Enhancing food sustainability in the acute hospital setting – a nurse-led study of patient food

The acute hospital food environment is a challenging one in terms of sustainability. The need to meet the complex, fluctuating demands of a dynamic patient population needs to be balanced against the provision of mass catering at a cost that is realistic for hospital finances. High patient turnover, an ageing population, often with small appetites (Pilgrim et al, 2015), and the high meat consumption of most patients at home (Stewart et al, 2021) all contribute to a high-waste, high-carbon footprint food environment that can be difficult to change.

The intensive farming that underpins global food supply causes soil degradation, loss of biodiversity, increased pesticide use and water pollution, all of which have a direct negative environmental impact (Gržinić et al, 2023) and which will have long-term consequences for planetary health, food production, nutritional value and price. Meat products, particularly beef, have a much higher carbon footprint than plant-based or vegetarian meal options (Poore and Nemecek, 2018). The packaging, processing and transport of food cause environmental pollution as the result of their reliance on fossil fuel-derived plastics, the unsustainable gas and electric energy supply required for cooking and refrigeration, and transport using diesel-powered lorries.

Currently, many hospitals use catering contractors who provide a meat-dominated food service on a budget, with sustainability remaining a low priority. Indeed, although catering accounts for 6% of NHS carbon emissions – around 1500ktCO₂e (kilotonnes of carbon dioxide equivalent) every year (NHS England, 2020) – when considering sustainability, the Independent Review of NHS Hospital Food (Department of Health and Social Care, 2020) focused on local/seasonal procurement and waste, rather than on reducing meat consumption and promoting plant-based menus.

Although healthcare sustainability has a growing profile, this is dominated by a reduction in the use of personal protective equipment (PPE) (Strain, 2022), aerosol inhalers (Woodcock et al, 2022) and anaesthetic gas (Lopes et al, 2021). At the time of writing, there are few published accounts of hospital food sustainability programmes, yet nurses are ideally placed to communicate the importance and wide-reaching consequences of a poor food sustainability profile, and to promote plant-based meals as a healthy option for patients, staff and planet. Initiatives, such as Greener by Default in the USA (https://www.greenerbydefault.com/healthcare), and, in the UK, projects such as the Hampshire Hospitals Green Team (Freeburn and Strachan, 2023) have shown that food sustainability can be improved in the healthcare setting.

The aim of the project described in this article was to understand the ward-level drivers of food waste, to quantify the carbon footprint associated with patient food choices and to develop a recycling stream for food packaging.

Methods

Food waste quantification

Ten adult wards over a range of medical and surgical specialties (respiratory, gastroenterology, hepatology, renal, trauma including silver trauma (trauma in the elderly), and orthopaedics) were approached and agreed to be trial wards. For 4 days in December 2023, each ward was supplied with two labelled lidded buckets for breakfast and lunchtime. Any food waste left at the end of service was segregated as ‘trolley waste’, and any food that had been given to a patient and not eaten was designated as ‘patient waste’.

The two buckets were weighed and recorded after breakfast and again after lunch. Multiple days were assessed to ensure that the data were not skewed by unforeseen incidents on the trial wards had only one day been sampled. The data were extrapolated from the 10 trial wards to the 50 that make up the hospital, to provide an estimation of overall food waste. Paediatric areas were excluded from the study because they operate from a separate menu, and intensive care units were not included because they do not have the standard menu and have a very small meal service.

The project focused on lunchtime waste because it is the greater contributor to overall meal waste and has a much higher carbon footprint than the porridge and toast that are typically served for breakfast.

Ward level drivers for food waste

To understand the potential reasons for food waste at ward level, informal conversations took place with ward hosts across the hospital. The hosts are contracted staff who take the food orders, plate up and deliver the meals, and collect waste. All hosts verbally consented to talk about food waste in the context of the project. Open questions were used to allow disclosure of any factors not yet considered, and to reduce bias from the interviewer. Notes were taken during the conversation and these were written up immediately afterwards. These data were then analysed for frequency of particular topics being mentioned.

Estimating the carbon-equivalent footprint of lunchtime waste

The meal ordering data for the trial wards were analysed and the number of portions of individual food components supplied over the 4 days quantified. This figure was extrapolated from the 10 trial wards to the hospital's 50 adult wards/areas. Each portion has a standard weight.This, combined with the estimated weight of the hospital-wide food waste, enabled estimation of the carbon footprint of food waste to be made (Poore and Nemecek, 2018). These data were then used to model the impact that changing meat consumption would have on the environmental footprint of the hospital food service.

Estimating packaging waste associated with the lunchtime service

The meal ordering data from the trial wards were used to calculate the number of plastic and foil containers supplied to the wards each lunchtime. In addition, the contractor supplied figures for milk bottle usage. This amounted to three 4-pint milk bottles per ward per day. These data were extrapolated from the 10 trial wards to the hospital's 50 wards. The empty packaging was weighed to enable estimation of the weight of packaging waste being produced. All food packaging waste went to the black bag waste stream at the start of the project so investigating the feasibility of recycling food packaging was addressed as part of the project.

