It is generally recognized that fruit and vegetables (F&V) form a part of a healthy diet. High daily F&V intakes are associated with reduced risk of becoming overweight and contracting diet-related chronic diseases, such as hypertension, CHD, type 2 diabetes and some types of cancer( Reference Boeing, Bechthold and Bub 1 ). Given these beneficial effects, continued efforts are made to promote F&V intakes. However, these efforts rarely target students aged between 18 and 25 years. This is unfortunate because F&V intakes of students are far below recommendations( Reference Keller, Maddock and Hannöver 2 – Reference Nour, Sui and Grech 6 ). The Dutch Nutrition Center recommends to eat two portions of fruit (200 g) and 250 g of vegetables daily( 7 ), but only 9 % of the Dutch population aged 19–50 years meets the fruit guideline and only 11 % meets the vegetable guideline( Reference van Rossum, Buurma and Vennemann 8 ). Students’ dietary behaviours, such as low F&V intakes, are likely to continue throughout adult life( Reference Mikkila, Rasanen and Raitakari 9 , Reference Craigie, Lake and Kelly 10 ). Promoting students’ F&V intakes might thus decrease their risk of becoming overweight and contracting diet-related chronic diseases during their adult life.
Insights into the determinants of students’ F&V intakes could inform future F&V interventions to improve students’ consumption. Previous studies have shown that students of Caucasian or Asian ethnicity, who are older, living with their parents or in houses where food is provided (e.g. campus housing) consume more F&V( Reference Alsunni and Badar 3 , Reference Peltzer and Pengpid 5 , Reference Nour, Sui and Grech 6 , Reference Mirabitur, Peterson and Rathz 11 , Reference Adams and Colner 12 ). In addition to these demographic characteristics, F&V intakes are also associated with lifestyle characteristics. Students who are more physically active, do not smoke, sleep more hours and drink less alcohol consume more F&V( Reference Alsunni and Badar 3 , Reference Peltzer and Pengpid 5 , Reference Adams and Colner 12 ). These previous studies included samples of college and university students from various countries( Reference Alsunni and Badar 3 , Reference Peltzer and Pengpid 5 , Reference Mirabitur, Peterson and Rathz 11 , Reference Adams and Colner 12 ). As far as we know, these associations have not been studied in a Dutch university student sample.
Next to demographic and lifestyle characteristics, it is well established that the physical food environment plays a key role in dietary behaviours( Reference Graham, Pelletier and Neumark-Sztainer 13 ). As university students spend a lot of time in and around their university environment( Reference Horne 14 ), this is an optimal environment for F&V interventions( Reference Nelson, Story and Larson 15 ). An effective F&V intervention strategy might be to enhance the availability of F&V( Reference Graham, Pelletier and Neumark-Sztainer 13 , Reference Roy, Kelly and Rangan 16 , Reference Hartman, Wadsworth and Penny 17 ), which could be done in many ways. Strategies that have received scant attention, but which might improve students’ consumption levels, are the presence of farmers’ markets, vegetable parcels or vegetable gardens( Reference McCormack, Laska and Larson 18 , Reference Pitts, Gustafson and Wu 19 ). Before developing F&V interventions suitable for the university environment, it is valuable first to investigate students’ perceptions of F&V availability and how they perceive the effectiveness of various F&V interventions in the university environment. This optimizes interventions by tailoring them to students’ needs, which may reinforce the effectiveness of the intervention.
The present study set out to investigate: (i) the F&V intakes of Dutch university students and (ii) the associated demographic and lifestyle characteristics; (iii) students’ perceptions of F&V availability in the current university environment; and (iv) students’ perceptions of F&V intervention strategies in the university environment. Results from the present study can guide the development of interventions to promote university students’ F&V intakes.
Materials and methods
The current cross-sectional study was part of the Green Healthy Student Research Programme, which is designed to create green and healthy study environments for students in the Netherlands( Reference van den Bogerd, Dijkstra and Seidell 20 ). The Medical Ethical Committee of the Vrije Universiteit (VU) Medical Center Amsterdam, the Netherlands, approved the study protocol. Participation was voluntary and all participants provided active informed consent.
