Mothers’ guilt responses to children’s family history-based obesity risk feedback

Abstract

This study explored the influence of family health history (FHH) obesity risk feedback for their child on 147 overweight mothers’ guilt related to children’s lifestyle behaviors and passing down a genetic propensity for overweight. Mothers were randomized to receive, or not, FHH risk feedback for their 4-5 year old child, then made food choices for them using a virtual reality-based buffet. Receipt of risk information increased lifestyle- and genetics-related guilt. Choosing fewer unhealthful foods for the child attenuated both types of guilt. Work in this area may aid in development of obesity risk feedback strategies that enhance child feeding.

There is great enthusiasm for targeting obesity prevention to children as an approach to reducing the high rates of obesity in the United States (Frieden, Dietz, & Collins, 2010). Family-based interventions have had limited success (Wang et al., 2013), and methods for improving these interventions are needed. It has been suggested that family health history (FHH) assessments describing a child’s risk might be a suitable way to evaluate whether inherited risk could motivate preventive behavior among parents (Tarini, Singer, Clark & Davis, 2008). This approach uses the weight status of a child’s relatives to predict the risk for that child of becoming overweight or obese. Indeed, research has demonstrated that a child’s obesity risk is increased if one of the child’s biological parents is overweight or obese, and increases even more if both parents are overweight or obese (Magarey, Daniels, Boulton, & Cockington, 2003; Whitaker, 1997).

The inherently interdependent, shared risk represented by FHH may prompt family-wide prevention strategies (Lyons, 1998; McBride, Sanderson, Kaphingst, & Koehly, 2010). As such, parents’ role and responsibility to protect their children from health risks may be made salient. If parents feel unable to adequately protect their child from the risk, it could engender feelings of guilt (Baumeister, 1994). Guilt is an undesirable outcome in that it is an aversive emotional experience. However, individuals are typically motivated to take steps to reduce these negative emotions (Tangney, Stuewig, & Mashek, 2007). To the extent that reducing negative child feeding habits is able to assuage parental guilt, such behavior changes may be reinforced.

Studies have found that identifying a parent as “genetically responsible” for a child’s inherited health risk can elicit guilt (Shostak, Zarhin, & Ottman, 2011; Lewis, Skirton, & Jones, 2011). Guilt is especially likely to be felt among mothers whose children are affected by conditions passed down via the mother’s genetic material alone (James, Hadley, Holtzman, & Winkelstein, 2006). However, conditions studied thus far largely have been rare, inherited disorders where the risk conferred is almost exclusively genetic (e.g., muscular dystrophy, fragile X). It is unknown to what extent these processes occur in the context of obesity, a complex condition that is common, and jointly influenced by behavior, environment and genetics.

With respect to pediatric obesity, parents may also experience heightened feelings of guilt in reaction to the shared environment component of FHH. Previous research has found that some parents report feeling guilt or shame related to their child’s weight and/or their child feeding practices (Hughes, Sherman, & Whitaker, 2010; Jackson, Wilkes, & McDonald, 2007; Pescud & Pettigrew, 2012). Therefore, provision of FHH-based information related to a child’s obesity risk might prompt parents to blame themselves regarding their family home environment (Davison, Francis, & Birch, 2005). It may also increase parents’ perceptions of their child’s obesity risk, making any perceived shortcomings in their feeding and lifestyle behaviors feel more consequential.

If FHH-based risk information does elicit feelings of guilt among parents, it will be important to understand the potential interplay between these reactions and parenting behaviors. Conceptually, despite its negative valence, guilt is considered to be an adaptive and constructive emotion that motivates compensatory behaviors to alleviate the source of the guilt (Tangney et al., 2007). Therefore, if provision of FHH information promotes feelings of guilt, parents may compensate by making positive child feeding choices (e.g., limiting portion sizes of unhealthy foods). In turn, making such changes may reduce parents’ experiences of guilt.

