The nutriRECIPE-Index - Development and Validation of a Nutrient-Weighted Index for the Evaluation of Recipes

doi:10.21203/rs.3.rs-22021/v1

Download PDF

Research article

The nutriRECIPE-Index - Development and Validation of a Nutrient-Weighted Index for the Evaluation of Recipes

https://doi.org/10.21203/rs.3.rs-22021/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 18 Nov, 2021

Read the published version in BMC Nutrition →

You are reading this latest preprint version

Background: Our objective was to develop a nutrient-based index for evaluating and improving menus in public catering. The nutriRECIPE-Index comprises 24 nutrients/nutrient groups, and in developing the index, the following steps were included: setting the goals of the index, nutrient selection, target metrics and scaling, weighing, proof of concept and the validation of the index. Furthermore, a unique database was created to integrate bioactive plant compounds in the assessment. An assessment of standard recipes and supposedly healthy recipes should show a significant difference in the results of the nutriRECIPE-Index. Finally, the nutriRECIPE-Index should prove to generate similar or more specific results than existing indices like the nutriScore and the Healthy Meal Index.

Methods : A whole meal cycle (comprising 6 weeks, 106 recipes and including different menu lines, partially with different side dishes) at a university canteen was analysed with the Federal Food Code (BLS) and the nutriRECIPE-Index. The Healthy Meal Index (comprising 3 nutritionally relevant items) and the nutriScore-algorithm (comprising 7 items) were used to validate the nutrient composition and the results of the nutriRECIPE-Index.

Results : The resulting scores of the recipes and menu lines show substantial differences, while the meals of a health-promoting menu line usually received higher scores than the standard recipes. A correlation between the scores of the nutriRECIPE-Index and the scores of the Healthy Meal Index (0.604) and the Nutri-Score (0.591) was observed. The nutriRECIPE-Index was better in identifying the worst menus, and in separating mediocre menus from the good ones.

Conclusion : The nutriRECIPE index is a useful and comprehensive tool to evaluate the nutritional value of recipes and is the first to consider bioactive plant compounds. Further adjustments to different target populations, settings, and cultural backgrounds are possible.

Nutrition & Dietetics

diet quality

public catering

Healthy Meal Index

Nutri-Score

public health nutrition

In addition to undernourishment, an unbalanced diet, also known as malnutrition, is one of the main disease risks globally. In 2017, malnutrition led to 9.5 million premature deaths from cardiovascular diseases, 0.9 million cancer deaths and 0.5 million deaths due to diabetes and chronic kidney disease^{(1, 2)}. In Europe, almost 25% of all premature deaths are caused by diet-related cardiovascular diseases⁽³⁾. A large part of this situation is due to the limited food offerings in the market, poor nutritional knowledge and lack of relevant information at the point of sale^{(4, 5)}. Several empirical studies have shown that improved information on the health value of food can support consumers in their purchasing decisions^{(6, 7)}.

In 2017, 7.4 billion euros were spent in Germany on company catering – an increase of 200 million euros and therefore almost 3% over the previous year⁽⁸⁾. In the EU-28, the food services sector recorded value-added of approximately EUR 175.5 million in 2016⁽⁹⁾. This demonstrates the growing importance of out-of-home catering. High consumption of meals outside the home has been shown to be associated with an increased body mass index (BMI)^{(10, 11)}; thus, improvement in the nutrient quality of these meals could make a substantial contribution to promoting the health of the population. Although policy makers have recognised the importance of nutrition for public health, activities to improve public nutrition are (mostly) limited to food labelling⁽¹²⁾. Moreover, the advertising of health-promoting services, including in canteens, is subject to legal limits within the EU ⁽¹³⁾. However, numerous reviews have shown that measures such as recipe changes, changes in portion sizes, adjusting prices, and more sophisticated labelling allow consumers to make more health-promoting choices^{(14, 15, 16)}.

Scope of existing models for evaluating nutrition and menus

Currently, a large number of a priori defined models (indices) for the assessment of nutrition (and nutrition patterns) exist; these tools are food- or food group-based or nutrient-based, or contain elements of both approaches⁽¹⁷⁾. They also differ in factors such as the objective of the application and scaling.

For individual menus, in particular, there are only a few models for evaluation and/or optimisation; examples include the Healthy Meal Index⁽¹⁸⁾, the Nutri-Score⁽¹⁹⁾, the susDISH method⁽²⁰⁾, NutriScale⁽²¹⁾, the Menu Sustainability Index⁽²²⁾, the NAHGAST method ⁽²³⁾ and the vegan checklist.⁽²⁴⁾ The association between index and health outcomes is often low^{(25, 26, 27)} or has not been explicitly evaluated.

