The increase in prevalence of obesity in children and adolescents is of major public health concern(Reference Bundred, Kitchiner and Buchan1–Reference Ogden, Flegal and Carroll4). Obesity in childhood is associated with medical and psychosocial complications(Reference Ebbeling, Pawlak and Ludwig5) and is likely to track into adulthood(Reference Dietz2, Reference Ebbeling, Pawlak and Ludwig5–Reference Guo and Chumlea7). Prevention and intervention programmes for children have focused on family- and school-based initiatives, but long-term weight loss is difficult to achieve when adverse environmental factors overwhelm behavioural and educational techniques designed to modify diet and increase participation in physical activity(Reference Ebbeling, Pawlak and Ludwig5).
In light of the obesity epidemic, it is perhaps not surprising that dieting is such a prevalent behaviour among children(Reference Crow, Eisenberg and Story8–Reference Currie, Nic Gabhainn and Godeau11) some of whom employ healthy dieting strategies to lose weight(Reference Eaton, Kann and Kinchen9, Reference Neumark-Sztainer, Hannan and Story12). Self-reports of ‘dieting’ can, however, include unhealthy dietary practices and restrictive eating(Reference Nowak10, Reference Neumark-Sztainer, Hannan and Story12, Reference Middleman, Vazquez and Durant13), with negative impacts on nutrient intakes(Reference Neumark-Sztainer, Rock and Thornquist14) and dieting often clusters with other risk behaviours(Reference Crow, Eisenberg and Story8, Reference Nic Gabhainn, Nolan and Kelleher15, Reference French, Story and Downes16). There are also psychosocial risks associated with dieting, such as low self-esteem and depressive symptoms(Reference Crow, Eisenberg and Story8, Reference French, Story and Downes16–Reference Canpolat, Orsel and Akdemir19). In addition, dieting is associated with weight gain(Reference Field, Austin and Taylor20–Reference Neumark-Sztainer, Wall and Guo23) and is also a risk factor for the development of a partial or full eating disorder(Reference Neumark-Sztainer, Wall and Guo23–Reference Chamay-Weber, Narring and Michaud26). The risks associated with dieting during childhood and adolescence, although normative, have led many to believe that it is not a harmless behaviour(Reference Field, Austin and Taylor20, Reference Stice, Cameron and Killen22, Reference Patton, Selzer and Coffey25).
Dieting is not confined to those who are overweight and obese, but is also prevalent among children who are not overweight and who are the focus of the current study(Reference Crow, Eisenberg and Story8, Reference Vander Wal27–Reference Strauss35). Indeed, perception of being overweight and body dissatisfaction, irrespective of weight status, are considered key factors in weight loss attempts(Reference Neumark-Sztainer, Paxton and Hannan36). Prevalence rates of dieting among non-overweight children vary between countries, which in part relates to the range of questions used to define dieting status, and the age range and time frame under investigation. In a cross-national study, current dieting among non-overweight girls ranged from 9 % (Netherlands) to 28 % (Denmark)(Reference Ojala, Vereecken and Valimaa37). Studies from Spain(Reference López-Guimerà, Fauquet and Portell38) and Portugal (including boys)(Reference Fonseca and Gaspar de Matos33) provide estimates of 4·3 % and 6 %, respectively. Rates among girls in the United States range from 30 %(Reference Crow, Eisenberg and Story8) to 42 %(Reference French, Perry and Leon29); outside of the European Union and the United States estimates vary from 47 % (Qatar)(Reference Bener and Tewfik30) to 38 % (India)(Reference Chugh and Puri32) for dieting among non-overweight girls. Notably, the primary focus of these studies has been the behaviours of overweight participants, often employing the non-overweight group as a reference point(Reference Vander Wal27, Reference Boutelle, Neumark-Sztainer and Story28), or the behaviours of dieters in general(Reference Koff and Rierdan31, Reference López-Guimerà, Fauquet and Portell38, Reference Brugman, Meulmeester and Spee-van der Wekke39). Other studies, with data on non-overweight youth, explored eating disorder development among overweight children(Reference Vander Wal27, Reference López-Guimerà, Fauquet and Portell38, Reference Vander Wal and Thelen40) and among restrained eaters(Reference French, Perry and Leon29).
