Mealtime family interactions in home environments of children with loss of control eating

Appetite 56 (2011) 587–593

Contents lists available at ScienceDirect

Appetite journal homepage: www.elsevier.com/locate/appet

Research report

Mealtime family interactions in home environments of children with loss of control eating§ Julia Czaja a,*, Andrea Sabrina Hartmann b, Winfried Rief b, Anja Hilbert b,c a

Department of Psychosomatic Medicine and Psychotherapy, University Hospital of Luebeck, Ratzeburger Allee 160, 23538 Luebeck, Germany Philipps University of Marburg, Department of Psychology, Gutenbergstr. 18, 25032 Marburg, Germany c Department of Psychology, University of Fribourg, Rue P.-A. de Faucigny 2, 1700 Fribourg, Switzerland b

A R T I C L E I N F O

A B S T R A C T

Article history: Received 5 October 2010 Received in revised form 20 January 2011 Accepted 24 January 2011 Available online 1 February 2011

Experimental and self-report studies have shown that parents have a strong influence on their normal or overweight children’s eating behavior, i.e. through parental feeding behavior or communication. Studies in children with loss of control (LOC) eating that have investigated this relationship are scarce, and ecologically valid observational studies are missing. This study examined family functioning at mealtimes in home environments in 43 families of a child with LOC eating and 31 families of a child without LOC eating; the children were 8–13 years old. Familial interactions, child eating behavior, and parental mealtime behavior were assessed using the Mealtime Family Interaction Coding System, observation of bite speed of the child, and self-report questionnaires. Less healthy patterns of communication (U = 201.53, p < .01) and interpersonal involvement (U = 207.54, p < .01) and more maladaptive overall family functioning (U = 233.52, p < .05) were observed but not self-reported in families of a child with LOC eating compared to those without LOC eating. Children with LOC eating (M = 4.73, SD = 1.88) ate faster than controls (M = 3.71, SD = 1.19; p < .05), with highest bite speed in a group with high recurrent LOC eating (p < .01). The results indicate that maladaptive patterns of family functioning during family mealtimes are present in LOC eating in children and are associated with the child’s eating behavior. Parent–child communication training should be tested as an intervention for children with LOC episodes. ß 2011 Elsevier Ltd. All rights reserved.

Keywords: Naturalistic test meal Family interactions Loss of control eating Children Ecological validity

Introduction Parental behavior is associated with the development of the eating behavior of their children (for review see Scaglioni, Salvioni, & Galimberti, 2008). Experimental studies (Fisher & Birch, 2000; Fisher, Mitchell, Smiciklas-Wright, & Birch, 2002) and self-reports (Klesges, Malott, Boschee, & Weber, 1986) have indicated that a number of problematic parent–child interaction patterns are associated with fruit intake, physical activity, and overweight in children. Whether families of children with overeating or binge eating display more negative interaction patterns during regular family mealtimes requires further exploration. Binge eating is the main criterion of binge eating disorder (BED; Diagnostic and Statistical Manual of Mental Disorders, DSM-IV-TR;

§ This research was part of a project supported by grant HI 1111/1-1 awarded to A. Hilbert from the German Research Foundation and by grant 01GP0491 from the German Ministry of Education and Research. * Corresponding author. E-mail addresses: [email protected], [email protected] (J. Czaja), [email protected] (A.S. Hartmann), [email protected] (W. Rief), [email protected] (A. Hilbert).

0195-6663/$ – see front matter ß 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.appet.2011.01.030

American Psychiatric Association, APA, 2000) and is defined as eating a large amount of food accompanied by a sense of loss of control (LOC) over eating. Although relatively few children fulfill the diagnostic criteria of BED (Glasofer et al., 2007; Levine, Ringham, Kalarchian, Wisniewski, & Marcus, 2006), in an expert interview, 9.3% of children ages 6–13 years reported LOC eating independent of the amount of food consumed (Tanofsky-Kraff et al., 2004). Therefore, recent criteria proposed for children of 12 years and younger therefore also focus on the loss of control rather than the amount of food consumed (Tanofsky-Kraff, Marcus, Yanovski, & Yanovski, 2008). LOC eating is associated with increased eating disorder and general psychopathology, overweight, and obesity in youth (Glasofer et al., 2007; Goldschmidt et al., 2008; Goossens, Braet, & Decaluwe, 2007; Hilbert & Czaja, 2009; Levine et al., 2006; Shapiro et al., 2007) and predicts further weight gain (Tanofsky-Kraff et al., 2009). A number of experimental studies have suggested that a child’s self-regulation of his or her feeding behavior, especially overeating, is influenced by the degree of parental control and restriction of food (Drucker, Hammer, Agras, & Bryson, 1999; Fisher & Birch, 1999, 2000). In the laboratory, children with obesity ate faster in the presence of their mother compared to when alone (Laessle, Uhl,

