by Johann Steszgal (Steszgal Informationstechnologie GmbH), Peter Kieseberg (St. Pölten UAS) and, Andreas Holzinger (University of Natural Resources and Life Sciences Vienna)

Reduction of food waste is an important target for reducing the human footprint and achieving better utilisation of natural resources. The healthcare sector especially offers a lot of potential for more sustainable handling of food. In this article we present major challenges for food waste reduction in real-world healthcare environments, as well as a solution approach.

The “Necta against Foodwaste” project calculates the sustainability costs for wasted food as a basis for reducing the ecological footprint of catering in the health sector. Image from Pixabay.
The “Necta against Foodwaste” project calculates the sustainability costs for wasted food as a basis for reducing the ecological footprint of catering in the health sector. Image from Pixabay.

The drastic reduction of food waste at the retail and consumer level has been identified by the United Nations as a key target (Target 12.3) to achieve the Global Sustainability Goals (SGDs) and is also closely related to SDG 2 ("Zero Hunger") [1]. A key challenge is the structured recording of food waste, its impact on the environment, as well as the adaptation of user behaviour to a more sustainable approach to food consumption.

Especially in mass catering, and in particular in the care sector, two special situations come together: on the one hand, food is distributed in a highly standardised way, especially with regards to portion sizes and the composition of the individual menus, which creates a great potential for waste; on the other hand, the actual consumption can be measured quite precisely in the case of inpatient admission. Therefore, this area in particular would be predestined for a pioneering role in food waste reduction, as first research results demonstrate: for example, a study of Swedish hospitals shows a very high level of waste, but this picture is very heterogeneous when looking at the details, as the actual impact in terms of sustainability was not measured, i.e., they made no distinction whether, e.g., meat or vegetables were left over. Furthermore, most studies do not take non-served food, e.g., due to patients being discharged after early rounds, into consideration. It should be noted that no intersection with sustainability scores was made in any of the studies known to us.

In the “Necta against Foodwaste” project [L2] we therefore focus on the utilisation of healthcare records related to food consumption for in-house patients and calculate the sustainability costs for wasted food as a basis for reducing the ecological footprint of catering in the health sector. This requires harnessing and integrating different data masters, especially intelligent unification of item master data to allow integration with external information like nutrition protocols that map the actual use of food in the hospital and care sector, and sustainability scores that allow ratings of individual foods in terms of sustainability, especially the one provided by Eaternity [2]. This leads to the following research challenges:

  • Applicability in real-world environments: In order to make the data usable and derive measures to reduce waste, it must be made analysable and evaluated. As real-world data is typically not flawless, this means that standard techniques for data cleansing, as well as their impact on the final result, need to be analysed in order to guarantee a certain accuracy [3]. In addition, the impact of privacy measures needs to be taken into account.
  • Data normalisation and integration: Currently, there exists no standard for data collections regarding food consumption of residents in the health sector, more precisely, often the data is collected in free text form. In order to make this information usable, the data has to be transferred into a structured form and requires normalisation between the different environments.
  • Establishing comparability: A major problem for meal scoring is the heterogeneity of the ingredient landscape and its comparability. In particular, establishing item master data comparability is a common problem, as equivalence is very difficult to define in general terms and is based on the application purpose, number of servings (package sizes), and many other side parameters that must be dynamically selected and integrated accordingly.
  • Scoring: Just measuring the amount of wasted food is not good enough to estimate the ecological impact of food waste: wasting 1 kg of potatoes, while not good, does far less harm than wasting 1 kg of beef. Furthermore, for more fine-grained analysis, sustainability needs to be calculated with respect to different parameters, e.g., water consumption. Introduction of the Eaternity score is especially interesting here, as it provides a wide variety of different parameters and is still actively developed.
  • Researching measures to reduce food waste: Depending on the operational scenario and organisation, different measures need to be applied – starting with the detour of unused food, to the adjustment of portion sizes, to the redesign of recipes with regard to unpopular parts.
  • Sustainable recipes: By integrating sustainability scores, not only ingredients, but also the sustainability of recipes and portions can be calculated through the knowledge of recipes stored in necta. Thus, these can be made more sustainable, e.g., by replacing or removing problematic ingredients or searching for alternatives.

However, all data related to individual patients has high privacy and data security requirements, especially since anonymisation in dynamic data sets is extremely problematic from a data quality point of view. Thus, while anonymisation is applied where possible, intelligent and flexible pseudonymisation of data that cannot be anonymised due to subsequent evaluation options needs to be designed, including a high degree of flexibility in the implementation of access controls in de-pseudonymisation in order to be able to implement concepts such as the 4-eyes principle and the like.


[1] A. Holzinger, et al., “Digital Transformation for Sustainable Development Goals (SDGs) – A Security, Safety and Privacy Perspective on AI”, in Proc. of CD-MAKE 2021 (pp. 1-20), LNCS vol 12844, Springer, 2021.
[2] L. Eymann, M. Stucki and E. Hirsiger, “Nose to tail: how to allocate the environmental burden of livestock production systems to different meat cuts?”, in 7th Int. Conf. on Life Cycle Management, 2015, 2015.
[3] K. Stöger, et al., “Legal aspects of data cleansing in medical AI”, Computer Law & Security Review, vol. 42, p.105587, 2021.

Please contact:
Peter Kieseberg, St. Pölten UAS, Austria
This email address is being protected from spambots. You need JavaScript enabled to view it.

Next issue: October 2024
Special theme:
Software Security
Call for the next issue
Image ERCIM News 132
This issue in pdf


Image ERCIM News 132 epub
This issue in ePub format

Get the latest issue to your desktop
RSS Feed