Comparison of the carbon footprint of different patient diets in a Spanish hospital

Abstract

Objectives

Mitigating climate change requires management strategies to reduce greenhouse gas emissions in any sector, including the health system. Carbon footprint calculations should play a key role in quantifying and communicating these emissions. Food is among the categories with low accuracy because the carbon footprint for food is still under development. We aimed to quantify the carbon footprint of different diets.

Methods

Average carbon footprint for a normal diet was based on detailed composition data in Juan Ramón Jiménez Hospital (Huelva, Spain). In addition, the carbon footprints of 17 other therapeutic diets were estimated using a streamlined variation of each diet published by Benidorm Clinical Hospital (Spain).

Results

The carbon footprint was calculated for 18 hospital diets for a variety of patients. The reference menu corresponds to the normal diet provided to patients who do not have special dietary requirements. This menu has a low carbon footprint of 5.083 CO₂ eq/day.

Conclusions

Hospital diets contribute to the carbon footprint of a hospital. The type of diet has a significant impact on the greenhouse gas emissions. A Mediterranean diet is associated with lower environmental impact than diets with more meat, in particular red meat.

Keywords

carbon footprint hospital diets Mediterranean diet

Introduction

Most food production gives rise to greenhouse gases (GHGs) contributing 15–31% of global emissions. In developed countries, it contributes 15–28% of national emissions.^1,2 Agricultural production contributes 80–86% in the food chain, while the remainder comes from pre-production (predominantly fertilizer manufacture) and post-production activities such as processing, packaging and transport.³

Food supply should therefore be a focus for health care professionals who wish to improve the environmental profile of health care. A barrier to this is that health care managers do not usually have an environmental background or training and only a few hospitals employ environmental managers.⁴

The urgency of climate change⁵ demands management strategies to reduce GHG emissions. Based on the principle that ‘what is not measured is not managed’, carbon footprint calculators should play a key role in quantifying and effectively communicating these emissions. Kim and Neff¹ provide a critical review of general purpose calculators that account for carbon footprint from food consumption and conclude that substantial opportunities exist to improve upon these tools: expanding the variety of diets and adopting more rigorous and transparent methodologies.

Climate change is gaining increasing importance in the health care sector.^6–8 Furthermore, some easy-to-use carbon footprint calculators have been developed to help health care professionals overcome the gap in awareness of environmental impacts.⁹

Our aim was to assess the carbon footprint of hospital menus. The average footprint for a normal diet was based on detailed composition data provided by Juan Ramón Jiménez Hospital (Huelva, Spain). From the normal diet, the carbon footprints of 17 therapeutic diets were estimated. In addition, a streamlined variation of each diet published by the Benidorm Clinical Hospital was assessed. The results can be used to calculate the carbon footprint of hospital menus in countries with a Mediterranean diet or similar diets.¹⁰

Methods

The carbon footprint expresses the amount of GHGs emitted in terms of equivalent quantities of carbon dioxide (i.e. kilograms of CO₂ eq). The quantities of CO₂ eq emitted are generally estimated on the basis of the Global Warming Potentials (GWPs) provided by the Intergovernmental Panel on Climate Change (IPCC) for each GHG.⁵ The GWP for a gas is a measure of the magnitude of climate change caused by one unit of that gas over a particular time period (usually 100 years) relative to that caused by one unit of CO₂ (which is assigned a GWP of one as it is used as the reference gas). For example, if methane (CH₄) has a GWP of 25, it means that 1 kg of CH₄ has the same impact on climate change as 25 kg of CO₂, and thus 1 kg of CH₄ is counted as 25 kg of CO₂ eq. The magnitude of climate change caused by different GHGs can thus be aggregated into a single unit of measure.

The carbon footprint can be quantified at a product level as in the life cycle assessment (LCA) described by the International Organization for Standardization.^11,12 LCA is a methodology aimed at assessing the environmental impacts associated with all the life cycle stages of a product from cradle to grave. The carbon footprint is an indicator of climate change amongst a set of environmental impact indicators that are quantified as part of LCA (e.g. ozone depletion, toxicity, acidification, eutrophication and so forth). IPCC guidelines⁵ represent the most formalized and widely accepted reference for the quantification of GHG emissions.

