PEDIATRICS Vol. 106 No. 6 December 2000, pp. 1429-1435

Improving Preschoolers' Self-Regulation of Energy Intake

Susan L. Johnson, PhD

From the Department of Pediatrics, The Children's Hospital of Denver, Denver, Colorado.

	ABSTRACT

Top Abstract Methods Procedures Results Discussion Conclusion References

Background.  Children exhibit individual differences in their ability to self-regulate energy intake. Feeding strategies that focus on external signals, like the time of day or amount of food left on a plate, tend to diminish children's ability to respond to internal cues of hunger and fullness.

Objectives.  We investigated whether children could be taught to focus on internal cues of hunger and satiety, and consequently improve their self-regulation of energy intake. We explored whether parents' eating behaviors and adiposity were related to their children's self-regulation skills and adiposity.

Design.  In a pretest and posttest design, preschoolers participated in single-meal protocols to assess their individual ability to self-regulate food intake. During a 6-week intervention period, children took part in individual and group activities designed to help them recognize internal cues. Parents completed questionnaires regarding adult dietary restraint and disinhibition.

Results.  At baseline, we found a large individual variability in children's regulation: some children overate, some regulated accurately, and others underate. At baseline, children's eating related to their adiposity and to mothers' disinhibition: heavier children and children whose mothers' reported difficulty controlling food intake showed less evidence of self-regulation. Both overeaters and undereaters responded to the intervention, improving their ability to self-regulate, and children's eating was no longer significantly related to mother's eating.

Conclusions.  Children's disregulated energy intake is related to mothers' weight status and mothers' perceived control over eating. Cues can be provided that help children to focus on internal signals and improve their ability to self-regulate energy intake.  Key words:  children, compensation, food intake, parenting, internal cues.

The incidence of pediatric feeding disorders has grown remarkably over the last 2 decades with both childhood obesity and nonorganic failure to thrive being included as primary public health concerns.^1-3 The dramatic increase in incidence of childhood obesity has been attributed to pervasive environmental changes. Included among the many factors related to this shift are: 1) an exponential increase in food variety, particularly highly palatable, energy-dense foods; 2) increases in restaurant and fast-food dining; and 3) a decline in family meals.

During the same period that childhood obesity rates have risen, nonorganic failure to thrive has continued to be a public health concern. Approximately 5% to 10% of children 2 to 5 years old exhibit poor growth⁴ that is unrelated to an organic cause. Pediatric undernutrition is associated with an array of health problems that are far-reaching, including decreased immune function, diminished physical activity, and long-term impairments in cognition and social development.^5-8 The causes of nonorganic failure to thrive have been intensely investigated, yet the relative importance of the various mechanisms studied has not been clearly agreed on. Environmental variables related to nonorganic failure to thrive include poverty, teenaged pregnancy, emotional deprivation, child neglect and abuse, interactional disorders between parents and children, and the current social climate with respect to parenting.^9-11

Paradoxically, there are consistent data to support that in noncontrolling, noncoercive conditions, children have the ability to self-regulate the amount of food and energy consumed. Birch and colleagues have performed numerous laboratory studies revealing that children can adjust energy intake, both in short-term, single-meal protocols and over longer 24-hour periods, to reflect the changes in energy density of the diet.^12-14 However, large individual differences in children's ability to self-regulate energy intake have been noted and these individual differences relate to children's weight status.¹⁵ In one study, children with greater adiposity showed less evidence of self-regulation than children of normal weight. Furthermore, very thin children were categorized as hyperresponders, showing a reluctance to eat after the consumption of a small energy load.

Variation in preschoolers' ability to self-regulate energy intake and in their food preferences has been related to various child-feeding practices. The family environment has a significant impact on children's consumption patterns and food preferences.¹⁶ A recent study by Fisher and Birch¹⁷ reported that restricted access to palatable foods (where the child knows certain foods are in the environment but that they are unavailable to the child) can result in focusing the child's attention on these items and consequently may promote their overconsumption. Moreover, we previously reported that controlling child-feeding strategies were associated with a decrease in children's ability to self-regulate energy intake.¹⁵

Parents are clearly concerned about how much their children should be eating (both on the enough and the too much ends of the spectrum) and they are frustrated by children's reluctance to consume healthy foods.¹⁸ Because of these concerns, some parents view decisions about children's eating and nutrition as residing under adult purview rather than being shared with the personal domain of the child. In attempts to improve children's eating patterns, parents often use child-feeding strategies that are power-assertive and coercive. Although they are effective in the short-term, these strategies are counterproductive to building desirable long-term eating habits.¹⁹ Satter has long argued that parents must understand and practice a critical division of responsibility in child-feeding. According to her model, it is the parents' job to provide a healthy array of foods at regularly scheduled opportunities; it is the child's choice to decide how much and whether they will consume the food that is offered.²⁰ As much as health professionals have advocated for this division of responsibility in child-feeding, parents remain unclear regarding their adult tasks and their young child's abilities.

