The environmental impact of a food, be that carbon footprint, water use, land use or some other factor, can be estimated by life cycle analysis (LCA). With the environmental impact of food an increasingly important consideration for many consumers, industry and policymakers, the FAO have recently published a report on the challenges and opportunities of nutritional LCAs – those that attempt to capture the nutritional value of food alongside its environmental impact.
LCAs are strongest when used to identify hotspots or areas for improvement within the supply chain for a single item. They can be used to answer industry questions like: where should we act first to lower the footprint of our product? They can also be used in comparisons between two otherwise identical products for consumers: which one should I buy? However, challenges arise when LCAs are used to compare the impacts of very different products.
Take Energy Rating labels on electrical appliances as an example. Analogous to LCAs, these are an indication of the relative energy usage of a particular model compared with other appliances of the same type. These are useful for comparing two refrigerators, but do not really help when comparing refrigerators with freezers. They are even less useful when comparing a refrigerator with a washing machine: the appliances have completely different functions, and a purchaser would be unlikely to use them to choose which of the two to take home.
Even within the category of refrigerators, ratings become less relevant when comparing different size models, as they provide a different level of service. Without considering the service or benefit provided by the product we do not have a fair basis on which to compare the footprint or cost of providing that service.
The same problem exists when comparing foods. When we look at the footprint of food products and start making comparisons, we need to be clear on the service or benefit being provided by the products to ensure we are making a valid comparison. However, the service provided by a food item depends on the purpose for which it is consumed.
Food is consumed for a variety of reasons: as a source of nutrition, for sensory experience or pleasure, or for social and cultural purposes. Accounting for these different purposes is not straightforward. For example, from a nutritional perspective, alcoholic beverages provide very little benefit, but many consumers may still place high value on their sensory or social purposes.
The FAO report focuses on nutrition, rather than the other services provided by food, and looks at how nutritional information can be combined with environmental impact data.
One approach is to try and bring together the “benefit” and “cost” into a single analysis: the development of a nutritional LCA (nLCA), a life cycle analysis that includes nutrition.
There are two different methods by which this can be done:
- As part of the definition of the functional unit (e.g., land use per 100 kcal)
- As part of the human impact assessment, what is often thought of as the cost side of the analysis (e.g., likely impact on human health)
Neither of these approaches is easy.
Shifting functional units
Often, an LCA uses mass as the functional unit. For example, if considering the water use needed to grow rice, an LCA might report results as “litres of water used per kg of rice”. In this case, the functional unit is “1 kg of rice”.
Putting nutrition into the functional unit moves away from just using mass. In the simplest form, this may be evaluating a set of foods based on the amount of a particular nutrient they contain. Protein is often used for this purpose. Our rice example would then change to “litres of water used per kg protein in rice”.
However, protein is not a single nutrient needed by the body, but rather a collection of amino acids, which are the essential nutrients. Not all proteins are created equal, having both different concentrations of these amino acids and varying in their digestibility. Rice protein is therefore different to soy protein, for example. Thus, comparing water use per kg protein does not capture this information. Sophisticated methods that include protein quality exist, but are challenging and rarely used.
Most food items provide more than one nutrient, and we need a broad range of nutrients to remain healthy. The DELTA Model® estimates the ability of the global food system to supply a basket of 29 nutrients, and would include more given suitable data. Evaluating a food item based on only one target nutrient misses this complexity.
An alternative to selection of a single nutrient as the functional unit is to use a basket of nutrients to create some form of nutrient reference score. The intention of this score would be to provide a more “balanced” view of the nutrition provided by foods. However, what nutrients should make up this score? Do they all have equal weighting? Or are some more important than others? And how does this relate to the needs of an individual? The scientific literature contains many different suggestions, each with their strengths and weaknesses. Each is at risk of introducing some form of bias into the assessment.