Organizations may report on two types of GHG emissions: direct emissions from sources which are owned or controlled by them (e.g. direct emission from fuel combustion in hospitals) and indirect emissions derived from their activities but occur from sources that are not owned or controlled by them (e.g. emissions from extraction, production and transportation of products purchased by hospitals). GHG emissions associated with food consumption in hospitals are mainly indirect emissions since most of them are produced in the food production stage.³

Existing LCA databases (e.g. Ecoinvent®, ELCD, U.S. LCI Database, or LCA Food Database) include life cycle inventories of resource use, energy demand and environmental releases associated with each life cycle stage for a wide variety of products and production processes. Moreover, several methods for life cycle impact assessment (e.g. Eco-indicator 99, CML 2002, or ReCiPe) have been developed to convert the inventory data into their corresponding environmental impacts. These methods have an impact category ‘climate change’ that expresses GHG emissions in terms of CO₂ eq, and they all use the GWPs developed by the IPCC.

Reliable data from LCA databases and literature, mainly about Spanish products^13,14 were used to obtain carbon footprint factors for the different food products in the hospital menus. Environmental burdens associated with the cooking of food and disposal of food waste from hospitals were not included to avoid double counting, since these are usually not considered in the food category but are included in energy and waste categories. Once the inventories were completed, the impact assessment method IPCC 2007 GWP 100 a, which is incorporated into the LCA software SimaPro®, was applied to quantify the carbon footprint of every food product. A large set of carbon footprint factors were thus obtained, which express the total emissions of CO₂ eq associated with 1 kg (or unit) of each food product without accounting for emissions from land use change (see online Appendix).

Meat is the foodstuff responsible for much of the carbon footprint in most hospital menus. Nijdam et al.¹⁵ has reviewed 15 LCA studies on beef, covering a variety of cattle farming systems. Carbon footprint ranged from 9 to 129 kg CO_2eqkg⁻¹ carcass weight (CW). The main reason why beef scores relatively high is that the process of fermentation in the rumen of ruminants produces the greenhouse gas methane. Differences in the studies analysed are mainly caused by differences in farming systems. Production of 1 kg of extensively farmed beef results in roughly three to four times as many greenhouse gas emissions as the equivalent amount of intensively farmed beef. Moreover, the carbon footprint of different parts of the cattle varies by price allocation.¹⁶

Pork shows a medium carbon footprint. Nijdam et al.¹⁵ has reviewed eight LCA studies on pork. Most of the eight studies reported values of around 5 kg CO_2eqkg⁻¹ CW (ranging from four to 11), which are reportedly mostly due to the N₂O emissions from feed production.

Carbon footprints CW for Spanish meat were obtained from Leip et al.¹³ Conversion factors of CW to bones free meat (BFM) were provided by Cederberg et al.¹⁷ for beef meat, and Sonesson et al.¹⁸ for pork and poultry meat (see online Appendix).

Carbon footprints of different parts of the cattle were differentiated by typical retail cuts in steaks, roasts and ground beef and stew meat. Spanish average prices for these cuts were used as allocation rules (see online Appendix).

The carbon footprint factors for food products were multiplied by the corresponding amounts to prepare each meal to quantify the total carbon footprint of each menu. Juan Ramón Jiménez Hospital provided detailed composition of their menus. The normal diet is given to patients who do not have special requirements. The daily menu corresponding to this diet has a calorific content of about 2.000 kcal.

Normal menus corresponding to one week in winter and one week in summer were analysed to obtain an average carbon footprint. The menus of the two weeks analysed correspond to the 50% of the total number of normal menus served in the hospital, as each week analysed was repeated every two weeks. For each lunch and dinner, there were two options for the first dish, second dish and side and four options for deserts. So, the total number of menus analysed were 448 lunches and 448 dinners.

