Towards dietary assessment and interventions for patients with Inflammatory Bowel Disease
Peters, Vera
DOI:
10.33612/diss.159023461
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2021
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Peters, V. (2021). Towards dietary assessment and interventions for patients with Inflammatory Bowel
Disease. University of Groningen. https://doi.org/10.33612/diss.159023461
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
8
CHAPTER 8
SUMMARY
In this thesis we clarified the role of nutrition as a pathophysiological factor in IBD, optimized dietary assessment tools and tried to increase high quality evidence for dietary interventions in IBD.
In Chapter 1 Inflammatory Bowel Disease (IBD) is described with a focus on the role of nutrition in its etiopathogenesis and its management. This thesis had three main goals: to aid clarification of how nutrition is involved in IBD (Part I), to develop an IBD-specific FFQ which will contribute to future diet studies in this population (Part II), and to provide high-quality research in the form of a systematic review on types of fibers and to present the set-up of a randomized controlled trial which tests the role of dietary interventions in disease susceptibility and its course (Part III). Furthermore, this chapter endorsed the necessity of multidisciplinary approaches to eventually develop and implement clinically effective dietary guidelines for IBD patients and to prevent IBD development in the general population.
Part I – Pathophysiological factors
In the absence of evidence-based clinical dietary guidelines for IBD patients, they follow self-defined “unguided” (i.e. without the guidance of a professional) dietary habits to alleviate symptoms. These unguided diets might lead to nutritional deficiencies and unintentional detrimental effects on disease course. Whereas differences in dietary intake were a-priori expected, Chapter 2 compared habitual dietary intake of 493 IBD patients (207 UC, 286 CD) to intake of 1,291 healthy individuals who participated in the LifeLines DEEP cohort study. Indeed, we found that 38.6% of patients in remission had protein intakes below the recommended intake of 0.8 g/kg per day. Of patients with active disease, 86.7% of patients had intakes below the recommended level of 1.2 g/kg per day. Statistical analyses showed that compared to controls, UC patients consumed more meat and spreads, but less alcohol, breads, coffee and dairy; while CD patients consumed more non-alcoholic drinks, potatoes, savory snacks and sugar and sweets but less alcohol, dairy, nuts, pasta and prepared meals. Patients with active disease consumed more meat, soup and sugar and sweets but less alcohol, coffee, dairy, prepared meals and rice, whereas patients in remission consumed more potatoes and spreads but less alcohol, breads, dairy, nuts, pasta and prepared meals. To conclude, the dietary behavior of patients raises concern as they avoid potentially favorable foods and gourmandize potentially unfavorable foods and often do not meet recommended protein intake levels.
The findings of Chapter 2 showed differences in habitual dietary intake between IBD patients and healthy individuals. These results contributed to the hypothesis that dietary patterns could impose a significant risk to the development and progression of IBD. Subsequently, in Chapter
3, dietary patterns were examined in the participants of the LifeLines cohort. Of 124,885
included participants, 190 developed UC and 73 CD within maximum 11-year follow-up. Five dietary patterns were identified explaining 41.8% of the dietary variation. The second pattern, characterized by high intake of condiments and sauces, non-alcoholic beverages, prepared meals, snacks and along with low intake of vegetables and fruits, which can be seen as an “Western” pattern, was associated with participants reporting to have developed CD (OR: 1.17, 95% CI:
8
1.02-1.32, p=0.017). However, this association was not confirmed when analyzing the second
robust dietary pattern (OR: 1.20, 95% CI: 0.93-1.54, p=0.164). The third pattern which includes
high intake of poultry, processed meat and red meat and can be classified as a “carnivorous” pattern, was associated with the risk of reporting to have developed UC (OR: 1.11, 95% CI: 1.01-1.22, p=0.024). This association was confirmed when analyzing the reversed third robust pattern (OR: 0.82, 95% CI: 0.71-0.94, p=0.004). All analyses were corrected for age, gender, BMI and smoking status. Furthermore, a higher LifeLines Diet Score (LLDS) which reflects a healthier eating pattern, was associated with a lower risk on reporting to have developed UC (OR: 0.97, 95% CI: 0.95-1.00, p=0.03). The same effect was seen for CD, although significance was lost after covariates correction (OR: 0.97, 95% CI: 0.93-1.01, p=0.10). Other dietary patterns were not associated with self-reporting disease development among participants. To conclude, whereas higher adherence to a Western or carnivorous dietary pattern was associated with CD and UC development respectively, a higher adherence to the LLDS was associated with a decreased risk of UC development. Future studies are needed to confirm these results. These findings show that it is important to consider a role for dietary advice in the general population to potentially aid prevention of IBD.