Participants
In February and March 2016, self-reported data were collected by means of an online questionnaire and an identical paper version. Both versions were available in Dutch and English. The researchers contacted the communication divisions of all thirteen (public) universities in the Netherlands to ask if they were willing to distribute an online questionnaire among their students. Eight universities were willing to participate and distributed the online questionnaire via a recruitment text with an Internet hyperlink on their student webpage. These universities included: VU University, Erasmus University Rotterdam, Eindhoven University of Technology, Leiden University, University of Amsterdam, University of Twente, Radboud University of Nijmegen, and Wageningen University and Research (WUR). Student unions of these eight universities were also asked to spread the recruitment text with the Internet hyperlink via their social media accounts. These eight universities vary in academic fields and in locations. All included universities are located in urban areas in the Netherlands, of which four are situated in the conurbation of cities in the western part of the country (Randstad). Generally, there is good access to healthy foods in the Netherlands( Reference Helbich, Schadenberg and Hagenauer 21 ); hence, the availability of F&V between the areas where students live or where the universities are located probably does not differ a lot. Some Dutch universities have campuses offering multiple amenities such as sports facilities, bars and supermarkets. However, some inner-city universities do not have a campus and the amenities on-site are limited, but every university offers at least canteen facilities where healthy foods are available. It is not common for Dutch universities to offer student housing on campus. To increase the number of participants, a researcher (N.v.d.B.) and four students visited six of the eight participating universities to distribute the paper version of the questionnaire among students at the university canteens and restaurants.
Inclusion criteria were the ability to understand the Dutch or English language and enrolment in a Dutch research university programme. Students of applied sciences were not included in the present study. In total, 1069 students accessed the questionnaire of whom 70 % (n 749) completed the questionnaire. A total of thirty-two students were excluded because they were not enrolled at one of the eight participating universities (n 23), were a PhD candidate or an applied sciences student (n 4) or had missing values on both fruit intake as well as vegetable intake (n 5), leaving 717 students in the analyses.
Questionnaire
The questionnaire consisted of four clusters: (i) background information; (ii) evaluation of the university environment and perception of intervention strategies; (iii) evaluation of greenery in the university environment; and (iv) lifestyle information (see online supplementary material). For the purposes of the present study, data from clusters (i), (ii) and (iv) were used.
Fruit and vegetable intakes
F&V intakes were measured using a food frequency tool, which is also used in the Dutch Health Monitor( Reference van den Brink, Ocke and Houben 22 ). This tool provides a reasonably accurate representation of F&V intakes( Reference van den Brink, Ocke and Houben 22 – Reference Block, Gillespie and Rosenbaum 24 ). To measure fruit intake, students indicated how many days per week they usually consume fruit on an eight-point scale ranging from 0 (less than once per week) to 7 (seven days per week). Students then indicated the number of portions they usually consume on the days that they consume fruit on a seven-point scale ranging from 0 (less than one portion) to 6 (more than five portions). A portion of fruit was illustrated as a medium-sized apple, two tangerines or a handful of small fruit such as cherries or grapes. To measure vegetable intake, students indicated how many days per week they usually consume vegetables. A definition was given: ‘Examples are vegetable side dishes or salads. Vegetables as part of a large dish that also contains non-vegetable ingredients also count. Lettuce on a sandwich, however, does not’. In accordance with the guidelines of the Dutch Nutrition Center( 7 ), we did not count potatoes as a vegetable. Students then indicated the number of serving spoons of 50 g they usually consume on the days that they eat vegetables on a seven-point scale ranging from 0 (less than one serving) to 6 (more than five servings). Fruit intake (portions/d) and vegetable intake (g/d) were calculated by multiplying the days of intake by the number of portions or serving spoons of 50 g/d, divided by 7. The Dutch Nutrition Center recommends to eat two pieces of fruit and 250 g of vegetables daily( 7 ). Adherence (yes/no) to the Dutch F&V guidelines was calculated.
Demographic characteristics
Demographic characteristics were assessed by items from the Dutch Student Monitor Questionnaire( Reference van den Broek, Brink and van Vugt 25 ) and included: gender (female or male); age (22 years and older or younger than 22 years); ethnicity (Dutch or non-Dutch); current study discipline; and housing situation. Study discipline was assessed by an open-ended question and then categorized into: health-related studies such as medicine and biomedical science; humanities and social science studies; economics and law studies; and technical studies such as architecture, engineering and biology. Housing situation was established by asking: ‘What is your housing situation?’ The answer possibilities were: ‘I live in a shared household’, ‘I live independently’ and ‘I live with my parents/guardians’.