Why do parents engage in seemingly unproductive strategies? Parent concern, worry, and frustration often converge in the feeding domain. Costanzo and Woody suggest that parents exhibit greatest control over domains that: 1) they believe their child has difficulty with; 2) they, themselves, struggle with; and 3) they believe children do not have the capability to self-control.²¹ According to their model, parents who are highly invested in weight status or for whom weight and eating have been personal issues are more likely to focus on their children's eating and weight. This would include much of the US population because 30% of males and >50% of females report dieting in the last 12 months and 20% to 50% of Americans are overweight, depending on their sex and ethnicity.²² Furthermore, parents who view their child as deviant, or who simply do not believe that children can learn to self-regulate energy intake, would also monitor and control their children's food intake closely. Ironically, parent concerns may become self-fulfilling prophecy as parental over-control is theorized to limit children's opportunities for learning self-control.

Parents also serve as role models for appropriate and normative behavior. Epstein et al²³ demonstrated the potential power of parental role modeling in the eating domain in a family-based treatment for obese children. Children assigned to the treatment group where both parent and child were reinforced for weight loss and behavior change achieved greater change in percent overweight than groups where only the children were reinforced for weight loss or where the family was reinforced for attending treatment sessions. At the 10-year follow-up, the children in the parent/child group had maintained greater weight loss than children in the other groups. The authors suggest that parent participation and modeling of behavior change was instrumental in promoting skills acquisition by children.

In contrast, when children have repeated opportunities to observe out-of-control eating, they may internalize these behaviors as normal and adopt them as viable strategies. Deficits in children's self-regulation of energy intake as well as childhood overweight have been linked to self-reports of parent disinhibition (out-of-control eating) and cognitive restraint (dieting).^{16, 24} If parents are dieting and exhibiting uncontrolled eating while imposing control over their children's eating in misplaced efforts to help their children avoid future weight issues, this could lead to heightened deterioration of children's self-control of energy intake. What remains to be determined is whether children's ability to self-regulate energy intake, once derailed, can be reestablished.

Birch and colleagues demonstrated in preschoolers that use of external cues related to feeding, such as time on the clock, prompts children to clean their plate and rewards for consumption, resulted in lack of self-regulation. The external cues used in their study were selected to be typical of tactics used by parents trying to help their child to establish healthy eating patterns. In that same study, a separate group of preschoolers who were given cues related to internal signals of hunger and satiety demonstrated evidence of learning and self-regulation.¹⁹

In the experiment reported below, we used a within-subject design to investigate whether an intervention based on identification of, and attention to, internal cues of hunger and satiety would be effective in improving preschool children's ability to self-regulate energy intake. At baseline, preschool children were characterized with respect to their individual ability to self-regulate energy intake using a previously established compensation index (COMPX)¹⁵ as an indicator of ability. After 6 weeks of intervention, the children's COMPX score was again tested. We were interested in relating children's compensation scores to their weight status and in determining whether a relation existed between children's eating behaviors and their weight outcome. We were also interested in examining the relationships between parent weight and eating patterns and children's weight status and eating behaviors.

We hypothesized that at baseline, large individual differences would exist with some children showing evidence of regulation, whereas others would show little or counterregulation of energy intake (overeating). We expected that children who were more precise at energy intake regulation would be leaner than children showing limited or counterregulation. We further hypothesized that parents who reported being overweight or reported disregulated eating (dieting or disinhibition) would have children who showed less evidence of regulation. We expected that children's ability to self-regulate energy intake would improve as a result of the intervention that would help them focus on internal cues of hunger and satiety.

	METHODS

Top Abstract Methods Procedures Results Discussion Conclusion References

Overview

The study used a within-subject treatment paradigm, testing for treatment effects preintervention and postintervention. Families were recruited from 2 Denver metropolitan area day care facilities. All procedures for children took place in their day care classroom. At baseline, preschool children were invited to participate in a pair of energy compensation trials and children's self-regulation of energy intake was assessed by measuring how much energy (food) was consumed during a 2-course lunch. A percentage compensation was calculated for each child and subsequently used as an index (COMPX) of self-regulation. Anthropometric measures were obtained after COMPX trials were completed.