Next to the finding that diet is associated with reporting IBD occurrence (Chapter 3), diet is likely to play a role in the progression of IBD too. Up to half of IBD patients believe that diet contributes to flares. Hence, in Chapter 4, we studied the association between dietary patterns and flare occurrence in two geographically distinct Dutch cohorts. In a longitudinal study, 724 IBD patients, including 486 patients of the Northern cohort and 238 patients of the Southern cohort, were included and followed for two years. We identified three dietary patterns explaining 28.8% of the total variance of diet consumption. The most pronounced dietary pattern which explained 11.6% of total dietary variance, was characterized by intake of grain products, oils, potatoes, processed meat, red meat, condiments and sauces, and sugar, cakes and confectionery. Of the 427 patients in remission at baseline, 106 (24.8%) developed an exacerbation during follow-up. Higher adherence to this dietary pattern is significantly associated with increased risk of flare occurrence (HR: 1.51, 95% CI: 1.04-2.18, p=0.029). To conclude, we found that a dietary pattern characterized by grains, oils, potatoes, processed meat, red meat, condiments, and confectionery which corresponds with a “traditional (Dutch)” or ”Western” diet pattern, is associated with increased risk of flare occurrence in IBD.
Part II – Dietary assessment tools
Most existing nutritional assessment tools neglect the intake of important foods specifically consumed or omitted by IBD patients or incorporate non-Western dietary habits, making the development of appropriate dietary guidelines for (Western) IBD patients difficult. Hence, in Chapter 5, we described the development of a food frequency questionnaire (FFQ), the Groningen IBD Nutritional Questionnaires (GINQ); suitable to assess dietary intake in IBD patients. Multiple steps were taken, including identification of IBD specific foods, a literature search and evaluation of current dietary assessment methods. Expert views were collected and this semi-quantitative FFQ was developed. Next, the GINQ-FFQ was digitized into a secure web-based environment which also embeds additional nutritional and IBD related questions. The GINQ-FFQ is an online self-administered FFQ evaluating dietary intake taking
the previous month as a reference period. It consists of 121 questions on 218 food items. Validation of the GINQ-FFQ is still needed and therefore data is currently being collected.
