Dilemma between health and environmental motives when purchasing animal food products: sociodemographic and nutritional characteristics of consumers

The NutriNet-Santé Study (https://​www.​etude-nutrinet-sante.​fr) is an ongoing web-based prospective observational cohort study launched in France in May 2009 with a scheduled follow-up of 10 years. It aims to investigate the relationship between nutrition and chronic disease risk, as well as the determinants of dietary behavior and nutritional status. The study was implemented in the general French population (internet-using adult volunteers, age ≥ 18 years). The rationale, design and methodology of the study have been fully described elsewhere [21]. In brief, volunteers are recruited via multimedia campaigns (television, radio, national and regional newspapers, posters, internet) and invited to visit the NutriNet-Santé website. In order to be considered included in the study, participants complete a baseline set of self-administered, web-based questionnaires assessing dietary intake, physical activity, anthropometric characteristics, lifestyle, socioeconomic conditions and health status. As part of the follow-up, participants are requested to complete the same set of questionnaires every year. Moreover, each month, participants are invited by e-mail to fill in optional questionnaires related to dietary intakes, determinants of eating behaviors, nutritional and health status. This study is conducted in accordance with the Declaration of Helsinki, and all procedures were approved by the Institutional Review Board of the French 75 Institute for Health and Medical Research (IRB Inserm n° 0000388FWA00005831) and the Commission Nationale de l’Informatique et des Libertés (CNIL n° 908,450 and n° 909,216). All participants provided informed consent with an electronic signature. This study was registered in EudraCT on 25th February 2013 (n°2013–000929-31).

Food choice motives during purchasing in general and for specific food groups (meat, fish, fruits and vegetables, dairy products) were assessed in September 2013 using an optional web-based questionnaire which was internally validated using factor analyses [11]. The present analyses focused on the following food groups: meat, fish and dairy products. In addition, dilemma for choice of meat, fish and dairy products was assessed with two approaches.

The sociodemographic, economic and behavioral data that were the closest to the food choice motives questionnaire were used. In particular, data were collected on sex, age (years), BMI (kg/m²), educational level (primary, secondary, university (≤ 3 years), university (> 3 years)), household composition (one adult, two adults, household with at least one child, household with at least one teenager, household with at least one child and one teenager), smoking status (never smoker, former smoker, current smoker) physical activity (low, medium, high) and monthly household income (<1200, 1200–1800, 1800–2700, >2700 euros per household unit). The monthly household income is calculated per household consumer unit (CU). One CU is attributed to the first adult in the household, 0.5 CU - for other persons aged 14 or older, and 0.3 CU - for children under 14. Physical activity was assessed using a short form of the French version of the International Physical Activity questionnaire (IPAQ) [22]. The weekly energy expenditure expressed in metabolic equivalent task minutes per week was estimated, and 3 scores of physical activity were constituted [ie, low (<30 min/d), moderate (30–59 min/d), and high (≥60 min/d)] according to the French guidelines for recommended levels of physical activity [5].

At inclusion and once a year thereafter, participants are invited to complete three non-consecutive 24-h dietary records, randomly assigned over a 2-week period (2 week days and 1 weekend day). The accuracy of web-based 24 h dietary records has been assessed by comparing to interviews by trained dietitians [23] and against 24 h urinary biomarkers [24, 25]. For the present analysis, participants who had completed at least three 24 h dietary records during the two years preceding the questionnaire on food choice motives were selected. Participants reported all foods and beverages consumed at each eating occasion. They estimated the amounts eaten using validated photographs of portion sizes [26] using household measures or by indicating the exact quantity (g) or volume (mL). For each participant, daily mean intakes were calculated from the 24-h dietary records, weighted for the type of day of the week. Nutrient intakes were estimated using the published NutriNet-Santé composition table including more than 2000 foods [27]. Dietary underreporting was identified on the basis of the method proposed by Black [28].

