Nutrition †Obesity Essay

Objective: To assess the association between the consumption of fast food (FF) and body mass index (BMI) of teenagers in a large UK birth cohort. Methods: A structural equation modelling (SEM) approach was chosen to allow direct statistical testing of a theoretical model. SEM is a combination of confirmatory factor and path analysis, which allows for the inclusion of latent (unmeasured) variables. This approach was used to build two models: the effect of FF outlet visits and food choices and the effect of FF exposure on consumption and BMI. Results: A total of 3620 participants had data for height and weight from the age 13 clinic and the frequency of FF outlet visits, and so were included in these analyses. This SEM model of food choices showed that increased frequency of eating at FF outlets is positively associated with higher consumption of unhealthy foods (b ? 0. 29, Po0. 001) and negatively associated with the consumption of healthy foods (b ? A1. 02, Po0. 001). The SEM model of FF exposure and BMI showed that higher exposure to FF increases the frequency of visits to FF outlets (b ? 0. 61, Po0.001), which is associated with higher body mass index standard deviation score (BMISDS; b ? 0. 08, Po0. 001). Deprivation was the largest contributing variable to the exposure (b ? 9. 2, Po0. 001). Conclusions: The teenagers who ate at FF restaurants consumed more unhealthy foods and were more likely to have higher BMISDS than those teenagers who did not eat frequently at FF restaurants. Teenagers who were exposed to more takeaway foods at home ate more frequently at FF restaurants and eating at FF restaurants was also associated with lower intakes of vegetables and raw fruit in this cohort. International Journal of Obesity (2011) 35, 1325–1330; doi:10. 1038/ijo. 2011. 120; published online 28 June 2011 Keywords: fast food; overweight; ALSPAC Introduction Childhood obesity prevalence have risen dramatically in the last 30 years in the Western world with the most recent figures for England and Wales show that 17% of boys and 16% of girls are obese. 1 An increase in the availability of calorie dense foods is implicated as one of the factors in the aetiology of the obesity epidemic. Fast food (FF) is one section of the food market that has grown steadily over the last few decades and it was worth d8. 9 billion in the United Kingdom in 2005. 2 FF is typically quick, convenient, cheap and Correspondence: Dr LK Fraser, School of Geography, University of Leeds, University road, Leeds LS2 9JT, UK. E-mail: l. k. fraser@leeds. ac. uk Received 6 February 2011; revised 21 April 2011; accepted 12 May 2011; published online 28 June 2011 uniform in its production,3 but FF is often high in saturated fats, energy dense and has low micronutrient content. 4–9 Studies from the United States of America have shown that children who consume FF (when compared with children who do not eat FF) have higher energy intake and higher fat intakes9,10 as well as lower vegetable and milk intake. 10,11 Therefore, the consumption of such foods could possibly result in a positive energy balance; and hence, weight gain. There is some evidence from longitudinal studies in the United States of America that consuming FF as a teenager can result in weight gain in both early12 and middle adulthood. 13 FF is often marketed to children and adolescents through television, internet and movie advertising,14–17 with brand recognition being present from an early age. 18 The addition of toys as gifts with FF meals also attracts children. There is growing body of literature that has assessed the location of FF outlets and has found that areas of higher deprivation Fast food and body mass index LK Fraser et al 1326 have more FF outlets19–21 and that FF outlets are often located close to schools. 22–24 The majority of research to date has been undertaken in the United States of America, but a study that analysed the fat content of a FF meal in McDonald’s and Kentucky Fried Chicken outlets in 35 countries showed that the amount of fat varied considerably between countries, within the same FF outlet. 