Although dieting among non-overweight children is well recognised and the types of weight control practices employed by this group have been explored to some extent(Reference Boutelle, Neumark-Sztainer and Story28, Reference Ojala, Vereecken and Valimaa37, Reference Neumark-Sztainer, Story and Hannan41), the simultaneous investigation of diet and the context of eating, in addition to perception of health, emotional well-being and somatic symptoms among non-overweight dieting children, has not been reported. Therefore, the current study examined non-overweight schoolchildren who were attempting to lose weight, aged 10–17 years in Ireland, with the aim of exploring outcomes such as food habits and food-related behaviour, in addition to the risks of poor self-reported health, negative body image and subjective health complaints associated with weight-reduction behaviours.
The current study utilised data from the 2006 Irish Health Behaviour in School-aged Children study, a part of the WHO collaborative study (WHO–HBSC; www.hbsc.org). In the Republic of Ireland, children in primary (aged 10–12 years) and post-primary (12–18 years) schools were randomly selected, and individual classrooms within these schools were subsequently randomly selected for inclusion. The overall student response rate was 83 %. Ethical approval was obtained from the Research Ethics Committee of the National University of Ireland, Galway. Data on 3599 students are presented here, aged 10–17 years, in which self-reports of weight, height and demographic information were available.
Data were collected using a self-completion questionnaire, administered by teachers, which was designed as part of the WHO–HBSC study involving researchers from all countries (www.hbsc.org).
To identify participants who were trying to lose weight, students were asked ‘At present are you on a diet or doing something else to lose weight?’, with response options ‘Yes’, ‘No, because I need to put on weight’, ‘No, but I should lose some weight’ and ‘No, my weight is fine’. Students who answered yes were defined as ‘those engaged in weight reduction behaviour’ and the remainder as those not engaged in weight loss attempts in the current study. Perceived body image was assessed by the question: ‘Do you think your body is …? Much too thin; A bit too thin; About the right size; A bit too fat; Much too fat’. Body dissatisfaction was categorised as ‘a bit’ and ‘much too fat’. Adequate validity and reliability of the weight reduction and body image questions have been reported(Reference Nemeth and Ojala42).
Information on height and weight were collected by asking, ‘How much do you weigh without clothes?’ and ‘How tall are you without shoes?’ which was used to calculate BMI (kg/m2) and which are considered reliable for population studies(Reference Strauss43). Students’ weight status was categorised by cut-offs corresponding to the international age- and gender-specific BMI reference values defined by Cole et al.(Reference Cole, Bellizzi and Flegal44) In the current study, the group of overweight adolescents includes those who are obese.
Students were asked how often they usually eat or drink the following items: fruit, vegetables, sweets, soft drinks, crisps, chips/fried potatoes and fish. They were provided with a seven-point scale with the following response options: ‘never’, ‘less than once a week’, ‘once a week’, ‘2–4 days a week’, ‘5–6 days a week’, ‘once a day, every day’ and ‘every day more than once’. Fruit and vegetable variables were dichotomised into more than once daily v. daily or less. Fish was dichotomised into weekly v. less than weekly and the remaining food variables into daily v. less than daily.
To assess weekday breakfast consumption, adolescents were asked to indicate how many days in a week they had breakfast (defined as ‘more than a glass of milk or fruit juice’). There were six response categories: ‘never’, ‘1 day’, ‘2 days’, ‘3 days’, ‘4 days’ and ‘5 days’. The responses were dichotomised into ‘daily breakfast consumption’ v. ‘less than daily’.
Students were asked how often do you ‘have breakfast together with your mother or father?’, ‘have an evening meal together with your mother or father?’, ‘eat a snack whilst you watch television (TV; including videos and DVDs)?’, ‘eat a snack while you work or play on a computer or games console?’ and ‘watch TV while having a meal?’ Response options were ‘never’, ‘less than once a week’, ‘1–2 days a week’, ‘3–4 days a week’, ‘5–6 days a week’ and ‘every day’. Responses were categorised as daily v. less than daily. The validity of the food frequency questions, the breakfast item and food-related lifestyle questions have been reported previously(Reference Vereecken and Maes45–Reference Vereecken, Rossi and Giacchi47).
Self-rated health was assessed by the question ‘Would you say your health is?’ and the response options were dichotomised at excellent v. good, fair or poor. Self-reported happiness was measured by the question ‘In general, how do you feel about your life at present?’ and the responses were dichotomised at very happy v. quite happy, don’t feel very happy and not happy at all. Children were also asked to rank themselves from 0 to 10 on a life satisfaction ladder(Reference Cantril48). This scale was used to identify those with high life satisfaction (response >7).
Subjective health complaints were assessed by asking children to report the frequency, in the 6 months before the survey, that they experienced a variety of emotional (feeling low) and physical symptoms (headache, stomach ache). Response options were ‘about every day’, ‘more than once a week’, ‘about every week’, ‘about every month’ and ‘rarely or never’. Responses were categorised at weekly or more v. less often. Items within the scale have shown an adequate content validity and test–retest reliability(Reference Haughland and Wold49).
