replication. This study explained obesity-specific behavior by 33% of socioeconomically vulnerable families with parenting self-efficacy and authoritarian styles. Future studies may be needed to identify more antecedent parenting determinants rather than self-efficacy and parenting styles than those used in the present study, which may explain parenting behaviors with a greater magnitude.
Among socioeconomically vulnerable families, we revealed a conceptual framework of parenting self-efficacy and authoritarian parenting styles with obesity-specific parenting behaviors directly associated with children’s obesity status. For future research, a childhood obesity prevention and management program targeting socioeconomically vulnerable families should include behavioral strategies to improve parenting behaviors by increasing parenting self-efficacy and alleviating authoritarian parenting styles.
XML Download