Background/Objectives:

To observe fruit and vegetable consumption as it relates to body mass index (BMI) and other variables, by analyzing the health surveys of two countries—the United States and France—with traditionally distinct diets, and identifying factors that may explain the differences.

Subjects/Methods:

Two nationally representative surveys that assess food intake via 24-h diet recalls. Respondents include 2126 women and 1911 men from the US National Health and Nutrition Examination Survey, and 1572 women and 1141 men from the French Nutrition Barometer Survey. Standard linear regression models and t-tests of both simple and predictive marginal means were run employing the software package SUDAAN.

Results:

Americans appear to consume fruits and vegetables less often than their French counterparts (1.04 vs 1.33 times per day fruits; 1.98 vs 2.29 times per day vegetables). American men consume fruits and vegetables least often (0.98 times per day fruits; 1.88 times per day vegetables) than do American women or French men (1.10 vs 1.25 times per day fruits; 2.07 vs 2.18 times per day vegetables). French women consume fruits and vegetables most often (1.41 times per day fruits; 2.41 times per day vegetables). The French have lower mean BMI than Americans (23.34 vs 28.22 women; 25.20 vs 28.02 men). Regression analyses showed that age, BMI and educational attainment are strongly associated with frequency of consumption.

Conclusions:

These results support our hypothesis that the French tend to eat fruits and vegetables more often than Americans. This study proves to be an important first step in determining some of the influential factors that may affect various populations' consumption of fruits and vegetables.

Introduction

The World Cancer Research Fund estimates that up to 30–40% of global cancers could be prevented through a change in dietary patterns (World Cancer Research Fund, 1997; Holdsworth et al., 2000). Fruits and vegetables contain many phytochemicals, vitamins and nutrients that are essential for good overall health. Moreover, studies reveal that high consumption of these foods may be associated with lower body mass index (BMI) and a decreased prevalence of obesity, select forms of cancer and other chronic diseases (Vecchia, 2004).

Research indicates that individuals who adhere to a Mediterranean diet consume more fruits and vegetables than those eating a more conventional Western diet (Cordain et al., 2005). While a Mediterranean diet is characterized by an abundance of fruits and vegetables, legumes, fish, olive oil and other unsaturated fats, as well as whole-grain foods (Vecchia, 2004; Perrin et al., 2005), the Western diet is typified by dairy products, refined sugars and cereals, saturated oils, fatty meats and a low intake of fruits and vegetables (Cordain et al., 2005). The disparities in intake may also account for lower rates of diet-related conditions such as heart disease and diabetes among certain Mediterranean groups (Irala-Estevez et al., 2000; Vecchia, 2004) and higher rates in countries where a Western diet is the norm (Cordain et al., 2005).

Though France is one such Mediterranean country where more healthful eating habits are mainstream (particularly in the southern regions), a typical French meal can be rich in fat, especially saturated fats derived from various cured meats, butter, cheese and pastries (Volatier and Verger, 1999). Still, France is known for its copious year-round outdoor markets that sell fresh local products, especially fruits and vegetables (Volatier and Verger, 1999).

France's first national dietary survey (1993–1994), conducted by l'Association Sucre et Produits Sucrés Consommation et Communication, revealed that fruit and vegetable intake was low, especially with regard to young people, labor workers and individuals residing in northern parts of the country (Volatier and Verger, 1999). In an effort to combat this growing problem, l'Institut National de Prévention et d'Éducation pour la Santé (INPES) and le Ministère de la Santé created a campaign called Programme National Nutrition Santé (PNNS). PNNS' mission is to help improve the overall health of the French population by stressing energy balance, while placing particular focus on enhanced dietary habits to include more fruits and vegetables and less fat and alcohol (Le Programme National Nutrition Santé, 2007). PNNS recommends that the French eat at least five fruits and vegetables each day (Le Programme National Nutrition Santé, 2007). These recommendations have become critically imperative at a time when, according to l'Institut National de la Santé et de la Recherche Médicale, one out of 10 children will be obese by the age of 10 years. This number has doubled over the past two decades (Institut National de la Santé) and evidence suggests that 16% of French children are currently overweight (Institut National de la Santé et de la Recherche Médicale, 2006).