588

J. Czaja et al. / Appetite 56 (2011) 587–593

Lindel, & Muller, 2001). Parental encouragement to eat was also associated with the child’s relative weight (Klesges et al., 1983). While in children ages 3–7 years no relationship was found between specific parental practices during family meals and child overweight (Koivisto, Fellenius, & Sjoden, 1994), an observational study of family interactions during mealtimes in a naturalistic setting found that parents of overweight children ages 7–13 displayed less interpersonal involvement and more maladaptive control than parents of children who were not overweight (Moens, Braet, & Soetens, 2007). In children with LOC eating far less research has been conducted regarding familial interactions, parental mealtime behavior, and children’s eating behavior. A laboratory test-meal study showed that parents of children with LOC eating expressed more critical comments on shape, weight, and eating than parents of children without LOC eating, and these comments were predictive of energy intake during a snack-eating trial (Hilbert, Tuschen-Caffier, & Czaja, 2010). Results of retrospective risk factor studies on LOC eating in children (Czaja & Hilbert, 2008) and BED in women (Fairburn et al., 1998; Striegel-Moore et al., 2005) also indicated negative parent–child interactions before the onset of disturbed eating, including critical comments by the family on shape, weight, or eating, higher tension at mealtimes, or underinvolvement. Only retrospective and laboratory-based studies have focused on family interactions in childhood LOC eating; naturalistic observation studies that provide a concurrent and ecologically valid assessment of mealtime interactions in children with LOC eating are missing. The present study aims to examine parent–child interactions as well as children’s eating behavior during a mealtime in a naturalistic setting in families of a child with and without LOC eating. Based on the research reported above, especially Moens et al. (2007) and Hilbert et al. (2010), we hypothesized that overall family functioning (integration of all interaction dimension) would be less adaptive in families with a child with LOC eating compared to families without LOC eating children. In particular, we expected that interactions would be characterized by less emotional involvement and more maladaptive communication patterns in families of a child with LOC eating compared to families of a child without LOC eating. In addition, we hypothesized that children with LOC eating would eat faster than children without LOC eating, which is a diagnostic criterion of BED in DSM-IV-TR (APA, 2000), and would report a higher sense of LOC over eating before and after the dinner as is the relevant criterion in the child-specific diagnostic criteria (Tanofsky-Kraff et al., 2008). Extreme group analyses to each research question should clarify whether clinically significant forms of the disorder differ from a lowrecurrent LOC eating episodes form of the disorder that has shown to be of low clinical validity (Hilbert & Czaja, 2009) regarding family functioning.

participating parent. The main inclusion criterion for LOC+ children was at least one episode of LOC eating in the past 3 months; the members of the LOC group should not have had a current or lifetime eating disorder or symptoms of disordered eating (e.g., LOC eating or history of dieting). For both groups, further inclusion criteria were age (8–13 years) and sufficient German language skills of the child and participating parent. Exclusion criteria were compensatory behaviors (vomiting, excessive dieting or exercising, or use of laxatives or diuretics as asked in the ChEDE see below), psychotic disorders of the child (screening from the Diagnostic Interview of Mental Disorders in Children, K-DIPS; Unnewehr, Schneider, & Margraf, 2008) or parent, medical conditions with an enduring effect on eating behavior, treatment for being overweight, attendance at special schools for learning deficits, a planned move, or a commute of more than 30 min to the laboratory site. A total of 44 LOC+ and 30 LOC children were recruited through schools, and 21 LOC+ and 32 LOC children were recruited from the advertising campaign. In a first diagnostic session, informed assent and consent were obtained from the child and the participating parent, and diagnostic status was ascertained using the German version of the Eating Disorder Examination adapted for Children (ChEDE; Bryant-Waugh, Cooper, Taylor, & Lask, 1996; German version, Hilbert, Hartmann, & Czaja, 2008), a semi-structured interview with good reliability and validity. Height and weight were measured using calibrated instruments, body mass index (BMI, kg/m2) was calculated, and BMI standard deviation scores were computed. Further, seven children were excluded following the diagnostic visit (no matching partner: 2; not interested: 2; no LOC: 3), leaving a total sample of 60 LOC+ children and 60 matched LOC children for participation. Participants were asked if in addition to being in the main laboratory study (Hilbert et al., 2010), they wanted to participate in a study including videotaping a family dinner at the family’s home 7 days later in order to get a better understanding of problematic eating behavior and associated factors in a naturalistic setting. For incentive, participants were offered s15. A total of 43 LOC+ children and 31 LOC children and their families agreed to participate, the remaining declining due to a feeling of intrusion into their privacy at home.

Methods

Measures

Recruitment and sample

Expert rating of family functioning. The Adapted Mealtime Family Interaction Coding System (MICS; Dickstein, Hayden, Schiller, Seifer, & San Antonio, 2004; Hayden et al., 1998) is an observational coding system for the assessment of family functioning during an unstructured, naturalistic situation. It is adapted from the McMaster Structured Interview of Family Functioning (McSIFF) and based on the McMaster Model of Family Functioning (Epstein, Baldwin, & Bishop, 1983). The dimensions of the MICS are listed and described in Table 1. Except for the roles dimension, which was not coded in the present study, all dimensions were rated on a 7point Likert scale ranging from 1 (very unhealthy) to 7 (very healthy). Ratings of 5 and greater indicate adequate-to-good functioning while those below 5 are considered unhealthy. The

Children with (LOC+) and without (LOC) LOC eating at ages 8– 13 were recruited from local schools and via an advertising campaign (public notices and newspaper articles) for participation in a larger project about LOC eating in children (for methodological detail see Hilbert & Czaja, 2009). The present sample represents a subsample of the participants in the laboratory test-meal intake study by our workgroup (Hilbert et al., 2010). The German Psychological Society’s Ethics Committee approved the present study. Inclusion and exclusion criteria for both recruitment avenues were checked in a telephone interview with the child and

Procedure At the families’ homes, a camera was set up by a research assistant before the beginning of dinner. The child was asked to rate his or her mood immediately before dinner (see Measures). The research assistant advised the family to have dinner as they would when not being recorded and left the room. After dinner, a family member came for the research assistant and the recording was stopped. The child then was asked to rate his or her current mood and the representativeness of the dinner.

