OBJECTIVES. Eating problems are a common cause of concern for the parents of toddlers, but few studies have examined the correlates of eating problems or the growth patterns associated with them in a large population-based sample. Our goal was to examine the distribution of eating behaviors in a large representative sample of toddlers and their mothers' approach to feeding. In addition, we describe the prevalence of parentally perceived eating problems and how they relate to specific behaviors, food preferences, and growth in the child.
METHODS. We conducted a cross-sectional analysis of data from a United Kingdom population-based birth cohort, the Gateshead Millennium Baby Study, which included 455 questionnaires completed by parents when their children were aged 30 months.
RESULTS. Eating was perceived to be a problem by 89 (20%) parents. Eating a limited variety (79 [17%]) and preferring drinks to food (57 [13%]) were the most prevalent problem behaviors. Thirty-seven children (8%) were described by parents as definitely “faddy” (picky), and these children liked fewer foods and had higher eating restriction scores than those described as not faddy. Children who were described as having an eating problem gained less weight over the first 2 years; 11.1% had weight faltering compared with 3.5% in children not described as having an eating problem. Being faddy was only weakly associated with poor growth, and simply eating a limited variety was unrelated to growth. High milk consumption was associated with lower appetite but not with poor growth.
CONCLUSIONS. Eating problems are common in toddlers and in the majority are associated with normal growth, although weight faltering is more common in such children. Excessive milk-drinking may be a cause of low appetite at meal times.
Parents often worry about their toddler's eating. Toddlers have relatively high energy requirements and are rapidly developing fine motor skills and progressively widening their repertoire of food types. However, parents have to be able to contend with neophobia (fear of new foods)1 and the oppositional behavior that is so characteristic of this age range. Thus, it is not remarkable that parents often present to family practitioners and pediatricians with concern about eating problems. US studies have found that 20% to 50% of children were described by their parents as at least sometimes “picky.”2–4 A population study in New Zealand5 found that 24% of parents of 2-year-old children described them as having eating and feeding problems, with 18% still doing so at 4 years old, and a population study in London, England, in 3-year-olds found 17% were described as having poor appetite and 12% “faddy” (picky) eating.6 These studies found such a high prevalence that it could reasonably be said that eating-behavior problems are a normal feature of toddler life.
When parents present these problems at primary care consultations, health workers conventionally reassure parents that children “do not starve themselves.” In clinical practice, some children with food refusal and pickiness (United States) or faddiness (United Kingdom) do gain weight poorly, and some children with slow weight gain from early infancy go on to manifest severe eating problems, but it could be argued that this was simply the chance occurrence of 2 common problems in the same child.
Few studies have followed up children with eating problems. In a clinical study,7,8 three quarters of children who had been “refusing to eat” early in the first year, when followed up to the age of 2 years, had persisting eating problems as well as poor weight gain. This might imply that a subgroup of problem eaters have long-standing problems with associated failure to thrive. Studies of younger children with failure to thrive suggest that they eat differently from normally growing peers from an early age with less hunger and enjoyment of foods9–11 and less self-feeding12 and disordered patterns of satiation.13 One small study in 4-year-old children with long-standing failure to thrive14 found meal times to be unhappy, stressful, and chaotic compared with meal times in normally grown controls.
Less is known about the normal variation in eating behavior, as opposed to whether parents find the behavior problematical. One small American study examined eating in children 2 years old and younger, and found that between a fifth and a half of all toddlers manifest each of a range of 4 difficult mealtime behaviors (refusal to eat, requesting then refusing specific foods, trying to end meal after a few bites, picky).15 However, only 50 children were in the 21 to 27 months' age range, and the study was not population based.
Thus, although eating problems are well recognized, few studies have examined the correlates of eating problems or the growth patterns associated with them in a large population-based sample.
The Gateshead Millennium Baby Study is a longitudinal program of research into feeding and growth in a cohort of 1029 Gateshead infants recruited at birth between June 1999 and May 2000. A wide range of data have been collected about the children and their families, from both parents and local National Health Services staff, via questionnaires mainly relating either to feeding and growth or to risk factors that may affect them. Weights from the first year are available for most children, as well as child and parental weights and heights collected at a health check at 13 months old. This cohort thus provides the opportunity to describe eating as reported by their mothers in 30-month-old children in a large, representative sample.
Specifically, the aims were to describe the range of eating behaviors seen in toddlers, the child feeding behavior of their parents, the prevalence of parentally perceived eating problems and how these relate to the child's behavior, food preferences, and growth.