25 This means that results from studies in the United States of America may not be generalisable to other countries. This study aims to assess the cross-sectional association between the consumption of FF and the body mass index (BMI) of teenagers in a large UK birth cohort. Methods The data for this study were obtained from the Avon Longitudinal Study of Parents and Children (ALSPAC),26 which is a birth cohort study where pregnant mothers who lived in the old Avon County in the United Kingdom (the Bristol region) were recruited in the early 1990s. A total of 14 541 mothers completed recruitment. Because of retrospective recruitment the total sample size was 15 224 fetuses and 14 610 live births. This paper presents data on the teenagers who attended the year 13 clinic and completed the year 13 questionnaire. Variables The food frequency data were collected from the questionnaires completed by mother (or carer) and separate questionnaires completed by the teenagers themselves at age 13 years. The data used from the carer questionnaire (collected at the same time point) referred to the questions ‘How often does s/he eat in a FF restaurant? The responses to this question were collected as never/rarely, once a month, once every 2 weeks, once or twice per week, 3–4 times a week, 5 or more times a week. The carers were also asked ‘In total, how many portions of vegetables does s/he eat in a week (do not include potatoes)’, ‘In total, how many portions of raw fruit does s/he eat in a week? ’ These were free numerical responses, which were retained as a continuous variable for analyses. In the food frequency part of the teenager completed questionnaire the teenagers were asked ‘If you ever buy food yourself from outside school, or from school vending machines, how often do you buy and eat each of the following things (include after school and weekends): chips, burger, pizza, sandwich, pies or pasties, chocolate, crisps, fruit and other food. ’ The height and weight data were collected at clinic visits at B13 years. The exact age, sex, height and weight were used to calculate a BMI standard deviation score (BMISDS) for each participant (1990 UK reference dataset). 27 The teenagers International Journal of Obesity were classified as obese if their BMISDS was greater than the 95th percentile (BMISDS41. 64). The physical activity data were collected via accelerometry at the age 13 clinic visit. 28 The participants wore an accelerometer for seven consecutive days and the measure used from this is mean counts per minute, which is a continuous variable. A deprivation score was assigned to each participant by matching the coordinates of their residential address (when carer questionnaire was completed) to the appropriate lower super output area. Each lower super output area has an index of multiple deprivation score (Index of Multiple Deprivation 2007 (IMD))29 assigned from the local census data. This is a continuous variable in which a higher number indicates an area of higher deprivation. Ethnicity was assigned as per the child’s ethnicity into a binary variable of ‘white British’ and ‘other’ ethnicity. Statistical modelling Descriptive statistics were performed in STATA version 10 (StataCorp LP, College Station, TX, USA). A structural equation modelling (SEM) approach was chosen to allow direct statistical testing of a theoretical model. SEM has many benefits over traditional regression techniques, which include the ability to model equations simultaneously and the incorporation of latent variables. 30 SEM is a combination of confirmatory factor and path analysis, which allows for the inclusion of latent (unmeasured) variables. 31 This approach was used to build two models: the effect of FF outlet visits and food choices and the effect of FF exposure on consumption and BMI. The SEM analyses were undertaken in AMOS version 17. 0 (IBM SPSS, USA). The hypothesised model for food choices is shown in the results section (Figure 2). The observed variables are displayed as boxes and latent variables as circles. Each observed variable has an associated random error term and each latent variable has an associated disturbance term, which represents the variance in the latent variable that has not been explained by the observed variables associated with that latent variable. Regression paths are shown by singleheaded arrows and covariances by double-headed curved arrows. The model fit was assessed by two indices; the comparative fit index (CFI) and the root mean square error of approximation (RMSEA). The CFI is a comparison of the hypothesised model compared with an independence model where all parameters are assumed to be independent. The RMSEA gives an indication of ‘how well would the model, with unknown but optimally chosen values, fit the population covariance matrix if it were available’. 