Across the United States, a Western diet is noted as the standard regimen (Cordain et al., 2005). As of 2005, based on a 2000-cal intake, the United States Department of Agriculture recommends that the Americans consume nine servings of fruits and vegetables per day (Office of Disease Prevention and Health Promotion, 2005). While there is some question whether this intake can be achieved, studies have shown that a low intake of fruits and vegetables combined with an increased intake of calories plays a prominent role in the high rates of obesity observed in the United States today (Cutler et al., 2003). According to the National Center for Health Statistics, 30% of US adults over the age of 20 are now obese, and 16% of children between the ages of 6 and 19 years are overweight (National Center for Chronic Disease Prevention and Health Promotion, 2007).

The purpose of this study was to examine adult fruit and vegetable consumption, as it relates to BMI and other variables in the United States and France, and to identify factors that may explain potential differences. This study is the first international comparison of fruit and vegetable intake between the two countries.

Materials and methods

Questionnaires

The National Center for Health Statistics: National Health and Nutrition Examination Survey

For the United States, we used the 2001–2002 National Health and Nutrition Examination Survey (NHANES), composed of 11 039 individuals, aged 6 months to 85+ years (National Center for Health Statistics, 2002). NHANES, established in the 1960s, is a series of reports that evaluates the health and dietary status of adults and children across the United States. This nationally representative survey includes both in-person and telephone interviews, as well as physical examinations. The diet intake interview portion includes 24-h diet recalls and a food frequency questionnaire.

L'Institut National de Prevention et d'Éducation pour la Santé: Nutrition Barometer Survey

For France, we used the 2002 Nutrition Barometer Survey (NBS), conducted by INPES, which comprised 3153 individuals, aged 12–75 years (Baudier et al., 2002). Launched in 1996, this questionnaire examines change and progression in behaviors, knowledge and attitudes with regard to food in the French population. NBS is a nationally representative telephone survey and includes a single 24-h diet recall, a dietary history, a food frequency questionnaire and a 2-week food recall.

Comparable coding

While there were 5575 distinct food choices available for the NHANES instrument, only 635 individual food choices were available for the NBS survey. Unlike NHANES, NBS does not query specific details of each food eaten or serving sizes. Instead, participants were asked to choose from a list of foods provided by the interviewer. The NHANES 24-h recall elicited participant responses for all foods eaten along with serving sizes, thus allowing for a much larger list of food codes.

Food codes of explicit interest were identified on both surveys, while primarily focusing on fruits and vegetables, and omitting most foods and mixtures deemed not principally fruit- or vegetable based. It must also be noted that vegetable sources with poor nutritional value such as French fries were excluded from both surveys. Accordingly, 322 NHANES food codes were labeled as fruits and 842 food codes were coded as vegetables; 47 NBS food codes were labeled as fruits and 71 food codes were coded as vegetables.

Study population and variables

We restricted our analysis to respondents aged 20–74 years for both surveys because we wanted to look at variables such as educational attainment and marital status. For the United States, height and weight were measured by the interviewers, whereas the French survey relied on self-report. Height and weight were used to determine BMI, defined as (weight in kilograms)/(height in meters)². Additional variables of interest were gender, age and 'day of intake'. Final study population resulted in 2126 women and 1911 men for NHANES, and 1572 women and 1141 men for NBS.

Consumption frequency

For this study, we included a single 24-h diet recall for both NHANES and NBS. To obtain fruit and vegetable consumption frequencies, we computed the number of times an individual mentioned consuming a food, summing up 24-h diet recall mentions of these foods for each person to achieve estimated daily frequencies of fruit and vegetable consumption.

One of the authors (TA-C), a registered dietitian, further collapsed the collections of selected NHANES and NBS fruit and vegetable food codes into smaller groups that cognitively corresponded to particular fruit and vegetable items; for instance, 'potato salad' became 'potato'.

Statistical analyses

Statistical analysis required the use of the software package SUDAAN (version 9.1) to properly account for the complex sampling designs used for both NHANES and NBS. Both surveys were treated as stratified cluster samples. NHANES provided stratum and cluster identification, whereas NBS did not. Thus, it was assumed that the entire NBS sample was derived from a single stratum; hence, each individual in NBS was treated as a separate cluster. Both surveys provided poststratification weights to ensure that the samples represented the populations from which they were drawn. Due to differences in the weighting schemes, and because the two instruments were carried out independently, each survey was analyzed separately.