Bioactive plant compounds (BPCs)

Although single, compound-specific dose-outcome curves are difficult to investigate in human trials (due to the abundance and diversity of BPCs in foods, in particular in fruits and vegetables), epidemiological studies show that BPCs – in addition to macro- and micronutrients – play a major role in the primary and secondary prevention of non-communicable diseases^{(28, 29, 30, 31)}. It has been shown that increased intake of lycopene – a carotenoid found in tomatoes and carrots – had beneficial effects on blood lipids, blood pressure, and endothelial function⁽³²⁾. Phytosterols – found in nuts and vegetable oils – have an LDL cholesterol-lowering effect⁽³³⁾. However, BPCs are currently not integrated into existing indices evaluating the nutritional quality of food.

Aim of the study

To overcome the limitations of existing nutrient-based evaluation models (narrow scope, arbitrary weighting of components, no explicit consideration of BPC), the nutriRECIPE-Index was developed and tested using 106 recipes from a university canteen. Therefore, a nutrient-weighted recipe evaluation model, which considers all relevant nutrients with an official recommendation status (according to the German, Austrian and Swiss Nutrition Societies), was developed. The model included BPC to more strongly consider the health value of secondary plant compounds. It was validated using the HMI and Nutri-Score.

Model development

Following the review of Waijers et al. ⁽²⁵⁾., the following steps were considered in the development of the nutriRECIPE-Index: definition of the aim of the index, choice of component type (e.g., nutrient, nutrient ratios or food groups), selection of the components (in this case various nutrients), target sizes and scaling, weighting, practical test and adjustment, and validation.

The aim of the nutriRECIPE Index is the assessment (and thus the possible optimisation) of individual recipes for meals, considering 19 macro- and micronutrients and five BPC groups with high public health relevance. Table 1 provides an overview of the nutrients and BPC groups included. Moreover, the nutriRECIPE-Index distinguishes between moderation and adequacy components as proposed by Thiele et al.⁽³⁴⁾ because a one-sided consideration of an advantageous or disadvantageous effect is less promising⁽³⁵⁾. While every nutrient can have beneficial or adverse effects, the nutriRECIPE-Index includes components with a minimum and maximum target value, considering current recommendations and nutrition patterns. The reference values for macro- and micronutrients of the German Nutrition Society (DGE), the Austrian Nutrition Society (ÖGE) and the Swiss Nutrition Society (SGE)^{(36, 37)} served as references. The reference value for protein was modified in line with more recent studies that found more beneficial effects when protein is ingested in amounts of 1.2 g per kg body weight per day^{(38, 39, 40, 41)}. The German Nutritional Society (DGE) and the World Health Organisation (WHO) have specified the minimum recommended protein intake as 0.8 g/kg body weight (BW)^{(36, 42)}, but the suggested optimal protein intake of 1.2 g/kg BW is used as the basis for calculating the nutriRECIPE-Index. It is also important to note that lunch is often the most protein-rich meal of the day. The maximum sugar in a meal followed the recommendation of the WHO⁽⁴³⁾, which corresponds with the recommendations of the consensus paper of the DGE on sugar consumption⁽⁴⁴⁾.

Table 1

**Nutrients integrated into the nutriRECIPE-Index, their target values and weighting factors**. Requirement values for macronutrients and micronutrients according to recommendations of DGE and WHO, Values for Bioactive Plant Compounds are calculated based on DGE-meal plans for one week, which incorporate the “Five-A-Day” recommendation for fruits and vegetables
	Nutrient	Daily recommendation	Unit	Average intake/day	Weighting factor
Favourable nutrients
Carbohydrates	Fibre	30	g	24.00	1.25
Proteins	Protein	84	g	78.75	1.07
Fats	Mono- and poly-unsaturated fatty acids	20	% of energy	20.00	1.00
Vitamins	Vitamin D	20	µg	3.35	3.00
	Vitamin E	14	mg	14.70	0.95
	Thiamine	1,2	mg	1.55	0.77
	Riboflavin	1,4	mg	1.95	0.72
	Vitamin B6	1,5	mg	2.30	0.65
	Folic acid	300	µg	314.00	0.96
	Vitamin B12	3	µg	5.40	0.56
	Vitamin C	110	mg	152.00	0.72
Minerals	Calcium	1000	mg	1581.00	0.63
	Magnesium	350	mg	412.50	0.85
	Iron	15	mg	13.70	1.09
	Iodine	200	µg	102.50	1.95
	Zinc	10	mg	10.90	0.92
Unfavourable nutrients
Carbohydrates	Added sugars	50	g	92.00	1.84
Fatty acids	Saturated fatty acids	10	% of energy	10.00	1.00
Minerals	Salt (sodium)	6	g	7.44	1.24
Bioactive Plant Compounds
	Carotenoids	20	mg	5–6	1.00
	Glucosinolates	40	mg	< 50	1.00
	Phenolic Acids	200	mg	200–300	1.00
	Polyphenols	300	mg	100–150	1.00
	Phytosterols	450	mg	170–440	1.00

While other indices (e.g., the Healthy Eating Index or the Menu Sustainability Index) usually assume linear correlations between the level of nutrient implementation and the health effect, the nutriRECIPE-Index is based on the concept of diminishing marginal utility and the logarithmic relationship of Bernoulli⁽⁴⁵⁾. Moreover, the degree of fulfilment is not observed in absolute terms but is related to the energy content of the menu, so that the nutrient density is evaluated as a result. Figure S1 in the supplementary material shows how the results differ when analysing the amount of nutrients in a recipe considering its energy content, in comparison with just assuming that a menu should include a third of the amount of nutrients in a daily recommendation.