J. Czaja et al. / Appetite 56 (2011) 587–593

589

Table 1 Scales and respective items of the Adapted Mealtime Family Interaction Coding System. Dimensiona

Description

Behavioral indicators

Task accomplishment

Structure of the meal and manner of resolution of disruptions

Communication Affect management

Exchange of interactions across family members Relevance and appropriateness of emotion expressed during the meal

Interpersonal involvement

Interest in and placement of value on each other’s activities and concerns Expression and maintenance of standards for the behavior of its members Overall sense of the family, including clinical impression

Safe, organized, and appropriate preparation, handling of disruptions, mealtime routine, attendance to individual needs Clear to masked conversation, direct to indirect conversation Production of affect (i.e. appropriateness, intensity) and responsiveness to others’ affect (sensitivity, contingency, congruence) Lack of or symbiotic to empathic involvement, lack of privacy

Behavior control Overall family functioning

Four control styles: chaotic, laissez-faire, rigid, flexible Integration of observations of family functioning

Note. Adapted Mealtime Interaction Coding System, MICS (Dickstein et al., 1994; Hayden et al., 1998). All dimensions rated on a 7-point Likert scale ranging from 1 (very unhealthy) to 7 (very healthy). a The roles dimension was not coded in this study.

MICS reliably discriminates between family interactions in children with chronic illnesses and healthy controls and has been applied to mealtime situations (Jacobs & Fiese, 2007; Janicke, Mitchell, & Stark, 2005; Moens et al., 2007). A rater blind to the hypotheses rated videotapes of the dinner from beginning to end. Interrater reliability was determined on the basis of a second rater’s ratings of a random subset of 28.4% (21/74) of the videotapes. Intraclass correlations (ICCs) for the dimensions were almost perfect (.96  ICC  .97; Landis & Koch, 1977). Self-rating of family functioning. The Family Assessment Device (FAD; Epstein et al., 1983; German version, Klann, Hahlweg, & Heinrichs, 2003) is a self-report questionnaire based on the McMaster Model of Family Functioning that was completed by the participating parent during the diagnostic session. The FAD contains seven subscales: a general functioning scale, which assesses overall health and pathology of the family, and a subscale for each of the six dimensions of the model: problem solving, communication, roles, affective responsiveness, affective involvement, and behavior control. The FAD consists of 53 items rated on a 4-point Likert scale with higher scores indicating poorer functioning. The internal consistencies of the FAD scales are good; adequate test– retest reliability and discriminant validity have been demonstrated. Child eating behavior. As a behavioral indicator of LOC eating, bite speed was determined. Recorded bites were counted and the number was divided by the duration of food consumption of the child (in minutes; Drabman, Cordua, Hammer, Jarvie, & Horton, 1979). Following Fisher (2007), all bites of solid food that were clearly discernable were counted. Taking multiple bites nonstop (i.e., nibbling) was counted as two bites. Picking up crumbs or eating tidbits was only counted when the quantities of foods were substantial, that is, when these bits of food were or could have been eaten with a fork. In a one-item self-report measure, the child selfrated the sense of LOC before and after the dinner on a scale of 1 (very slightly to not at all) to 5 (extremely). Context variables. The research assistant recorded the duration of the dinner (in min), number of participating family members, and location of the dinner (kitchen, dining room, elsewhere). Control variables. Since emotion regulation deficits have been shown in children with LOC eating (Czaja, Rief, & Hilbert, 2009), mood was used as a control variable. The child was asked to answer six mood items (distressed, irritable, afraid, determined, enthusiastic, alert) of the Positive and Negative Affect Schedule for Children (PANAS-C; Joiner, Catanzaro, & Laurent, 1996; Laurent, Catanzaro, & Joiner, 1999) before and after dinner. The items were selected on the basis of their factor loading on either the positive or negative affect scale. The PANAS-C is a measure with high reliability and validity. Additionally, three eating-related control variables were rated (hunger, satiety, and feeling full) as was done in other studies (Hilbert et al., 2010). Furthermore, the child was