During the subsequent 6-week period, children were introduced to the concepts of hunger and satiety via video and adult role play in group settings and through one-on-one interactive play with dolls. Our objective was to present concrete, developmentally appropriate definitions to the preschoolers for both hunger and fullness that could be linked to their own internal feelings of hunger and satiety. We instructed the children regarding the anatomy of eating; naming body parts and describing their functions. Further, children were instructed before, during, and after their regular snack periods, to attend to cues of hunger and fullness. After the intervention period, children again participated in one pair of compensation trials.

To investigate parental influences on children's eating patterns, we asked parents to complete: 1) demographic information; 2) self-reports of height and weight; and 3) the Eating Inventory²⁵ to measure adult dieting and impulsive (disinhibited) eating.

Participants

Forty families from 2 day care facilities serving primarily a high socioeconomic population were contacted for participation. Each family was informed that a study investigating children's energy intake regulation and parents' child-feeding strategies was being embarked on in their child's preschool. Children and families were eligible to participate if the parents spoke English, the child was 3 to 4 years old, and did not have any chronic illness related to gastrointestinal difficulties, developmental disabilities, or food allergies and was not on medications that might affect appetite. Informed written consent was obtained from all participating parents, and assent was obtained from children each day they participated in the study. Parents were nominally compensated for their participation ($10 per parent). The Colorado Multiple Institutional Review Board approved all consent and assent procedures.

Five families declined to participate (4 for unknown reasons and 1 family with a child with developmental disabilities), 2 children did not speak English, and 2 children refused to participate. Therefore, 31 children were included in the intervention. Of these 31 children, 6 children (2 girls, 4 boys) experienced repeated absences (because of illness or vacation) during the intervention and were excluded from the analyses. The children who were not included in final analyses did not differ significantly from the intervention group on any parameter (age, weight status, ethnicity, or COMPX). The final group was comprised of 25 children (15 girls, 10 boys) who were 54.6 ± 1.1 months of age and had a mean age adjusted weight-for-height percentile²⁶ of 53.4 ± 5.8 (2 children <5th percentile and 5 children 85th percentile). For the 25 children who completed the intervention, 17 mothers and fathers completed and returned parent information. Families were primarily middle- to high-income and white (1 Asian, 2 black, and 3 Hispanic). The mean education level for mothers and fathers was 17.6 ± 0.80 and 18.1 ± 0.55 years, respectively.

Materials for Compensation Trials

The children consumed 2 juice drink preloads that differed in carbohydrate content and energy density but not in flavor or appearance. The details of drink preparation have been previously described.¹⁵ One serving of either drink weighed 163 g; the final energy contents were 625 kJ (150 kcal) and 13 kJ (3 kcal) for the high- and low-energy drinks, respectively. Drinks were placed in clear plastic cups with lids and chilled to 6°C. Children consumed the drinks through straws, with the lids left in place to prevent spills.

The menu for all ad libitum lunches included turkey hotdogs (Louis Rich Turkey Hotdogs, Madison, WI, 33 g), enriched hotdog buns, American cheese slices (Kraft Deluxe Slices, Waukegan, IL, 37 g), unsweetened applesauce (Musselman's single serving, Biglerville, PA, 120 g), carrot sticks (20 g), Fig Newtons (Nabisco, 3 each, East Hanover, NJ, 47 g) and 2% milk (Robinson Dairy, Denver, CO, 154 g). The total energy of the foods initially served to each child was ~2815 kJ (672 kcal), as calculated from manufacturer's information. At children's requests, additional servings of any menu item were provided. The same menu was used throughout the study following both the low- and high-energy preloads.

	Procedures

Top Abstract Methods Procedures Results Discussion Conclusion References

COMPX Trials

Each child participated in consumption trials at baseline and postintervention with one trial pair consisting of 2 test days. A test day included a morning snack, a preload, and a lunch meal. On each test day, children were offered a small snack at their usual snacktime (~9:30 a.m.). The purpose of this snack was to make an effort to have all children start the compensation trials in the same relative state of energy balance. The energy intake at snack was not included in the intake analyses.