For the purpose of this study, analyses focused on intake (g/100 g) of meat (red meat, poultry and game, organ meat), fish (fish, shellfish) and dairy product (milk, cheese, yoghurt) groups. The level of adherence to French dietary guidelines for fish (≥ 2 servings/week), dairy products (intake between 2.5–3.5 servings/day (individuals <55 years) and between 2.5–4.5 servings/day (individuals ≥55 years) was also assessed [5]. Since there are no specific dietary guidelines available for meat, guidelines for meat/fish/eggs were considered (≥1–2 servings/day) [5]. Finally, the mPNNS-GS was calculated. The mPNNS-GS is an a priori score assessing overall adherence to the French nutritional recommendations, and therefore reflecting overall dietary quality. It is based on the PNNS-GS score [29] but accounts for dietary component only, excluding the physical activity component [30]. Briefly, the score has a range of 0–13.5 points, with a higher score indicating a better overall diet quality. It includes 12 components: eight refer to food-serving recommendations (fruit and vegetables; starchy foods; whole grain products; dairy products; meat, eggs and fish; seafood; vegetable fat; water vs soda) and four refer to moderation (added fat; salt; sweets; alcohol). Negative points apply for overconsumption of salt (>12 g/day), added sugars (>15% energy intake). A penalty also applies when energy intake exceeds the energy requirement – as assessed by physical activity level and basal metabolic rate calculated using Schofield eqs. [31] – by more than 5%.

Characteristics of included and excluded participants were compared using Student’s t tests or chi-square test, as appropriate. The degree of agreement between dilemma for meat, fish and dairy products was calculated using unweighted Kappa [32]. Sociodemographic, economic and lifestyle characteristics as well as food intake were compared across dilemma (yes, no) for meat, fish and dairy products. Continuous variables were presented as means ± SDs, and categorical variables as percentages. Multivariable logistic regression analyses were used to assess the associations between sociodemographic and economic characteristics and the composite variable of dilemma (yes, no) for meat, fish and dairy products (dependent variable). Analysis of covariance was performed to estimate the association between the composite variable of dilemma for meat, fish, dairy products and intake of these foods as well as adherence with food group guidelines (dependent variables). Logistic regression analysis was performed to estimate the association between dilemma for meat, fish and dairy products and mPNNS-GS score (dependent variable). Analysis of variance and logistic regression analyses were adjusted for, sex, age, education level, income, household composition, smoking status, physical activity, BMI and energy intake. Sensitivity analyses were performed on the implicit and explicit approaches separately. All tests of significance were two-sided, and a P value <0.05 was considered significant. All statistical analyses were performed using SAS software (version 9.3, SAS Institute Inc.).

From the initial 122,091 subjects who received the questionnaire measuring food choice motives regarding sustainable foods, a total of 46,958 (38.5%) completed it. Among them, 402 subjects (0.9%) were excluded due to missing data for BMI, educational level or household composition. An additional 18,056 participants (38.5%) who had less than three dietary records available were excluded and 5565 participants (11.9%) due to underreporting, leaving 22,935 subjects (48.8% of those who completed the questionnaire) available for analysis (17,247 women and 5688 men).

Compared with excluded individuals, included individuals were more often men, were older, had a higher income and were less likely to have children (all P < 0.0001). They also showed a higher educational level (P < 0.01).

Percentage of participants who had a dilemma when choosing meat, fish or dairy products are shown in Table 1. More participants showed dilemma in the case of meat and fish than in the case of dairy products. In addition, the unweighted kappa coefficients between dilemmas for meat, fish and dairy products were as follow: meat vs fish (kappa = 0.43 (95% CI: 0.41–0.45), fish vs dairy products (0.33 (0.32–0.36)), meat vs dairy products (0.27 (0.26–0.29). Among participants, 22.4% presented a dilemma for at least one of these foodstuffs and 2.0% presented a dilemma for all these foodstuffs.