Children were also asked to report the occupation of their parents through which a three-category social class (SC) scale was created using the UK Registrar General’s classification of occupations(50). SC was defined as high (SC 1 and 2), medium (SC 3 and 4) and low (SC 5 and 6).
Associations between weight reduction behaviour among non-overweight children (‘dieting’ among non-overweight children) and the outcome measures are expressed in OR from logistic regression models in the Statistical Package for the Social Sciences statistical software package version 15·0 (SPSS Inc., Chicago, IL, USA). All analyses were controlled for age, gender and SC, and were conducted separately for those who were not overweight, overweight and obese and the group as a whole. The reference group was the group of children who were not attempting to lose weight.
The current study presents data from 3599 children, 12·6 % of whom report that they are currently on a diet or doing something to lose weight. The percentage of girls who are engaged in weight reduction behaviour increases with age, from 7·8 % of 10–11-year-olds to 21·7 % of 15–17-year-olds, whereas the percentage of boys engaged in such behaviour remains relatively stable across age groups (9·2 % of 10–11-year-olds to 8·2 % of 15–17-year-olds).
Among non-overweight students, 10·3 % are currently engaged in weight reduction behaviours. Table 1 presents the children’s sociodemographic characteristics broken down by BMI and dieting status. Those who are not overweight and engaged in weight reduction behaviour are older (P < 0·01), more likely to be girls (P < 0·001), and from higher SC (P < 0·05). Those who are overweight and employing weight reduction behaviour are more likely to be male (P < 0·01) with no clear age or SC patterns.
Among non-overweight children, weight reduction behaviour was significantly associated with less frequent consumption of sweets, soft drinks, crisps and chips/fried potatoes and with more frequent consumption of diet soft drinks (Table 2). This group of children were less likely to have breakfast during the week than those not engaged in weight reduction behaviour within this BMI category, with adjusted OR of 0·62 (95 % CI 0·48, 0·80). Although there was a trend for not having family meals together, this was not statistically significant.
Reference group: those not engaged in weight reduction behaviour.
Odds ratio was significant: *P <0·05, **P <0·01; ***P <0·001.
†Adjusted OR, adjusted for age, gender and social class.
Engaging in weight reduction behaviour among non-overweight children was also found to be significantly associated with frequent headaches, stomach aches and reports of feeling low (Table 2). These children were significantly more likely to feel dissatisfied with their life and were less likely to report that they feel very happy. Perception of their health was poor with adjusted OR of 0·50 (95 % CI 0·37, 0·66). Body dissatisfaction, defined as feeling a little or too much fat, was positively associated with weight reduction behaviour with adjusted OR of 9·17 (95 % CI 6·99, 12·02). Corresponding figures for the group as a whole and the overweight group were 7·85 (95 % CI 6·13, 10·05) and 3·48 (95 % CI 2·03, 5·97), respectively. Gender differences in these patterns were explored by stratifying analyses; no consistent or substantial differences were identified, with two exceptions; the odds for girls engaged in weight reduction behaviour to report high life satisfaction and happiness were lower than those for boys, whereas girls engaged in weight reduction behaviour had higher odds for subjective health complaints than boys.
One-way analysis of covariance, controlling for age, gender and SC were conducted with all dependent variables listed in Table 2. Significant differences emerged between overweight/obese and non-overweight ‘dieters’ on two dependent variables; ‘dieters’ who were overweight/obese were less likely to report snacking while watching TV every day than those who were non-overweight (F = 3·89; df = 1.423; P < 0·05), and were more likely to report feeling a little or too much fat (F = 19·05; df = 1.420; P < 0·001).
Dieting is now considered normative among adolescents, which increases the need to investigate the various weight control strategies employed and psychosocial risks associated with dieting, by BMI status. To date, fewer analyses have focused on non-overweight children engaged in dieting for whom dieting may be associated with different risks than those who are overweight and dieting. The current study aimed to explore weight reduction behaviour among non-overweight schoolchildren in Ireland, investigating diet and family meal habits, but with a particular focus on subjective health complaints, perceived health, life satisfaction and happiness.