Part III – High-quality evidence
Conflicting practice-based dietary recommendations on fiber intake are given to patients with IBD; whereas intake should be increased during remission, it should be avoided during relapse. Moreover, European countries set daily requirements of total fiber but do not specify any types. Therefore, in a systematic review in Chapter 6, data from randomized clinical trials (RCTs) on types of fibers for patients in the treatment of IBD is appraised to enable better dietary fiber advice on types of fibers in the future. The PubMED database was searched following PRISMA guidelines including quality assessment of the selected full-text articles using the Cochrane Risk of Bias Tool. Eventually, of the 186 found articles, 8 studies could be included, reporting on 4 types of fibers. We concluded that, although sparse, evidence on GBF and inulin seems propitious and merits further exploration. Furthermore, evidence on wheat bran and psyllium is still too limited. Adequately powered long-term human RCTs with objective outcomes are needed to improve dietary advice on types of fiber in IBD. Dietary guidelines are needed to better support patients with IBD. To develop those guidelines, appropriate evidence-based research is warranted. Hence, in Chapter 7, the study design of a randomized placebo-controlled trial (RCT) will be described. This multicenter RCT is set-up to study the effects of 1) dietary components transformed into the Groningen anti-inflammatory diet (GrAID), 2) the effect of a colon-delivered supplement containing a combination of vitamin B2, B3, and C (ColoVit) and 3) placebo (ColoPulse consisting microcrystalline cellulose) place on the course of CD, changes in the microbiome and inflammatory markers in a) patients and b) healthy controls. In study group 1 participants will have to adhere to an anti-inflammatory diet, in group 2 participants will have to take supplements with a mixture of vitamin B2, B3, and C twice daily and in group 3 participants will ingest a placebo twice daily. All dietary interventions have a duration of three months. Patients and physicians are blinded to the ColoVit and placebo. In total 510 participants will be included of which 255 are patients and 255 controls. The main study endpoints (effect of a dietary intervention) will be assessed in three research areas, namely clinical, mechanical and societal and differ for patients and controls. The clinical aim is to study whether those dietary interventions reduce the number of flares (defined as fecal calprotectin >200 µg/g and Crohn’s Disease Activity Index (CDAI)
>220) and improves quality of life of IBD patients. The mechanical aim is to investigate the
effect of those dietary interventions on the microbiome composition and diversity (primary outcome in healthy controls), fecal parameters such as fecal calprotectin (primary outcome in CD patients) and inflammatory and oxidative stress parameters. The societal aim is to assess the (post-intervention) adherences to those dietary interventions. As IBD patients are urgently seeking for proper dietary advice, the GrAID and/or ColoVit, when proven effective, could function as complementary therapy. In Chapter 8 (this chapter), the results of this thesis are summarized, discussed, and put into perspective by deliberating future directions.8
DISCUSSION & FUTURE PERSPECTIVES
A holistic view - Integration of all beneficial pathophysiological factors
The pathophysiology of IBD is a multifactorial process in which factors such as environment influences, genetics, immune system and gut microbiome play a role. The hypothetical model of pathophysiology as presented in the introduction describes that when risk factors accumulate and interact, disease or flares develop. Here, the focus of this thesis, nutrition, was highlighted and the conducted research chapters and future perspectives are integrated in the model, Figure 1.
Figure 1. Focus of this thesis, conducted research chapters and future perspectives in the context of the proposed IBD
model.
Differences in dietary intake were found between IBD patients and healthy individuals (Chapter 2). Besides, dietary patterns were identified and associated with IBD development in healthy participants (Chapter 3) and occurrence of flares in IBD patients (Chapter 4). Although these associations fall within the logical and pathophysiological expectations, an alternative explanation is reversed causality due to the fact that dietary data was only collected once during these studies. To exemplify, associations between dietary patterns and disease occurrence or disease relapse could be causal on one hand. On the other hand, changes in the dietary patterns could be caused due to IBD itself. In Chapter 2, we found changes in dietary habits which might be implemented after or due to the diagnosis of IBD. However, comparable dietary patterns were observed before (Chapter 3) and after disease development (Chapter 4).
In a complex multifactorial disease like IBD, one should not focus on a single pathophysiological factor, the therapeutic approach should address all factors involved. Our model shows that focus needs to be put on balancing all factors in play in an individual patient. Precision nutrition is part of precision medicine. On this path, gaining sufficient knowledge is needed to integrate dietary habits into patient-tailored therapy. Whereas, integration of all pathophysiological beneficial factors is the ultimate goal, every improvement noticeable to a patient is already a positive outcome. This approach might lead to an improvement for patients by increasing quality of life and it will save a lot of costs, healthcare related, or otherwise. Patients often ask their physicians what to eat to prevent flares and a patient-tailored approach gives patients a tool to self-manage their disease symptoms. So where some physicians see only little value for diet inducing remission, to some patients this little effectiveness of diet can be the difference between being able to participate in society or not. Thus, we should focus on combining all feasible improvements in a holistic patient-tailored manner, instead of focusing on one promising treatment.