For calculating the degrees of fulfilment, the following formulas are used:

For favourable nutrients: f(x) = ln(x) + 1

For unfavourable nutrients: f(x) = -ln(x).

If the density of a favourable nutrient is high enough that the menu includes the daily requirement of that nutrient, the function gets the value "1". If the maximum recommended daily intake of an unfavourable nutrient is not exceeded, the function gets the value "0". However, if the density for a specific nutrient is tripled, so that one-third of the energy is already sufficient to cover the daily requirement, the logarithmic function returns the value "2.1", which is also the unweighted maximum value for favourable nutrients (upper cut-off). On the other hand, exceeding the recommended daily amounts of unfavourable nutrients leads to malus points, whereby the unweighted minimum value is "-2.1" (lower cut-off).

According to Waijers et al. ⁽²⁵⁾, de facto non-weighting the various compounds of a score is also a weighting, namely, where all components are weighted equally. To avoid this bias, the nutriRECIPE-Index applies a moderate weighting taking into account the degree of supply within the target group (in this publication, the population of Germany) based on officially documented intake data from the last nutrition survey in Germany⁽⁴⁶⁾ and based on the method of ecological scarcity. ⁽⁴⁷⁾. The more inadequate the supply in the overall population with regard to the nutrient is, the lower the weighting of the nutrient and the higher the impact in the assessment. On the other hand, if the supply in the general population is in accordance with the corresponding recommendation, the weighting factor is 1. At the same time, the supply level of the nutrient supply can change the number of bonus points by multiplying the unweighted maximum value of "2.1" or minimum value of "-2.1" by the corresponding weighting factor for the particular nutrient. Table 1 gives an overview of the weighting factors used.

The formula of nutriRECIPE scores for nutrients with beneficial or adverse effects was as follows:

N_menu – nutrient content in a menu

N_rec– recommended nutrient intake per day

E_rec – recommended energy intake per day

E_Menu – energy content of one menu

N_act – average nutrient intake per day

The sum of the single nutrient scores results in the nutriRECIPE-Index. The higher the index value, the more nutrients are contained in the dish in a balanced ratio. Given the weightings, a total score of 100% is possible, even if not all respective dietary target values are met. With bonus points, it is also possible for a menu (e.g., if the nutrient reference values are exceeded) to have a nutriRECIPE score higher than 100%.

Considering bioactive plant compounds:

A particular focus during the index development was the inclusion of BPC. As no food-specific reference databases exist, a separate database has been created containing information on the occurrence and content of BPC in fruits, vegetables, nuts and oil seeds⁽⁴⁹⁾. The eBASIS database⁽⁵⁰⁾ served as a basis. Data for phytosterols in bread and cereals were added from Normen et al. ⁽⁵¹⁾. Retention factors of BPC for different cooking methods were taken from the review of Palermo et al.⁽⁵²⁾ and for phytosterols from Thanh et al.⁽⁵³⁾. The nutriRECIPE score ultimately includes aggregated sum values for the five main classes of BPC: carotenoids, phenolic acids, polyphenols, glucosinolates, and phytosterols.

Recipe origin and validation

Using the nutriRECIPE Index, a complete 6-week menu cycle of a university canteen (Chemnitz in Saxony, Germany) was evaluated, which allows adjustments to the model to review its usability. This menu included standard as well as MensaVital recipes. MensaVital® is a trademark used for certain dishes in German student canteens that claim to be physiologically balanced⁽⁴⁸⁾. Validation was carried out by comparing the results of the analysis of the recipes using two established and validated tools for the assessment of nutritionally balanced meals in canteens, the Healthy Meals Index⁽¹⁸⁾ and the Nutri-Score⁽¹⁹⁾.

NutriRECIPE database

For effective data acquisition and processing, the recipes were recorded in a self-generated MS Access® database. The data of the German Nutrition Database Federal Food Code (BLS) Version 3.02 and the data of the BPC database were stored and linked via the BLS code number.

The results of the recipe evaluation using the nutriRECIPE-Index are shown in Fig. 1 as well as in Figures S2 and S3 in the supplementary material. The arithmetic mean of the nutriRECIPE-Index of all 106 individual recipes is 65.2% [confidence interval (CI) 95%: 61.6–68.8%]. The lowest value observed was 16.8%, and the highest value calculated was 120.8% due to bonus points. A separate evaluation, distinguishing between standard recipes (93) and MensaVital® recipes (13), resulted in higher mean and median values for MensaVital® (arithmetic mean of standard recipes: 62.8% [CI 95%: 59.1%; 66.6%], MensaVital®: 82.2% [CI 95%: 75.3%; 89.0%]).