Assessing potential for a more plant-based menu

The patient food survey is collected across the hospital and collated each month. These data were used to calculate the average percentage of the hospital population who stated that they were vegan or vegetarian.

Results and Discussion

Patient food waste: biggest contributor to ward-level food waste

Food waste was collected and segregated into patient waste and trolley waste on the 10 trial wards for 4 days at breakfast and lunchtime (Figure 1). This represented a total of 266 beds, with an average of 26 patients per ward. The variations between wards were discussed with the ward hosts on the trial wards and they attributed it to differences in patient populations. Wards with low levels of food waste had ward hosts who, during the course of the study, vocalised awareness of food waste and concern that it was anecdotally known to be high. These ward hosts were likely to consider portion size for specific patients when plating from multi-portion packs and described engaging family members to help the ordering process for patients who were unable to make menu choices independently. (Only carbohydrate and vegetables come in multi-portion packs; the protein element of meals, the key contributor to the carbon footprint, is a set size.)

As shown in Table 1, based on the data from the 10 trial wards, average ward-level food waste across the hospital could be estimated at both daily and annual scales. Lunchtime generated nearly double the food waste (patient and trolley waste combined) than breakfast time. The total ward-level food waste from breakfast and lunchtime over 1 year was estimated at over 73 metric tonnes.

Portion size identified as a key driver of food waste

Patient satisfaction with the food served was not considered to be a cause of the high level of waste as the patient feedback survey (conducted monthly) suggested that over 95% of patients thought that their dietary needs were catered for and over 95% rated the food as acceptable, good or very good.Therefore, to understand the drivers behind ward-level food waste informal conversations with ward hosts across 10 different areas of the hospital were initiated. This was sufficient to discover that new themes were not being raised, suggesting that saturation point had been reached. These conversations captured a wider picture than the trial wards, and housekeepers (who are involved in the meal service on the wards) and members of the clinical team would often join the conversation to give their thoughts.

The questions were open to encourage novel ideas and themes to be explored. Figure 2 shows that the most frequent topic raised was portion size, which was mentioned in 9 out of 10 conversations. This was sometimes in the context of needing to portion a multi-portion pack to ensure that patients' nutritional needs were met and that those with bigger appetites were given sufficient food. In 6 out of 10 conversations it was mentioned that portion size was too big, and that this was a major contributor to patient waste. In response to this feedback salad sizes have been reduced (while retaining the same added protein content, eg the same amount of egg but less salad) because salads were noted as being significant contributors to waste, and in addition energy-dense mini-meals have also been introduced.

One of the themes was the consumption of fruit, summed up by the comment from one clinical member of staff:‘Nobody ever eats an orange.’ Naval oranges were being supplied for the ward fruit bowls but were rarely eaten. This was down to the lack of sharp knives available to cut them up, meaning that they were inaccessible to most patients. Each week, 45kg of oranges were ordered, with more than 1.1 tonnes CO₂e emissions associated with their production and transport over the year. (This figure uses an estimation of 0.5kg CO₂e per kg oranges based on work published by Bell and Horvath (2020); however, the actual environmental footprint will vary with season, country of origin and transport.) As a result of this work, naval oranges were replaced with easy peelers, and the ordering practice changed so that patients were able to order oranges. Today, 45kg of oranges continue to be ordered each week, but ward hosts report that they now rarely go in the waste.

Estimated carbon footprint of lunchtime waste

Estimating the carbon equivalent footprint (CO₂e) of food waste is challenging and requires some assumptions that should be explicit when interpreting data. In this project, the weight of each meal was known, enabling calculation of the carbon footprint for a standard meal (Figure 3a). The food choice data for 3 days from the trial wards was used to calculate the proportion of each type of meal that was ordered and this was then used to calculate how much of the 126kg of waste per day was represented by each meal type (Figure 3b).

The carbon footprint of lunchtime waste across the hospital was calculated to be over 1500kg CO₂e per day, with this high figure driven by high beef consumption (Figure 4a).

The potential carbon footprint impact of changing patient food choices was also modelled. As shown in Figure 4b, a 25% reduction in beef ordering (with assumed concomitant rise in uptake of vegetarian meals) would deliver nearly 300kg CO₂e saving each day, while a 50% reduction in beef ordering (again, assuming vegetarian meals are chosen instead) would deliver nearly 600kg CO₂e saving each day (Figure 4c).

The changes introduced to enable this reduction were:

• Replacing roast beef (60kg CO₂e per kg associated with production) with roast pork (which has a much lower carbon footprint of 7kg CO₂e per kg)
• Removing some beef options from the menu
• Increasing vegetarian or plant-based options and including these in the ‘Specials’ menu each day
• The menu design was changed to position vegetarian and plant-based options at the top.

