William Irwin’s supervisors had a simple request. They just wanted him, a scientist who assesses the safety of chemicals at the Environmental Protection Agency, to sign off on a report that would give a chemical the agency’s OK to enter the market. It would have been easy for him to affix his signature to the document, which would have allowed him to finally get the project he had worked on for months off his desk.

The only problem? The science didn’t show that the chemical was safe.

Irwin, who has a doctorate in biochemistry and molecular biology and three board certifications in toxicology and has worked at the EPA for 12 years, was asked to greenlight the chemical in May 2020. But he had already found reason to be concerned about the chemical’s effects on developing fetuses. Because the manufacturer hadn’t submitted any health studies, he had found an analog — a structurally similar chemical to help predict its effects. That compound was bisphenol A, or BPA, an additive to plastics that is well known to cause reproductive problems as well as immune and neurological effects. Irwin had noted his concerns and a lack of studies that could allay them. His work on the assessment, a document that determines if and how a chemical is allowed to be used, also pointed out that workers faced risks from both breathing and touching the chemicals.

Hair on Fire

But because the chemical fell into a category known within his division as “hair-on-fire” cases, high-priority situations in which manufacturers object to scientists’ findings that their chemicals pose hazards, the conclusions Irwin reached in the assessment were scrutinized and ultimately changed. After he refused to sign off on the altered findings, a higher-ranked employee in the division took over the case and used a different — and, according to Irwin, scientifically inappropriate — approach that allowed the agency to avoid calculating the developmental risks posed by the chemical. On June 15, 2020, the agency posted the final risk assessment of the chemical, which minimized risks to the general population and didn’t calculate the risks to workers that Irwin had highlighted.

Because of confidentiality claims by the manufacturers, Irwin and his colleagues who spoke with The Intercept felt that they were legally bound not to reveal the names of this and other chemicals mentioned in this story.

As The Intercept reported in the first installment of this series, managers within the EPA have pressured health assessors to minimize and downplay the risks posed by chemicals for use even when there is clear evidence that they present harms. That reporting was based on extensive information provided by four whistleblowers — Elyse Osterweil, Martin Phillips, Sarah Gallagher, and William Irwin — who worked as health hazard assessors in the new chemicals division of the agency. This article is based on new information from those whistleblowers along with one additional EPA risk assessor, who has recently come forward but has chosen to remain anonymous because of fear of retaliation.

Through documents, interviews, and a written disclosure that they provided exclusively to The Intercept, the EPA inspector general, and select members of Congress, these government scientists have detailed some extreme examples of interference with the scientific evaluation of chemical safety. Working with Public Employees for Environmental Responsibility, also known as PEER, an environmental group that helps whistleblowers, they laid out the process through which industry exerts particular pressure in “hair-on-fire” cases — sometimes through members of Congress — to expedite the approval of chemicals or minimize their restrictions.

email-EPA2

Internal EPA email referencing “hair-on-fire” cases.

Image: Obtained by The Intercept

In these cases, some EPA managers have worked hand-in-hand with industry to fend off the safety concerns raised by the agency’s own scientists about new chemicals entering the marketplace. While not all of the instances of manipulated chemical assessments were designated “hair on fire,” these high-priority cases have become so common that the division of new chemicals has arranged to have someone always available to oversee them. Since at least January 2021, managers have been on a rotating schedule for “HOF Duty,” as it’s called on internal calendars.

calendar-EPA

An internal calendar for human health assessors within the EPA’s new chemicals division includes rotating HOF — or “hair-on-fire” — duty.

Screenshot: Obtained by The Intercept

EPA managers use a variety of tactics to persuade staff scientists to change their assessments in these cases, according to the whistleblowers, including emphasizing the displeasure and impatience of the companies that make the chemicals; criticizing the scientists for taking too long to approve chemicals in their performance reviews; and even insinuating that individual assessors will be sued by the chemical manufacturers if they insist on standing by their findings. In several cases in which the whistleblowers refused to make changes, other EPA employees deleted or altered the information in the documents without their approval or knowledge.