Table 2 shows socio-demographic, economic, lifestyle and food intake characteristics of participants who demonstrated dilemma vs no dilemma for each of these foods. Table 3 shows results of the multivariable logistic regression analysis evaluating the association between sociodemographic and economic characteristics and dilemma toward food choice. Men were less likely to report dilemma, while individuals belonging to low income class and aged >50 years were more likely to report dilemma. Dilemma also increased with age. In the case of dairy products, individuals aged >30 years (and in particular >50 years), with a low education level and a low income were more likely to report dilemma compared with those without dilemma. Sensitivity analyses were performed on the implicit and explicit approaches separately. Analyses using the implicit question showed additional associations in the case of meat: individuals with a greater educational level and household including only one adult were more likely to report a dilemma. No major differences were observed between the composite variables and the explicit approach.

Table 4 shows results of differences in food intake, adherence to dietary guidelines and the mPNNS-GS between individuals who reported dilemma and those who did not. Participants who reported a dilemma toward meat, fish and dairy products purchase consumed less of these foods overall and had greater mPNNS-GS scores compared with those without dilemma. In the case of fish, differences were small, although significant. In addition, participants who presented a dilemma toward meat showed a better adherence to meat/fish/eggs guidelines. Additional analyses focusing on the explicit and implicit approaches separately showed similar results. Observed differences concerned intake of fish which did not differ significantly using the implicit approach (P > 0.05), and adherence to food group guidelines for meat/fish/eggs which was not significantly different in individuals with and without dilemma (both implicit and explicit) for meat (P > 0.05).

Our study showed that a significant number of individuals were torn between health and environmental considerations when purchasing meat or fish (around 10%), while it was the case for 5% only in the case of dairy products. In addition, around one-fifth of the participants had a dilemma for at least one of these foods. To our knowledge, this study is the first to investigate potential dilemmas between health and environmental considerations when purchasing animal food. Occurrence of dilemmas is expected to be greater when there is a contradiction between health and environmental guidelines, but also when a product has a strong perceived environmental impact. In agreement, meat and in particular the one from ruminants present the highest greenhouse gas emissions together with fish, while dairy products score lower [33]. However, it has been demonstrated that it is possible to create a diet that is healthy and sustainable [16, 34, 35] without performing drastic dietary changes [16, 36] or increasing the cost to the consumer [16]. The public is aware of a number of dietary guidelines [6] which are widely disseminated [5], while awareness of what constitutes a low environmental impact diet is likely to be lower given the lack of information on this topic. In addition, consumer difficulty to understand what constitutes a good, balanced diet from a health, environmental, ethical or other perspective has been acknowledged [18] and can lead to dilemmas. More and more authors suggest that dietary guidelines should integrate recommendations for environmental sustainability [17, 37] as it has been done in other countries such as Germany [19]. It is likely that the level of dilemma will be lower in these countries.

In our study, women declared more dilemmas in the case of meat and fish compared with men, while no difference was observed for dairy products. In a previous study, women were also more likely to agree that they were ‘not sure whether to buy farmed fish’ [20]. Women are known to be more interested in healthy eating [38‐41] and to be more often concerned by environmental issues [9, 42‐44]. In particular, they were more willing to reduce their meat consumption [45]. However, women have been shown to rate almost all eating motivation as more frequently influencing eating behavior than men [38, 46], which may be due to women being more preoccupied with food in general [9, 47]. This could also partly explain why they declare more dilemma in general.