In contrast, parents who feel guilty for passing down risk-related genetic susceptibility may feel that there is no way to compensate, or may feel a lack of control over outcomes associated with genetic risk. Guilt can lead to maladaptive or defensive behavior, particularly when there are no perceived options for compensatory action (Miceli & Castelfanchi, 1998; Nelissen & Zeelenberg, 2009). Therefore, if parents are made to feel guilty as a result of FHH feedback about genetic transmission of obesity, they might not engage in positive, compensatory feeding behavior changes. Regardless of whether they engage in these behaviors, they also may not experience a corresponding reduction in this type of guilt.

The current analysis is a first attempt at exploring these processes using data from the Mothers’ TAKE project (McBride, Persky, Wagner, Faith, & Ward, 2013). The Mothers’ TAKE project assessed the influence of personalized FHH-based risk feedback related to a child’s obesity risk on overweight mothers’ feeding behavior.

Hypotheses

1. Mothers’ guilt related to their child’s lifestyle behavior (hereafter: “lifestyle guilt”) will increase when they receive FHH-based obesity risk information for that child.

2. Mothers’ guilt related to genetic transmission of obesity risk to their child (hereafter: “genetic guilt”) will increase when they receive FHH-based risk information for that child.

   • 2a.
     Mothers’ genetic guilt will be highest when only the mother’s (and not the father’s) weight status contributes to the child’s increased risk.

3. Choosing a smaller amount of unhealthful food for their children from a virtual reality-based buffet will reduce mothers’ lifestyle guilt.

4. Mothers’ food choice behavior will not influence genetic guilt.

Method

Participants

Participants were enrolled in the Mothers’ TAKE project, a randomized controlled trial that assessed mothers’ feeding behavior in response to FHH-based feedback regarding their child’s obesity risk (McBride et al., 2013). The current analysis includes only those participants who were randomly assigned to receive information about behavioral risk factors for childhood obesity alone (n=73), and those who received the information plus FHH-based obesity risk feedback for their child (n=75), but not those from a control group who did not receive information about childhood obesity.

Primary inclusion criteria included: 1) being the biological mother of a child between the ages of 4 and 5; 2) having a self-reported body mass index of greater than or equal to 25. Additional criteria are reported elsewhere (McBride et al., 2013). Participants were compensated. The activities were approved by the Institutional Review Board of the National Human Genome Research Institute.

Procedure

Informed consent was obtained. An ‘index child’ between the ages of 4-5 was identified during eligibility screening. Participants were asked to consider this child when completing study tasks and responding to questions. Participants deemed eligible visited a secure website where they provided consent and filled out a pre-test questionnaire. Participants visited the lab for an in-person session. They completed a training session with the virtual reality buffet, watched a computer-based information module, completed a questionnaire, and then completed the virtual buffet measure. Participants then completed a final questionnaire.

Risk Information Presentation

All participants included in the current analysis watched a computer-based module that provided behavioral risk education regarding childhood obesity. Some participants were additionally randomly assigned to watch a module that provided a personalized FHH-based obesity risk assessment for the index child. The assessment was preceded by information about the role of genetic factors and gene-environment interaction in weight and obesity. The assessment was based on the mother’s and the index child’s biological father’s current weight status. Those participants whose child had a single overweight parent (i.e., the mother) received information indicating that their child had 28% risk of obesity as an adult. For children with two overweight parents (i.e., mother and biological father), the participant received information that their child had a 58% chance of obesity as an adult (Magarey et al., 2003). This figure was presented in comparison to risk associated with having no overweight parents (9%).

Measures

Guilt.

Mothers’ guilt reactions were measured by items created for this study because no relevant measures existed. Guilt was measured only at post-test, following the buffet feeding measure. Guilt related to the child’s lifestyle was measured with two items (e.g., I feel guilty about [CHILD’S NAME]’s current eating habits). Guilt related to passing down a genetic predisposition for obesity was measured with two items (e.g., I feel guilty about the part of [CHILD’S NAME]’s risk for obesity that is due to genetics). Responses were collected on a 1-7 scale.

Potential predictors of guilt.