On the first test day of the trial pair, ~35 minutes before their usual lunchtime, children were asked to drink the high-energy or low-energy density drink as a first course or preload. Children were randomly assigned to preload sequence (high or low energy on test day 1 and the alternate version on test day 2), an equal number of boys and girls were assigned to each sequence and researchers and classroom personnel were blinded to the experimental condition. Thirty minutes later, children were given a variety of foods from which they could self-select and eat ad libitum. Children were instructed to consume what they wished until they were no longer hungry. One week later, children ate a snack, received the alternate version of the preload, and again consumed a self-selected lunch.

Food intake was measured by preweighing and postweighing all food offered. Energy intake of foods consumed was calculated using manufacturers' information. A measure of children's self-regulation or compensation, COMPX, was derived from the extent to which they adjusted ad libitum intakes: children who were sensitive to energy density of the drinks would compensate by eating less lunch after consuming a high-energy drink. An index of the percentage of compensation was devised to characterize children on a continuum from negative compensation (eating more lunch after drinking the high-energy drink) to overcompensation (consuming little after the high-energy drink). This COMPX was calculated for each child using the following formula: Perfect compensation (100%) would be shown if a child ate 612 kJ more after the low-energy preload than after the high-energy preload. The COMPX typically ranges from 200% for children who overeat, to 200% for children who respond to the preload by eating very little at lunch. Because neither end of this distribution is ideal, we also established a deviation score (COMPX_dev) calculated using the following formula:

The COMPX_dev is the absolute value of the child's COMPX score minus the ideal COMPX (100%). In this manner, we could examine disregulation of energy intake looking at overeating that is associated with greater adiposity, as well as hyperresponsiveness associated with extreme leanness or undernutrition (<10th percentile for age-adjusted weight-for-height).

Anthropometrics

Anthropometric measures obtained on the children included height (cm), weight (kg), and triceps and subscapular skinfold thicknesses (mm). All measures were collected by one trained professional from The Center for Human Nutrition Energy Balance Laboratory. Height and weight were measured to the nearest 0.5 cm and 0.1 kg, respectively. Body mass index (BMI) was computed from these measures. Skinfold thicknesses were measured on the right side of the body by Harpenden calipers and were recorded to the nearest 0.1 mm. Three consecutive skinfold measurements were taken at each site, and a mean of three measurements was used for the analyses. Because anthropometric data were not expected to change significantly over the 6-week period, all measures were collected at the end of the study intervention to avoid the possibility that children might associate our intervention and the weighing procedures.

Parent Information

Both mothers and fathers were asked to give self-reports of height and weight, from which BMI was derived. Parents also completed information regarding education, occupation, and relationship to child (biological or adoptive parent) and the Eating Inventory. Stunkard and Messick's Eating Inventory²⁵ is a 51-item questionnaire that assesses 3 aspects of eating behavior: 1) cognitive restraint, or the cognitive control of food intake (dieting); 2) disinhibition, or uncontrollable, impulsive eating; and 3) perceived hunger. We used factor 1 (dieting) and factor 2 (impulsive eating) in our analyses as factor 3 (perceived hunger) has been reported to co-vary with disinhibition.

Intervention

Each research assistant visited the preschools at least 3 times before the start of the study to diminish stranger effects. On the first day of the intervention, researchers performed a skit, the themes of which centered on hunger (rumbling in the stomach), eating to fullness (stomach extension, satisfaction) and the signals associated with overeating (stomach distention and discomfort). At this point we also introduced the terms: mouth (where you chew food), esophagus (where food goes when you swallow), and the stomach (where swallowed food goes to). One week later, children watched the children's video Winnie the Pooh and the Honey Jar and researchers discussed hunger, overeating, and the results of chronic overeating with the preschoolers.

During the next week, we used doll play to help children identify internal cues of hunger and satiety and to relate these feelings to decisions about how much to eat at snack and lunch. We elaborated on a previous study by Birch and colleagues in which a doll prototype (a glass jar stomach with a doll's head attached that children could feed) was used to acquaint children with hunger and fullness. In our experiment, we developed dolls that had a mouth, esophagus, and a place for a stomach on the exterior of the doll abdomen. Dolls were constructed to be androgynous and to represent different hair, eye, and skin colors. In addition, several sets of stomachs were made from nylon material filled to varying degrees with salt to represent 1) an empty stomach that was hungry, 2) a stomach that was a little full, and 3) another stomach that was very full. These stomachs could be attached to the dolls by Velcro strips.