The study received approval from the Gateshead Local Research Ethics Committee. Study staff approached the mothers of infants born to Gateshead resident mothers shortly after birth, in 34 prespecified weeks, June 1999 through May 2000, and 82% consented to join the study. Basic demographic information was collected at recruitment and parents received postal questionnaires at 6 weeks and at 4, 8, 12, and 30 months, when they completed questions about feeding and other issues, as well as transcribing routinely collected weights.10,16 Data from the 30-month questionnaire and weights from the first year were used for this analysis.
The questionnaire addressed 4 general areas.
Eating problems: In a section headed “Difficulties with your child,” parents were asked “Do you see your child as having eating problems at present?” Possible answers were “yes,” “sometimes,” and “no.” Unless they answered no they were asked about specific sorts of eating problem (poor eater, faddy, or behaves badly at mealtimes) with possible responses “definitely,” “maybe,” and “no.” They were then asked what, if any, external advice they had sought. Children were classified as having an eating problem if parents answered “yes” to the first question, or if they said their child definitely had 1 of the 3 problems or if they stated they had sought any professional help for an eating problem.
Eating and feeding behavior: In a separate section, all parents were asked about their child's eating behavior and about their own feeding methods, including the mealtime strategies they used to encourage their child to eat and their management of food refusal. For the child, an avoids feeding score was constructed from the sum of responses to 6 questions (does your child push food away; gag on food; hold food in mouth; spit food out; throw food; cry/scream during meals: possible responses were never, sometimes, or often). Each “often” was given a score of 2 and each “sometimes” a score of 1. Parents were also asked to rate their child's appetite on a 5-point scale from very good to very poor.
Preferences for food and food types: The questionnaire included a modified version of a food preferences questionnaire based on the Survey of the Diets of British School children (Department of Health, 1989) already used in older children.17 This questionnaire listed 89 specific foods and 9 types of food that parents rated on a 5-point scale from “dislikes a lot” to “likes a lot” or 6 “never tried.” Parents also rated if their child refused food because it was not a particular brand or was damaged (eg, a broken cookie) on a 4-point scale from never to often. An eating restriction score was constructed by counting the number of food types (out of slimy, crunchy, chewy, sloppy, messy, mixed, strong, unfamiliar), which each child was described as strongly disliking, adding 1 point each for often rejecting damaged or wrong brand foods. One other type (colorful or bright) was omitted as it had low prevalence and was uncorrelated with the other variables. The eating restriction score had a Cronbach's α of .67.
Drinking pattern: Parents were asked the type and frequency of different drinks given. The total number of milky and other drinks consumed per day was estimated by summing relevant responses.
Parents were asked to supply an up-to-date clinic weight and height on their child; routine weights from infancy and measurements had already been collected at 13 months old. All measurements were converted into standard deviation (z) scores compared with the United Kingdom 1990 growth reference.18 Weight gain was assessed using the Thrive Index (TI), a measure of change in weight over time, conditional on starting weight (initially birth weight), which adjusts for regression to the mean19 (see Fig 1). Weight faltering was defined for 2 time points, from birth to 2 years and between 1 and 2 years, as conditional weight gain (TI) below the 5th percentile in the cohort as a whole.
Comparisons were made between categorical data by using Pearson's χ2 test, between normally distributed data by using the Student's t test and analysis of variance (F for linear trend), and between nonnormally distributed continuous data by using either the Mann-Whitney or the Kruskal-Wallis test. Spearman's or Pearson's correlations were used for univariate associations.
For this analysis, preterm infants (<37 weeks' gestation) were excluded, as were 14 ultraorthodox Jewish and 2 Muslim respondents, because of their very different feeding and growth characteristics.10
Questionnaires were returned for 455 (49%) of 923 eligible children. The mean age of the child when the questionnaire was completed was 30 months (range: 29–33). Of these, 385 (85%) also included weights at a mean (SD) age of 2.4 (0.23) years and 270 (59%) included heights at mean age 2.4 (0.15) years. All the respondents were white; 86% of households were waged, but only 22% of mothers had received higher education (college or equivalent); 59% infants had been breastfed at some point, and 52% were firstborn. The respondents were slightly less likely to live in the more deprived part of the borough and more likely to have college education than all those in the original cohort. None of the children had illnesses likely to impact on eating behavior.