We analyzed mean daily consumption frequency in each survey for fruits, vegetables and their sum (overall, and by gender). We used two-sample t-tests to determine whether the mean frequencies differed between subgroups in the same populations and also to test for differences between the two countries.

We then examined the estimated distribution of categorized demographic and other study variables in the two countries. Since there is evidence that individuals consume differently on weekends than on weekdays (Baranowski et al., 2000; Haines et al., 2003), we controlled for the effect of 'day of intake' in subsequent analyses.

Linear regression models were used to determine if observed variations in the distribution of demographic covariates (that is, age, marital status and education) explained part of the noted differences in fruit and/or vegetable consumption frequency, while controlling for 'day of intake' (weekend/weekday) and BMI (<25, normal weight; 25–29.9, overweight; 30, obese). Separate regression models were run for fruit and vegetable consumption frequency as well as the sum.

For each fitted model, an adjusted mean frequency of consumption was estimated using the method of predictive margins (Korn and Graubard, 1999) that was obtained by directly standardizing the predictive means from the regression model to the estimated joint distribution of covariates in the analysis.

Results

Demographic and other variables in the target populations were comparable, as shown in Table 1. The proportion of women is approximately 50% for both countries and the mean ages are 42.8 (United States) and 44.6 (France). The distribution of BMI in the two countries is strikingly different with 30.92% of the United States with a BMI 30 kg m^-2, compared to 8.12% of the French population. Although the French have lower mean BMI than the US population, the education distribution is reversed; 56.87% of the US population has an education above high school, which is roughly the same percentage of the French population with an education below high school. Finally, while not a demographic variable per se, it must be noted that nearly 65% of NHANES participants' 24-h diet recall fell on a weekend (Friday, Saturday or Sunday); a more even split between weekend and weekday was observed in NBS.

Table 1 - Sample characteristics by country.

Full table

We identified significant differences in consumption frequency for fruits, vegetables and their combination across populations and genders. Table 2 illustrates that French women consumed the most fruits and vegetables (1.41 times per day fruits; 2.41 times per day vegetables), followed by French men (1.25 times per day fruits; 2.18 times per day vegetables) and American women (1.10 times per day fruits; 2.07 times per day vegetables), whereas American men consumed the least (0.98 times per day fruits; 1.88 times per day vegetables).

Table 2 - Average consumption frequency of fruits and vegetables by country and gender.

Full table

When examining fruit intake solely (data not shown), number of mentions increased with age in both the US and French populations and between both the genders. BMI was a predictor for the US population (P<0.01) only, while education was highly significant for all four combinations, with number of mentions rising with level of education. Number of vegetable mentions also increased with age in US women and among the French (data not shown). BMI was a predictor for US women (P<0.05), and education was highly significant for US women and men.

Finally, the predictive marginal for overall fruit and vegetable consumption (Table 3) increased with age in both the US and French populations and between both genders. BMI was highly significant for US women (P<0.01), significant for French women and US men (P<0.05) and not significant for French men. Marital status had a significant effect for US women and men, and education was significant for all four combinations. Lastly, 'day of intake' was not significant except for French men.

Table 3 - Predictive marginal of fruit and vegetable consumption by gender and country.

Full table

In general, when controlling for covariates and demographic subgroups, the predictive marginal of fruit frequency, vegetable frequency and the sum of both, shows significant differences between the United States and France, with the French having significantly higher consumption frequencies (P<0.05). Exceptions include women and men in the 70–74 years age bracket and BMI <25 kg/m² category, men in the divorced/widowed cluster and both genders in the above high school level of education group.

Discussion

Numerous studies have examined dietary patterns in the United States and France separately (Volatier and Verger, 1999; Perrin et al., 2002; Stables et al., 2002; Serdula et al., 2004). However, not only have few studies endeavored to compare nutrition-related factors between these two specific countries (Drewnowski et al., 1996; Rozin et al., 2003), but also, to our knowledge, none has focused exclusively on fruit and vegetable consumption. Consequently, little is known about both common and distinct factors that influence intake in these two populations.