32 A combination of CFI40. 95 and a RMSEA of o0. 50 is a sign of good model fit. The w2-test of overall fit is very sensitive to large sample size so has not been used in these models. 30 The two models were constructed a priori using previous research. The nutritional content of chips, burgers, pizza and Fast food and body mass index LK Fraser et al 1327 pies are known to be high in saturated fat and energy and therefore are ‘unhealthy’,4–9,33 whereas fruit and vegetables are known to contain fibre and vitamins and so are classified as ‘healthy’. Exposure to FF outlets is known to be higher in areas of higher deprivation. 19–21 In the food choices model, unhealthy consumption (latent variable) was modelled from the frequency of consumption of chips, burger, pizza and pies (reported by the teenagers themselves), and the healthy consumption was modelled from the number of pieces of vegetables and raw fruit consumed by the teenager (maternal report). The number of times that the teenager visited a FF outlet (maternal report) was regressed on the unhealthy and healthy consumption variables. The model for the effect of FF exposure on consumption and BMISDS is shown in Figure 3. Here exposure is a latent variable modelled from maternal and paternal takeaway frequency and deprivation score. The exposure is regressed on the number of visits to FF outlet. The BMISDS at age 13 years is the main outcome of this model. Ethical approval for the study was obtained from the ALSPAC Law and Ethics Committee and the local research ethics committees. Results A total of 3620 participants have data for height and weight from the age 13 clinic and the frequency of FF outlet visits, and were included in these analyses (SEM cannot use individuals with missing data). A total of 1711 (47. 3%) were boys and 456 (12.6%) obese. The descriptive statistics are shown in Table 1. Frequency of visiting FF outlets and food consumption frequencies are shown in Figure 1. The results of model 1 are shown in Figure 2 with regression weights shown in Table 2. This model showed that increased frequency of eating at FF outlets was positively associated with higher consumption of unhealthy foods (b ? 0. 29, Po0. 001) and negatively associated with the consumption of healthy foods (b ? A1. 02, Po0. 001). The CFI for model 1 was 0. 98 and the RMSEA was 0. 05 (90% confidence interval 0. 044, 0. 058). These represent good approximate model fit. Table 1. The results of model 2 are shown in Figure 3 with regression weight shown in Table 3. This model showed that increased exposure to FF increased the frequency of visits to FF outlets (b ? 0. 61, Po0. 001), which in turn was associated with higher BMISDS (b ? 0. 08, Po0. 001). Deprivation was the largest contributing variable to the exposure (b ? 9. 2, Po0. 001). The CFI for model 2 was 0. 98, and the RMSEA was 0. 021 (90% confidence interval 0. 009, 0. 033). These represent very good approximate model fit. Discussion This study shows that teenagers who are exposed to more unhealthy foods at home are more likely to eat at FF restaurants and have a higher BMISDS. The negative association of increased visits to FF outlets on consumption of healthy foods (fruit and vegetables) has also been demonstrated. The FF restaurant use in this analysis was reported by the mother or main carer of the teenager and showed that nearly 60% of all the teenagers ate at a FF restaurant at least once a month. This appears to be less frequently than in the United States of America, where studies showed that 60% of older children and adolescents ate FF more than once per week34 and that B30% of children ate at a FF restaurant on any typical day. 9. As one part of the SEM this study showed that eating at a FF outlet was associated with a higher BMISDS. There were no previous UK studies to compare these results with, but previous studies from the United States of America have not found consistent results. Boutelle et al. 11 found no association between frequency of FF consumption and adolescent BMI or weight status, and an Australian study Descriptive statistics Mean BMISDS Deprivation (IMD 2007)29 Physical activity (c. p. m. ) Raw fruit (portions per week) Vegetables (portions per week) s. d. Median IQR 0. 