BEHAVIORAL CHANGE
It is expected that pathophysiological mechanisms will be clarified further in the coming years based on the currently growing body of nutritional knowledge. Besides, efforts are being made to further clarify the role of other pathophysiological factors such as the microbiome, environmental and lifestyle factors and genetics (outside the scope of this thesis). In the future, a conduct of composite risk scores calculated based on each single pathophysiological factor, might be used not only to guide secondary or tertiary prevention in patients but also for primary prevention in the general population. A real-life example is MyIBDCoach.1 This is an application in which patients can fill out questionnaires allowing physicians to monitor patients’ disease course from a distance. As a follow-up of this thesis, it is planned to implement the Food Frequency Questionnaire of the Groningen IBD-specific Nutritional Questionnaires (GINQ-FFQ) into this system together with the Food Related Quality of Life (FR-QOL) questionnaire. Patients with dietary limitations will be identified and referred to dieticians specialized in IBD, who can utilize these tools (GINQ-FFQ and FR-QOL) to start-off a patients’ counseling session. They can discuss the dietary intake and identify problems together. These measures hopefully attribute to improvement of the dietary intake of IBD patients, i.e. preventing overly restrictive diets and nutritional deficiencies, or prevent IBD symptoms and flares.
In Chapter 3 we focused on the general population as represented by the participants of the LifeLines cohort. A Western and a carnivorous dietary pattern were associated with (self-) reporting development of IBD in these participants. An inverse association was found between the LifeLines score (a high score reflects a healthier diet) and risk on reporting UC development. In Chapter 7, healthy individuals are included in the study design and will be asked to adhere to Groningen Anti-Inflammatory Diet (GrAID) and their compliance will be evaluated. Can we enable these individuals to apply dietary modification when no direct result (i.e. prevention of illness) is visible? We often refer to diet as an “easy” to modify pathophysiological factor but that is of course relative. Though diet is a modifiable risk factor, changing patients’ habits is one of the hardest things to effectuate. An example,
8
almost everybody is aware that one should eat less meat and more vegetables to aid their personal health and the environment, yet many still consume too much meat and not enough vegetables. This underlines that multiple factors play a role in behavioral change; capability, opportunity, and motivation interact to generate behavior and vice versa.2 Therefore, we should keep these behavior influencing factors in mind when designing interventions or dietary measures for patients and the general population. Solely providing knowledge is not sufficient to motivate individuals to change their behavior, more measures are needed. In IBD, I believe that sustainable dietary behavioral change could be achieved since patients will experience benefits from a proven anti-inflammatory diet within a limited time frame.
DIETARY ASSESSMENT TOOLS
The LifeLines FFQ consists of a baseline questionnaire and multiple follow-up questionnaires. The baseline questionnaire tries to capture overall diet with as few questions as possible, whereas the follow-up questionnaires provide additional information on a limited number of food items (e.g. vegetables). This baseline questionnaire was not optimal to assess IBD patients since it was developed for the general population. When using such a questionnaire in the IBD population, you neglect the fact that patients modify their food intake to alleviate symptoms. Food items that IBD patients alternatively consume, might not be part of a general FFQ (e.g. non-dairy alternatives like soymilk). Such a general FFQ does not depict the overall intake of IBD patients correctly when too many regularly consumed alternative products are missing. Another reason why this FFQ is not optimal to assess IBD patients is due to the questionnaire’s structure, as described the FFQ consists of a baseline and multiple follow-up questionnaires. It assumes that dietary habits do not change much over time. This assumption is reasonable (unless someone develops a disease) and is the reason that FFQ are often seen as the best tool to assess long-term intake.3,4 Due to this assumption, the LifeLines questionnaire was designed to only cover the overall diet in the baseline questionnaire. Thus, some food items are only briefly questioned in the baseline questionnaire because these items are questioned in more detail in the follow-up questionnaires. However, in practice, the baseline LifeLines questionnaire is used without its follow-up questionnaires; the follow-up questionnaires were never administered to IBD patients and the habitual intake is calculated solely on the brief baseline questionnaire. Moreover, although the follow-up questionnaires were administered to healthy participants of the LifeLines cohort, so far, no strategy is developed to incorporate the follow-up questionnaires into calculating the overall diet (internal communication). Hence, in Chapter 3, only the baseline questionnaire was used to calculate dietary intake for healthy individuals too. Although these facts confirm that this questionnaire might not be the most suitable tool to assess habitual dietary intake in the IBD population, it was the best-available tool to compare dietary intake between our patient population and the LifeLines participants.
