Purpose and limitations of carbon footprinting for healthcare

How can carbon footprinting within healthcare be helpful?

We suggest four key ways that carbon footprinting can be informative in healthcare (Figure 1).

OPEN IN VIEWER

Figure 1.

Four purposes of conducting a carbon footprint.

First, it can be used to define a baseline and, by applying a consistent methodology, monitor progress. For example, in 2019, the carbon footprint of the NHS in England was estimated, with a backcast to 1990 and a forecast to track progress against targets. ⁴

Second, carbon footprinting can be used to identify the products or processes that contribute the most to greenhouse gas emissions, thereby helping to define where efforts could be targeted. For example, analysis of septic shock management in intensive care found that energy use (primarily for heating, ventilation and air conditioning) was responsible for 76% (Australia) to 87% (USA) of emissions, ⁵ and a UK study of five common operations identified some of the products with the greatest carbon footprint reports, including single-use drapes, gowns and clip appliers. ⁶

Third, carbon footprinting can be used to compare alternative products or processes if consistent methods are applied. For example, an analysis of hand decontamination found that alcohol-based techniques had a lower carbon footprint than soap (1060–1460 million vs. 2300–4240 million kg carbon dioxide (CO₂) per year for UK population). ⁷

Fourth, carbon footprinting can help to optimise processes, using scenario analysis to model changes. For example, a US study found that strategies to mitigate the carbon footprint of laparoscopic hysterectomy included using total intravenous anaesthesia, reducing single-use instruments and re-processing single-use devices. ⁸

Using carbon footprints for these four purposes may provide useful insights to inform change in practice and policy, but with caveats as outlined below.

What questions should we ask when looking at carbon footprint studies?

There are limitations to carbon footprint estimation, and pitfalls in interpretation. There may be misconceptions that carbon footprints provide a definitive value for environmental impact, but instead they should be considered an estimate. Many factors can heavily influence calculations, including a methodological approach, what is included in analysis, assumptions on missing data and sources of conversion factors.

The first questions to ask are where and when was the study conducted and whether the findings are applicable to your setting. For example, in 2022, the proportion of fossil fuel-derived energy in China was 82% but in Sweden it was 26% (ourworldindata.org/energy), and an increase in the proportion of renewable energy meant that emission factors for UK electricity dropped from 0.35 kg CO₂e per kWh in 2018 to 0.28 in 2023. ⁹ Hence, a similar product manufactured in Sweden will likely have a lower carbon footprint than one manufactured in China, and a product manufactured in the UK in 2023 will likely have a lower carbon footprint than one manufactured in 2018. When applying findings to a specific clinical setting, important factors affecting carbon footprint beyond country energy mix include how a product is manufactured, distributed (modes of transportation and distances), re-processed (e.g. how they are cleaned or sterilised) and disposed. Our analysis of personal protective equipment used in England during the coronavirus disease 2019 pandemic found that altering overseas transportation from shipping to air freight increased the carbon footprint by 50%. ¹⁰

The second question to ask is what is included in the analysis. A line will always be drawn around what is included and excluded, causing a gap between estimated and real-world greenhouse gas emissions. The decision on what to incorporate is usually driven by the research question but should include all relevant associated material and energy. An example of how incomplete inclusion criteria can result in misleading results is studies of remote patient consultation, which often measure the reduction in carbon associated with a consultation without examining changes to the entire patient pathway. A review of remote consultations in ear, nose and throat surgery found that 13–72% of patients required subsequent face-to-face consultations, ¹¹ which could limit or even reverse any apparent carbon benefit.

The third question is whether a comparison between alternative products or processes is fair. For example, a German study that examined approaches to spinal fusion surgery compared six reusable instrument sets (weighing 45 kg) to two single-use instrument sets containing just a few instruments (weighing 2 kg). ¹² The study found that single-use sets had a lower carbon footprint. However, if reusable sets (with fewer instruments) had been compared, they would almost certainly have had a lower carbon footprint than the equivalent single-use versions. Comparisons must be made with consideration to functional equivalence.

What are some other limitations of carbon footprinting?

There are some wider risks and pitfalls to be aware of in carbon footprinting.

First, carbon footprints focus on greenhouse gas emissions because these are key drivers of climate change and the primary focus of national and international targets. However, there are wider environmental impacts, which can be evaluated using full life cycle assessment (LCA). For example, the previously mentioned study on hand hygiene techniques found that isopropanol-based hand sanitiser had lowest environmental impact in 14 of 16 impact categories. However, ethanol-based hand sanitizer had the lowest impact on fossil fuel use, and bar soap and water had the lowest impact on photochemical ozone formation. ⁷ Focusing solely on carbon footprint provides simplicity, while calculating and interpreting multiple environmental impacts require greater expertise, especially where results are discordant across different impact categories. Comparing normalised results, whereby environmental impacts are converted to values relative to the average human’s emissions, may reduce subjective preferencing of different impacts. Those accessing such studies should also be aware that existing LCA methods may fail to capture some environmental harms, such as microplastics generation, bioaccumulation of antibiotics or pharmaceutical exposure in wastewater.

Second, absolute results should be compared with caution between studies. For example, the carbon footprint for three laparoscopic products (single-use laparoscopic ports, clip appliers and scissors) was 37–44% lower in our studies when using carbon footprinting ⁶ compared with LCA methods. ¹³ Differences were due to what was included, granularity and specificity of data collected, and the source of conversion factors (methodological choices driven by the respective research questions). For example, when considering conversion factors, Ecoinvent (ecoinvent.org, a commonly used life cycle inventory database) reports 0.38 kg CO₂e per kWh of electricity for 2016, which is 9% higher than the UK government value for the same reference year (0.35 kg CO₂e per kWh). ⁹ While individual carbon footprinting studies can provide valuable insights into strategies to improve sustainability (in line with Figure 1), we caution against making contractual, financial or value-based decisions comparing absolute results across studies. We should also recognise that for studies funded by industry, there is incentive to make commercially favourable decisions, which increases the risk of false or incomplete reporting.

Third, it is important to recognise there may be alternatives not captured in studies. For example, a German study estimated that a remanufactured single-use electrophysiology catheter was associated with half the carbon footprint of a new catheter (1.75 vs. 0.87 kg CO₂e), ¹⁴ but did not model the impact of a catheter redesign to allow multiple re-use, which would likely reduce carbon footprint further.

Finally, our experience in the NHS indicates increasing sentiment among stakeholders (including healthcare professionals, procurement teams, managers and policy makers) that we should carbon footprint everything in healthcare. As outlined above, we should exercise caution in trying to use such data to inform procurement decisions or clinical product choices. Generating product-level carbon footprints for every healthcare product would be a protracted process and an opportunity cost, requiring sustained financial investment and expansion of carbon footprinting expertise (both to undertake and interpret such analysis). Instead, we recommend targeting high-level approaches, such as reducing and reusing equipment where possible, and once that opportunity is exhausted, targeting select product categories with greatest environmental impact using a standardised methodology, as explored in depth in our recent policy brief. ¹⁵

Appropriate use of carbon footprinting can provide valuable insight for improving environmental sustainability of healthcare, but we do not need to carbon footprint everything, and nor should we try.