29 13. 7 541 9. 5 9. 5 1. 14 11. 4 190 7 7 0. 024 10. 6 511 8 8 A0. 47, 1. 06 5. 9, 17. 0 404, 653 5, 14 5, 12 Abbreviations: BMISDS, body mass index standard deviation score for age and sex; c. p. m. , cycles per minute; IMD 2007, Index of Multiple Deprivation 2007; IQR, interquartile range. Figure 1 Food frequency data. International Journal of Obesity Fast food and body mass index LK Fraser et al 1328 Figure 2 Results of SEM model of food choices. Table 2 Results of SEM model of food choices Regression weights a Unhealthy’fast food Healthy’fast food Chips’unhealthy Burger’unhealthy Fruit’healthy Vegetables’healthy Pizza’unhealthy Pies’unhealthy Estimate s. e. CR P 0. 285 A1. 023 1. 000 0. 732 1. 000 1. 157 0. 774 0. 530 0. 021 0. 124 13. 439 A8. 274 o0. 001 o0. 001 0. 016 45. 243 o0. 001 0. 148 0. 018 0. 016 7. 802 42. 483 32. 720 o0. 001 o0. 001 o0. 001 Abbreviations: CR, critical ratio; SEM, structural equation modeling. aAll consumption variables units: never/rarely, once a month, once every 2 weeks, once or twice per week, 3–4 times a week, 5 or more times a week. showed that FF eaten at home (but not away from home) was associated with higher BMI in adolescents (MacFarlane). Two longitudinal studies using data from the CARDIA study found that higher FF intake in adolescence was associated with higher BMI in young adulthood12 and those who ate FF more than twice a week had put on an extra 4. 5 kg of weight 15 years later. 13 The teenagers who ate more frequently at FF restaurants were more likely to eat less fruit and vegetables, as well as consume more unhealthy foods (chips, burger, pizza, pies) than those teenagers who ate at FF restaurants less frequently. This is an indication that the consumption of unhealthy foods may displace healthy food choices. This is similar to previous research in the United States of America, International Journal of Obesity which showed that children who ate FF consumed 45 g less vegetables per day than children who did not eat FF. 10 At age 13 years the food frequency data were a combination of maternal and self-report from the teenagers, but the total macro- and micronutrient values could not be assessed in this study as these data were not yet available at the time of writing. Deprivation was the largest contributor to the FF exposure variable. This could be explained by the fact that those of higher deprivation eat more FF because of the relative cheapness of FF. It has also been shown in many studies in the United Kingdom and the United States of America that areas of higher deprivation have more FF outlets than more affluent areas therefore, FF is more readily available. 35 An interesting economics paper from the United States of America showed that increasing the cost of FF by $1 could decrease BMI by 0. 78 units. 36 The increased consumption of unhealthy foods (chips, burger, pizzas and pies) by those teenagers who ate more frequently at FF outlets was not surprising, but the associated negative effect of the consumption of fruit and vegetables by these participants is important. These teenagers will not only be consuming more of the saturated fat and salt from the burgers, and so on, but at the same time they are not consuming important nutrients from fruit and vegetables. Although many FF outlets now offer more healthy alternatives such as fruit and vegetables, the consumers may still be choosing the unhealthy foods. Fast food and body mass index LK Fraser et al 1329 The FF question completed by the carer did not specify what constituted FF so some respondents may only count large franchises as FF whereas others may use a broader definition that includes independent takeaways. Although the frequency of eating at a FF restaurant was asked, the carers were not asked about the food eaten from these establishments and many FF restaurants now offer more ‘healthy’ alternatives. Although the majority of FF items do not meet the Food Standards Agency nutrient standards for total fat, saturated fat, sugar and sodium there are wide variations in similar products from different FF outlets with sodium content varying by up to four times in fried chicken products. 