Another important consideration is portion size. Once we move away from a functional unit based on mass, we lose some of the context around the amount of food that needs to be consumed to deliver a particular nutrient or group of nutrients, and how that relates to the size of a normal serving. Functional units “per serving” have also been explored, but face the same problems as mass based units.
Bringing human health into the assessment
The alternative approach is to leave the functional unit as the mass of the food item and build the nutritional assessment into the impact side of the LCA. This requires having data on the expected impact of consuming a food for human nutrition or health. The main approach that has been considered to date uses epidemiological data on diets, health, and mortality. This is usually of the kind captured in the Global Burden of Disease (GBD) study, which calculates statistical links between consumption of food groups and expected lifespan or quality of life.
Unfortunately, this data is limited to comparatively coarse effects. The GBD study reports statistical measures for 15 health aspects related to diet and 3 related to nutrient deficiency. The statistical associations are the result of a complex analysis that attempts to isolate the impact of individual food factors on overall outcomes. Changes in assumptions used in the analysis between the 2017 and 2019 data sets resulted in significant changes in the apparent impact of several food groups. These have been highlighted in a recent letter to The Lancet, and would have a major effect on any nLCA employing this data.
In general, the benefits of consumption of food or nutrients follow a curve. Initially there is a positive impact on health, with increasing consumption providing nutrients essential to bodily functions and growth. This benefit is reduced once daily requirements are met, and, if consumption continues to increase, may eventually have negative health outcomes.
This is illustrated with the energy content of diets: eating insufficient calories leads to wasting, but eating too many leads to obesity and a range of related health conditions, and just how much is too few or too many depends upon the need of the individual. Sodium is another example: a diet deficient in sodium can have serious health consequences. However, many diets contain a considerable excess of sodium, carrying health risks for many individuals.
Putting food and nutrients in context
Food items are consumed as part of meals and diets, and it is at this level that we need to apply considerations of nutritional sufficiency. The relative nutritional benefit of consuming a food item varies based on the dietary context of the individual. For example, the protein or amino acid content of a food item may be of limited value in a diet that is otherwise oversupplied with this nutrient, but of immense value in a diet that is deficient.
Within the DELTA Model we have implemented a simple nutrient contribution measure for food items. This is based on the sum of the relative contribution the food item makes to each of the nutrients captured in the model. As such, it gives a higher weighting to nutrients that have low global availability and a lower weighting to nutrients that are abundant.
For example, the default 2018 DELTA Model scenario has a 34% deficiency for calcium against global requirements (achieving 66% of target), whereas phosphorous has a 150% excess (250% of target). Thus, a food that provides 33% of the daily target for calcium gets a score of 0.5, whereas 33% of the daily target of potassium scores only 0.13 – approximately ¼ the importance. A similar approach has recently been published for the individual dietary context.
The right use of nLCA
The challenges described above stem from trying to compare refrigerators with washing machines, and lead us to the fact that nutrition does not easily collapse into a single score.
The scope of comparisons, or the grouping of foods into groups becomes important. If food items are grouped with others that provide or purport to provide similar nutritional benefits, we can make more realistic comparisons that better reflect the real choices facing us.
As an example, we might compare the nutritional LCA of milk with that of a plant beverage and use a nutritional functional unit that reflects the role of these items within the overall diet. Milk products make a significant contribution to the global supply of calcium, phosphorous, and potassium, six indispensable amino acids, dietary fat, overall protein, and vitamins A, B2, B5, and B12. A nutritional functional unit could be designed that reflects this nutritional value to enable us to compare milks and milk-alternatives when consumed as a source of nutrients. However, this same approach would not necessarily be appropriate if the purpose of the product was simply to whiten a cup of coffee. The intended service or benefit of foods must be understood when deciding how to compare costs.
Whilst the concept of a universal nutritional LCA that provides all the information necessary to support a wide range of decisions is attractive in its apparent simplicity, the reality is that nutrition and environmental impacts are too complex, and too important, to be reduced to a single number.