Children were invited to go to a corner of their classroom with a researcher to play with the dolls. Each child was seen before and after the snack, 1 day per week, over the course of a 4-week period. The child was then introduced to the 3 doll stomachs. The child was subsequently asked to identify each of the doll stomachs. All but 1 child correctly identified the stomachs on the first occasion; this child answered correctly on the second day. Thereafter, all children answered correctly. Second, each child was asked to place their hand over their own stomach and tell us whether they were hungry, a little full, or very full. The child was then asked to choose the doll stomach that was most like the state of their own stomach and to place it on the doll. The child then returned to their normal classroom activities.

During their normal mid-morning snacktime, all children were prompted at least 2 times to check in and see if they were still hungry and whether they wanted to eat more. After children consumed their usual mid-morning snack, we asked the same child to play with the dolls. Using the doll they had previously chosen, each child was again asked to identify the doll stomachs and then was asked to place their hand over their own stomach to determine their hunger state. Next, the child chose the doll stomach that felt most like their own. Afterward, the child was given a hand stamp for participating and then returned to the classroom activity.

Statistics

Descriptive statistics, including means, standard error of the mean, and ranges were computed for all variables. Tests for normality of distribution for each variable were also completed. Dependent measures t tests were used to examine differences in mean values of children's eating measures at baseline and postintervention. ² analysis was used to determine whether a difference existed in the observed versus the expected frequency of children improving in self-regulation. Mothers' and fathers' and boys' and girls' variables were examined by Student t tests. There were no significant differences in mean values between any of girls' and boys' measures. Correlational analysis was used to test hypotheses regarding relations between measures within and between children and parents. Data were analyzed with SAS (version 6.12; SAS Institute Inc, Cary, NC). Significance was set at P < .05.

	RESULTS

Top Abstract Methods Procedures Results Discussion Conclusion References

Effects of the Intervention

Means and ranges for children's compensation (COMPX %) and deviation from ideal compensation (COMPX_dev %) are reported in Table 1. At baseline, children showed limited evidence of compensation for the difference in energy density of the drinks and exhibited a large degree of deviation (either above or below 100%) from ideal compensation. After 6 weeks of the intervention, children's mean compensation improved and a significant decrease in the deviation from ideal compensation (~40%) was noted; that is, children who both overcompensated and undercompensated improved in their ability to adjust their energy intake in response to energy density differences in the juice preload. Of the 25 children who completed the protocol, 17 improved (scored closer to ideal), 4 stayed the same (± 10%), and 4 showed decreases in compensation (² = 13.52, P < .001). Thus, the intervention seemed, over the short-term, to increase their ability to self-regulate energy intake as indicated by the change in deviation scores.

Relations Between Children's Compensation and Adiposity

Children's anthropometric data, reported in Table 2, reveal that most children (84%) fell between the 5th and 90th percentile for weight-for-height. Only 2 children (1 boy, 1 girl) fell at or below the 5th percentile and 3 children (2 boys, 1 girl) were above the 90th percentile.

As indicated in Table 3, there was a trend at baseline for children's COMPX to be negatively related to their age (P = .06). In other words, older children, particularly older girls, were not as adept as younger children in self-regulation of energy intake. Children's compensation scores were negatively related to adiposity, indicating that children with greater adiposity showed less evidence of self-regulation. When examined for sex effects, these relationships were present only for girls and were highly significant. It should also be noted that the leanest children tended to show an overcompensation with COMPX scores that were well above the 100% ideal. After the intervention, neither age nor adiposity significantly related to children's compensation scores suggesting that the intervention was effective in improving all children's intake regulation.

Relations Among Parent Variables

Descriptive statistics for parent age, weight status, and Eating Inventory scores are reported in Table 4. Mothers were of normal weight and fathers' BMIs fell into the normal to overweight category (5 fathers' BMIs >28.0). As noted in many samples, mothers' mean restraint score was significantly higher than fathers' suggesting that mothers engaged in more dieting practices than did fathers. Parents' disinhibition scores correlated positively to their BMIs (r = 0.46, P < .08 and r = 0.47, P < .05 for mothers and fathers, respectively) indicating that parents who experienced episodes of out-of-control eating had higher BMIs. Dietary restraint was not significantly related to parent weight status.

