Can Scientists Calculate a Carbon Footprint for Drugs?

Anyone picking up a prescription from their pharmacy is used to the reams of paper that typically accompany it: information on safety, storing, and how to use the drug. What if that information also included your medicine’s carbon footprint?

Some scientists are trying to calculate those footprints, as part of a growing effort to understand biotech and pharmaceutical companies’ contributions to climate change. While overall estimates suggest this industry emits millions of tons of carbon dioxide each year, it’s often tough to pinpoint the exact sources of those emissions, presenting a challenge as companies seek to reduce their footprints and outside organizations seek to evaluate their efforts.

A recent STAT report found that the majority of large pharmaceutical and biotech companies aren’t publicly disclosing their emissions to a global organization that sets standards for climate transparency. One key reason, experts say, is that pharma and biotech emissions can be difficult to estimate, especially for midsize and smaller companies unable to devote teams of analysts to these calculations.

In other industries, most greenhouse gas emissions come from companies’ direct activities and energy use, referred to as Scope 1 and Scope 2 emissions, respectively. But for pharma and biotech, about 90% of emissions come from indirect sources, falling into the Scope 3 category. These sources include emissions from the raw materials that go into drugs and devices, chemical processes that turn those raw materials into products, transporting and storing the products, their use in medical settings, and disposal, often in a landfill.

Drugmakers, health care organizations, and outside research groups are increasingly focused on calculating carbon footprints for individual products. The analyses help to provide more accurate estimates of overall corporate emissions and show where there’s room for improving sustainability.

These calculations are “a priority for everyone,” said Nazneen Rahman, founder and CEO of Yewmaker, a startup that is working on medical carbon-footprint research. “It’s obviously a priority for the manufacturers, because they have to reduce their … emissions. And it is a real priority for health systems” because medicines make up a significant share of their own emissions, she said.

Some hospitals, for instance, want to use specific emissions data to inform which medical products they purchase, said Sean McGinnis, a professor in green engineering at Virginia Tech who specializes in these assessments. By buying more sustainable medicines, a hospital can reduce its own greenhouse gas emissions — responding to requests of administrators who are increasingly interested in “having a good carbon footprint,” he said.

In pharma, GSK aims to reduce the environmental impact of its products and packaging 25% by 2030, and footprint calculations help show opportunities to cut those emissions, said Claire Lund, an author of research on this topic and vice president for sustainability at GSK, one of the top-ranked firms in STAT’s recent report. “Having product-specific footprints allows us to better understand the environmental impact of our products across their entire lifecycle,” she said.

Global health agencies are also starting to ask for these data, said Neel Lakhani, senior director of strategy and innovation at the Clinton Health Access Initiative (CHAI), which has supported research into health care sustainability. For example, health organizations in Africa might want to know if manufacturing COVID-19 vaccines locally would reduce the carbon footprint of those products, compared to shipping them from elsewhere in the world, he said.

While the data requests are growing, calculating the carbon footprint of an individual medical product can be complex. Researchers like McGinnis face challenges with obtaining data on a drug’s composition, accounting for inconsistencies in how companies report emissions data, and understanding what happens to products after they leave the factory.

While precise numbers may be hard to find, the carbon footprint information that we have suggests health products continue to be major contributors to climate change. For example, one recent paper from Adam Cimprich, a postdoctoral fellow at the University of Waterloo, and colleagues, found that the annual carbon emissions from treating the occupants of a single bed at a hospital in British Columbia, Canada is about the same as for five households. The medical products used to care for patients in that bed led to a significant share of emissions, Cimprich said.

Researchers interested in the carbon footprint of a medicine typically arrive at an estimate through a scientific process called life cycle assessment (or LCA). Matt Sawyer, a consultant specializing in environmental sustainability in health care, described the process as similar to a “cake recipe”: researchers add together many different ingredients and might arrive at an unexpected end result.

Each component in the recipe is a smaller-scale assessment in itself, answering questions such as: How much carbon dioxide is generated in collecting the raw materials for this medicine? How much carbon dioxide is generated from transporting those raw materials to a factory where they will be processed? How much carbon dioxide from the chemical procedures used to turn the raw materials into a medical product? How much carbon dioxide from packaging the resulting medicine, transporting it to a health care facility, storing it, giving it to a patient, discarding of the final product if some is left unused?

The list of questions can be endless, so one key step in the life cycle assessment process is setting boundaries. In a course he teaches on these assessments, McGinnis typically tells students, “Draw a box around what you consider your product.” Some assessments might go all the way from raw materials to disposal, while others could focus on the activities in a factory setting. There’s no standard procedure in the pharmaceutical industry for which activities are left in or out of the box, so these choices can vary widely from one research project to another.

For scientists analyzing medical products, one challenge can be the number of components and steps involved, said Amy Booth, a doctoral candidate at Oxford University who studies environmental impacts of health care. For example, she said, compare the life cycle of a drug to that of a tomato. The tomato’s life cycle may include its growth on a farm or in a greenhouse (requiring water, maybe some pesticides, maybe heating), followed by packaging and distributing it for consumption, and typically ends in a human stomach. It’s a fairly straightforward, easy-to-measure process with readily available data, she said.

Drugs, on the other hand, require a variety of active pharmaceutical ingredients and other materials involved in their research, development, and manufacturing. “You go through that whole chemical process with the manufacturing, and then there’s the packaging and distribution,” Booth said. Drugs also may require different considerations for transportation and storage than food products, such as if they need to be stored in a special freezer. And drugs tend to create significant waste: All the pill bottles discarded, expired or simply unused add up.