Acceptability of a plant-based menu

Any change to healthcare meal provision must be balanced against the needs of patients. To assess the dietary needs of the inpatient population, the self-reported patient feedback survey data for the 3 months leading up to and including weighing week were used. As shown in Figure 5a, the majority of patients declared no specific dietary needs, with about 3.5% indicating that they were vegan or vegetarian. Analysis of the food ordering habits of the trial wards revealed that 7–15% of patients were already choosing vegetarian or plant-based options, suggesting that increasing this provision had a good likelihood of acceptability.

The percentage range was dependent on whether Friday was included in the analysis, as far more patients ordered vegetarian or fish-based meals on Friday. No other days showed such a distinct ordering pattern, but there was always a significant proportion of orders from the ‘Specials’ menu suggesting that removing beef from this menu would further reduce the carbon footprint associated with the meal service.

To further investigate this, the food ordering data for the same wards were analysed in June 2024, 2 months after the introduction of the new, more sustainable menu. This analysis showed that vegetarian and plant-based meals now accounted for over one-third of patient orders, and this figure did not drop significantly when Friday orders were included (Figure 5b). This represented a two- to four-fold increase in vegetarian meal choices compared with the baseline data. This suggested that the new menu was successful in promoting a much lower carbon footprint diet among our inpatients.

The patient feedback data for June 2024 showed that the number of vegan or vegetarian patients dropped from 3.7% to 2.2%, suggesting the menu is the key driver for the difference in ordering habits, rather than a change in population attitude. In addition, the patient survey data for June showed that 97% of patients thought the choice of food was very good, good or acceptable, and 99% said that the food flavours were very good, good or acceptable. Since the introduction of the more sustainable menu, both of these outcome measures have increased, which suggests strong patient support for the changes.

Introduction of milk bottle recycling

At the start of the project all food packaging was being disposed of in the black bag waste stream. This consisted of foil and plastic food trays, and 4-pint milk bottles made of high-density polyethylene (HDPE), with the latter having a carbon footprint of 2.6 kg CO₂e per kg. The hospital-wide consumption of milk was about three 4-pint bottles per ward per day, so by introducing a recycling stream 1.7 tonnes HDPE per year could be recycled. This will save over 4 tonnes of waste from being incinerated and generating air pollution each year.

Further food packaging recycling is a challenge because food residue needs to be washed off prior to removal from the hospital site. This cannot be done in the catering dishwashers, which means that the current option is for the packaging to be rinsed by hand, which is not currently feasible for a catering service of this size (approximately 1100 lunches per day). Work is ongoing with the main supplier to recycle both plastic and foil food packaging.

Conclusion

The effective changes made in the study hospital could be adopted by any healthcare setting that has a food service – from a large acute hospital, through community hospitals, to care homes. There are no negative financial implications – rather, these changes may save providers money in terms of the cost of meat-based meals and the cost of disposal of food and packaging. Factors that may impact ease of implementation include the regularity of menu review, the engagement of dietitians, and the waste contract of the healthcare setting for recycling food packaging.

Reducing food waste and promoting a more sustainable plant-based menu meet the triple bottom line of being better for:

• The environment (due to decreased need for intensive farming, reduced packaging, reduced CO₂e emissions associated with meat consumption)
• The patient (due to reduced pollution associated with food production and potentially healthier menu choices), and
• Hospital finances (due to reduced food requirements and reduced waste disposal).

Increasing pressure from NHS trusts to supply plant-based meals will generate market pressure on suppliers, even if, at this point in time, there is no NHS England standard for plant-based meals to make up the majority of menu options. Although this project focused on the lunchtime service, the same menu is used for suppertime, so the results here can probably be applied to both. Further research is needed to understand the patient perspective on the food service and whether the same ordering and waste trends apply to the children's hospital.

Overall, this was a simple project, with significant environmental impact and potential to change the food behaviours of both patients and staff.

KEY POINTS

• This project showed that portion size is a key driver of food waste in the acute hospital setting
• Reducing the number of beef options on a menu is key to reducing the greenhouse gas emissions associated with hospital food
• Increasing plant-based and vegetarian meal options has a high probability of being acceptable to patients
• Recycling food packaging can be challenging, but there are routes to achieve this

CPD reflective questions

• Consider how you could promote vegetarian or plant-based menu options to your patients/residents
• How confident do you feel about explaining the factors around food sustainability to patients/residents? What could you do to educate yourself further on food sustainability?
• How often do we talk to our patients/residents about their food waste? Do we know what is driving it? (For example, portion size, food temperature, menu choices, difficulty using standard cutlery.)
• Consider the feasibility of healthcare organisations reducing their food waste. How could this be achieved?