The goal of the second part of this thesis was to improve tools to measure dietary intake of IBD patients. As result, the development process of an IBD-specific FFQ was presented (Chapter
5). It is a questionnaire (GINQ-FFQ) that also includes alternative food items often consumed
by IBD patients. It consists of 121 questions on 218 food items and takes about 50 minutes to fill out. This tool will be an improvement to the beforementioned LifeLines FFQ. However, the GINQ-FFQ still has some limitations. The NEVO table (Dutch Food Composition Database)
does not incorporate every alternative food item these patients often consume. Moreover, because it concerns an FFQ, it should be concise since data quality decreases when patients spend too much time filling out a survey.3 Hence, a selection of items had to be made. Furthermore, general limitations of using an FFQ still exist; the GINQ-FFQ still relays on the memory of patients who tend to focus on the changes they have made in their diet and are often inclined to under- or over-report. There might be a difference in how patients under- or over-report compared to the general population.
The strength of this new FFQ is that it was developed specifically for the IBD population. Since IBD patients do not refrain from experimenting with their habitual intake to alleviate symptoms, it is necessary to develop tools that enable capturing their overall intake. This will improve the accuracy of the assessed food intake of this specific group and gives better insight into how diet plays a role in the pathophysiology of IBD. However, as said before, this tool is not limited to use solely in the IBD population since it incorporates alternative food items and still it reflects the overall intake of the general population.
Since the overall intake of IBD patients is based on a slightly different set of products, we could have potentially taken another approach to come to a tool that captures their intake. Many products included in Dutch FFQs are based on consumption in the general population as measured by the Dutch National Food Consumption Surveys5 (“Voedsel consumptie peiling” - VCP). The data is collected using two non-consecutive 24hour recalls and in some age groups these records are complemented with food diaries. We could have followed this approach and administer 24hour recalls specifically to IBD patients to identify what products are often consumed and include those in the GINQ-FFQ. This could still be done in the future to further optimize the GINQ-FFQ. Besides, the GINQ-FFQ needs to be validated in the IBD patients before it can be used in this population. Therefore, we have set-up a validation study in which all patients of the 1000IBD cohort (>1200 patients) will be invited to collect fecal samples (for potential identification of new biomarkers), fill out the GINQ online and keep a diet record for three non-consecutive days. The overall diet as captured by the diet records and the GINQ-FFQ will be compared as validation of the GINQ-FFQ. Furthermore, as many as possible alternative products were incorporated in the GINQ-FFQ and similar dietary modifications are often used by patients affected by a variety of disorders. Therefore, the GINQ-FFQ might be useful in such populations too and needs to be validated is these settings as well.
When efforts are made to improve current-available FFQs and develop population-specific FFQs, the underlying analyses should be optimized as well. Since the NEVO table is a summarization of commonly used food products, additional data is needed. A variety of multiple databases is currently being assessed to see what food composition data can be included. Besides, missing information of recently marketed products or products that have gotten a more prominent place in society should be considered to be complemented in the NEVO table. The Netherlands Organization for applied scientific research (TNO) is currently working on collecting (more detailed) additional food composition data to integrate in the NEVO table.