Overall, “the more steps you have in that production process, the bigger [environmental] impact it’s going to have,” Booth said. Emissions can also vary widely depending on where a drug is made, which adds complexity. The carbon footprint of a medicine produced in a factory powered by coal would be significantly higher than the footprint of the same medicine produced in a factory powered by solar.

When adding up those production steps, there’s little public data describing how pharmaceutical processing leads to emissions. For life cycle assessments in other industries, researchers can rely on open databases that provide standard values, called conversion factors, translating from common materials to the greenhouse gases emitted in their production. These databases can be used to analyze medical devices and other products made of metal or plastic, such as masks and gowns, said Xiang Zhao, a doctoral student at Cornell University who has worked on these assessments.

But there’s no database for the active pharmaceutical ingredients used in drugs, and pharma companies tend to keep that information secret. While proprietary data is an issue for life cycle assessments across industries, McGinnis said, the medical industry tends to be “less willing” to share. Any attempt to estimate emissions from a drug that uses proprietary chemicals is “where it really gets hard,” he said.

Some researchers try to ask companies for their data, with mixed results. Sawyer’s attempts typically lead to no reply, or a reply simply linking to the company’s public sustainability report (which usually has limited details), or — in the best case scenario — a total carbon footprint figure that fails to share any methodology behind the number.

Even pharma companies themselves are limited in their ability to calculate carbon footprints, because their internal data miss a key part of the life cycle: what happens to drugs after they enter the health care system. To capture those emissions, companies would need data from health providers; and those emissions, too, can vary by location. The same medicine might have a lower footprint in a big city, where patients have a short trip to their pharmacy, compared to rural areas where more driving is required.

Scientists still persist in this research because the results can be incredibly informative for health organizations. For example, a paper by Zhao and colleagues found that hospital gowns marketed as biodegradable are actually less environmentally friendly than their conventional counterparts, due to carbon dioxide and methane released after the biodegradable gowns are placed in landfills. Another paper by the same group at Cornell and Lakhani at CHAI identified major sources of emissions — and potential options for improving sustainability — in the production of a common HIV drug.

These assessments show the pros and cons of choosing one medical product over another, or adjusting aspects of the production process, Zhao said. Otherwise health organizations are kept guessing about which option is the most sustainable.

To better understand the carbon emissions generated by medical products, the pharma and biotech industry needs to develop more transparency around sharing their existing data and standards for calculating life cycle assessments, experts say.

In one step towards data transparency, several Canadian health institutions have supported HealthcareLCA, an online library of academic papers estimating the environmental impacts of different health products and processes. This project is a helpful starting point, said Cimprich, the University of Waterloo researcher who studies health care life cycles. But different papers in the library use “different methods, assumptions, and qualities of data,” he added — there’s no standardization in the results.

“Health care is playing catch-up here with other sectors” when it comes to pooling data on products’ carbon emissions, Cimprich said. Other industries like food, construction, and different manufacturing sectors have more extensive data available.

Experts like Cimprich and Booth, at Oxford, would like to see leaders in health, pharma, and biotech companies look to these other industries as models for building public databases that help researchers translate from chemical building blocks to emissions and other environmental impacts. Companies shouldn’t “reinvent the wheel,” but should rather “draw on other industries that have done product footprinting already,” Booth said.

Researchers can look forward to one such database later this year. Rahman and colleagues at Yewmaker are working on a scientific paper and open-access database that will provide carbon footprint estimates for medicines made with small molecules, a type of drug that accounts for about 90% of pharma products. The estimates are based on data science models, incorporating different chemicals’ molecular structures and standard manufacturing processes, Rahman said.

Yewmaker’s database is set to offer more comprehensive information than scientists could previously access about drugs’ emissions: It will have an “internal rigor and comparability” unlike prior papers that evaluate one drug at a time, Rahman said. But these estimates will still be less accurate than information that pharma companies might provide from internal research. Rahman hopes any companies that notice inaccuracies in Yewmaker’s data will be motivated to publicly correct the record.

Another potential source for standardized data might be the Sustainable Markets Initiative’s Health Systems Task Force, a collaboration of executives from top pharma companies including AstraZeneca, GSK, Merck, Novo Nordisk, and others. By working together, top companies could develop emissions measurement standards for the rest of the industry. Such standards may be particularly helpful for smaller companies like those that produce generic medicines, Rahman said: these companies have fewer resources for internal measurement but still make a lot of drugs.

However, some scientists are critical of pharma companies’ ability to develop their own standards and regulations. Sawyer, the consultant, would prefer to see regulations come from government agencies, such as the U.K.’s National Health System — which is already ahead of U.S. agencies on sustainability commitments. Governments may consider incorporating environmental assessments into their standards for approving new drugs, Sawyer said, perhaps on a parallel track to existing standards for safety and effectiveness in clinical trials.

At the same time, large health organizations are starting to put pressure on companies to share carbon footprint data. Lakhani at CHAI sees the increased pressure as a classic “carrot or the stick” situation: health organizations could offer a “carrot,” by telling companies that they’d be more likely to buy medicines that are more environmentally friendly; government agencies could offer a “stick,” by only allowing companies that disclose emissions data to sell their products. “Sometime, hopefully in the near future, [environmental disclosure] becomes the standard,” Lakhani said.

This story is part of coverage of climate change and health, supported by a grant from The Commonwealth Fund.

This story was originally published by STAT, an online publication of Boston Globe Media that covers health, medicine, and scientific discovery.