asked to rate on a 5-point Likert scale the representativeness of his or her own eating behavior at the family dinner, his or her overall behavior, the behavior of family members, and the kinds of food eaten (with 1 being very slightly to not at all representative and 5 being extremely representative), and the taste of the food (with 1 being good and 5 being bad). Classificatory variable for extreme group analyses. For extreme group analyses of differences family functioning and child eating behavior between a subthreshold and a clinical cluster of children with LOC eating were compared. Clusters were derived by hierarchical cluster analysis based on frequency of LOC over eating and associated eating disorder psychopathology using DSMIV-TR diagnostic criteria of BED and undue influence of weight or shape on self-evaluation (APA, 2000) as well as provisional childspecific research criteria of BED (Tanofsky-Kraff et al., 2008; for a detailed description of hierarchical cluster analysis and clusters see Hilbert & Czaja, 2009). Statistical analyses All analyses were conducted using SPSS 15.0. An a priori power calculation with G*Power (Erdfelder, Faul, & Buchner, 1996) determined a total sample size of 62 to detect medium-sized effects in the below-mentioned multivariate analysis of variance (MANOVA), Power (1  b) = .80, a < .05, response variables: 6. Preliminary analyses of control variables included a two-way MANOVA with the factor group (LOC+, LOC) and the repeated measures factor time (pre-, post-dinner) with subsequent univariate analyses of variance (ANOVAs) to examine group differences in change of subjective ratings over time, and Mann–Whitney U tests to investigate group differences in ratings of representativeness. The preliminary analyses of context variables incorporated analyses of group differences in duration of dinner in an ANOVA, in number of participating family members in a Mann–Whitney U test, and in location of dinner in a Chi-square test. Variables that yielded significant group differences were controlled for in further analyses. If they account for group differences, this will be reported. The expert rating of family functioning was analyzed with Mann–Whitney U tests for the dimension of the MICS, and with a MANOVA with subsequent ANOVAs in case of significance over the dimensions of the FAD. Child eating behavior was analyzed in an ANOVA for bite speed, and in Mann–Whitney U tests for sense of LOC over eating before and after the dinner. To determine differences between children with clinical and subthreshold levels of LOC eating, analyses were repeated with the clusters described above. To analyze group differences in family functioning and child eating behavior, Mann–Whitney U tests were computed for MICS and FAD scales as well as bite speed. In order to control for pubertal maturation (Shomaker et al., 2010), the variables sex, weight, and

J. Czaja et al. / Appetite 56 (2011) 587–593

590

Family functioning

age (as proxy) were introduced into the above analyses (ANCOVAs). For nonparametric analyses of the MICS scales above parametric multivariate analyses with and without inclusion of covariates as well as correlations between covariates and dependent variables (Kendall-Tau-b coefficients) were presented. As effect size for ANOVAs, partial h2 was reported (interpretation according to Cohen, 1988; small: .01  partial h2 < .09; medium: .09  partial < .25, large: partial h2  .25). For nonparametric tests, r was reported and interpreted according to Cohen (1992); small effect: .10  r < .30; medium: .30  r < .50; large: r  .50). A twotailed a level of .05 was applied to all statistical tests.

Expert rating. Figure 1 displays the average ratings on each MICS dimension for families of a LOC+ child compared to families of a LOC child. Families with LOC+ children showed less healthy patterns of interpersonal involvement, U = 207.54; p < .01, r = .38, less adequate communication patterns, U = 201.53; p < .01, r = .41, and more maladaptive overall family functioning, U = 233.52; p < .05, r = .33. Extreme group analyses did not show differences between high and low recurrent LOC eating clusters in Mann–Whitney U tests (p > .05). Self-rating. A multivariate analysis revealed no significant main effect of group for FAD scales, F(6, 67) = 2.02, p = .11, h2 = .24. Extreme group analyses did not yield significant differences in Mann–Whitney U tests in FAD scales (p > .05).

Results Sample characteristics According to the ChEDE, mean LOC eating frequency over the past 3 months was 9.12 episodes (SD = 16.67; range 1–81). All children reported at least one LOC eating episode in the past 3 months. Sample characteristics are presented in Table 2. The groups did not differ in weight status. Of the total sample, 20 children (14.8%) were classified as overweight (90th percentile), 14 (10.4%) as obese (97th percentile), and a further 6 children (4.4%) as extremely obese (99.5th percentile).

Child eating behavior LOC+ children ate faster than LOC children F(1, 72) = 5.23; p < .05, h2 = .09 (see Fig. 2). The extreme group analysis yielded a significant cluster effect, U = 9.01; p < .01, r = .34, with the cluster with high recurrent LOC eating (M = 32.30, SD = 4.87) showing a greater bite speed than the cluster with low recurrent LOC eating (M = 21.37, SD = 9.99). In contrast with the hypothesis, Mann–Whitney U tests did not reveal significant group differences for the children’s reported sense of LOC over eating (before dinner: U = 247.51; p = .18, r = .07; after dinner: U = 200.02; p = .75, r = .09). This finding could be due to the low incidence of sense of loss of control in both groups. Only one child from the LOC+ group and two children from the LOC group reported a sense of LOC over eating before the dinner, and one child from the LOC group reported LOC over eating after the dinner (a rating of 4 [medium to extreme] on the 5-point Likert scale).

Preliminary analyses Control variables. A significant main effect of time, F(9, 65) = 30.89; p < .001, and an interaction effect Group  Time, F(9, 65) = 2.19, p < .05, but no main group effect, F(9, 65) = 0.86, p = .56, were detected in a MANOVA. Subsequent ANOVAs of subjective ratings revealed less hunger, F(1, 72) = 29.27, p < .01, and more satiety, F(1, 72) = 17.63, p < .001, joy, F(1, 72) = 8.69 p < .01, and calmness, F(1, 72) = 4.83, p < .05, after compared to before dinner. In univariate analyses no significant interaction effects were observed (p > .05). By trend, however, the LOC+ group became less irritable and the LOC group more irritable from before to after dinner (p = .09). No group differences were observed for ratings of representativeness (eating behavior child: U = 499.00, behavior child: U = 506.00, behavior others: U = 408.00, food: U = 532.00, taste U = 498.00, all p > .05). Context variables. The mean duration of dinners was 14.00 min (SD = 6.57). The number of family members attending the dinner ranged from 2 to 7 persons (M = 3.43, SD = 1.20). Most families were seated in either the kitchen (37/74; 50.0%) or the dining room (35/74; 47.3%), one family sat outside on the terrace (1/74; 1.4%), and another family in the living room (1/74; 1.4%). Groups did not differ in context variables (dinner duration: F(1, 72) = 0.42, p = .75; number of members: U = 607.50; p = 45, and location of dinner: U = 646.00, p = .80).