37 Therefore, having data on which food items were consumed from which FF outlet would further enhance future studies. There was no information on why the teenagers ate at FF restaurants, and key questions for the future include; was there no alternative eating establishments in their neighbourhood? Did they prefer FF to other meals or was the cost of food important? Conclusions This study has shown that the teenagers who ate at FF restaurants consumed more unhealthy foods and were more likely to have higher BMISDS than those teenagers who did not eat frequently at FF restaurants. Teenagers who were exposed to more takeaway foods at home ate more frequently at FF restaurants. Eating at FF restaurants was also associated with lower intakes of vegetables and raw fruit in this cohort. Figure 3 The SEM model of FF exposure and BMI. Table 3 Results of SEM model of FF exposure and body mass index Regression weights Fast food ’exposure. Maternal fast food’exposure Deprivation’exposure Paternal fast food’exposure BMISDS’fast food BMISDS’c. p. m. a Estimate s. e. CR 0. 61 1. 000 9. 20 0. 66 0. 08 0. 00 0. 07 8. 654 1. 07 0. 08 0. 02 0. 00 8. 605 8. 680 3. 586 A3. 351 P o0. 001 o0. 001 o0. 001 o0. 001 o0. 001 Abbreviations: BMISDS, body mass index standard deviation score for age and sex; c. p. m. , cycles per minute; CR, critical ratio; FF, fast food; IMD 2007, Index of Multiple Deprivation 2007; SEM, structural equation modeling. a All consumption variables units: never/rarely, once a month, once every 2 weeks, once or twice per week, 3–4 times a week, 5 or more times a week. Strengths/limitations This is a large dataset with good-quality height and weight data taken at clinic visits by trained staff using validated equipment. There were food consumption data about the teenagers available from both the teenagers and their carers, but this is a cross-sectional study so causation cannot be implied from this data. As expected in a longitudinal study there is attrition and the subcohort used in this study may not be truly representative of the whole cohort. Conflict of interest The authors declare no conflict of interest. Acknowledgements. We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting and the whole ALSPAC team, which include interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses. The UK Medical Research Council (grant ref: 74882), The Wellcome Trust (grant ref: 076467) and the University of Bristol provide core support for ALSPAC. LKF was funded by ESRC/MRC studentship. References 1 Craig RS. Health survey for England 2007, 2008. Available from http://www. natcen. ac. uk/study/health-survey-for-england-2007. 2 Keynote.UK fast food and home delivery outlets, 2006. International Journal of Obesity Fast food and body mass index LK Fraser et al 1330 3 DeMaria AN. Of fast food and franchises. J Am Coll Cardiol 2003; 41: 1227–1228. 4 Astrup A. Super-sized and diabetic by frequent fast-food consumption? Lancet 2005; 365: 4–5. 5 Brown K, McIlveen H, Strugnell C. Young consumers and the hospitality spectrum. Appetite 1998; 31: 403. 6 Harnack LJ, French SA, Oakes JM, Story MT, Jeffery RW, Rydell SA. Effects of calorie labeling and value size pricing on fast food meal choices: results from an experimental trial. Int J Behav Nutr Phys Act 2008; 5: 63. 7 Lewis LB, Sloane DC, Nascimento LM, Diamant AL, Guinyard JJ, Yancey AK et al. African Americans’ access to healthy food options in South Los Angeles restaurants. Am J Public Health 2005; 95: 668–673. 8 Paeratakul S, Ferdinand DP, Champagne CM, Ryan DH, Bray GA. Fast-food consumption among US adults and children: dietary and nutrient intake profile. J Am Diet Assoc 2003; 103: 1332–1338. 9 Schmidt M, Affenito SG, Striegel-Moore R, Khoury PR, Barton B, Crawford P et al. Fast-food intake and diet quality in black and white girls – the national heart, lung, and blood institute growth and health study. Arch Pediatr Adolesc Med 2005; 159: 626–631. 10 Bowman SA, Gortmaker SL, Ebbeling CB, Pereira MA, Ludwig DS. Effects of fast-food consumption on energy intake and diet quality among children in a national household survey. Pediatrics 2004; 113: 112–118. 11 Boutelle KN, Fulkerson JA, Neumark-Sztainer D, Story M, French SA. Fast food for family meals: relationships with parent and adolescent food intake, home food availability and weight status. Public Health Nutr 2007; 10: 16–23. 