Mothers’ BMIs and the index children’s BMIs were based on mother-reported height and weight for herself and her child. These were collected at eligibility screening and pre-test, respectively. We determined whether the biological father was overweight at pre-test using mothers’ assessment of fathers’ body size (Pulvers, Lee, & Kaur, 2004). During the eligibility screening phone call we collected: index child’s gender, mother’s education level (high school or less versus some college versus college graduate) and race/ethnicity (white or non-white). At pre-test, we assessed: mothers’ dieting history using a single item [“Have you taken part in a formal weight loss program (like Weight Watchers or the South Beach Diet) in the last 12 months?”] and mothers’ beliefs about genetics as a cause of obesity using a single item (“Obesity is almost never affected by one’s genetic makeup”, reverse scored). We also assessed mothers’ guilt and shame about their own weight using the Weight- and Body-Related Shame and Guilt Scale (Conradt, 2007). Finally, we determined whether the biological father resided in the household with the mother and the index child at post-test.

Child Feeding Behavior.

We used a virtual reality-based behavioral assessment of mothers’ feeding behavior. Mothers were asked to make a lunch plate for their index child in a 3-dimensional simulation of a buffet restaurant. They could choose as many servings of as many foods as they wanted as long as there was room on the plate. Mothers interfaced with the buffet using a head-mounted display and a pointing device. Several foods, condiments and beverages were available in the simulation. We categorized each item using the Coordinated Approach to Child Health guide (“We Can! GO, SLOW, and WHOA Foods,” 2002). Foods were categorized as “go” (healthful), “slow” (less-healthful) and “whoa” (unhealthful). The measure used in the current study was the number of servings of “slow” and “whoa” foods and beverages chosen by the mother (Figure 1). A “serving”, in this measure, was equivalent to a single spoonful or single piece of food. This measure was an overall metric of the amount of less-healthful and unhealthful food that mothers chose for their children.

Figure 1:

Screenshot of the buffet and names of go, slow and whoa foods included in buffet. Based on Coordinated Approach to Child Health guide (“We Can! GO, SLOW, and WHOA Foods,” 2002)

Data Analysis

A Pearson correlation was performed to assess the relationship between lifestyle and genetic guilt. Hypotheses 1 and 2 were addressed using one regression model for each type of guilt. We first ran the model with two indicator variables to represent the three information-type groups. Next, preliminary variables representing potential predictors of guilt were entered as independent variables in the model. To maximize the power to detect an effect of the FHH risk message, potential covariates were dropped from the final model if p-value of >0.20. To address hypotheses 3 and 4, we ran two separate mediation models using bootstrapping methods (Preacher & Hayes, 2008) with 5000 resamples.

Results

Demographics and Guilt Relationship

On average, mothers were 37 years old (SD=5.7) and had a body mass index of 30.0 (SD=3.6). Fifty percent of mothers reported their race as white, and 74% had a college degree. Thirty-nine percent of the index children were overweight, and 45% were male. None of the demographic variables differed significantly by condition. Lifestyle and genetic guilt measures were not correlated, r=.064, p=.44.

Influence of FHH Risk Information on Lifestyle Guilt (Hypothesis 1)

Means and standard deviations of guilt are available in Table 1. FHH feedback receipt influenced mothers’ reports of their guilt related to their child’s lifestyle behavior in the unadjusted model, F(2,145)=4.48, p=.013. Mothers who received FHH information implicating one parent (i.e., the mother herself) and those who received information implicating two parents in heightening the child’s obesity risk reported more guilt than mothers who received only general childhood obesity information; F(1,145)=5.10, p=.025 and F(1,145)=6.62, p=.011 respectively. In the adjusted model (Table 2), lifestyle guilt was significantly higher when mothers received FHH-based risk information, when they endorsed genetics as a cause of childhood obesity, and when they felt more shame and guilt about their own bodies.