Relations Between Parent's and Children's Variables

To determine whether parent BMI and Eating Inventory Scores were related to children's weight status and eating behaviors, mothers' and fathers' measures were correlated to their daughters' and sons' data. Mothers' BMI was significantly related to children's BMI (r = 0.91, P < .05 and r = 0.69, P < .05 for girls and boys, respectively), although this result should be interpreted with caution as none of the mothers in this sample fell into the overweight category (BMI >25.0). Mothers' disinhibition was also significantly correlated with children's COMPX_dev at baseline (r = 0.52, P < .05) indicating that mothers who have difficulty controlling their own food intake have children who do not show evidence of good self-regulation. Further, mothers' level of dietary restraint, or dieting, related significantly to daughters' COMPX_dev (r = 0.73, P < .01) revealing that mothers who practiced greater cognitive control over food intake had daughters who were less likely to show precise regulation of energy intake. These relations between mothers' and children's eating patterns were no longer significant after intervention suggesting that perhaps the intervention was effective in ameliorating the effects of parents' role modeling of negative eating behaviors. No such relationships were found between fathers' BMIs and children's weight status or compensation scores at baseline or postintervention.

	DISCUSSION

Top Abstract Methods Procedures Results Discussion Conclusion References

Our findings support the concept that internal cues of hunger and satiety, when presented in a developmentally appropriate manner, can be used by young children to make decisions regarding how much they wish to eat. At the onset of this study, preschoolers showed a limited ability to adjust the amount they were eating to reflect the energy density of the preload; that is, they showed little evidence of self-regulation of energy intake. The degree to which children made these adjustments was reflective of their weight status: children with greater adiposity undercompensated (COMPX scores) for the energy difference in the preload whereas the leanest children overcompensated (COMPX scores >100% ideal). After the intervention, the majority of children's COMPX scores improvedwhether they were initially over- or undercompensating. This suggests that the intervention was successful in reaching both the finicky eater who overresponds to satiety cues as well as the heavier child who may not have previously related internal cues to the act of eating.

That significant improvements in children's ability to self-regulate occurred as a result of the intervention is suggested both by their change in compensation scores and also by anecdotal evidence that indicates children began to relate the doll play to their own eating during snacktime. After several weeks, during snacktime when we would prompt the children to tune in and see how hungry/full they were, children spontaneously began to respond with comments like: "I'm not hungry any more so I'm going to stop eating" or "My stomach's getting full." This provides further confirmation that children understood the concepts and were able to apply them to their snacktime eating. These research findings substantiate Ellyn Satter's long-held clinical opinion that when children's internal cues are encouraged and attended to by parents, and when control over eating is shared, children are capable of controlling the how much of eating. This evidence does not give any credence, however, to the notion that children are capable of making healthy food choices without adult guidance. Parents and caregivers must still be gatekeepers of the kinds of foods that are offered and it is the parent's responsibility to ensure a healthy variety of foods are available, to provide a structured eating environment, and to serve as role models for appropriate eating habits.

The responsibility of acting as positive role models, or do as I do, may be more important than previously recognized. We found evidence that parents, mothers in particular, may be serving as role models for eating behaviors in ways that they do not anticipatevia dieting and impulsive eating behaviors. Mothers with self-reports of higher levels of dieting and impulsive (disinhibited) eating had children who experienced greater difficulty in self-regulating energy intake. This suggests that mothers may be serving as role models for maladaptive eating strategies and behaviors. Despite the limitations of a small sample size, our results relate well to previously published findings from our colleagues' laboratory. Cutting et al²⁴ reported a relationship between maternal and daughters' weight status that was mediated by maternal self-reports of disinhibition. One alternative possibility exists that this relationship between maternal and children's eating and weight status is mediated via genetic components; however, we are unable to determine this from our data. Irrespective of their origins, dieting and disinhibited eating have been increasing in incidence in recent decades, especially in women. Our results, as well as those published by other investigators, suggest that these adult eating behaviors may have adverse and unintended effects on the development of very young children's eating styles.

A positive finding was that our intervention seemed to alter the negative relationships between mothers' and daughters' eating styles, as there were no significant relations between children's postintervention compensation scores and mothers' reported eating behaviors. We suggest that helping children attend to internal cues of hunger and satiety should be promoted as a productive child-feeding strategy and as an alternative to coercive or restrictive practices.

Parenting practices in the child-feeding domain merit further examination. In general, today's parents have become less authoritarian and more likely to encourage individual expression and the development of the self.^27,28 However, for very young children, the development of self is enabled only in areas where parents consider children capable of performing the necessary tasks and in areas where parents believe children should have a personal choice.²⁹ The extent to which parents hold the belief that children are capable of self-regulating how much they eat is highly variable among parents and relates to their personal concerns and struggles with weight and eating.²¹

In a study investigating mothers' influence on preschoolers' food intake, Drucker et al³⁰ noted that a measure of general maternal control was positively correlated with both encouragement and discouragement of children's eating and that prompts to eat increased children's energy intake. They also noted a negative relationship between preschoolers' BMI and the amount of maternal discouragement of eating. This study provides evidence that parents impart instruction that can override children's ability to self-regulate. Future research in the home setting should investigate parent knowledge and beliefs about children's eating and the usefulness of promoting methods that encourage parents to enable their child to attend to internal cues. The results of our study suggest that focusing on physiologic, internal cues would be a viable strategy to promote optimal growth.