About half of the 585 daily menus served were normal diets and the remaining had one of 17 therapeutic diets. The carbon footprints of the 17 therapeutic diets were estimated using a streamlined variation of each diet from the normal diet published by the Benidorm Clinical Hospital expressed in final weights (see online Appendix).

Results

Table 1 shows the carbon footprint for the different options offered for lunch in Juan Ramón Jiménez Hospital during one week in winter. Carbon footprints for dinner and for summer menus are in the online Appendix. The inclusion of red meat in some first and second courses for lunch leads to the carbon footprints being much higher. This result is more clearly observed in Table 2, which shows the significant difference in the average and standard deviation for lunch and dinner.

Table 1.

Carbon footprint for winter lunches during one week.

	First courses		Second courses		Side dishes		Desserts
Monday	Stew of lentils	1.691	Beef stew with tomato sauce	4.382	Chips	0.174	Yogurt	0.222	Apples	0.020
Monday	Chickpeas with spinach	0.348	Baked perch	0.489	Salad beet	0.146	Syrup fruit	0.143	Kiwis	0.069
Tuesday	Noodle soup	1.013	Roasted chicken with mushrooms	1.117	Mixed salad	0.062	Vanilla custard	0.239	Oranges	0.010
Tuesday	Macaroni with tomato sauce	0.465	Veal with mashed potatoes	2.151	Chips	0.174	Bananas	0.033	Chocolate custard	0.239
Wednesday	Rice ‘Mariniere’	1.280	Grilled pork chop	1.790	Mixed salad	0.062	Yogurt	0.222	Pears	0.053
Wednesday	Stewed red beans	1.003	Slice of pink ling (fish)	0.483	Baked potatoes with scrambled eggs	0.136	Vanilla custard	0.239	Rice pudding	0.275
Thursday	Chickpeas stew	1.054	Dogfish in tomato sauce	0.753	Mixed salad	0.062	Syrup fruit	0.143	Apples	0.020
Thursday	Three delights rice	1.194	Veal stew with vegetables	3.809	Vegetable stew	0.171	Cornstarch	0.275	Vanilla custard	0.239
Friday	Broad bean with squids stew	0.649	Marinated dogfish	0.797	Slices of tomatoes with dressing	0.101	Yogurt	0.222	Oranges	0.010
Friday	Stew of lentils	1.691	Roasted chicken breast	0.492	Mixed salad	0.062	Vanilla flan	0.239	Vanilla custard	0.239
Saturday	Noodle soup	1.013	Baked pork ‘Castellana Style’	1.501	Mixed salad	0.062	Yogurt	0.222	Bananas	0.033
Saturday	Italian pasta with tomato sauce	0.473	Grilled hake	1.341	Boiled new potatoes and green beans	0.110	Oranges	0.010	Apples	0.020
Sunday	Paella rice	1.329	Chicken breast with breadcrumbs	0.703	Mixed salad	0.062	Chocolate custard	0.239	Apples	0.020
Sunday	Beans casserole	1.003	Boiled green vegetables	3.804	Slices of tomatoes with dressing	0.101	Pears	0.053	Vanilla custard	0.239

Table 2.

Carbon footprint for lunch and dinner.

	Lunch		Dinner
	Mean	SD	Mean	SD
First course	0.987	0.495	0.336	0.207
Second course	1.620	1.328	0.728	0.489
Side	0.148	0.206	0.142	0.207
Desert	0.136	0.099	0.160	0.177
Bread	0.047		0.047
Total	2.939	1.436	1.414	0.596

Mann-Whitney U test was used to compare carbon footprint differences between winter and summer menus. The null hypothesis was that there is no seasonality in carbon footprint. The results of the test (p = 0.667) conclude the distribution of carbon footprint does not vary with seasonality. Therefore, all menus can be taken together to calculate the average value of carbon footprint of the normal diet (Table 2).

Average daily carbon footprint is 5.083 kg CO₂ eq (SD 1.555) comprising four meals (breakfast, lunch, afternoon snack and dinner) and a daily calorific intake of 2000 kcal.