Lifestyle characteristics
Based on the literature( Reference Alsunni and Badar 3 , Reference Peltzer and Pengpid 5 , Reference Adams and Colner 12 ), six lifestyle characteristics were identified as potential characteristics associated with F&V intakes: physical activity; smoking (‘Do you smoke?’ yes/no); alcohol intake; stress; sleep duration; and BMI. Physical activity was assessed by two items: students first indicated the number of days per week they were physically active at a moderate intensity such as brisk walking or cycling for at least 30 min/d. Second, students indicated the number of days they were physically active at vigorous intensity such as playing soccer, tennis or exercising at the gym for at least 20 min/d. Adherence (yes/no) to the Dutch physical activity guidelines was achieved in the case of 30 min of moderate physical activity during a minimum of five days per week, or 20 min of physical activity at a heavily intense level during a minimum of three days per week( Reference Hildebrandt, Bernaards and Hofstetter 26 ).
To measure alcohol intake, students indicated how many days per week they usually consume alcohol and how many glasses they then usually consume on a seven-point scale ranging from 0 (less than one) to 6 (six or more glasses). In accordance with definitions given by Statistics Netherlands( 27 ), alcohol intake was then categorized as none, moderate (<21 glasses/week for males and <14 glasses/week for females) and heavy to excessive. Alcohol intake was considered heavy to excessive if it exceeded moderate consumption or if females consumed four glasses and males consumed six glasses or more at least once per week( 27 ).
Stress was assessed by two items( Reference Littman, White and Satia 28 ). Students indicated the amount of stress they usually experience on a five-point scale ranging from 0 (no stress) to 4 (extreme stress), and how well they can cope with stress on a five-point scale ranging from 0 (excellently) to 4 (very poorly)( Reference Littman, White and Satia 28 ). An average score (Cronbach’s α=0·7) was created and then dichotomized around the mean, with scores of 2 and higher reflecting high stress.
Sleep duration was calculated using four items that assessed the time students usually go to bed and wake up on weekdays and during the weekend. Based on recent recommendations, sleep duration was categorized as inadequate (less than 7 h per night), adequate (7 to 9 h per night) and excessive (more than 9 h per night)( Reference Hirshkowitz, Whiton and Albert 29 ). Because only twenty-nine students slept for less than 7 h per night, this subgroup was not included in the analyses.
BMI was calculated by taking students’ self-reported weight in kilograms and dividing it by the square of their self-reported height in metres. BMI was categorized as healthy weight (18·5–24·99 kg/m2), underweight (<18·5 kg/m2) and overweight or obese (≥25·0 kg/m2)( Reference Garrow and Webster 30 ).
Fruits and vegetables in the university environment and intervention strategies
Perceptions towards the availability of F&V in the university environment were assessed by asking students if they agreed with two statements: (i) ‘There are sufficient healthy foods available in the university environment’; and (ii) ‘There are sufficient fruits and vegetables available in the university environment’. Answer possibilities ranged on a five-point Likert scale from 0 (strongly disagree) to 4 (strongly agree).
Students’ perceptions of the effectiveness of various intervention strategies of increasing their F&V intakes were assessed by asking: ‘Would you eat more fruits and vegetables if you had access to the following facilities in your university environment (no or yes)? Five possible intervention strategies were asked based on literature( Reference Roy, Kelly and Rangan 16 , Reference McCormack, Laska and Larson 18 , Reference Pitts, Gustafson and Wu 19 ): (i) affordable F&V in the university canteen; (ii) affordable F&V in the university supermarket; (iii) a weekly local farmers’ market; (iv) vegetable parcels that can be picked up at the university; and (v) a university vegetable garden.
Statistical analyses
Descriptive statistics are presented as numbers with percentages for categorical variables and as means with standard deviations for continues variables. Univariate and multivariable linear regression analyses were used to analyse the association between demographic and lifestyle characteristic and F&V intakes. The final multivariable models were established through a backwards selection procedure. This selection procedure is preferred over a forward selection procedure because it is less likely to make a type II error( Reference Field 31 ). Characteristics were excluded from the multivariable model if the P value was higher than 0·10. Assumptions were checked, including normal distributions of the outcome variables, linearity and collinearity. Collinearity of the saturated and final multivariable linear regression models was checked with collinearity statistics; in all models, variance inflation factor values were below the threshold of 10 and all tolerance statistics were above 0·1, indicating little likelihood of biased data due to collinearity( Reference Field 31 ). The R 2 statistic represents the model fit. Effect sizes are presented as regression coefficients (β) with 95 % confidence intervals. All analyses were performed with the statistical software package IBM SPSS Statistics version 23, and P<0·05 was considered to identify statistical significance.