Control for pubertal maturation In order to control for pubertal maturation in the analyses conducted, sex, weight, and age were included as covariates in the analysis of variance bite speed and FAD scales. While inclusion of the covariates led to a decrease of significance in the difference of bite speed (F(4, 69) = 9.11, p = .07), no difference in FAD scales was observed after inclusion (F(9, 64) = 1.89, p = .11) which is comparable with the findings without covariates. Concerning the extreme group analyses, inclusion of the covariates did not change the findings reported in the manuscript (bite speed: F(4, 20) = 6.92, p < .05; FAD: F(9, 15) = 2.14, p = .13). With regard to the MICS scales groups did trendwise differ in multivariate analyses (F(6, 67) = 2.21, p < .10), with univariate analyses illustrating group differences mentioned in nonparametric

Table 2 Sample characteristics of children with (LOC+) and without (LOC) loss of control eating. LOC+ (N = 43)

Sex, female SES of mother Low High

Age, years Body mass index BMI SDS

LOC (N = 31)

Test

n

%

n

%

x2 (df = 1)

p

26

60.5

16

51.6

0.57 0.38

.45 .54

28 15

65.1 34.9

18 13

58.1 41.9

M

SD

M

SD

F(1, 72)

10.58 21.68 1.03

1.44 4.61 1.06

11.00 22.97 1.30

1.44 4.53 0.93

1.53 1.45 1.21

.22 .23 .28

Note. SES, socioeconomic status: low = no school degree, or degree with less than 13 years of school; high = degree with 13 years of school or university degree; BMI SDS, body mass index (kg/m2) standard deviation scores.

[()TD$FIG]

J. Czaja et al. / Appetite 56 (2011) 587–593

591

Figure 1. Means and standard deviations of the Mealtime Interaction Coding System (MICS) ratings for families of a child with loss of control eating (LOC+) compared to families of a child without LOC eating (LOC). Range of possible scores of dimensions ranging from 1 (very unhealthy) to 7 (very healthy). Significant findings for communication and interpersonal involvement remain unchanged when controlled for multiple testing according to Bonferroni–Holm. Overall family functioning remains by trend significant. Bars denote  2 standard errors. *p < .05, **p < .01.

tests (interpersonal involvement p < .01; communication p < .01; overall family functioning p < .05). Inclusion of covariates led to a non-significant multivariate group difference (F(9, 64) = 1.83, p = .12). Univariate analyses, however, highlighted the same significant dimensions (interpersonal involvement p < .01; communication p < .01; overall family functioning p < .05). Multivariate extreme group analyses with included covariates did not yield significant between group differences (F(9, 15) = 1.75 p = .18). Additionally, MICS scales did not significantly and substantially correlate with the variables (j.030j  t  j.150j, all p > .05), except for the small association of behavior control and weight (t = .27, p > .01). Discussion In this naturalistic study of family interactions and children’s eating behavior families with children with LOC eating engaged in less healthy communication and dysfunctional interpersonal involvement, and they displayed a more maladaptive overall family functioning at mealtime than families with children without LOC eating. Retrospective studies in adults also found parental underinvolvement to be associated with LOC eating

[()TD$FIG]

Figure 2. Higher bite speed (bites/min) in children with LOC eating episodes (LOC+) compared to children without LOC eating episodes (LOC). Bars denote  2 standard errors. *p < .05, **p < .01.

(Ackard, Neumark-Sztainer, Story, & Perry, 2006; Fairburn et al., 1998). In the present study, both groups were rated in the ‘‘healthy’’ range for the dimension task accomplishment, that is, structure of the meal and manner of resolution of disruptions. Families had the chance to plan the dinner in advance, possibly eliminating common disturbances, which could account for the lack of difference between groups. In contrast to the observations in expert-rated family interactions, no group differences were observed in the participating parent’s self-reported family interactions. This is in line with findings that objective and subjective appraisals often diverge in eating-related studies, for example, on evaluation of exercise (Jakicic, Polley, & Wing, 1998) or weight and BMI (Gorber, Tremblay, Moher, & Gorber, 2007), but also in the assessment of LOC eating episodes in youth with self-report measures and expert ratings (Field, Taylor, Celio, & Colditz, 2004; Tanofsky-Kraff et al., 2003). Maladaptive interaction patterns during mealtime are relevant given the fact that in children, the energy intake is mainly influenced by the family context, and therefore contribute to a specific high-risk environment and higher likelihood to booster the development of LOC eating in children. Different pathways from dysfunctional interaction patterns to LOC eating in children are imaginable. Several laboratory studies have shown that parental restriction and control are associated with lower later self-control in children (Drucker et al., 1999; Fisher & Birch, 1999), thus parent–child interactions might also lead to the development of loss of control against a background of a generally deficient selfcontrol. Another thinkable pathway could be that overweight has also shown to be associated with parental involvement and control (Moens et al., 2007). Overweight children in turn have been reported to be more likely to show LOC eating. Against the background of the present findings and associations reported earlier, longitudinal studies should illustrate pathways of the impact of familial interactions on LOC eating in children. No difference in subjective ratings of sense of LOC before and after the dinner was found. This finding needs to be interpreted carefully as the frequency of LOC eating is based on a single answer on a questionnaire rather than on detailed expert interview as was performed to form the groups. The laboratory test-meal study of our group (Hilbert et al., 2010) also found only a marginally