Table 3 summarizes the carbon footprint for 18 different hospital diets. The average daily carbon footprint for the 17 therapeutic diets is 4.217 CO₂ eq. The menu for a high-protein diet has a higher carbon footprint (8.112 kg CO₂ eq) than the menu for normal diet and its calorific intake is also higher at 3.300 kcal. Four other diets (residue-free, hepatobiliary protective, low-protein diet with 60 g of protein and diet for bowel inflammation) have slightly higher carbon footprints but the rest have lower footprints. The most extreme case is the liquid anti-diarrhoea diet, which produces only 0.473 kg CO₂ eq/day, since it does not contain food but is only intended to rehydrate the patients.

Table 3.

Carbon footprint for daily hospital diets.

Menu	Carbon footprint (kg CO₂ eq/daily diet)
Normal or basal diet	5.083
Salt-free normal diet	5.081
Liquid diet	1.652
Semi-soft diet	2.781
Soft diet	3.839
Gastroprotective diet	4.696
Liquid anti-diarrhoea diet	0.473
Broad anti-diarrhoea diet	2.385
Residue-free diet	5.143
Residue-rich diet	4.909
Hepatobiliary protective diet	5.389
Low-protein diet with 20 g of protein	3.028
Low-protein diet with 40 g of protein	4.179
Low-protein diet with 60 g of protein	5.304
High-protein diet	8.112
Hyperuricaemia diet	4.718
Diet for bowel inflammation	5.684
Gastrectomy diet	4.386

Discussion

CO₂ eq emissions associated with food consumption in a hospital as a function of the number of menus consumed can easily be calculated. The values are based directly on purchased food weights and, as a result, are more reliable than other factors based on food costs, which are used in some carbon footprint calculators (e.g. DHMC Eco-Health Footprint Calculator⁹). Since the carbon footprint values for hospital menus do not depend on the fluctuation in price of food products, these do not require continuous updating.

The choice of suitable carbon footprint factors is critical to get reliable results. Peer-reviews are difficult, if not impossible, for calculators that do not provide clear, accurate, specific and transparent methods and scope.¹ Therefore, factors that consider local conditions have been chosen (available in online Appendix).

One frequent source of error (about 22%) is self-reported food intake diaries that are known to suffer from systematic under-reporting both of total food intake and of specific food types.¹⁹ In this paper, this error is minimized since carbon footprint of each diet relates to the amounts of purchased food.

Special attention needs to be paid to the carbon footprint of red meat as it has a very significant effect. Frequently in the literature, this factor is expressed by CW,¹³ but it is important that it is expressed by BFM¹⁷ and price allocation rules for different cuts are used.¹⁶

It should be noted that hospital menus considered here were based on detailed data from a Spanish hospital with a Mediterranean diet, but significant differences may exist in the carbon footprint of food consumption in other countries due to differences in both agricultural practices and dietary patterns. As an example of the total GHGs embodied in diet, Weber and Matthews²⁰ or Kim and Neff¹ estimated that average per capita food consumption in the US has embodied emissions of 8.5 kg CO₂ eq/day and 8.8 kg CO₂ eq/day, respectively. In the UK, Berners-Lee et al.¹⁹ reported 7.4 kg CO₂ eq/day for the average diet. In these two countries, the average calorific intake is also about 2000 kcal/day.

The average carbon footprint of these two countries is therefore much higher than the normal diet in the Spanish hospital analysed. While these differences may be partly attributed to uncertainty in the results, it is mainly due to differences in diet. The consumption of meat, or specifically red meat, is the greatest contributor to GHG emissions from food consumption, and the US is among the top countries in the world in terms of red meat consumption per capita.²¹ Conversely, the Mediterranean diet is characterized by abundant plant foods whilst red meat is consumed in low amounts.¹⁰ Duchin,²² who studied diets from multiple viewpoints of sustainability, concluded that the Mediterranean diet has a lower environmental impact than the average US diet and is also closer to public health recommendations issued by the World Health Organization.^10,23–26

References

Kim

Neff

. Measurement and communication of greenhouse gas emissions from US food consumption via carbon calculators. Ecol Econ 2009; 69: 186–196.