Eating Behavior and Parents' Feeding Methods
Eighty-nine parents (20%) perceived their child as having eating problems, and 58 (13%) had sought some sort of professional help for this, most commonly from their health visitor (children's public health nurse). Of those with eating problems, 37 (42%) parents described their child as definitely faddy (35 [39%] maybe) and 27 (30%) as definitely a poor eater (47 [47%] maybe), of whom 12 (13.5%) were definitely both. All but 4 of the remainder were described as maybe faddy or a poor eater or both. Only 6 (7%) said their child definitely behaved badly at meals (31 [35%] maybe).
Difficult mealtime behaviors were fairly common throughout the cohort, with eating a limited variety of foods and preferring drinks to food the most prevalent (Table 1). A wide range of strategies were generally adopted by parents to encourage children to eat, ranging from distraction and alternative foods for the majority, to force or physical punishment in a minority (Table 1).
As would be expected, all difficult mealtime behaviors were reported significantly more often among children perceived as having eating problems, but the relative distribution of these was similar, with the most common problem again being eating a limited variety of food, followed by preferring drinks to food. Most mealtime strategies were used significantly more with children with eating problems, with the exception of strategies involving food withdrawal (not giving dessert and not offering anything else until the next meal), although more parents of children with eating problems did threaten to take food away. Strategies involving force or physical punishment were used slightly more for children with eating problems, although these differences did not reach statistical significance (Table 1).
Children in the sample as a whole were described as having tried a median (range) of 70 (20–90) foods and as liking 52 (14–82). Children described as having eating problems or as faddy liked a significantly smaller number of food items than the remainder as did those with high eating restriction scores and those described as having poor appetites. The biggest difference in items liked was for children described as faddy, who liked on average 15 items less than those described as not faddy. However, children described as definitely faddy or often eating a limited variety still ate a median of 39 and 44 foods, respectively (Table 2).
The 2 most commonly disliked food types were “slimy” (n = 198 [46.7%]) and “unfamiliar” (n = 123 [30.5%]); some children often rejected foods on the grounds that they were “damaged” (n = 40 [8.9%]) but few often rejected foods that were the “wrong brand” (n = 7 [1.6%]). Over half the children had an eating restriction score of 0 (no dislike of any of the food types) and only 11.4% scored 3 to 6 points (Table 2). As might be expected, the eating restriction score was positively associated with both parent-described faddiness (r = 0.2; P < .001) and negatively associated with number of food items liked (r = −0.27; P < .001), that is, a child with a higher score was more likely to be described as faddy by parents and liked fewer food items. Both being described as faddy by parents and a high eating restriction score were negatively associated with appetite (Table 2).
The median number of drinks taken in 1 day was 8 (range: 2–21); 421 (93%) children had a milky drink at least once a day, 438 (97%) had sweet drinks, and 329 (73%) had at least 1 sugar-free drink a day. The number of milk drinks was associated negatively with appetite, whereas the number of other drinks was associated positively (Tables 3 and 4). Children said to often prefer drink to food drank more milk but there was no relationship with amounts of other drinks (Tables 3 and 4).
Associations With Growth
Children identified by their parents as having an eating problem were significantly lighter and shorter than other children, and had gained weight significantly less well since birth, but not since the age of 1 year (Table 5). Only 8 (11%) of the children with eating problems met the conventional criteria for weight faltering (<5th percentile for weight gain, birth to 2 years) but this was 3 times the proportion in those without (11 [3.5%]; P = .01).
Children described as faddy by parents were slightly lighter and shorter and had grown less well than the remainder, but these differences did not reach significance at the 0.05 level (Table 5). Growth was unrelated both to the eating restriction score and to being described as eating a limited variety of foods (Table 5). The number of food items liked was also unrelated to weight gain both from birth to 2 years (r = −0.05; P = .3) and from 1 to 2 years (r = −0.07; P = .1). Children with good appetites were heavier and had gained more weight since birth (Table 5). The number of milk drinks taken was unrelated to growth (correlation with weight r = 0.02, P = .8; weight gain birth to 2 years r = 0.001, P = 0.99).
We found no evidence of variation in any of the key variables by socioeconomic or educational status (data not shown).
This study found that one fifth of the parents perceived their child as having an eating problem, with eating a limited variety of foods and preferring drinks to food the most common behaviors associated with reported eating problems. Children described by their parents as having eating problems grew significantly less well over the period from birth to 2 years, and 3 times as many met a clinical criterion for weight faltering (failure to thrive). A low appetite was the only 1 of the more specifically defined problems consistently related to growth. Milk consumption was associated with lower appetite but it was not associated with poor growth.