592

J. Czaja et al. / Appetite 56 (2011) 587–593

significant difference in sense of LOC over eating between the groups during a regular parent–child test meal. In the present study, children with LOC eating showed a greater bite speed than children without LOC eating, thus scoring higher in a behavioral diagnostic criterion for BED (DSM-IV-TR, APA, 2000). The fact that children from the high recurrent LOC eating cluster showed an even greater bite speed than those from the low recurrent LOC eating cluster suggests that bite speed can be used as a valuable behavioral indicator, especially in children with high recurrent LOC eating. It could be argued that behavioral indicators for LOC eating are more useful in naturalistic compared to laboratory settings (Hilbert et al., 2010). Recently, it was shown that children with LOC eating episodes experience a certain ‘‘numbing out’’ while eating that is a certain dissociative quality to the binge eating episodes (Tanofsky-Kraff et al., 2007). A new study, additionally, showed that youth with LOC eating have more trouble recalling amount of carbohydrates and dessert foods consumed (Wolkoff et al., 2010). This finding together with the higher bite speed found despite the missing difference in feeling of LOC in children with LOC eating might underline the validity of this feeling of numbness reported. It would be worthwhile to investigate family interactions in regard to its impact on the experience of this feeling. Interestingly, proxy variables of pubertal maturation did not account for group differences in family interaction dimensions or bite speed. Influence of these variables on dependent variables could have been expected, in particular for bite speed, since, e.g. Shomaker et al. (2010) have shown that especially sex significantly influences food intake. It could be argued, however that participants of the present study were mainly in pre-pubertal stage, and that patterns of food intake cannot easily be translated in family interactions during mealtime and bite speed of the child. The following strengths and limitations should be noted when interpreting the present results. To examine parent–child interactions during a family dinner we used observational coding in addition to self-report and rated it using established coding systems with high interrater reliability. Furthermore, the study assessed behavior of parents and children at their homes and should therefore be considered ecologically valid. However, the presence of a camera might have influenced the participants’ behavior toward more socially desirable behavior and selfconsciousness which is an issue raised by several studies recording mealtimes (Moens et al., 2007; Haycraft & Blissett, 2008; Kiser, Medoff, Black, Nurse, & Fiese, 2010). It needs to be considered though, that Gardener found in his review no indications for the fact that the presence of an observer, i.e. a camera, does interfere with the natural interactions but that it is rather structural or artificial settings that lead to non-representative interactions (Gardener, 2000). Nevertheless, tendencies toward socially desirable behavior and self-consciousness could potentially be eliminated if repeated mealtimes are recorded, which was beyond the scope of this initial study but should be considered in a replication. Nevertheless, ratings of children in the present study and ratings of parents in a study with a comparable design (Jacobs & Fiese, 2007) indicated high representativeness of behaviors of all family members. A further limitation is that only a few children reported LOC episodes during dinner. However, given that children, as adults (APA, 2000), often experience LOC eating episodes in secrecy (Tanofsky-Kraff et al., 2008), this is not surprising. Nevertheless, family functioning can affect eating behavior of the child in general and also influencing eating during the LOC eating episode. Nevertheless, the children’s ratings of the representativeness of their family members’ behavior and also of their own behavior suggested that a typical dinner was recorded. The findings of the present study have important implications for the understanding of maintenance of LOC eating in children. Unhealthy family interactions during mealtimes, characterized by