However, we must also acknowledge the limitations of our findings. First, the COMPX measure is not a perfect indicator of regulation of energy intake; rather, it is a good short-term snapshot of children's ability to exhibit self-control over food intake. There are other aspects of energy regulation that are not accounted for by the COMPX measure that occur over larger time frames than a single meal or that are related to energy expenditure. To address long-term energy intake regulation will require studies of sufficient duration to follow energy intake and expenditure as well as growth parameters. Second, our sample size for this intensive intervention was necessarily small and, although conducted in a naturalistic day care setting, we cannot be certain that these findings are transferable to the family setting. Additionally, our sample consisted of mostly normal weight upper socioeconomic status white children and this intervention must be conducted with ethnically diverse groups of both overweight and underweight children that cross socioeconomic strata to determine its effectiveness for a broader population.

	CONCLUSION

Top Abstract Methods Procedures Results Discussion Conclusion References

We found evidence that children differ in their ability to self-regulate energy intake and these differences are related to parental eating behaviors and adiposity, perhaps via role-modeling of dieting and impulsive eating. Our findings suggest that it may be productive to help children focus on internal cues of hunger and satiety and to encourage adults to serve as positive role models for lifelong healthy eating behavior.

We suggest that one strategy to enable parents and caregivers to create optimal feeding environments is to give them clear information that helps them recognize their responsibilities as caregivers and that provides acknowledgment of children's capabilities. In a supportive, structured environment, most children can regulate the how much of eating. Parents must, however, provide structured meals and snacks consisting of a variety of healthy choices, repeated opportunities to try new foods, and gentle, firm limit-setting regarding appropriate mealtime behavior. The ideal is for control over children's eating to be shared by parents and children according their unique capabilities and responsibilities.

	ACKNOWLEDGMENTS

This work was supported by the National Institutes of Health (Grant R29 HD33511) and a grant from the American Dietetic Association Children's Health Campaign.

I would like to thank Carol Mahoney for excellent technical assistance and Jennifer O. Fisher for her generous support in data analysis and her suggestions for manuscript development.

	FOOTNOTES

Received for publication Mar 13, 2000; accepted May 16, 2000.

Reprint requests to (S.L.J.) University of Colorado Health Sciences Center, Department of Pediatrics, The Children's Hospital of Denver, Center for Human Nutrition, C225, 4200 E 9th Ave, Denver, CO 80262. E-mail: susan.johnson{at}uchsc.edu

	ABBREVIATIONS

COMPX, compensation index; BMI, body mass index; COMPX_dev, deviation from ideal compensation index.