So great are the tensions in the new chemicals division that managers recently sought to have their internal computer system revised so that they could automatically override staff scientists in some “hair-on-fire cases,” according to audio obtained by The Intercept. In a November 18, 2020, meeting with information technology contractors who were working on the new chemical review tracking system, managers of the division suggested adding a new button to the interface that with a single click would enable these high-priority cases to bypass scientific review from staff and go straight to management.

“Is there a way to override it all?” a division manager asks at the meeting. “So for example, say there are nine open discussions but then we suddenly had to expedite a case due to some hair-on-fire situation, uh, do we need nine people to go in and close out these discussions for it to move forward or, uh, does, is there an all-powerful person like [name removed] can come in and override everything?”

Ultimately, the division manager and the consultant did settle on just this plan at the end of the meeting. “Yes, we’d, we’d like a button or something, and then we’ll assign whoever that person or persons may be to override everything,” the division manager says. It’s not clear whether the override button has been implemented.

“Okay. Uh, override meaning that this turns to green?” the consultant clarifies.

“Yes.”

“OK, perfect.”

The Chocolate Factory

The EPA has the authority to prevent new chemicals from entering the market if they present an unreasonable risk to human health or the environment. But the agency’s division of new chemicals didn’t do that for any of the 3,835 new chemical applications submitted to the agency between June 22, 2016, and July 1, 2021. In response to questions, the EPA pointed out that, during that period, 3,256 cases were reviewed; of those, 266 were determined to be invalid or incomplete and 11 were required to undergo further testing. But not a single chemical was “prohibited from commercializing,” according to the EPA’s own data. Instead, whether in “hair-on-fire” or regular priority cases, health assessors who identified hazards caused by chemicals routinely faced intense pressure to dismiss or downplay them. According to the whistleblowers, there is a strong expectation that assessors should not just deem chemicals safe but also do so quickly.

The law gives the EPA 90 days to respond to a new chemical submission. But determining whether chemicals are safe can be time-consuming, especially since the assessors often receive very little information about their health effects. In some cases, companies provide only the chemical name and structure without any information on health and safety. To properly vet them, assessors need to find and read through relevant studies and sometimes seek opinions from outside experts. Yet the whistleblowers said they were regularly encouraged to pass even these mysterious chemicals so quickly that they were often unable to thoroughly review them.

“We’re not being given enough time,” said Gallagher, a toxicologist who worked as a human health assessor in the Office of Pollution Prevention and Toxics. “We need time to figure out if there are experts we need to talk to. But we’re being told to get cases done within two days of our hazard meeting.” In addition to evaluating the safety of new chemicals, the assessors say they were increasingly asked to make changes to cases they had already completed. “I got cases I finished four months before, and they would want me to make substantial revisions that were deviations from our process, and it’d be thrown on top of my current work,” said Gallagher, who likened the experience to the “I Love Lucy” episode set in the chocolate factory. “It was speeding up and speeding up, and we had more and more, and we just couldn’t keep up with it.”

Irwin said he regularly had to work on weekends to keep up with the workload and was even asked to hand in two complex assessments by the end of the day immediately after telling his supervisor that he had to go home because he was experiencing severe chest pain.

All four whistleblowers have been reprimanded for not moving quickly enough when reviewing new chemicals. Gallagher’s supervisor told her that she had difficulty “letting go when a decision was made” about a particular chemical. In that case, someone else had changed her assessment, removing information she had included about the chemical’s carcinogenicity.

Irwin’s supervisor cited his refusal to sign off on the assessment of the chemical that was similar to BPA as a reason to downgrade his performance rating in an annual review. And Martin Phillips, another of the whistleblowers, was also criticized for the timeliness of his assessments in his performance review.

Similarly, in a June 21 performance review shared with The Intercept, Elyse Osterweil’s supervisor accused her of having a “pattern of providing late and incomplete work products.” According to Osterweil, these delays were the result of her refusal to capitulate to the pressure to deem certain chemicals as safe based on evidence that they may cause specific harms, including birth defects and cancer. Yet her supervisor pointed out that the missed deadlines “resulted in the Agency being contacted by several outside submitters, on several occasions, to express their dissatisfaction with the timeliness of cases assigned to you.”