dysfunctional communication and lack of involvement, may reduce the frequency of family mealtimes so that children more often eat on their own, which in turn raises the risk for LOC eating episodes (Tanofsky-Kraff et al., 2007). Therefore, treatment studies should examine the effectiveness of parent–child communication trainings in children with LOC eating episodes. Given the crosssectional design of the study, we can only assume that the identified interaction styles have a prognostic value for the development of LOC eating in children, which should be verified in longitudinal studies. References Ackard, D. M., Neumark-Sztainer, D., Story, M., & Perry, C. (2006). Parent–child connectedness and behavioral and emotional health among adolescents. American Journal of Preventive Medicine, 30, 59–66. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author. Bryant-Waugh, R. J., Cooper, P. J., Taylor, C. L., & Lask, B. D. (1996). The use of the eating disorder examination with children. A pilot study. International Journal of Eating Disorders, 19, 391–397. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Mahwah, NJ: Erlbaum. Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159. Czaja, J., & Hilbert, A. (2008). Psychosoziale Risikofaktoren von Essanfa¨llen im Kindesalter. In P. Warschburger, W. Ihle, & G. Esser (Hrsg.), Seelisch gesund von Anfang an (S. 253). Potsdam: Universita¨tsverlag. Czaja, J., Rief, W., & Hilbert, A. (2009). Emotion regulation and binge eating in children. International Journal of Eating Disorders, 42, 356–362. Dickstein, S., Hayden, L., Schiller, M., Seifer, R., & San Antonio, W. (1994). Providence Family Study. Family mealtime interaction coding system. Unpublished coding manual. East Providence, RI: Brown University School of Medicine: Bradley Hospital. Bradley Research Center. Drabman, R. S., Cordua, G. D., Hammer, D., Jarvie, G. J., & Horton, W. (1979). Developmental trends in eating rates of normal and overweight preschool children. Child Development, 50, 211–216. Drucker, R. R., Hammer, L. D., Agras, W. S., & Bryson, S. (1999). Can mothers influence their child’s eating behavior? Journal of Developmental and Behavioral Pediatrics, 20, 88–92. Epstein, N., Baldwin, L., & Bishop, D. (1983). The McMaster family assessment device. Journal of Marital and Family Therapy, 9, 171–180. Erdfelder, E., Faul, F., & Buchner, A. (1996). GPOWER. A general power analysis program. Behavior Research Methods, Instruments, & Computers, 28, 1–11. Fairburn, C. G., Doll, H. A., Welch, S. L., Hay, P. J., Davies, B. A., & O’Connor, M. E. (1998). Risk factors for binge eating disorder. A community-based, case–control study. Archives of General Psychiatry, 55, 425–432. Field, A. E., Taylor, C. B., Celio, A., & Colditz, G. A. (2004). Comparison of self-report to interview assessment of bulimic behaviors among preadolescent and adolescent girls and boys. International Journal of Eating Disorders, 35(1), 89–92. Fisher, J. O. (2007). Effects of age on children’s intake of large and self-selected food portions. Obesity, 15, 403–412. Fisher, J. O., & Birch, L. L. (1999). Restricting access to palatable foods affects children’s behavioral response, food selection, and intake. American Journal of Clinical Nutrition, 69, 1264–1272. Fisher, J. O., & Birch, L. L. (2000). Parents’ restrictive feeding practices are associated with young girls’ negative self-evaluation of eating. Journal of the American Dietetic Association, 100, 1341–1346. Fisher, J. O., Mitchell, D. C., Smiciklas-Wright, H., & Birch, L. L. (2002). Parental influences on young girls’ fruit and vegetable, micronutrient, and fat intakes. Journal of the American Dietetic Association, 102, 58–64. Gardener, F. (2000). Methodological issues in the direct observation of parent–child interaction. Do observational findings reflect the natural behavior of participants? Clinical Child and Family Psychology Review, 3(3), 185–198. Glasofer, D. R., Tanofsky-Kraff, M., Eddy, K. T., Yanovski, S. Z., Theim, K. R., Mirch, M. C., et al. (2007). Binge eating in overweight treatment-seeking adolescents. Journal of Pediatric Psychology, 32, 95–105. Goldschmidt, A. B., Jones, M., Manwaring, J. L., Luce, K. H., Osborne, M. I., Cunning, D., et al. (2008). The clinical significance of loss of control over eating in overweight adolescents. International Journal of Eating Disorders, 41, 153–158. Goossens, L., Braet, C., & Decaluwe, V. (2007). Loss of control over eating in obese youngsters. Behavior Research and Therapy, 45, 1–9. Gorber, S. C., Tremblay, M., Moher, D., & Gorber, B. (2007). A comparison of direct vs. self-report measures for assessing height, weight and body mass index. A systematic review. Obesity Reviews, 8, 307–326. Haycraft, E. L., & Blissett, J. M. (2008). Maternal and paternal controlling feeding practices. Reliability and relationships with BMI. Obesity, 16(7), 1552–1558. Hayden, L., Schiller, M., Dickstein, S., Seifer, R., Sameroff, A., & Miller, K. (1998). Levels of family assessment I. Family, marital, and parent–child interaction. Journal of Family Psychology, 12, 7–22. Hilbert, A, & Czaja, J. (2009). Binge eating in primary school children. Towards a definition of clinical significance. International Journal of Eating Disorders, 42, 235–243.