Results
Sample characteristics
The study sample comprised 717 students of whom 459 were female (Table 1). Half of the sample (50 %) was younger than 22 years, the greater part of students were of Dutch ethnicity (72 %) and about a third lived with their parents or guardians (31 %). A large percentage of students adhered to the physical activity guidelines (64 %), did not smoke (78 %) and reported high levels of stress (63 %).
Characteristics of the study population: Dutch university students (n 717) from eight (public) universities in the Netherlands, February–March 2016
About half of the sample (n 368, 51 %) filled out the online questionnaire. Univariate regression analyses showed no statistically significant difference between the paper version and the online questionnaire for fruit intake and vegetable intake. Of the total sample, 28·6 % studied at the VU University, 20·5 % studied at Erasmus University Rotterdam, 18·4 % studied at Eindhoven University of Technology, 11·0 % studied at Leiden University, 9·1 % studied at the University of Amsterdam, 7·5 % studied at the University of Twente, 3·3 % studied at Radboud University Nijmegen and 1·5 % studied at WUR.
Fruit intake
Fruit intake ranged from 0 to 4 portions/d with a mean intake of 1·37 (sd 1·00) portions/d. Adherence to the fruit guideline was met by 27·9 % of the students (Table 2). Univariate linear regression analyses showed a statistically significant association between students’ fruit intake and gender, age, ethnicity, housing situation, study discipline, physical activity and alcohol consumption (Table 3). All characteristics were included in the saturated multivariable linear regression model for fruit intake (results not shown). The model fit of this saturated multivariable model was R 2=0·12. The final multivariable linear regression model included eight variables: gender, age, ethnicity, housing situation, study discipline, physical activity, alcohol consumption and smoking (Table 3). Adjusted for all the variables in the final model, fruit intake was statistically significantly lower among students who were male, living independently compared with those living in a shared household, enrolled in a technical study compared with students following a health-related study, not adhering to the physical activity guidelines, and heavy to excessive alcohol drinkers compared with students who do not drink alcohol. The model fit of the final multivariable model was R 2=0·11.
Fruit and vegetable intakes and adherence to guidelines among Dutch university students (n 717) from eight (public) universities in the Netherlands, February–March 2016
† Mean fruit intake in portions per day.
‡ Mean vegetable intake in grams per day.
Univariate linear models and the final multivariable linear model for the association between demographic and lifestyle characteristics and fruit intake (portions per day) in Dutch university students (n 717) from eight (public) universities in the Netherlands, February–March 2016
β, regression coefficient; Ref., reference group.
*P<0·05, **P<0·001.
† Shows the regression coefficients of the variables included in the final model.
Vegetable intake
Vegetable intake ranged from 0 to 300 g/d with a mean intake of 126·18 (sd 64·89) g/d. Adherence to the vegetable guideline was met by 6·8 % of the students (Table 2). Univariate linear regression analyses showed a statistically significant association between students’ vegetable intake and ethnicity, housing situation and physical activity (Table 4). All characteristics were included in the saturated multivariable linear regression model for vegetable intake (results not shown). The model fit of this saturated multivariable model was R 2=0·06. The final multivariable linear regression model included six variables: age, ethnicity, housing situation, study discipline, physical activity and alcohol consumption (Table 4). Adjusted for all the variables in the final model, vegetable intake was statistically significantly lower for students who were non-Dutch, living with their parents compared with those living in a shared household, not adhering to physical activity guidelines, and moderate alcohol drinkers and heavy to excessive alcohol drinkers compared with students who do not drink alcohol. The model fit of the final multivariable model was R 2=0·05.
Univariate linear models and the final multivariable linear model for the association between demographic and lifestyle characteristics and vegetable intake (grams per day) in Dutch university students (n 717) from eight (public) universities in the Netherlands, February–March 2016
β, regression coefficient; Ref., reference group.
*P<0·05, **P<0·001.
† Shows the regression coefficients of the variables included in the final model.