Figure 1 and Figure S2 show an evaluation of the 24 individual nutrients and nutrient groups that are included in the nutriRECIPE score for each dish. A value of 100% for a favourable nutrient means that the average content of all foods (Fig. 1 arithmetic mean; Figure S2 in the supplementary material median values) contains exactly the amount of that nutrient required to meet the recommendations of the D-A-CH reference values. In the case of unfavourable nutrients, a value of 100% corresponds to the fact that the maximum recommended intake has been strictly adhered to.

The protein assessment for both menu lines was above 90%, whereby the higher values of the MensaVital® recipes were reached because they contain on average significantly fewer calories (759 kcal) than the standard recipes (996 kcal) with a similar amount of protein. For dietary fibre, the recommendations for the standard recipes are only partially fulfilled, with an average of 60.7% [CI 95%: 53.4%; 68.0%]. MensaVital® recipes with mean values of 102.0% [CI 9%: 85.1%; 118.9%] achieved dietary fibre scores according to the recommendations.

The supply of vitamins E, C, and B is between 60% and 90% for standard recipes. The MensaVital® recipes show all values above 90% (see Fig. 1). Especially for vitamin B2 and folic acid, the standard recipes drop to mean values of 74.4% [CI 95%: 65.0%; 83.8%] and 61.8% [CI 95%: 53.5%; 70.0%], clearly below the values of the MensaVital® recipes (vitamin B2: 101.6% [CI 95%: 77.5%; 125.6%]; folic acid: 95.4% [CI 95%: 75.7%; 115.1%]). As expected, the menus did not contribute to an adequate vitamin D supply, as the values were 0% for almost all dishes that were analysed.

The supply of minerals is comparable to that of vitamins. In standard recipes, the nutriRECIPE nutrient score for iodine, magnesium, and zinc was always above 75% on arithmetic average and above 60% for iron. For MensaVital® recipes for iodine, magnesium, and zinc, the score was above 95% and for iron above 75% (see Fig. 1). The calcium supply was insufficient in both menu lines, so the nutriRECIPE nutrient score for standard recipes arithmetically averages 35.8% [CI 95%: 26.1%; 45.6%] and for MensaVital® recipes 62.1% [CI 95%: 35.8%; 88.4%].

In the case of unfavourable nutrients, sodium intake in both menu lines was markedly higher than recommended, whereby the recommendations for MensaVital,® with a score of 66.3% [CI 95%: 33.1%; 99.4%], were better fulfilled compared to the standard recipes, with 36.0% [CI 95%: 25.7%; 46.2%]. The standard recipes contained more saturated fatty acids than recommended, so that the score reached values of 65.0% [CI 95%: 59.1%; 70.8%]. The total sugar content in both recipe lines was not problematic for most menus, although the two standard dishes "semolina porridge with sour cherries" and "rice pudding with applesauce" exceeded the daily WHO recommendations for sugar by almost three times.

The MensaVital® recipes were characterised by higher contents of BPC than the standard recipes. Because lunchtime meals usually contain little fruit, the two main classes of phenolic acids and polyphenols, which mainly occur in fruit, are contained in significantly lower amounts compared to carotenoids, glucosinolates, and phytosterols (see Fig. 1).

Validation

The first validation was performed using the food group and macronutrient-based Healthy Meal Index ⁽¹⁸⁾, comprising the three items: “fruits and vegetables”, “fat quantity and quality”, “wholegrain and potatoes”. According to the Healthy Meal Index, a dish receives points ranging from “0” to “6”, see Table 2.

Table 2

**Evaluation scheme according to the Healthy Meal Index**⁽¹⁸⁾ **(simplified)**.
Compound	Rating	Criteria	Remarks
fruits and vegetables	0	< 1 unit	< 75 g
	1	>= 1 unit	>= 75 g
	2	> 2 units	>= 150 g
fat quantity and quality	0	amount fat units > amount starch units	fat unit: 5 g; starch unit: >=400j = 23,5 g carbohydrates
	1	amount fat units = amount starch units	(minus sugar and sweets)
	2	amount fat units < amount starch units	or: amount fat units = amount starch units and the fat is from plants
wholegrain and potatoes	0	< 0,5 units wholegrain/potatoes	< 37,5 g pasta/rice or 25 g bread or 75 g potatoes
	1	>= 0,5 units wholegrain/potatoes	>= 37,5 g pasta/rice or 25 g bread or 75 g potatoes
	2	>= 1 units wholegrain/potatoes	>= 37,5 g pasta/rice or 25 g bread or 75 g potatoes

The Healthy Meal index suggests, as seen in Table 2, one out of two points if the amounts of fat and starch are equal. In the following analysis, a deviation of 25% in the difference between carbohydrates and fat is tolerated and in that case, the two are considered equal. Otherwise, the more precise carbohydrate and fat quantities of the nutritional analysis are used for the analysis, as opposed to the simplified calculation according to food groups by Lassen et al.⁽¹⁸⁾. Figure 2 illustrates the correlation of the scores of the 6-week menu cycle (divided into standard and MensaVital recipes) determined by the nutriRECIPE index and the Healthy Meal Index, which has a correlation coefficient of r = 0.604.