Our results show a negative association between fruit and vegetable intake and BMI. There is an unmistakable difference between the two groups, with the French eating fruits and vegetables more often and maintaining lower levels of BMI. Frequency of consumption among the French seems to be affected by the same factors that impact American consumption frequency. Unsurprisingly, stark disparities in intake by gender can be observed in both countries.

Education, arguably a proxy for income/socioeconomic status, has been shown to be one of the strongest influences on fruit and vegetable intake in the United States (Adler and Newman, 2002; Stables et al., 2002; Gregory-Mercado et al., 2006). In France, we found this relationship holds true to a lesser degree. In effect, some purport that one of the most pressing and shaping determinants of dietary habits is, in fact, culture (Holdsworth et al., 2000; Perrin et al., 2002). In France, as in many other nations, food practices are seen as an important distinctive aspect between diverse populations and are heavily molded by cultural and societal influences, which help set apart different groups (Ghadirian et al., 1995; Holdsworth et al., 2000). These cultural differences, environment and lifestyle variances often impact nutrition patterns. This might occur more so in France, where food appears to be so readily ingrained in the population's cultural values, identification and pride (Holdsworth et al., 2000; Perrin et al., 2005). As such, French citizens may place more emphasis on the experience of eating (variety, moderation and quality vs quantity), and a higher priority on food-related ecological factors such as what they eat, how much they eat and how much time they spend eating, than their US counterparts (Rozin et al., 2003; Rozin, 2005). The extent to which these cultural and environmental imprints are so in either country remains to be seen, as relatively few epidemiologic studies have yet to characterize, measure and analyze such issues, owing to a lack of defined metrics to truly measure culture. It is our contention that until such metrics are identified, we have, at best, an incomplete picture of what influences fruit and vegetable consumption in the United States and France.

Based on our findings, the consumption frequency of fruits and vegetables of the least formally educated group in France is comparable to even the most formally educated group in the United States. This detail, coupled with the differences in education distribution observed, might lead one to think that if a more equivalent division were noted, and if education were as much of a driving force in France as it appears to be in the United States, overall differences in intake might be even greater.

Among study limitations, NBS only permitted us to examine mentions of fruits and vegetables rather than portion sizes for foods consumed. Although this method is not ideal, using the frequencies allowed for a general comparison between the two countries. The NHANES 24-h recall elicited participant responses for all food eaten, thus allowing for a much larger list of food codes than observed in the NBS. Since the US survey had more food codes, it should have favored higher fruit and vegetable frequencies when compared to France; however, since the opposite was noted, we believe this methodological bias did not affect our results. Another limitation is the cross-sectional design used for both surveys and the utilization of one 24-h diet recall for comparison. As a result of this, only a single snapshot in time comparing fruit and vegetable frequency for each country is provided.

In addition, a misclassification for overweight and obesity may exist due to differing methods used to assess height and weight. For the United States, height and weight were measured directly by the interviewer, whereas the French survey relied solely on self-report. Still, measured and self-reported weights and heights have strong correlations (Field et al., 2007; McAdams et al., 2007); self-report appears to be reliable with younger adults (Kuczmarski et al., 2001), and is a valid measure in epidemiological studies (McAdams et al., 2007). Since a substantial difference was discerned in that roughly 31% of the US NHANES sample was found to be obese compared to only 8% of the NBS French participants, we believe that the discrepancy is unlikely due to reporting differences.

As previously noted, the method in which both surveys was carried out differs. NBS was a telephone survey, while NHANES used a face-to-face interview approach. It is possible that characteristics of individuals with telephones might be dissimilar in France compared to the United States. Nonetheless, we believe there are very little variations between the two populations in this respect since 95% of French households subscribe to a fixed telephone line (Moreau et al., 2004), which is similar to the United States (Brogan et al., 2001).

Another limitation is that a large majority of the NHANES 24-h recalls was conducted on a weekend, whereas a more even split between weekday and weekend was observed for the NBS interviewees. This could have potentially introduced a comparability issue in that a fair representation of fruit and vegetable intake between all days of the week was not achieved. In order to minimize this concern, we controlled for 'day-of-the-week'.

