This study was conducted as part of the NutriNet-Santé study, which is a large ongoing web-based prospective cohort started in France in May 2009.The rationale, design and methods of the study have been described elsewhere . Its overall aim is to explore the relationships between nutrition and health and the determinants of eating behavior and nutritional status. Participants are adult volunteers (age ≥ 18 years) of the general French population with a scheduled follow-up of at least 10 years. At inclusion, participants have to complete several self-reported web-based questionnaires to assess their diet, their physical activity, anthropometric measures, lifestyle characteristics, socioeconomic conditions and health status. Participants complete this set of questionnaires every year after inclusion. Finally, another set of optional questionnaires related to determinants of eating behaviors, nutritional status, and specific aspects related to health are sent to every participant each month. A flowchart of the participants included in this study is available as Additional file 1.
This study was conducted in accordance with the guidelines of the Declaration of Helsinki, and all procedures were approved by the International Research Board of the French Institute for Health and Medical Research (IRB Inserm n° 0000388FWA00005831) and the Commission Nationale Informatique et Libertés (CNIL n° 908450 and n° 909216). Electronic informed consent was obtained from all participants.
Consideration of future consequences
Consideration of Future Consequences was assessed with the French version of the CFC-12 questionnaire  over a 6-month period from June to November 2014. The CFC-12 is a 12-item self-report questionnaire  developed to measure the extent to which individuals consider distant versus immediate consequences of their behavior. Each item is measured on a 5-point Likert scale ranging from “extremely uncharacteristic” to “extremely characteristic”. An example of the items of the CFC-12 is as followed: I consider how things might be in the future, and try to influence those things with my day to day behavior. The total score is obtained by summing each item ratings leading to a possible range from 12 to 60 (higher scores indicating greater consideration of future consequences). Participants were divided into 4 categories determined by quartiles of the total score (Q1, Q2, Q3, and Q4). A good internal consistency was obtained in our sample with a Cronbach’s α of 0.79.
To assess their organic food consumption, participants completed a semi-quantitative organic food frequency questionnaire (Org-FFQ) by providing the frequency and portion sizes of consumed foods and beverages. The Org-FFQ was administered over a 5-month period from June to October 2014. This questionnaire was based on a validated food frequency questionnaire  supplemented by a section pertaining to the frequency of organic food consumption. More precisely, participants were asked to report their frequency of consumption and the quantity consumed over the past year for 264 items allowing to assess total food intakes (g/d). In addition, the frequency of organic food consumption for each item was assessed with a 5-point Likert scale ranging from never to always. Organic food intake (g/d) was obtained for each item by applying a weight of 0, 0.25, 0.5, 0.75 and 1 to the five respective categories of frequency (never, rarely, half the time, often and always). A full description of the Org-FFQ as well as sensitivity analyses pertaining to weighting can be found elsewhere .
Beverage and food items were aggregated into 17 food groups: fruits and vegetables (including juices and soups); seafood; meat, poultry and processed meat; eggs; dairy products; starchy refined foods; whole-grain products; legumes; fats (oil, butter, and margarine); fatty sweets (including cake, chocolate, ice cream, and pancakes); non-fatty sweets (including honey, jelly, sugar, and candy); alcoholic beverages; non-alcoholic beverages; fast food; snacks (including chips and salted biscuits); dressings and sauces; and dairy products and meat substitutes (including soya-based products). For each food group, contribution of organic food consumed was estimated by computing the organic food intake of the food group (g/d) out of the total food intake of the food group (g/d) multiplied by 100. Total energy intake (kcal/day) was also calculated using a validated composition table . Participants with unlikely estimates of energy intake were identified as under- and over-reporting participants against estimated energy requirement. Basal metabolic rate (BMR) was calculated according to age, gender, weight and height using Schofield’s equations . The ratio between energy intake and estimated energy requirement (physical activity level x BMR, with physical activity level set by default at 1.55) was calculated and individuals with ratios below the 1st percentile (0.35) or above the 99th percentile (1.93) were excluded. These cutoffs were calculated on the validated FFQ for usual dietary intake used in the NutriNet-Santé cohort .
Socio-demographic, economic, anthropometric and lifestyle characteristics
Potential confounders of the relationship between CFC and organic food consumption were collected based on information provided yearly by the participants after their inclusion: age (years), gender, education level (primary, secondary, undergraduate, and postgraduate), occupational status (unemployed, student, self-employed and farmer, employee and manual worker, managerial staff and intellectual profession, intermediate profession, and retired), monthly income per household unit, place of residence (rural community, urban unit with a population < 20,000 inhabitants, urban unit with a population between 20,000 and 200,000 inhabitants, and urban unit with a population > 200,000 inhabitants), and BMI (kg/m2). More precisely, monthly income per household unit was calculated with information about income and composition. The number of people of the household was converted into a number of consumption units (CU) according to a weighting system: one CU is attributed for the first adult in the household, 0.5 for other persons aged 14 or older and 0.3 for children under 14 . Categories of income were defined as followed: < 1200; 1200–1799; 1800–2299; 2300–2699; 2700–3699; and > 3700 euros per household unit as well as “unwilling to answer”.
The Programme National Nutrition Santé Guidelines Score (PNNS-GS), which is an a priori nutritional diet quality score reflecting the adherence to the French nutritional recommendations of the participants , was considered as a confounder in the analyses. The original score includes 13 components: eight refer to food serving recommendations, four refer to moderation of nutrients or food, and one refers to physical activity. Points are deducted for overconsumption of salt and sweets. Points are also deducted from the total when energy intake exceeds the energy needs by more than 5%. A modified version of the PNNS-GS (mPNNS-GS) that did not include the physical activity component was used in this study. The score has a range of 0 to 13.5 points, with a higher score indicating a better overall nutritional quality of the diet.
The characteristics of the sample across quartiles of the CFC-12 were compared with linear contrast tests for continuous variables, and with Mantel-Haenszel chi-square tests for categorical variables. Logistic regression models were performed between organic food consumption as a dependent variable (organic food consumer versus non-organic food consumer (reference)) for each of the 17 food groups) and the four categories (quartile, Q) of the CFC-12 as the main independent variable (Q1 as reference). The strength of the association was estimated by calculating odds ratios (ORs) and 95% confidence intervals (95% CI). Furthermore, adjusted means of proportions of the contribution of organic food to the total food intake by food group were compared across categories of the CFC-12 for the 17 food groups among organic food consumers only. A percentage of the relative difference between adjusted means of Q4 and Q1 was calculated to estimate the effect size of the differences. For every analysis on each food group, participants who did not report to consume at least one food item of the group (organic or non-organic food intakes) were excluded from the analysis of this food group. Since socio-economic positions are associated with CFC  and dietary intakes, all adjusted models included the following confounders: age, gender, education level, occupational status, monthly income per household unit, and place of residence. In addition, it has been suggested that time perspective can predict or be predicted by health behaviors . Moreover, BMI, energy intake, mPNNS-GS (diet quality), and total food intake of the food group all predict the level of organic food consumption and were thus taken into account. No significant interaction terms were found between the CFC-12 and confounders. Missing data on confounding variables were handled with multiple imputation by chained eqs. (20 imputed datasets) .
All tests of statistical significance were 2-sided and significance was set at 5%. A Hochberg procedure was applied to correct for multiple testing. Statistical analyses were performed using SAS software (SAS Institute Inc., version 9.4).