8
Next to the FFQ, there are other tools to measure dietary intake which were briefly mentioned in the introduction: screeners, 24hour recalls and food records. It should be considered if these tools also need improvement and if new tools should be developed. There are multiple innovative perspectives on what these new tools should look like. For example, a group in Switzerland is currently developing software which uses artificial intelligence to calculate the nutrients as visualized in a picture of your plate/dinner that you have taken with your smartphone.6 I speculate that we will see multiple tools pop-up that enable individuals to easily collect data. However, I do not expect an app using artificial intelligence to overcome the difficulties distinguishing between the constituents of similar products of two different brands.
Funding for prevention research
In my opinion, studies focusing on the prevention of diseases in the general population and in specific patient populations should be better supported. Setting up our RCT (Chapter 7) was limited by funding opportunities. It took more than 5 years to arrange funding. At first glance, it does not seem profitable to invest in a diet that promotes home cooking, preparing products from raw or base ingredients or components, and using whole foods. It does not fit in the current society since individuals seek convenience in every aspect of their lives. Companies respond to this culture by developing as many products as possible that shorten preparation time or limit preparation effort and subsequently serve consumers in their search for convenience. Companies seem to overlook that marketing such a diet can potentially save costs for society if it is shown to be helpful for primary, secondary or tertiary prevention. Corporate social responsibility becomes a more urgent theme for investors and stockholders7 and for that reason alone, companies should consider investing in prevention research and market such diets. In addition, governments start to see the potential of research focusing on prevention as for example demonstrated in the 2020-2030 strategic plan for Nutrition Research by the National Institute for Health (NIH)8 and the European food and nutrition action plan 2015–2020.9 It would be ideal if governments start supporting these prevention studies (even) more. Because Immune-Mediated Inflammatory Diseases (IMIDs) share common inflammatory pathways, the GrAID and ColoVit could potentially be utilized in disorders other than IBD. This means that it can have an even larger potential. Therefore, I also see a role for insurance companies since in the long run, less healthcare costs will be incurred if people are sick for a shorter period of time or don’t get sick at all.
Cultural change
We have evolved into a society in which meat is consumed daily,10 obesity is a common disorder,11 young children consume energy drink regularly12 and convenience food is the new standard.13 As mentioned, behavioral change is hard to effectuate, let alone cultural change. Luckily, we have come a long way since the first woman (Aletta Jacobs) was permitted to study at the University of Groningen (1871) and women were allowed to vote (around 1920). Nevertheless, the patriarchate is still present in our culture and women have not completely broken through the glass ceiling.14 This example demonstrates how complicated it is to accomplish cultural change. Although it might be hard to achieve nowadays, I advocate for a culture in which we are socially connected in real life (offline) and everybody has access to a healthy diet which we mainly consume to optimize body functions to prevent diseases,
and to food products which are produced with the goal to keep the planet a fruitful place. In other words, in an “ideal” world we all (women and men) put in time and effort into preparing home-cooked meals using whole plant-based ecologic foods and consume these meals with friends and family. As the Roseto effect showed, community social cohesion might be an important factor in the context of promoting health15 and could perhaps play a role in this “ideal” world.