	REFERENCES

Top Abstract Methods Procedures Results Discussion Conclusion References

1. Ogden CL, Troiano RP, Briefel RR, Kuczmarski RJ, Flegal KM, Johnson CL. Prevalence of overweight among preschool children in the United States, 1971 through 1994. Pediatrics. 1997;99(4). URL: http://www.pediatrics.org/cgi/content/full/99/4/e1
2. Federation of American Societies for Experimental Biology, Life Sciences Research Office. Prepared for the Interagency Board for Nutrition Monitoring and Related Research. Third Report on Nutrition Monitoring in the United States: Volume 1. Washington, DC: US Government Printing Office; 1995
3. Steward DK, Garvin GJ Nonorganic failure to thrive: a theoretical approach. J Pediatr Nursing. 1997; 12:342-347
4. Drotar D, Sturm L Prediction of intellectual development in young children with early histories of nonorganic failure to thrive. J Dev Behav Pediatr. 1998; 13:281-296
5. Drotar D, Sturm L Personality development, problem solving, and behavior problems among preschool children with early histories of non-organic failure to thrive: a controlled study. J Dev Behav Pediatr. 1992; 13:266-273 [Medline]
6. Pollitt E, Eichler A Behavioral disturbances among failure-to-thrive children. Am J Dis Child. 1976; 130:24-29 [Abstract/Free Full Text]
7. Listernick R, Christoffel K, Pace J, Chiaramonte J Severe primary malnutrition in US children. Am J Dis Child. 1985; 139:1157-1160 [Abstract/Free Full Text]
8. Greenspan SI, Lieberman AF. Infants, mothers, and their interaction. In: Greenspan, SI, ed. The Course of Life: Vol 1. Infancy and Early Childhood. Washington, DC: US Department of Health; 1980
9. Drotar D, Eckerle D, Satola J, Pallotta J, Wyatt B Maternal interactional behavior with nonorganic failure-to-thrive infants: a case comparison study. Child Abuse Negl. 1990; 14:41-51 [CrossRef][Medline]
10. Taitz LS, King JM Growth patterns in child abuse. Acta Paediatr Scand. 1988; 343S:62-72
11. Kessler DB. Failure to thrive and pediatric undernutrition: historical and theoretical context. In: Kessler DB, Dawson P, eds. Failure to Thrive and Pediatric Undernutrition: A Transdisciplinary Approach. Baltimore, MD: Paul H. Brookes Publishing Co; 1999:3-17
12. Birch LL, Deysher M Caloric compensation and sensory specific satiety: evidence for self-regulation of food intake by young children. Appetite. 1986; 7:323-331 [Medline]
13. Birch LL, Johnson SL, Andersen G, Peters JC, Schulte MC The variability of young children's energy intake. N Engl J Med. 1991; 324:232-235 [Abstract]
14. Johnson SL, McPhee L, Birch LL Conditioned preferences: young children prefer flavors associated with high dietary fat. Physiol Behav. 1991; 50:1245-1251 [CrossRef][Medline]
15. Johnson SL, Birch LL Parents' and children's adiposity and eating styles. Pediatrics. 1994; 94:653-666 [Abstract/Free Full Text]
16. Birch LL, Zimmerman S, Hind H The influence of social-affective context in preschool children's food preferences. Child Dev. 1980; 51:856-861 [CrossRef]
17. Fisher JO, Birch LL Restricting access to foods and children's eating. Appetite. 1999; 32:405-419 [CrossRef][Medline]
18. Birch LL, Fisher JO Development of eating behaviors among children and adolescents. Pediatrics. 1998; 101:539-549 [Abstract/Free Full Text]
19. Birch LL, McPhee L, Shoba BC, Steinberg L, Krehbiel R "Clean up your plate": effects of child feeding practices on the conditioning of meal size. Learning Motivation. 1987; 18:301-317
20. Satter E The feeding relationship: problems and interventions. J Pediatr. 1990; 117:S181-S189 [CrossRef][Medline]
21. Costanzo PR, Woody EZ Domain-specific parenting styles and their impact on the child's development of particular deviance: the example of obesity proneness. J Soc Clin Psychol. 1985; 4:425-445
22. Federation of American Societies for Experimental Biology, Life Sciences Research Office. Prepared for the Interagency Board for Nutrition Monitoring and Related Research. Third Report on Nutrition Monitoring in the United States: Executive Summary. Washington, DC: US Government Printing Office; 1995
23. Epstein LH, Valoski A, Wing RR, McCurley J Ten-year follow-up of behavioral, family-based treatment for obese children. JAMA. 1990; 264:2519-2523 [Abstract/Free Full Text]
24. Cutting TM, Fisher JO, Grimm-Thomas K, Birch LL Like mother, like daughter: familial patterns of overweight are mediated by mothers' dietary disinhibition. Am J Clin Nutr. 1999; 69:608-613 [Abstract/Free Full Text]
25. Stunkard AJ, Messick S The three-factor eating questionnaire to measure dietary restraint, disinhibition, and hunger. J Psychosom Res. 1985; 29:71-83 [CrossRef][Medline]
26. Hamill VV, Drizd TA, Johnson CL, Reed RB, Roche AF, Moore WM Physical growth: National Center for Health Statistics percentiles. Am J Clin Nutr. 1979; 32:607-629 [Abstract/Free Full Text]
27. Baumrind D Parental disciplinary patterns and social competence in children. Youth Society. 1978; 9:239-276
28. Youniss J. Rearing children for society. In: Smetana JG, ed. Beliefs About Parenting: Origins and Developmental Implications. San Francisco, CA: Jossey-Bass Publishers; 1994:37-50
29. Smetana JG. Parenting styles and beliefs about parental authority. In: Smetana JG, ed. Beliefs About Parenting: Origins and Developmental Implications. San Francisco, CA: Jossey-Bass Publishers; 1994:21-36
30. Drucker RR, Hammer LD, Agras WS, Bryson S Can mothers influence their child's eating behavior? J Dev Behav Pediatr. 1999; 20:88-92 [Medline]