J. Czaja et al. / Appetite 56 (2011) 587–593 Hilbert, A., Hartmann, A. S., & Czaja, J. (2008). Child Eating Disorder Examination Questionnaire. Evaluation der deutschsprachigen Version des Esssto¨rungsfragebogens fu¨r Kinder. Klinische Diagnostik und Evaluation, 1(4), 447–463. Hilbert, A., Tuschen-Caffier, B., & Czaja, J. (2010). Eating behavior and familial interactions of children with loss of control eating. A laboratory test meal study. American Journal of Clinical Nutrition, 91, 510–518. Jacobs, M. P., & Fiese, B. H. (2007). Family mealtime interactions and overweight children with asthma. Potential for compounded risks? Journal of Pediatric Psychology, 32, 64–68. Jakicic, J. M., Polley, B. A., & Wing, R. R. (1998). Accuracy of self-reported exercise and the relationship with weight loss in overweight women. Medicine and Science in Sports and Exercise, 30, 634–638. Janicke, D. M., Mitchell, M. J., & Stark, L. J. (2005). Family functioning in school-age children with cystic fibrosis. An observational assessment of family interactions in the mealtime environment. Journal of Pediatric Psychology, 30, 179–186. Joiner, T., Catanzaro, S., & Laurent, J. (1996). Tripartite structure of positive and negative affect, depression, and anxiety in child and adolescent psychiatric patients. Journal of Abnormal Psychology, 105, 401–409. Kiser, L. J., Medoff, D., Black, M. M., Nurse, W., & Fiese, B. H. (2010). Family Mealtime QSort. A measure of mealtime practices. Journal of Family Psychology, 24(1), 92–96. Klann, N., Hahlweg, K., & Heinrichs, N. (2003). Diagnostische Verfahren fu¨r die Beratung. Materialien zur Diagnostik und Therapie in der Ehe-, Familien- und Lebensberatung. Go¨ttingen: Hogrefe. Klesges, R. C., Coates, T. J., Brown, G., Sturgeon-Tillisch, J., Moldenhauer-Klesges, L. M., & Holzer, B. (1983). Parental influences on children’s eating behavior and relative weight. Journal of Applied Behavior Analysis, 16, 371–378. Klesges, R., Malott, J., Boschee, P., & Weber, J. (1986). The effects of parental influences on children’s food intake, physical activity, and relative weight. International Journal of Eating Disorders, 5, 335–346. Koivisto, U. K., Fellenius, J., & Sjoden, P. O. (1994). Relations between parental mealtime practices and children’s food intake. Appetite, 22, 245–257. Laessle, R. G., Uhl, H., Lindel, B., & Muller, A. (2001). Parental influences on laboratory eating behavior in obese and non-obese children. International Journal of Obesity and Related Metabolic Disorders, 25, 60–62. Landis, J., & Koch, G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159–174. Laurent, J., Catanzaro, S., & Joiner, T. (1999). A measure of positive and negative affect for children. Scale development and preliminary validation. Psychological Assessment, 11, 326–338. Levine, M. D., Ringham, R. M., Kalarchian, M. A., Wisniewski, L., & Marcus, M. D. (2006). Overeating among seriously overweight children seeking treatment. Results of the

593

children’s eating disorder examination. International Journal of Eating Disorders, 39, 135–140. Moens, E., Braet, C., & Soetens, B. (2007). Observation of family functioning at mealtime. A comparison between families of children with and without overweight. Journal of Pediatric Psychology, 32, 52–63. Scaglioni, S., Salvioni, M., & Galimberti, C. (2008). Influence of parental attitudes in the development of children eating behaviour. British Journal of Nutrition, 99, S22–S25. Shapiro, J. R., Woolson, S. L., Hamer, R. M., Kalarchian, M. A., Marcus, M. D., & Bulik, C. M. (2007). Evaluating binge eating disorder in children. Development of the children’s binge eating disorder scale (C-BEDS). International Journal of Eating Disorders, 40, 82–89. Shomaker, L. B., Tanofsky-Kraff, M., Savastano, D. M., Kozlosky, M., Columbo, K. M., Wolkoff, L. E., et al. (2010). Puberty and observed energy intake. Boy, can they eat. American Journal of Clinical Nutrition, 92, 123–129. Striegel-Moore, R. H., Fairburn, C. G., Wilfley, D. E., Pike, K. M., Dohm, F. A., & Kraemer, H. C. (2005). Toward an understanding of risk factors for binge-eating disorder in black and white women. A community-based case–control study. Psychological Medicine, 35, 907–917. Tanofsky-Kraff, M., Goosens, L., Eddy, K. T., Righam, R., Goldschmidt, A., Yanovski, S. Z., et al. (2007). A multisite investigation of binge eating behaviors in children and adolescents. Journal of Consulting and Clinical Psychology, 75, 901–913. Tanofsky-Kraff, M., Marcus, M., Yanovski, S. Z., & Yanovski, J. A. (2008). Loss of control eating disorder in children age 12 years and younger. Proposed research criteria. Eating Behavior, 9, 360–365. Tanofsky-Kraff, M, Morgan, C. M., Yanovski, S. Z., Marmarosh, C., Wilfley, D. E., & Yanovski, J. A. (2003). Comparison of assessments of children’s eating-disordered behaviors by interview and questionnaire. International Journal of Eating Disorders, 33(2), 213–224. Tanofsky-Kraff, M., Yanovski, S. Z., Schvey, N. A., Olsen, C. H., Gustafson, J., & Yanovski, J. A. (2009). A prospective study of loss of control eating for body weight gain in children at high risk for adult obesity. International Journal of Eating Disorders, 42, 26–30. Tanofsky-Kraff, M., Yanovski, S. Z., Wilfley, D. E., Marmarosh, C., Morgan, C. M., & Yanovski, J. A. (2004). Eating-disordered behaviors, body fat, and psychopathology in overweight and normal-weight children. Journal of Consulting and Clinical Psychology, 72, 53–61. Unnewehr, S., Schneider, S., & Margraf, J. (2008). Kinder-DIPS - Diagnostisches Interview bei psychischen Sto¨rungen im Kindes- und Jugendalter. Go¨ttingen: Hogrefe. Wolkoff, L. E., Tanofsky-Kraff, M., Shomaker, L. B., Kozlosky, M., Columbo, K. M., Elliott, C. A., et al. Self-reported vs. actual energy intake in youth with and without loss of control eating. Eating Behaviors (epub ahead of print).