As shown, high nutriRECIPE scores are correlated with high Healthy Meal Index scores and vice versa. The best-rated dish according to the Healthy Meal Index (HMI value of 6 = 100%), "vegetable plate with Dutch sauce" and potatoes, received a nutriRECIPE score of 24.9 = 112%. The recipe with the highest nutriRECIPE score of 26.9 = 120.8%, is only rated as mediocre by the Healthy Meal Index (4 = 66.66%). After applying both methods as the basis for calculations, the MensaVital® menu line reached higher values on average than the standard menu line. However, because the correlation was only 0.604, there were some deviations. The differences can be explained, on the one hand, by the focus of the HMI on only three criteria, which are therefore very important, and on the other hand by a low differentiation within the categories compared to the nutriRECIPE-Index. The vast majority, 84%, of the dishes were high in fat and received zero points in the fat content category. While 46% of the MensaVital® recipes received one or two points in the fat content category, only 2% of the standard recipes received more than zero points. Furthermore, in the fruit and vegetable category, less than 10% of all dishes failed to receive the maximum possible value of 2 points. This could also be an indication of the cultural specificity of the HMI because in the Scandinavian countries, the quantities of fruit and vegetables consumed are generally lower than in southern European countries (although these differences are getting smaller)⁽⁵⁴⁾. It would be desirable to produce a more pronounced differentiation similar to the evaluation of the dishes considered here.

The second validation was performed using the Nutri-Score, which is used as a front-of-package label in Europe, particularly in France⁽¹⁹⁾. The Nutri-Score is food and nutrient-based, comprising seven items. In general, the Nutri-Score first calculates points, which are split into five different point ranges with a letter code from “A” to “E”. For additional illustration, the background of the letter code has the traffic light colours. The unfavourable food components included in this calculation are calories, total sugars, saturated fatty acids, and sodium. As favourable food components, protein and dietary fibre, as well as the portion of fruit, vegetables, and nuts, are evaluated. All calculations performed are based on 100 g with defined limits. For beverages, cheese, and fatty spreads such as margarine, there are specific adjustments⁽¹⁹⁾. The lower the Nutri-Score, the better the nutritional profile of the food or drink.

Figure 3 shows that the high values of the nutriRECIPE-Index correlate with low values of the Nutri-Score (inverse plot). The majority of MensaVital® dishes received an A rating and a few received a B rating. The standard dishes range from A to D, with the majority of the dishes rated C. The coefficient for the correlation of Nutri-Score and NutriRECIPE-Index is 0.591.

Although the evaluation results of both indices seem similar, a closer look reveals the weaknesses of the relatively simple Nutri-Score. The Nutri-Score does not adequately assess the two worst dishes according to the nutriRECIPE index, "semolina with cherries" (16.8%) and "rice pudding with applesauce" (25.4%). Both dishes receive a C rating, although they contain 164 g and 128 g of sugar, respectively, per serving. Both dishes contain no vegetables, very little dietary fibre, and only small amounts of micronutrients because canned fruits are not an adequate replacement for vegetables.

The assumption that low sugar and fat content and high protein and fibre content may lead to a sufficient micronutrient supply is likewise disproved several times. For example, the dish "Chicken BBQ" receives an A-rating with the Nutri-Score. Taking the nutriRECIPE-Index as a basis, the rating of “BBQ Chicken” is only 53%. This difference is caused by the fact that the nutriRECIPE index included micronutrients and BPC.

Here, we successfully developed a new food quality assessment tool, the nutriRECIPE-Index, which has several strengths: i) consideration of all nutrients with an official recommendation status, ii) nutrient-weighting based on representative supply data, and iii) inclusion of BPC. The nutrient-based approach allows an evaluation independent of dietary style. For example, in other models, aspects of vegetarian nutrition, such as the avoidance of meat or fish (see Healthy Eating Index or Mediterranean Diet Score), are generally devalued in most cases. The adjustments at the recipe level are a further advantage in comparison to other indices that measure nutritional behaviour over a long period. Additionally, by referring to the energy value of the menus to be evaluated, it is possible to adjust the quantities individually and therefore enable personalised nutrition. Existing models (e.g., ONQI, Healthy Eating Index), for example, systematically disadvantage low-calorie meals. In perspective, the model can also be adapted to the needs of individual groups, e.g., the special nutritional needs of target groups such as older people, children, and physically active people. Limitations of the nutriRECIPE-Index result from the complex assessment of data and the availability of high-quality data.