Our data showed that older subjects were more likely to perceive dilemma compared with younger subjects, for all food groups considered. Older age groups are known to be more interested in healthy eating compared with younger subjects [20, 40, 41, 46, 48]. It is however still unclear whether environmental concerns vary with age, with studies showing more pro-environmental attitudes in younger individuals [49, 50], while older subjects have been shown to be more familiar with the “meatless meals” topic [45] and to have greater ethical concern [9, 20, 48]. The greater health concerns observed in older subjects might partly explain why dilemma also increases with age. It has also been suggested that older consumers may have had a bias to give higher ratings on scales generally [48]. In our study, the risk of dilemma occurrence increased linearly with age in the case of meat and fish while dairy products showed a threshold effect starting at 50 years, with particularly high odds ratio. French dietary guidelines advise an increase of dairy products to four portions per day from the age of 55 years due to the high prevalence of osteoporosis in the elderly population [51]. This may increase dilemmas in older subjects also concerned by the environment.

Level of education had an influence on dilemma occurrence in the case of dairy products only, with less educated individuals showing more dilemmas. In a previous study, level of education had no influence on health motives to select food [38, 39]. In addition, attitude towards fish was not modified by a composite variable of socio-economic characteristics (educational level, occupation, household income and individual food expenditure) [20]. However, better educated individuals have been shown to express more pro-environmental attitudes compared with less educated individuals [49, 50]. People with higher education levels have more nutrition knowledge [52, 53] and might therefore be more able to compile nutritional and environmental information.

Finally, our data showed that individuals with low income had more dilemma compared with higher income individuals. The literature showed no influence of income on health motives underlying the selection of food [38] or intention to purchase sustainably sourced food [54]. Only minor effects for income was observed when it came to purchasing [43] or consuming [55] organic food. It is possible that individuals with low income perceive more dilemma in general when purchasing foods, since they have to take into account cost barriers, and as a consequence also perceived more health vs environmental dilemmas. In the case of meat, symbolic role of this food, such as its contribution to physical strength and energy, and social norms in low socioeconomic categories could also weight on the decision to maintain its important status in meals in this socioeconomic group [56], and may lead to dilemma related to the knowledge of the high environmental impact of this food.

Individuals with dilemma for meat and dairy products showed lower intake of these foods. These results suggest that dilemmas are translated into actual change in behavior and that barriers preventing changes can, at least partially, be overcome [15]. Our results showed limited association in the case of fish. In the literature, participants were less likely to report purchase of fish from sustainable source if they were confused about which type of fish they should be eating to protect fish stocks [20]. Dilemma for fish might potentially impact the type of fish purchased rather than the quantity consumed. Individuals reporting dilemmas for meat and dairy products had a diet of better quality overall. Our data therefore suggest that having environmental concerns is not inconsistent with a diet of adequate nutritional quality as previously shown [16, 34, 35]. In the literature, other studies showed that individuals with higher level of interest in health [57‐60], or environment [57] had healthier food choice behavior.

The major strength of our study is its large sample size providing high statistical power. Also, this study includes a vast heterogeneous sample of volunteers in whom a wide range of socio-demographic and lifestyle characteristics were assessed so as to effectively control for potential confounding factors [61]. Further strength of the current study was the use of repeated 24 h records to assess nutritional intake that were validated against urinary biomarkers [24, 25]. Another strength is the use of a composite variable of dilemma including both implicit and explicit approaches. In the case of meat, an explicit measure of dilemmas led to 30% of individuals with dilemmas whereas an implicit led to 20% of individuals with dilemmas only. These results highlight the importance to use an instrument with various items to measure the adhesion to a theory [62]. The choice of a composite variable might also have influenced the results. However, additional analyses were performed on implicit and explicit approaches separately and showed very similar results. The main limitation of the study was its cross-sectional design, preventing inference of causality. Caution is also needed when generalizing our results, since the NutriNet-Santé study is a long-term cohort focusing on nutrition and participants are recruited on a voluntary basis, implying that they might have increased health consciousness and interest in nutritional issues as well as healthier lifestyle. Individuals with dilemma may also have a healthier profile overall, which was partially accounted for in the models by adjusting on smoking, physical activity, BMI and energy covariables. Finally, the sample size can also be a constraint since it produces significant results even though differences are small, but it enables highly accurate estimates.