It is also true that fruit and vegetable consumption patterns may vary according to season; thus, the effect of seasonality would be important to explore. Unfortunately, though the NHANES survey is carried out over the course of an entire year (and so the 24-h noted are approximately equally distributed along the year) we do not have access to geographic locations by calendar time. Likewise, we do not have seasonality information for the NBS, as the survey is only carried out from early February to late March.

Finally, we were unable to examine disparities in income due to a lack of comparability between the two countries. Nevertheless, we feel this study may be an important first step in determining some of the influential factors that may affect various populations' fruit and vegetable consumption frequency.

In summary, while both groups fall short of their respective country's minimum fruit and vegetable recommendations, we find that the French consume fruits and vegetables more frequently than their US counterparts, even when demographic and BMI differences are taken into account. Our findings underscore the need for additional research to verify these results and to investigate those country-specific underpinnings that could serve to better understand the disparities found. A variety of factors such as culture, public health campaigns and initiatives and other relevant lifestyle factors not considered in this study could provide further insight. In turn, this knowledge may be helpful in promoting long-term behavior change through increased fruit and vegetable intake among populations.

References

1. Adler NE, Newman K (2002). Socioeconomic disparities in health: pathways and policies. Health Aff 21, 60–76. | Article |
2. Baranowski T, Davis M, Resnicow K, Baranowski J, Doyle C, Lin LS et al. (2000). Gimme 5 fruit, juice, and vegetables for fun and health: outcome evaluation. Health Educ Behav 27, 96–111. | Article | PubMed | ChemPort |
3. Baudier F, Carel D, Delamaire C, Gautier A, Guilbert P, Le Bihan G et al. (2002). Baromètre Santé Nutrition. Paris: Institut National de Prévention et d'Éducation pour la Santé.
4. Brogan DJ, Denniston MM, Liff JM, Flagg EW, Coates RJ, Brinton LA (2001). Comparison of telephone sampling and area sampling: response rates and within-household coverage. Am J Epidemiol 153, 1119–1127. | Article | PubMed | ChemPort |
5. Cordain L, Eaton SB, Sebastian A, Mann N, Lindeberg S, Watkins BA et al. (2005). Origins and evolution of the Western diet: health implications for the 21st century. Am J Clin Nutr 81, 341–354. | PubMed | ChemPort |
6. Cutler DM, Glaeser EL, Shapiro JM (2003). Why have Americans become more obese? J Econ Perspect 17, 93–118. | Article | ISI |
7. Drewnowski A, Henderson SA, Shore A, Fischler C, Preziosi P, Hercberg S (1996). Diet quality and dietary diversity in France—implications for the French paradox. J Am Diet Assoc 96, 663–669. | Article | PubMed | ISI | ChemPort |
8. Field AE, Aneja P, Rosner B (2007). The validity of self-reported weight change among adolescents and young adults. Obesity 15, 2357–2364. | Article | PubMed |
9. Ghadirian P, Shatenstein B, Lambert J, Thouez JP, PetitClerc C, Parent ME et al. (1995). Food habits of French Canadians in Montreal, Quebec. J Am Coll Nutr 14, 37–45. | PubMed | ChemPort |
10. Gregory-Mercado KY, Staten LK, Ranger-Moore J, Thomson CA, Will JC, Ford ES et al. (2006). Fruit and vegetable consumption of older Mexican-American women is associated with their acculturation level. Ethn Dis 16, 89–95. | PubMed |
11. Haines PS, Hama MY, Guilkey DK, Popkin BM (2003). Weekend eating in the United States is linked with greater energy, fat, and alcohol intake. Obes Res 11, 945–949. | Article | PubMed |
12. Holdsworth M, Gerber M, Haslam C, Scali J, Beardsworth A, Avallone MH et al. (2000). A comparison of dietary behaviour in Central England and a French Mediterranean region. Eur J Clin Nutr 54, 530–539. | Article | PubMed | ChemPort |