Fruits and vegetables in the university environment
Table 5 shows that most students agreed or strongly agreed with the statements that their university environment offered sufficient healthy foods (60 %) and sufficient F&V (65 %). About 64 % indicated that they would consume more F&V if they had access to a university canteen with affordable F&V, and 60 % if they had access to a university supermarket with affordable F&V. Less than half of the students indicated that they would consume more F&V if the university environment offered a weekly local farmers’ market (44 %), vegetable parcels that could be picked up at the university (40 %) or a vegetable garden (29 %).
Perceptions towards the availability of healthy foods and fruits and vegetables (F&V) in their university environment and five intervention strategies aiming to increase their F&V intakes among Dutch university students (n 717) from eight (public) universities in the Netherlands, February–March 2016
Discussion
The current cross-sectional study was designed to provide insights into F&V intakes in Dutch university students. First, the study showed that F&V intakes were low and the vast majority of students did not adhere to Dutch F&V guidelines. Second, it identified risk groups for low F&V intakes based on demographic and lifestyle characteristics. Third, it showed that a large proportion of students perceived that there are sufficient healthy foods and F&V available in their university environment. Fourth, it showed that most students expected that affordable F&V in the university canteen or university supermarket would increase their F&V intakes.
The results of the current study showed that students’ mean F&V intakes were far below the recommendations, with only 28 % consuming at least two portions of fruit and 7 % consuming at least 250 g of vegetables daily. Although guideline adherence rates found in other student populations are based on five servings of F&V per day, they are comparable with the low adherence rates found in our study. For example, F&V recommendations are met by only 5 % of first-year German university students( Reference Keller, Maddock and Hannöver 2 ), 16 % of Saudi Arabian university students( Reference Alsunni and Badar 3 ) and 34 % of university college students from the UK( Reference Dodd, Al-Nakeeb and Nevill 4 ). This suggests that low F&V intakes among students are a widespread problem, and more efforts are needed to increase F&V intakes of students. It should be noted that our study sample was relatively healthy compared with the Dutch population. We found a higher adherence rate to the fruit guideline (28 % v. 9 %) than in the general Dutch population aged 19–50 years and a comparable adherence rate to the vegetable guideline (7 % v. 11 %)( Reference van Rossum, Buurma and Vennemann 8 ). Our sample also had a higher a adherence rate to the physical activity guideline (65 % v. 57 %), and included more smokers (78 % v. 71 %) and students with a healthy BMI (71 % v. 67 %) compared with the Dutch population of the same age( 27 ). An explanation for this might be that university students are more highly educated than the general Dutch population as a whole and therefore adapt healthier lifestyle choices( Reference Giskes, Avendaňo and Brug 32 ).
An essential aspect of the present study is the identification of several risk groups for low F&V intakes. Multivariable linear regression analyses showed that fruit intake was lower among students who were male, living independently, enrolled in a technical study, not adhering to the physical activity guidelines, and heavy to excessive alcohol drinkers. Vegetable intake was lower among students who were non-Dutch, living with their parents or guardians, not adhering to the physical activity guidelines, and moderate and heavy to excessive alcohol drinkers. Some of these associations are consistent with previous studies, which showed that non-European students who are more physically active( Reference Alsunni and Badar 3 , Reference Peltzer and Pengpid 5 , Reference Adams and Colner 12 ) and drink less alcohol( Reference Peltzer and Pengpid 5 , Reference Adams and Colner 12 ) consume more F&V. Although in the current study the associations between a younger age and low F&V intakes were not statistically significant, studies in non-European low-, middle- and high-income countries have shown lower F&V intakes in younger students( Reference Peltzer and Pengpid 5 , Reference Nour, Sui and Grech 6 ). Contrary to other studies, the current study results imply that, in the Netherlands, living in shared housing has a beneficial influence on students’ F&V intake. Studies among students from Asia, Africa and South America have suggested that living with parents or family has a positive influence on students’ F&V intakes( Reference Alsunni and Badar 3 , Reference Peltzer and Pengpid 5 ). Consistent with these studies, European students who live with their parents reported healthier dietary behaviours than students who do not( Reference El Ansari, Stock and Mikolajczyk 33 ). This discrepancy may be explained by the fact that living in shared student housing in the Netherlands also involves shared grocery shopping, cooking with F&V and eating together. Students who live independently might have insufficient resources to purchase fruits, and students who live with their parents or guardians might not be involved or have little influence on the vegetable groceries or the content of the family meals. However, future research should investigate these differences in more detail, for example by performing focus group discussions with students from different housing situations to discuss underlying mechanisms. Furthermore, the results of our study suggest that there are differences in F&V intakes between subgroups based on study discipline. In particular, students from a technical study reported lower F&V consumption than students from a health-related study. It is not surprising that those with an interest in and knowledge of health and the determinants influencing health are more prone to make healthier food choices. Yet, this finding may suggest that students enrolled in a technical study need additional support.