How to pursue this “ideal” world? First, in this thesis efforts were made to further clarify the role of nutrition in the pathophysiology of IBD. These efforts should be continued and expended to other IMIDs. If we gain more knowledge, we will be able to better influence the behavioral change wheel.2 This wheel shows that there are multiple levels in which one can intervene to successfully change behavior. The levels in which you can intervene to achieve change are being called intervention functions and policy categories. An example of an intervention function is education. Perhaps, we could start small by teaching young children at ground school that they should consume fruits instead of cookies daily. Children attending high school could be taught how to prepare a “home”-cooked meal from raw or base ingredients. Another suggestion, future physicians studying at the University of Groningen spend only 5.5 hours on nutrition in the current medical curriculum.16 Modifications could be made to include more hours in the curriculum on nutrition and its effects on health and disease as wanted by 80% of the students.17 Furthermore, an example of a policy category is a fiscal measures. The government could decide to up taxes on meat and subsidize fruits and vegetables to influence the behavior of the population. Besides, it would be progressive if food producers take their social responsibility and produce products that are based on a circular economy instead of profits and are good for humans and the planet. To conclude, I suggest turning the behavioral change wheel until we find the optimal culture in which we have social cohesion and a healthy lifestyle as a standard especially for people at risk such as IBD patients.
8
REFERENCES
1. de Jong, M. J. et al. Telemedicine for management of inflammatory bowel disease (myIBDcoach): a pragmatic, multicentre, randomised controlled trial. Lancet 390, 959–968 (2017).
2. Michie, S., van Stralen, M. M. & West, R. The behaviour change wheel: A new method for characterising and designing behaviour change interventions. Implement. Sci. 6, (2011).
3. Keshteli, A. et al. A Dish-based Semi-quantitative Food Frequency Questionnaire for Assessment of Dietary Intakes in Epidemiologic Studies in Iran: Design and Development. Int. J.
Prev. Med. 5, 29–36 (2014).
4. Willett, W. C. Nutritional Epidemiology. Oxford
University Press (Oxford, 1998). doi:10.1093/ISBN
5. van Rossum, C., Nelis, K., Wilson, C. & Ocké, M. National dietary survey in 2012‐2016 on the general population aged 1‐79 years in the Netherlands. EFSA Support. Publ. 15, 1–25 (2018). 6. Schneider, T., Byun, J., Kim, S. & Fuchs, K. FoodCoach: Societal Implications of Digital Receipt-based Diet Monitoring & Interventions enabled by Graph Analytics Techniques. ETH
Zurich University of st. Gallen 1 (2020). Available at:
https://www.autoidlabs.ch/projects/foodcoach- societal-implications-of-digital-receipt-based- diet-monitoring-interventions-enabled-by-graph-analytics-techniques/. (Accessed: 26th October 2020)
7. McWilliams, A., Siegel, D. S. & Wright, P. M. Corporate social responsibility: Strategic implications. J. Manag. Stud. 43, 1–18 (2006).
8. NIH Nutrition Research Task Force. 2020
– 2030 Strategic Plan for NIH Nutrition Research.
(2020).
9. WHO. European food and nutrition action plan 2015 – 2020. Eur 24 (2014).
10. Rijksinstituut voor Volksgezondheid en Milieu. Consumptie van suikers in Nederland. (2019). 11. Wright, S. M. & Aronne, L. J. Causes of obesity. Abdom. Imaging 37, 730–732 (2012). 12. Owens, J. A., Mindell, J. & Baylor, A. Effect of energy drink and caffeinated beverage consumption on sleep, mood, and performance in children and adolescents. Nutr. Rev. 72, 65–71 (2014).
13. Fischler, C. [Culinary art and social change: some remarks]. Ann. Nutr. Aliment. 30, 415—425 (1976).
14. Schulpen, T. W. J. The glass ceiling: A biological phenomenon. Med. Hypotheses 106, 41– 43 (2017).
15. Hanibuchi, T. et al. Place-specific constructs of social capital and their possible associations to health: A Japanese case study. Soc. Sci. Med. 75, 225–232 (2012).
16. Dam-Nolen, D. H. K. van. Voeding en leefstijl in de opleiding Geneeskunde. 56 (2017). 17. van Dam-Nolen, H. K. Factsheet: Voeding en leefstijl in de opleiding (huisarts)geneeskunde Genoeg of te weinig? 2 (2017).