Table 1:

Means and SDs of Guilt by Type of Risk Information Provided

	All		Type of Obesity Risk Information Provided
			Behavioral Information only (n= 73)		Behavioral Information plus Family Health History Risk Information (n=75)
					1 Parent1 (n=22)		2 Parents2 (n=53)
	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Lifestyle Guilt	2.39	1.45	2.04	1.17	2.82	1.63	2.70	1.62
Genetic Guilt	3.24	1.96	2.71	1.86	4.70	1.97	3.36	1.79

All scales ranged from 1-7 (strongly disagree to strongly agree)

¹Received behavioral information plus family health history-based obesity risk feedback based on one overweight parent

²Received behavioral information plus family health history-based obesity risk feedback based on two overweight parents

Table 2:

Final Multivariate Models for Lifestyle Guilt and Genetic Guilt after Backward Entry and Removal

Lifestyle Guilt
	df	F	p
FHH Information	2	6.50	0.002*
Education	2	1.78	0.17
Child BMI	1	2.12	0.15
Dieting history	1	2.07	0.15
Father in HH	1	3.09	0.081
Genetic causal beliefs	1	7.65	0.007*
Mother’s body shame and guilt	1	18.52	<.0001*
Genetic Guilt	1	3.00	0.086
Genetic Guilt
FHH Information	2	10.28	<.0001*
Education	2	2.84	0.062
Mother is white	1	5.12	0.025*
Mother’s BMI	1	5.89	0.017*
Dieting history	1	3.78	0.054
Genetic causal beliefs	1	1.86	0.17
Mother’s body shame and guilt	1	9.01	0.003*
Lifestyle guilt	1	2.25	0.14

^*p<.05

Influence of FHH Risk Information on Genetic Guilt (Hypotheses 2 and 2a)

FHH feedback condition influenced mothers’ reports of genetic guilt in the unadjusted model, F(2,145)=9.96, p<.0001. Mothers who received information about FHH risk implicating one overweight parent (i.e., the mother herself) reported feeling more guilt than those who received general behavioral obesity information and those who received FHH risk information implicating two parents, F(1,145)=8.22, p=.005 and F(1,145)=19.58, p<.0001 respectively. In the adjusted model, mothers’ genetic guilt was higher when mothers received FHH-based risk information, when they were white, had higher BMIs, and when they felt more shame and guilt about their own bodies.

Association between FHH Risk Information, Unhealthful Food Choice, and Lifestyle and Genetic Guilt (Hypotheses 3 and 4)

The mediation models for lifestyle and genetic guilt are illustrated in Figure 2. Bootstrapping revealed that the relationship between FHH risk information and both lifestyle and genetic guilt were mediated by unhealthful food choices; M_B=−0.10, 95% CI = −0.28 – −0.008 and M_B=−0.16, 95% CI = −0.42 – −0.03, respectively. The size and significance of the effect of FHH information on each type of guilt increased when the association of unhealthful food choice with guilt was statistically controlled. This pattern of effects is indicative of an indirect effect known as suppression (MacKinnon, Krull, & Lockwood, 2000). Controlling for the effect of a suppressor variable increases the predictive validity of the initial variable. Part of the effect of the FHH risk message on guilt is suppressed by the beneficial effect of the risk message on decreasing unhealthful food choices.

Figure 2:

Mediation Models Predicting Lifestyle and Genetic Guilt with Unstandardized Coefficients

*=p<.05; **=p<.01; ***=p<.001

Discussion

Pediatric organizations recommend that FHH approaches be used in risk stratification for common health conditions like obesity (Barlow, 2007). The current report demonstrates that use of FHH-based approaches for obesity risk provision can elicit guilt among mothers. This guilt may be especially strong when it relates to conferring genetic risk for obesity to one’s child, and when feedback suggests that it is the mother’s weight alone that confers this risk. The current work has also demonstrated that enacting healthier feeding behaviors can reduce feelings of guilt related to the child’s lifestyle, and also, unexpectedly, reduce guilt related to passing down genetic susceptibility for obesity.

The positive association between provision of FHH feedback and both types of guilt persisted even when other factors likely to influence guilt were included in the model. This suggests a robust association of FHH information provision with guilt. That FHH feedback increased mothers’ guilt about their children’s lifestyle behavior suggests that it may have reminded mothers about less-healthful elements of the shared family environment that influence mothers’ own weight and also confer risk to their child. The feedback indicating increased risk also may have caused mothers to critically evaluate their child’s lifestyle and their role in influencing that lifestyle.