Garnett

. Cooking up a storm: food, greenhouse gas emissions and our changing climate, Guildford: University of Surrey, 2008.

Vermeulen

Campbell

Ingram

. Climate change and food systems. Annu Rev Environ Resour 2012; 37: 195–222.

Kaiser

Eagan

Shaner

. Solutions to health care waste: life-cycle thinking and “green” purchasing. Environ Health Perspect 2001; 109: 205–207.

Intergovernmental Panel on Climate Change (IPCC). Climate change 2007: the physical science basis. Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change, Cambridge: Cambridge University Press, 2007.

Grose

Richardson

. Managing a sustainable, low carbon supply chain in the English National Health Service: the views of senior managers. J Health Serv Res Policy 2013; 18: 83–89.

Lynch

. Greening health care: how hard can that be? J Health Serv Res Policy 2011; 16: 247–248.

NHS Sustainable Development Unit. Carbon footprint update for NHS in England 2012, Cambridge: NHS Sustainable Development Unit, 2012.

Dartmouth-Hitchcock Medical Center (DHMC). Eco-health footprint calculator, https://sites.google.com/site/dhmccalculator/home (2009, accessed 11 June 2013).

10.

Willett

Sacks

Trichopoulou

. Mediterranean diet pyramid: a cultural model for healthy eating. Am J Clin Nutr 1995; 61: 1402S–1406S.

11.

ISO 14040:2006. Environmental management – life cycle assessment – principles and framework.

12.

ISO 14044:2006. Environmental management – life cycle assessment – requirements and guidelines.

13.

Leip A, Weiss F, Wassenaar T, et al. Evaluation of the livestock sector's contribution to the EU greenhouse gas emissions (GGELS)–final report. Ispra: European Commission, Joint Research Centre, 2010.

14.

Iribarren

Vázquez-Rowe

Hospido

. Estimation of the carbon footprint of the Galician fishing activity (NW Spain). Sci Total Environ 2010; 408: 5284–5294.

15.

Nijdam

Rood

Westhoek

. The price of protein: review of land use and carbon footprints from life cycle assessments of animal food products and their substitutes. Food Policy 2012; 37: 760–770.

16.

Weidema BP. Market information in life cycle assessment. Vol. 863. Copenhagen: Miljøstyrelsen, 2003.

17.

Cederberg C, Sonesson U, Henriksson M, et al. Greenhouse gas emissions from Swedish production of meat, milk and eggs 1990 and 2005. Gothenburg: SIK-Institutet för livsmedel och bioteknik, 2009.

18.

Sonesson U, Davis J and Ziegler F. Food production and emissions of greenhouse gases. Gothenburg: SIK-Institutet för livsmedel och bioteknik, 2010.

19.

Berners-Lee

Hoolohan

Cammack

. The relative greenhouse gas impacts of realistic dietary choices. Energy Policy 2012; 43: 184–190.

20.

Weber

Matthews

. Food-miles and the relative climate impacts of food choices in the United States. Environ Sci Technol 2008; 42: 3508–3513.

21.

Food and Agriculture Organization of the United Nations (FAO). FAOSTAT: Food supply: livestock and fish primary equivalent, http://faostat.fao.org/site/610/DesktopDefault.aspx (2009, accessed 19 December 2013).

22.

Duchin

. Sustainable consumption of food: a framework for analyzing scenarios about changes in diets. J Ind Ecol 2005; 9: 99–114.

23.

World Health Organization (WHO). Diet, nutrition and the prevention of chronic diseases: report of a joint WHO/FAO expert consultation. WHO technical report series 916, Geneva: WHO, 2003.

24.

Willett

. The Mediterranean diet: science and practice. Public Health Nutr 2006; 9: 105–110.

25.

Trichopoulou

Lagiou

. Healthy traditional Mediterranean diet: an expression of culture, history, and lifestyle. Nutr Rev 1997; 55: 383–389.

26.

Willett

. Diet and health: what should we eat? Science 1994; 264: 532–537.