The explicit integration of BPC into the nutriRECIPE-Index is thus far unique in the field of nutrition indices. At the same time, the difficulty here lies in the fact that it has not been possible to access established systems and databases. First, the concentrations of BPC in plants vary markedly. Second, the role of food processing in bioavailability cannot be fully considered due to a lack of data. Investigations on the content of BPC and their modification during processing are still incomplete and should be improved in the future. Nor can it be assumed that all BPC contained in our edible plants have already been found and characterised. Here, future studies will complete the mosaic with ever-advancing analytical methods. For these reasons, the amount of BPC is integrated as the sum of the five main classes (carotenoids, phenolic acids, polyphenols, glucosinolates, phytosterols) and has been included as one score in the nutriRECIPE-Index with no differentiated subdivision according to subclasses or even individual compounds. Our calculations show that lunch menus usually contain a maximum of three of these five main classes. Thus, with these extra points, the importance of a large portion of vegetables, fruit, legumes, nuts, and vegetable oils in nutrition is emphasised, without overestimations of the micronutrients.

The limitations in the application of the nutriRECIPE index stem primarily from data quality. The existence of exact recipes is necessary for nutrient-based analysis; such recipes are not always available in public catering facilities, and methodological limits may be exceeded in the context of, for instance, convenience products and salt. Expansion and improvement of the nutrient databases are also of high relevance here. As general reference values and supplies of ingredients are used, individual requirements are not taken into account. The intercorrelation of nutrients cannot be considered due to the complexity of those interactions.

Other factors that play an essential role in healthy eating behaviour cannot be incorporated into the nutriRECIPE-Index. These include, among other things, shelf life and differences in food quality, pricing, and availability of food, actual recipe preparation, as well as nutritional culture and social desirability. For the future, it would be advisable to validate health outcomes through human trials to obtain more accurate information on the relevance of a priori indices.

The nutriRECIPE-Index allows an extensive evaluation and optimisation of complex dishes, taking into account 24 nutrients and nutrient groups as well as supply-dependent weighting factors. The nutriRECIPE-Index permits structural adjustments, e.g., concerning the nutrients included or the weighting depending on the supply status of the population group under consideration. Therefore, an adaptation to individual target groups such as older people, children, and physically active people is possible and desirable. Through these adaptations, other reference values could be applied and different supply situations could be addressed. Thus, a nutrient-specific assessment is possible, which represents a decisive advantage of the nutriRECIPE index over other generalising nutrition indices. The nutriRECIPE Index offers the opportunity for caterers to optimise their menus, which in turn can contribute to improving public health.

Ethics approval and consent to participate

N/A

Consent for publication

N/A

Availability of data and materials

All data generated or analysed during this study are included in this published article and its supplementary information files. Any additional information on the datasets used and/or analysed during the current study, are available from the corresponding author on reasonable request.

Competing interests:

The authors declare that they have no competing interests.

Funding:

This work was funded by the German Federal Ministry of Education and Research (funding number 01EA1411C).

Authors’ Contributions:

FF was a major contributor in writing the manuscript, created the database for bioactive plant compounds, calculated the nutriRECIPE-Index and the comparison with the NutriScore. IV was a major contributor in writing the manuscript, calculated the recipes and the comparison with the Healthy Meal Index. TM gave the idea for the nutriRECIPE-Index and helped with writing and reviewing the manuscript. OC reviewed the final manuscript. GS reviewed the final manuscript. All authors read and approved the final manuscript.

Acknowledgements:

nutriCARD - Competence cluster for nutrition and cardiovascular health – Universities of Halle, Jena, and Leipzig