12 Duffey KJ, Gordon-Larsen P, Jacobs DR, Williams OD, Popkin BM. Differential associations of fast food and restaurant food consumption with 3-y change in body mass index: the Coronary Artery Risk Development in Young Adults Study. Am J Clin Nutr 2007; 85: 201–208. 13 Pereira MA, Kartashov AI, Ebbeling CB, Van Horn L, Slattery M, Jacobs DR et al. Fast-food habits, weight gain, and insulin resistance (the CARDIA study): 15-year prospective analysis. Lancet 2005; 365: 36–42. 14 Sutherland LA, MacKenzie T, Purvis LA, Dalton M. Prevalence of food and beverage brands in movies: 1996–2005. Pediatrics 2010; 125: 468–474. 15 Powell LM, Szczypka G, Chaloupka FJ. Trends in exposure to television food advertisements among children and adolescents in the United States. Arch Pediatr Adolesc Med 2010; 164: 794–802. 16 Hillier A, Cole BL, Smith TE, Yancey AK, Williams JD, Grier SA et al. Clustering of unhealthy outdoor advertisements around child-serving institutions: a comparison of three cities. Health Place 2009; 15: 935–945. 17 Lingas EO, Dorfman L, Bukofzer E. Nutrition content of food and beverage products on Web sites popular with children. Am J Public Health 2009; 99(Suppl 3): S587–S592. 18 Robinson TN, Borzekowski DLG, Matheson DM, Kraemer HC. Effects of fast food branding on young children’s taste preferences. Arch Pediatr Adolesc Med 2007; 161: 792–797. International Journal of Obesity. 19 Cummins SCJ, McKay L, MacIntyre S. McDonald’s restaurants and neighborhood deprivation in Scotland and England. Am J Prev Med 2005; 29: 308–310. 20 Fraser LK, Edwards KL. The association between the geography of fast food outlets and childhood obesity rates in Leeds, UK. Health Place 2010; 16: 1124–1128. 21 Macdonald L, Cummins S, Macintyre S. Neighbourhood fast food environment and area deprivation-substitution or concentration? Appetite 2007; 49: 251–254. 22 Neckerman KM, Bader MDM, Richards CA, Purciel M, Quinn JW, Thomas JS et al. Disparities in the food environments of New York City public schools. A J Prev Med 2010; 39: 195–202. 23 Davis B, Carpenter C. Proximity of fast-food restaurants to schools and adolescent obesity. Am J Public Health 2009; 99: 505–510. 24 Seliske LM, Pickett W, Boyce WF, Janssen I. Density and type of food retailers surrounding Canadian schools: variations across socioeconomic status. Health Place 2009; 15: 903–907. 25 Stender S, Dyerberg J, Astrup A. Fast food: unfriendly and unhealthy. Int J Obes 2007; 31: 887–890. 26 Golding J, Pembrey M, Jones R, Team AS. ALSPAC-The Avon Longitudinal Study of Parents and Children – I. Study methodology. Paediatr Perinat Epidemiol 2001; 15: 74–87. 27 Cole TJ, Freeman JV, Preece MA. Body-mass index reference curves for the UK, 1990. Arch DisChild 1995; 73: 25–29. 28 Riddoch CJ, Leary SD, Ness AR, Blair SN, Deere K, Mattocks C et al. Prospective associations between objective measures of physical activity and fat mass in 12–14 year old children: the Avon Longitudinal Study of Parents and Children (ALSPAC). Br Med J 2009; 339: b4544. 29 Index of Multiple Deprivation 2007 (IMD 2007). 30 Kline R (ed) Principles and Practice of Structural Equation Modeling. The Guildford Press: New York, 2005. 31 Tomarken AJ, Waller NG. Structural equation modeling: strengths, limitations, and misconceptions. Annu Rev Clinic. Psychol 2005; 1: 31–65. 32 Byrne BM (ed). Structural Equation Modelling with AMOS. Lawrence Erbaum Associates: London, 2001. 33 Astrup A, Dyerberg J, Selleck M, Stender S. Nutrition transition and its relationship to the development of obesity and related chronic diseases. Obes Rev 2008; 9: 48–52. 34 Taveras EM, Berkey CS, Rifas-Shiman SL, Ludwig DS, Rockett HRH, Field AE et al. Association of consumption of fried food away from home with body mass index and diet quality in older children and adolescents. Pediatrics 2005; 116: E518–E524. 35 Fraser LK, Edwards KL, Cade J, Clarke GP. The geography of fast food outlets: a review. Int J Environ Res Public Health 2010; 7: 2290–2308. 36 Powell LM. Fast food costs and adolescent body mass index: evidence from panel data. J Health Econ 2009; 28: 963–970. 37 Dunford E, Webster J, Barzi F, Neal B. Nutrient content of products served by leading Australian fast food chains. Appetite 2010; 55: 484–489. Copyright of International Journal of Obesity is the property of Nature Publishing Group and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder’s express written permission. However, users may print, download, or email articles for individual use.