Strengths and Limitations
The children used for this analysis represent 50% of the original cohort, a response rate fairly typical for a longitudinal study of this kind. As a result, the most deprived subjects were relatively underrepresented, so these data will not be entirely representative of the general population. On the other hand, the large number of children allowed the relationships between the various aspects of feeding and behavior to be explored in some detail and we found no evidence of social patterning in parents' responses, which is in keeping with our previous work.16 All data were collected through parent report, so we do not have an independent description of the child's behavior, but the sections on eating behavior and eating problems were deliberately separated in the questionnaire to ensure that parents considered the questions separately, and we have been able to provide some evidence for face validity by cross relating descriptions of behavior with reported food preferences and intake of drinks. Because these questions were asked on only 1 occasion, we can know little about the persistence of these problems in the long-term, but future work in the cohort may be able to provide more information.
Approximately one fifth of parents reported an eating problem, which is remarkably similar to previous estimates5,6 and two thirds of these had sought professional help, most commonly from their health visitor. Parents worried most that their child was eating a limited variety of foods. Although these children did like significantly fewer foods, they still ate a substantial number and eating a limited variety of foods itself was not associated with poor growth. Tactics to encourage children to eat were more commonly employed by mothers who reported eating problems, but such mothers were less willing to use tactics that involved taking food away. Only a minority of children with parent-reported eating problems had significant weight faltering but as a group they were slightly smaller. It may be that eating problems trigger greater parental concern in children who are already relatively small.
Faddiness and Pickiness
In this population, 1 in 12 parents described their child as faddy. This perceived quality related to both the number and type of food items liked. Pickiness in the United States has similarly been shown to relate to lower food variety2 and lower nutrient intake, as well as a tendency to a relatively low weight for age.3 In this cohort, faddiness related only weakly to growth. The prevalence of dislikes for whole types of food was relatively low, as was the liking of very small numbers of foods, both problems that will be commonly seen in children referred with severe eating problems.
As in previous survey rounds in this cohort,10 the majority of children were rated as having good appetites. This demonstrates that although parents may often worry about toddler eating habits, the majority still eats well and enthusiastically. Poor appetites were rare, and appetite was significantly related to weight gain. Preferring drinks to food was common; it is widely believed in clinical practice that frequent drinking in toddlers suppresses appetite. In this cohort, no such relationship was found for other drinks, but milk-drinking did relate inversely to appetite but not to weight gain. Thus in some children low appetite as perceived by the parents probably reflects substitution by the child of milk for solid food, with no effect on weight gain.
These results offer reassurance that the majority of children with parent-reported eating problems will suffer no adverse consequences in terms of growth. Parents who worry about the range of foods that their children eat may be worrying more about the difficulty of getting their child to eat “right” rather than getting them to eat enough food. Our results do suggest that milk is a significant suppressor of appetite and that cutting down on milk consumption could be a useful way to promote eating habits that are more acceptable to parents.
This work was supported by the Northern and Yorkshire National Health Services Research and Development Funding.
We thank Ann Pattison for work on this study and Liz Towner for hosting this phase of the study in her unit. Thanks also go to Jayne Kelly and Anne Trail for collecting the earlier health-check data. The study would not have been possible without the invaluable support of the midwives at the Gateshead and Newcastle maternity units, the health visitors of Gateshead, and the loyal participation of all the parents.
- Accepted March 14, 2007.
- Address correspondence to Charlotte M. Wright, MB, BCH, MSc, MD, PEACH Unit, QMH Tower, Yorkhill Hospitals, Glasgow G3 8SJ, United Kingdom. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Beautrais A, Fergusson D, Shannon F. Family life events and behavioral problems in preschool-aged children. Pediatrics.1982;70 :774– 779
- ↵Wright CM, Parkinson KN, Drewett RF. How does maternal and child feeding behavior relate to weight gain and failure to thrive? Data from a prospective birth cohort. Pediatrics.2006;117 :1262– 1269
- ↵Wilensky D, Ginsberg G, Altman M, Tulchinsky T, Yishay F, Auerbach J. A community based study of failure to thrive in Israel. Arch Dis Child.1996;75 :145– 148
- ↵Wright CM, Parkinson KN, Drewett RF. The influence of maternal socioeconomic and emotional factors on infant weight gain and weight faltering (failure to thrive): data from a prospective birth cohort. Arch Dis Child.2006;91 :312– 317
- ↵Freeman JV, Cole TJ, Chinn S, Jones PRM, White EM, Preece MA. Cross sectional stature and weight reference curves for the UK, 1990. Arch Dis Child.1995;73 :17– 24
- Copyright © 2007 by the American Academy of Pediatrics