Most students reported that their university environment offers sufficient healthy foods and F&V. Nevertheless, students expected that F&V interventions in the university environment could encourage their F&V intakes. The most popular proposed intervention strategies were offering affordable F&V in the university canteen and university supermarket. Similar results were found in a Dutch study which reported that students aged between 13 and 15 years expected that free fruit at school would stimulate healthy dietary behaviours( Reference Ridder, Koning and Visscher 34 ). Moreover, Australian university students reported that they would like an increased availability of fresh fruit on the university campus( Reference Tam, Yassa and Parker 35 ). With the availability of affordable F&V, the university environment can play a pivotal role in encouraging students’ F&V intakes. Increased availability of healthy foods has been shown to be effective in promoting healthy dietary behaviours( Reference Roy, Kelly and Rangan 16 , Reference Caspi, Sorensen and Subramanian 36 ) and offering F&V that are affordable or lowering the price of F&V in the university environment could be effective in increasing intakes( Reference Alagiyawanna, Townsend and Mytton 37 ). However, more research is needed to establish what kind of price promotions on F&V are required to increase students’ F&V intakes and to find out what the effects are of F&V interventions in the university environment on students’ intakes.
The present study has several strengths. To the best of our knowledge, it is the first study that has investigated demographic and lifestyle characteristics associated with F&V intakes and, although explorative, perceptions of F&V interventions in Dutch university students. Not only did the findings extend the current knowledge, they appear relevant. Between some subgroups, fruit intake differed by 0·5 portions/d and vegetable intake differed by 30 g/d. Moreover, fruit intake and vegetable intake were not associated with the same characteristics, which can be valuable information for health promoters because it might demonstrate that promoting fruit intake requires a different approach from promoting vegetable intake.
Limitations should be noted as well. Selection bias might have led to reduced generalizability because the response rate was low and not all thirteen Dutch public universities participated. Response bias, such as social desirability, is common in self-reported questionnaires, and might have led to underestimations or overestimations of the present associations. Nevertheless, F&V intakes were measured with a validated short FFQ( Reference van den Brink, Ocke and Houben 22 – Reference Block, Gillespie and Rosenbaum 24 ) that accurately identifies the F&V intakes in samples comparable to the student sample( Reference Slater, Enes and López 38 , Reference Lillegaard, Øverby and Andersen 39 ). Additionally, the R 2 of the multivariable regression models were low, indicating a large proportion of unexplained variance( Reference Twisk 40 ). This implies that other factors such as individual factors (e.g. personal barriers) or environmental factors (e.g. costs) also play a role in influencing students’ F&V intakes. We therefore recommend that future studies into determinants influencing students’ F&V intakes not only include demographic and lifestyle factors, but also environmental and individual factors.
Conclusion
In conclusion, given that the vast majority of Dutch university students did not meet the F&V guidelines, more efforts are needed to promote their F&V consumption. When promoting F&V intakes in students, health promoters should consider that F&V intakes differ between subgroups based on demographic and lifestyle characteristics; and, moreover, they should consider that fruit intake is associated with other factors than vegetable intake. Affordable and accessible F&V in the university environment could potentially increase students’ F&V intakes. Nevertheless, more research is needed to establish the effectiveness of such strategies. The findings of the present study could guide the design of interventions that aim to increase students’ F&V intakes.
Acknowledgements
Acknowledgements: The authors would like to thank the students who participated in this study and the students who helped with data acquisition. They also thank Liz Watkins for critical reviewing the manuscript on grammar and vocabulary. Financial support: This study was funded by the Triodos Foundation. The Triodos Foundation had no role in the design, analysis or writing of this article. Conflict of interest: The authors have no conflicts of interest to declare. Authorship: The authors’ responsibilities were as follows. N.v.d.B. conducted the study and wrote the first draft of the manuscript. J.M., J.C.S. and S.C.D. supervised the execution of the study, gave significant advice concerning interpretation of the results and critically reviewed the manuscript. All authors approved this manuscript for publication. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Medical Ethical Committee of the VU Medical Center Amsterdam. Written informed consent was obtained from all subjects.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S136898001800174X