13. Institut National de la Santé et de la Recherche Médicale (2006). Obésité des Jeunes. Institut National de la Santé et de la Recherche Médicale: Paris, France.
14. Irala-Estevez JD, Groth M, Johansson L, Oltersdorf U, Prattala R, Martinez-Gonzalez MA (2000). A systematic review of socio-economic differences in food habits in Europe: consumption of fruit and vegetables. Eur J Clin Nutr 54, 706–714. | Article | PubMed | ChemPort |
15. Korn E, Graubard B (1999). Analysis of Health Surveys. New York: John Wiley.
16. Kuczmarski MF, Kuczmarski RJ, Najjar M (2001). Effects of age on validity of self-reported height, weight, and body mass index: findings from the Third National health and Nutrition Examination Survey, 1988–1994. J Am Diet Assoc 10, 28–34. | Article |
17. Le Programme National Nutrition Santé (2007). Campagnes de Communications. Institut National de Prévention et d'Éducation pour la Santé: Paris, France.
18. McAdams MA, Van Dam RM, Hu FB 2007)). Comparison of self-reported and measured BMI as correlates of disease markers in US adults. Obesity 15, 188–196. | Article | PubMed |
19. Moreau C, Bajos N, Bouyer J, The Cocon Group (2004). Access to health care for induced abortions. Eur J Public Health 14, 369–374. | Article | PubMed |
20. National Center for Chronic Disease Prevention and Health Promotion (2007). Overweight and Obesity. Centers for Disease Control and Prevention: Atlanta, GA, USA.
21. National Center for Health Statistics (2002). National Health And Nutrition Examination Survey 2001–2002. Centers for Disease Control and Prevention: Atlanta, GA, USA.
22. Office of Disease Prevention and Health Promotion (2005). Dietary Guidelines for Americans 2005. United States Department of Agriculture: Washington DC, USA.
23. Perrin AE, Dallongeville J, Ducimetiere P, Ruidavets JB, Schlienger JL, Arveiler D et al. (2005). Interactions between traditional regional determinants and socio-economic status on dietary patterns in a sample of French men. Br J Nutr 93, 109–114. | Article | PubMed | ISI | ChemPort |
24. Perrin AE, Simon C, Hedelin G, Arveiler D, Schaffer P, Schlienger JL (2002). Ten-year trends of dietary intake in a middle-aged French population: relationship with educational level. Eur J Clin Nutr 56, 393–401. | Article | PubMed | ChemPort |
25. Rozin P (2005). The meaning of food in our lives: a cross-cultural perspective on eating and well-being. J Nutr Educ Behav 37, S107–S112. | Article | PubMed |
26. Rozin P, Kabnick K, Pete E, Fischler C, Shields C (2003). The ecology of eating: smaller portion sizes in France than in the United States help explain the French paradox. Psychol Sci 14, 450–454. | Article | PubMed | ISI |
27. Serdula MK, Gillespie C, Kettel-Khan L, Farris R, Seymour J, Denny C (2004). Trends in fruit and vegetable consumption among adults in the United States: behavioral risk factor surveillance system, 1994–2000. Am J Public Health 94, 1014–1018. | Article | PubMed |
28. Stables GJ, Subar AF, Patterson BH, Dodd K, Heimendinger J, Van Duyn MA et al. (2002). Changes in vegetable and fruit consumption and awareness among US adults: results of the 1991 and 1997 5 A Day for Better Health Program surveys. J Am Diet Assoc 102, 809–817. | Article | PubMed | ISI |
29. Vecchia CL (2004). Mediterranean diet and cancer. Public Health Nutr 7, 965–968. | Article | PubMed |
30. Volatier JL, Verger P (1999). Recent national French food and nutrient intake data. Br J Nutr 81 (Suppl 2), S57–S59. | Article | PubMed |
31. World Cancer Research Fund (1997). Food, Nutrition and the Prevention of Cancer: A Global Perspective. World Cancer Research Fund: London, UK.

Acknowledgements

We acknowledge the invaluable assistance of l'Institut National de Prévention et d'Éducation pour la Santé (INPES). We also express gratitude to Andrea C. Dunlavy for critical review and scientific editing and to Josseline de Saint Just for translational aid. None of the authors had any personal, professional or financial links that compromised the scientific veracity of this study.