Weight reduction behaviour among non-overweight schoolchildren
The current study illustrates the extent of weight reduction behaviour among non-overweight schoolchildren; 10·3 % of children who are not overweight are currently attempting to lose weight. It is difficult to compare the extent of this problem to other studies due to methodological differences, but this is the first study to report this phenomenon among primary and post-primary schoolchildren in Ireland. In a cross-national survey conducted in 2001–2002 that included 13- and 15-year-old children, rates of current dieting to lose weight among non-overweight girls and boys in Ireland were 20 and 4 %, respectively(Reference Ojala, Vereecken and Valimaa37). The difference between these two studies is likely to reflect the increase in prevalence of dieting with age among girls. Non-overweight girls were reporting weight reduction behaviour more often than boys in the current study, which is consistent with the literature(Reference Pesa18, Reference Ojala, Vereecken and Valimaa37, Reference Vander Wal and Thelen40). More children from higher SC were engaged in such behaviours compared to those from lower SC, and although there is evidence that diet quality follows a socio-economic gradient(Reference Darmon and Drewnowski51), there appears to be inconsistency around the relationship between ‘dieting’ in general and SC. No clear consistent relationship between family affluence and dieting across countries(Reference Currie, Nic Gabhainn and Godeau11), or between dieting and SC(Reference Neumark-Sztainer, Rock and Thornquist14), have been reported for the children. However, others have reported a positive association between dieting and SC among 18-year-old college female students(Reference Drewnowski, Kurth and Krahn52). The current study appears to be the first report of a positive relationship between weight control behaviour and SC among non-overweight children.
Diet and meal habits
Non-overweight children who are trying to lose weight are not eating more health-promoting foods, such as fruit, vegetables and fish, but do report eating less high-fat, sugar-containing snack foods. Food choice of such ‘dieters’ has not been the focus of earlier work, unlike studies of overweight children and dieters per se, and so this finding is important to document. It is, however, consistent with reports of poor food choice among adolescent dieters generally(Reference Nowak10, Reference Neumark-Sztainer, Rock and Thornquist14), including children in Ireland who are dieting(Reference Nic Gabhainn, Nolan and Kelleher15).
Breakfast skipping was also prevalent among non-overweight ‘dieters’, which is consistent with the literature on weight loss practices employed by children generally(Reference Strauss35). This tends to result in an overall poorer quality diet(Reference Matthys, De Henauw and Bellernans53–Reference Rampersaud, Pereira and Girard56), coupled with the fact that these children are not eating more fruit, vegetables or fish, suggests that these children are at risk of nutritional inadequacies. The association between breakfast consumption and cognitive function at school(Reference Rampersaud, Pereira and Girard56) also suggests that these children may be at a disadvantage in terms of their participation at school.
Daily family meals (i.e. breakfast and evening meals) were found to be less common among non-overweight ‘dieters’ compared with ‘non-dieters’. Regular family meal times are associated with more healthful diets of children and adolescents(Reference Gillman, Rifas-Shiman and Frazier57–Reference Neumark-Sztainer, Hannan and Story60). Eating together is also an opportunity for parents to act as a model for healthy eating(Reference Neumark-Sztainer, Wall and Story61–Reference Neumark-Sztainer, Story and Perry63), to have discussions about food and health(Reference Gillman, Rifas-Shiman and Frazier57) and to contribute to the development of ‘regular’ meal patterns, all of which may also help the positive psychosocial development of children and adolescents(Reference Neumark-Sztainer, Hannan and Story60). Although the protective role of family meals against disordered eating among adolescents has been identified(Reference Neumark-Sztainer, Eisenberg and Fulkerson64), the reasons for less frequent family meals among non-overweight children who are engaged in attempts at weight loss has not been explored to date. It is possible that family structure, organisation within the family, communication and connectedness may influence the frequency and nature of family meals and may also pose as potential barriers to habitual family meal times.
Subjective health complaints
Somatic symptoms, such as stomach ache and headache, and psychological symptoms, such as ‘feeling low’, were more likely to be reported by non-overweight ‘dieters’. These psychosomatic symptoms can place an immense burden on the individual(Reference Roth-Isigkeit, Thyen and Stöven65) and can, in some cases, also impact on the health-care system(Reference Campo, Comer and Jansen-McWilliams66, Reference Campo, Jansen-McWilliams and Comer67). In most cases, it is thought that the presentation of such symptoms reflects not an organic disorder but an imbalance between the increasing educational, social, and perhaps sporting, demands on children and physiological ‘debts’ owed to rapid growth and sexual development(Reference Viner and Christie68). For this particular group, attempts at weight loss may present an additional ‘stress’ or demand. Subjective health complaints such as poor self-esteem, depressive symptoms(Reference Crow, Eisenberg and Story8, Reference French, Story and Downes16, Reference Neumark-Sztainer and Hannan17), increased irritability, problems with concentration and sleep disturbances(Reference Pesa18) have been associated with dieting, but the prevalence of somatic symptoms such as headaches and stomach aches among dieters has not been reported in the literature. It could be hypothesised that weight loss attempts may increase the likelihood of stomach aches, headaches and feeling low, with hunger playing a dominant role. The likelihood of ‘feeling low’ could also be increased by the desire, and perhaps the failure, to reach the thin ideal under pursuit.