GBD 2017 Risk Factor Collaborators. Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017 (2018). The Lancet Global health Metrics. 2018;392:1923–94.
Afshin A, Sur PJ, Fay KA, et al. (2019) Health effects of dietary risks in 195 countries, 1990–2017. A systematic analysis for the Global Burden of Disease Study 2017. The Lancet 393,1958–1972.
Meier T, Gräfe K, Senn F, et al. (2019) Cardiovascular mortality attributable to dietary risk factors in 51 countries in the WHO European Region from 1990 to 2016: a systematic analysis of the Global Burden of Disease Study. European journal of epidemiology 34, 37–55.
Jeruszka-Bielak M, Kollajtis-Dolowy A, Santoro A, et al. (2018) Are Nutrition-Related Knowledge and Attitudes Reflected in Lifestyle and Health Among Elderly People? A Study Across Five European Countries. Frontiers in physiology 9, 994.
Carson JAS, Gilliam MB, Kirk LM, et al. (2002) Enhancing self-efficacy and patient care with cardiovascular nutrition eduction. American journal of preventive medicine 23, 296–302.
Ni Mhurchu C, Eyles H, Jiang Y, et al. (2018) Do nutrition labels influence healthier food choices? Analysis of label viewing behaviour and subsequent food purchases in a labelling intervention trial. Appetite 121, 360–365.
Bandara BES, De Silva DAM, Madzwanthi BCH, et al. (2016) Impact of Food Labeling Information on Consumer Purchasing Decision: With Special Reference to Faculty of Agricultural Sciences. Procedia Food Science 6, 309–313.
Destatista (2018) Umsatz im Außer-Haus-Markt in Deutschland nach Segmenten in den Jahren 2016 und 2017 (in Milliarden Euro). (accessed September 2019).
Eurostat. (2019) Accommodation and food service statistics - NACE Rev. 2. , zuletzt aktualisiert am May 2019 (accessed September 2019).
Lachat C, Nago E, Verstraeten R, et al. (2012) Eating out of home and its association with dietary intake: a systematic review of the evidence. Obesity reviews 13, 329–346.
Nago ES, Lachat CK, Dossa RAM, et al. (2014) Association of out-of-home eating with anthropometric changes: a systematic review of prospective studies. Critical reviews in food science and nutrition 54, 1103–1116.
Lachat C, Roberfroid D, Huybregts L, et al. (2009): Incorporating the catering sector in nutrition policies of WHO European Region: is there a good recipe? Public health nutrition 12, 316–324.
Volkhardt I, Christen O, Stangl, et al. (2017) Legal aspects regarding product innovations in the foodsector. Ernahrungs Umschau 64, 158–165.
Skov LR, Lourenço S, Hansen GL, et al. (2013) Choice architecture as a means to change eating behaviour in self-service settings: a systematic review. Obesity reviews 14,187 – 96.
Reisch L, Eberle U, Lorek S. (2013) Sustainable food consumption: an overview of contemporary issues and policies. Sustainability: Science, Practice and Policy 9, 7–25.
Teitscheid P, Langen N, Speck M et al, editors (2018) Nachhaltig außer Haus essen. Von der Idee bis auf den Teller. Munich, Germany: oekom.
Burggraf C, Teuber R, Brosig S, et al. (2018) Review of a priori dietary quality indices in relation to their construction criteria. Nutrition reviews 76, 747–764.
Lassen AD, Biltoft-Jensen A, Hansen GL, et al. (2010) Development and validation of a new simple Healthy Meal Index for canteen meals. Public health nutrition 13, 1559–1565.
Julia C, Hercberg S. (2017) NutriScore: evidence of the effectiveness of the French front-of-pack nutrition label. Ernahrungs Umschau 64, 181–187.
Meier T, Gärtner C, Christen O. (2014): The susDISH analysis method - Sustainability in the catering industry. Taking account of both nutritional and environmental aspects in recipe planning. Institute of Agricultural and Nutritional Sciences, Halle-Wittenberg University, German Agricultural Society (DLG), Frankfurt (Main).
Bauer J, Volkhardt I, Michl M, et al. (2016) NutriScale. Key Figures for Daily Food Choices. Advanced Engineering Forum 19, 156–163.
Müller C, Stucki M, Zehnder P, et al. (2016) The “Menu Sustaina-bility Index”. Assessment of the environmental and health impact of foods offered in commercial catering. Ernahrungs Umschau 63, 198–205.
Engelmann T, Speck M, Rohn H, et al. (2018): Sustainability assessment of out-of-home meals: potentials and challenges of applying the indicator sets NAHGAST Meal-Basic and NAHGAST Meal-Pro. Sustainability 10, 562.
Volkhardt I, Semler E, Keller M, et al. (2016) Checklist for a vegan lunch menu in public catering. Ernahrungs Umschau 63, 176–184.
Waijers, Patricia MCM, Feskens EJM et al. (2007) A critical review of predefined diet quality scores. The British journal of nutrition 97, 219–231.
Kourlaba G, Panagiotakos DB. (2009) Dietary quality indices and human health: a review. Maturitas 62, 1–8.
Arvaniti F, Panagiotakos DB. ((2008) Healthy indexes in public health practice and research: a review. Critical reviews in food science and nutrition 48, 317–327.
Mahoney SE, Loprinzi PD. (2014) Influence of flavonoid-rich fruit and vegetable intake on diabetic retinopathy and diabetes-related biomarkers. Journal of diabetes and its complications 28, 767–771.
Nöthlings U, Murphy SP, Wilkens LR, et al. (2007) Flavonols and pancreatic cancer risk: the multiethnic cohort study. American journal of epidemiology 166, 924–931.
Rienks J, Barbaresko J, Nöthlings U. (2017) Association of Polyphenol Biomarkers with Cardiovascular Disease and Mortality Risk: A Systematic Review and Meta-Analysis of Observational Studies. Nutrients 9, 415.