In addition, lifestyle guilt was also predicted by the extent to which mothers attributed genetics as a cause of obesity. At first glance, it may seem odd that genetic causal beliefs were linked to increased lifestyle, rather than genetic, guilt. However, this association may indicate that mothers understand and consider the interactive nature of genes and behavior. As genetic risk increases, one’s behavior becomes all the more important. Lifestyle guilt was also predicted by mothers’ levels of shame and guilt about their own bodies. This may be due to the fact that guilt-proneness is considered to be a personality trait (Tangney, 1990), and is likely influential in both cases.

Obesity-related FHH risk provision had the most influence on guilt related to passing down genetic predisposition for obesity. As hypothesized, genetic guilt was highest among mothers who were told that their child’s risk was elevated because the child had one overweight parent (i.e., the mother). This was the case even though the numerical level of risk conveyed to mothers in this group was lower. This may have occurred because mothers could not share or diffuse responsibility between themselves and the child’s biological father. Because mothers were pinpointed as the source of their child’s health risk, they may have taken FHH feedback more personally. This could be important to consider when employing this feedback to encourage parental behavior change.

Mothers with higher BMIs felt more genetic guilt. This may be related to the level and degree of the obesity risk that mothers perceived they were passing down to their child. The finding that white mothers felt more genetic guilt is difficult to interpret but may be related to racial differences in body satisfaction and norms (Roberts, Cash, Feingold, & Johnson, 2006). Finally, increased genetic guilt among mothers with higher levels of body shame and guilt likely again reflects trait levels of guilt-proneness.

What is less clear is whether the propensity for FHH risk feedback to elicit guilt is purely negative in that it causes negative affect and potentially distress among mothers, or whether it is in part an influence pathway through which mothers may be prompted to improve child feeding behavior. The temporal order of data collection in the current study did not allow us to assess this possibility directly. However, the study did demonstrate that choosing fewer unhealthful foods for one’s child was associated with less guilt following that food choice episode. Because resisting unhealthful food choices even in a simulated, virtual exercise like this one was associated with reduction in guilt, it appears that food choices in real-world environments could offer mothers a way to compensate for or alleviate their guilt.

While we anticipated this outcome with respect to lifestyle guilt, we did not hypothesize that acting to reduce unhealthful food given to one’s child would be associated with reduced genetic guilt. This is because one’s genetic code is fixed, and therefore mothers could perceive that there are no direct avenues through which to ‘correct’ the source of that guilt. The current findings suggest that mothers appropriately consider their child’s risk holistically. That is, taking downstream action to reduce the health threat that is conveyed through genetic means seems to make mothers feel less guilty about passing down genetic susceptibility in the first place.

The current study had limitations. First, there were fewer mothers in the ‘1 overweight parent’ group than the other two groups. Although this reduces power to see effects, it reflects the distribution of parental couples (The & Gordon-Larsen, 2009). Additionally, assessment of BMI was based on self-report. Guilt was measured at post-test only. Participants were randomly assigned to condition and therefore differences in this outcome are expected to result primarily from the manipulation. Participants’ guilt responses may have also been influenced by their behavior within the buffet. The contribution of this behavior was assessed in analyses to better understand its influence. Finally, guilt measures were two-item measures created for this study. This is because no suitable measures existed, and the research questions forming the basis of this analysis were exploratory in nature. Future work in this area should develop and validate measures that explore the multiple facets of parental guilt and its psychological correlates.

The capacity for FHH obesity risk information to elicit feelings of guilt among mothers is an important process that until now has not been investigated. Further investigation is needed into the potential array of parental psychological and behavioral outcomes associated with guilt in this context; in particular, whether, how and under what conditions guilt leads to changes in feeding and other parenting behaviors. These issues are crucial to investigate because although guilt could result in beneficial parental behavior, the accompanying elicitation of aversive, negative emotions among parents is not a desirable route to this outcome. The ultimate goal should be to develop strategies for providing FHH-based risk feedback to parents in ways that both minimize negative reactions and enhance positive child feeding practices.