Perceptions of health and well-being
The likelihood of reporting high life satisfaction, happiness and excellent self-rated health was decreased among non-overweight children who were ‘dieting’, but whether this is as a direct result of dieting or adolescence more generally could not be deduced from these analyses. Nevertheless, children who are of normal weight and are engaged in ‘dieting’ tend to be more unhappy and rate their health and satisfaction with life poorly. Interventions that could improve these outcomes are likely to have positive influences on health behaviours generally, including engaging in dieting.
There was a strong positive association between weight reduction behaviour among non-overweight schoolchildren and body dissatisfaction. This is consistent with earlier work showing that body dissatisfaction is associated with dieting and unhealthy weight control behaviours, irrespective of weight status(Reference Neumark-Sztainer, Paxton and Hannan36). Studies indicate that between 25 and 80 % of adolescent girls report significant body dissatisfaction(Reference Kelly, Wall and Eisenberg69), with 59 % of 15-year-old girls from an urban city in Ireland expressing a desire to be thinner. Elevated perceived pressure to be thin, thin-ideal internalisation and body mass are the main risk factors for body dissatisfaction(Reference Stice and Shaw70). Socio-cultural pressures to be thin from the mass media(Reference Groesz, Levine and Murnen71–Reference Field, Austin and Camargo73), and the role of the family(Reference Byely, Archibald and Graber74) and friends on body satisfaction(Reference Eisenberg, Neumark-Sztainer and Story75–Reference Field, Camargo and Taylor77), have been documented previously. The current study illustrates that non-overweight ‘dieters’ are at an elevated risk of body dissatisfaction; yet despite work to date, the best approach to fostering a positive body image among children remains unclear. Future work could explore the impact of body dissatisfaction.
One limitation of the current study is the cross-sectional design, making it impossible to determine causality or to differentiate a precursor of weight loss attempts among non-overweight children from a consequence for any of the relationships described. Another potential limitation includes reliance on self-report for the variables under investigation. However, students’ responses were anonymous; therefore, participants had no reason to misreport their responses. BMI based on self-reported data can produce lower prevalence estimates of overweight than those based on objective measurements, both among adults(Reference Connor Gorber, Tremblay and Moher78) and adolescents(Reference Himes, Hannan and Wall79). Others have reported high accuracy for classification of youth as obese or non-obese based on self-reported data(Reference Strauss43), and BMI based on self-reports has been found to be fairly reliable and suitable for identifying valid relationships in epidemiological studies(Reference Strauss43). Strengths include the large sample size and inclusion of children at a national level. All measurements have also been piloted and tested before administration to the children.
In adolescence, the need for a diet of high nutritional quality is paramount, given the physiological changes taking place and the emotional challenges young people face. An unbalanced diet can result in obesity or indeed undernutrition. Although obesity has received considerable attention as a major public health challenge, dieting in young people is perhaps perceived as less of a problem. One of the possible reasons is that dieting is considered a normative behaviour in children, despite the documented risks associated with it. Second, the challenge of identifying and working with those who are dieting currently lies with the dietetic profession. However, this report and others suggest that ‘dieting’ among non-overweight children requires the attention of various health professionals and not dietitians alone. Frequent somatic symptoms and poor emotional well-being will ultimately affect primary care, schools and families. Thus, the need to highlight weight reduction behaviours among non-overweight schoolchildren as a public health problem is paramount. Families and schools are most likely to observe the consequences of dieting and may be the very avenues that can help with identification and addressing such challenges.
The current study was funded by the Department of Health and Children, Ireland. There are no conflicts of interest. C.K. participated in the design and data collection for the study and drafted the manuscript. M.M. participated in the design and data collection for the study and edited the manuscript. S.N.G. participated in the design and data collected for the study, conceived of the study questions, performed the statistical analysis and edited the manuscript. All authors have read and approved the final manuscript. We acknowledge all the parents and children who consented and participated, as well as the management authorities, principals and teachers in all schools who participated. HBSC is an international study carried out in collaboration with WHO/EURO. The international coordinator of HBSC is Professor Candace Currie (University of Edinburgh) and the data bank manager is Dr Oddrun Samdal (University of Bergen). For details, see http://www.hbsc.org/