Zamora-Ros R, Fedirko V, Trichopoulou A, et al. (2013) Dietary flavonoid, lignan and antioxidant capacity and risk of hepatocellular carcinoma in the European prospective investigation into cancer and nutrition study. International journal of cancer 133, 2429–2443.
Cheng HM, Koutsidis G, Lodge JK, et al. (2017) Tomato and lycopene supplementation and cardiovascular risk factors: A systematic review and meta-analysis. Atherosclerosis 257, 100–108.
Marangoni F, Poli A. (2010) Phytosterols and cardiovascular health. Pharmacological research 61, 193–199.
Thiele S, Mensink GBM, Beitz R. (2004) Determinants of diet quality. Public health nutrition 7, 29–37.
Burggraf C, Volkhardt I, Meier T. (2017) Benefits of a Modified Traffic Light Labelling System for Food Products. IAMO-Policy Letter 28. Halle, Germany: Leibniz Institute of Agricultural Development in Transition Economies (IAMO).
Deutsche Gesellschaft für Ernährung (DGE), editor. (2018) Referenzwerte für die Nährstoffzufuhr D-A-CH. 2nd edition, Bonn, Germany: DGE.
Deutsche Gesellschaft für Ernährung (DGE), editor. Umsetzung der D-A-CH-Referenzwerte in die Gemeinschaftsverpflegung. Bonn: DGE; 2013.
Humayun MA, Elango R, Ball RO, Pencharz PB. (2007) Reevaluation of the protein requirement in young men with the indicator amino acid oxidation technique The American Journal of Clinical Nutrition 86, 995–1002.
Tang M, McCabe GP, Elango R, et al. (2014) Assessment of protein requirement in octogenarian women with use of the indicator amino acid oxidation technique. The American journal of clinical nutrition 99, 891–898.
Rafii M, Chapman K, Elango R, et al. (2016) Dietary Protein Requirement of Men 65 Years Old Determined by the Indicator Amino Acid Oxidation Technique Is Higher than the Current Estimated Average Requirement. The Journal of nutrition 146, 681–687.
Rafii M, Chapman K, Owens J, et al. (2015) Dietary protein requirement of female adults 65 years determined by the indicator amino acid oxidation technique is higher than current recommendations. The Journal of nutrition 145, 18–24.
World Health Organization (WHO), editor. (2007) Protein and amino acid requirements in human nutrition. Report of a joint FAO/WHO/UNU Expert Consultation. WHO Technical Report Series 935), Geneva, Switzerland: WHO.
World Health Organization (WHO), editor. (2015) Sugars intake for adults and children. Guideline. WHO/NMH/NHD/15.2 (Executive summary) Geneva, Switzerland: WHO.
Ernst JB, Arens-Azevêdo U, Bitzer B, et al. (2018) Quantitative Empfehlung zur Zuckerzufuhr in Deutschland. Bonn, Germany: Deutsche Adipositas-Gesellschaft, Deutsche Diabetes Gesellschaft und Deutsche Gesellschaft für Ernährung.
Laux H, Schabel MM. (2009) Subjektive Investitionsbewertung, Marktbewertung und Risikoteilung. Grenzpreise aus Sicht börsennotierter Unternehmen und individueller Investoren im Vergleich. Berlin, Germany: Springer.
Deutsche Gesellschaft für Ernährung (DGE), editor. 12. Ernährungsbericht 2012. Bonn: DGE; 2012.
Frischknecht R, Knöpfel SB. (2013) Swiss eco-factors 2013 according to the ecological scarcity method. Methodological fundamentals and their application in Switzerland. Environmental studies (1330).
Deutsches Studentenwerk. (2019) Die mensaVital-Menülinie. (accessed September 2019).
Forner F, Meier T, Stangl G, Christen O. (2020, under preparation) Development of a Database for Bioactive Plant Compounds and Linkage to German Food Code Items. Journal to be determined.
EuroFIR AISBL. (2017) eBASIS. European Food Information Resource Network Project www.ebasis.eurofir.org (accessed september 2019).
Normén L, Bryngelsson S, K´Hohnsson M. The phytosterol content of some cereal foods commonly consumed in Sweden and in the Netherlands. J Food Compos Anal. 2002;6:693–704.
Palermo M, Pellegrini N, Fogliano V. (2014) The effect of cooking on the phytochemical content of vegetables. Journal of the Science of Food and Agriculture 94, 1057–1070.
Thanh TT, Vergnes MF, Kaloustian J, et al. (2006) Effect of storage and heating on phytosterol concentrations in vegetable oils determined by GC/MS. Journal of the Science of Food and Agriculture 86, 220–225.
Naska A, Fouskakis D, Oikonomou E, et al. (2006) Dietary patterns and their socio-demographic determinants in 10 European countries: data from the DAFNE databank. European journal of clinical nutrition 60 181–190.

Download PDF

Journal Publication

published 18 Nov, 2021

Read the published version in BMC Nutrition →

Review #4 received at journal
01 Apr, 2021
Reviewer #4 agreed at journal
24 Mar, 2021
Review #3 received at journal
25 Dec, 2020
Reviewer #3 agreed at journal
08 Dec, 2020
Reviewer #2 agreed at journal
07 Oct, 2020
Reviewer #1 agreed at journal
17 Jun, 2020
Editor assigned by journal
14 Apr, 2020
Reviewers invited by journal
14 Apr, 2020
Submission checks completed at journal
09 Apr, 2020
Editor invited by journal
09 Apr, 2020
First submitted to journal
07 Apr, 2020

You are reading this latest preprint version

The nutriRECIPE-Index - Development and Validation of a Nutrient-Weighted Index for the Evaluation of Recipes

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Scope of existing models for evaluating nutrition and menus

Bioactive plant compounds (BPCs)

Aim of the study

Methods

Model development

Results

Validation

Discussion

Conclusion

Declarations

References

Supplementary Files

Status:

Journal Publication

Version 1