Insufficient Data

Assessors can halt or delay approval if they don’t have enough information to declare a chemical as not likely to pose an unreasonable risk. But their superiors often push assessors to find chemicals safe even when they lack the data they need to know that. That was the case with a chemical that was granted conditional approval for manufacture in April 2020 despite the fact that both Gallagher and Phillips said a two-year carcinogenicity study was necessary to assess its cancer-causing potential.

Managers also reassigned a chemical that Gallagher was assessing after she made it clear that she believed it could be dangerous. Gallagher was concerned that the chemical, which is slated to be used as a spray coating for cars, could have respiratory effects on workers who use it. But a supervisor insisted that the assessment describe the chemical as “low risk.” In a March 2020 email, Gallagher was instructed to remove all references to hazards. Gallagher refused; hours later, she was told that the chemical, along with all but one of her other cases, had been reassigned to other assessors. Two months later, the chemical was approved for use and found “not likely to present an unreasonable risk.” Two months after that, Gallagher was removed from the program entirely.

Even when assessors do succeed in getting hazards listed in their assessments, the submitting companies often dispute them. In some cases, companies have resubmitted applications for new chemicals more than 10 times. While the EPA does need to communicate with chemical manufacturers, according to the whistleblowers, some managers regularly talk with company representatives by cellphone and have been heard encouraging their industry colleagues to avoid communicating by email because it is subject to public records requests.

While chemical assessors’ original recommendations to new chemical applications used to be publicly available, that policy changed under the Trump administration. Then-EPA Administrator Scott Pruitt removed information about the process of reviewing new chemicals, including records of meetings with manufacturers and concerns staff scientists have or studies they recommend, from the EPA website. Although President Joe Biden and current EPA Administrator Michael Regan have vowed to prioritize science and root out corruption, this policy remains in place under the current administration. And the whistleblowers contend that the pressure to minimize and remove risks in chemical assessments continues unabated.

In an emailed response to questions about this story, the EPA wrote that “This Administration is committed to investigating alleged violations of scientific integrity. It is critical that all EPA decisions are informed by rigorous scientific information and standards. As one of his first acts as Administrator, Administrator Regan issued a memorandum outlining concrete steps to reinforce the agency’s commitment to science.”

The statement also said that EPA leadership is reviewing these complaints and that any appropriate action will be taken. “EPA takes seriously all allegations of violations of scientific integrity. EPA’s scientific integrity official and scientific integrity team members will thoroughly investigate any allegation of violation of EPA’s scientific integrity policy that they receive and work to safeguard EPA science. Additionally, EPA is currently reviewing agency policies, processes, and practices to ensure that the best available science and data inform Agency decisions. EPA is committed to fostering a culture of evaluation and continuous learning that promotes an open exchange of differing scientific and policy positions. Additionally, retaliation against EPA employees for reporting violations alleged to have occurred will not be tolerated in this administration.”

Threats and Intimidation

In June 2020, Osterweil was working on a chemical that she believed could cause severe developmental harms. Her suspicion was based on the molecular structure of a chemical that forms as the body processes the chemical under review — which is associated with effects such as missing bones, clubfoot, and extra fingers. After she submitted her draft report, the manufacturer rebutted her finding. A consultant for the company made repeated calls to EPA about the urgency of changing her report. After one call, the manager of the case called to tell her that the company was considering a lawsuit over her scientific assessment and suggested that they may have to “lawyer up,” which Osterweil took to mean that she might be sued as an individual.

Osterweil, along with Phillips and Gallagher, had already heard vague threats about the wrath of companies angered over their science. In 2019, a manager who was frustrated that they wouldn’t delete a cancer hazard wrote in an email to the three of them that “you all can defend when the companies come in to discuss cancer hazard.”

“Regardless of how good of a scientist you are, they can still ruin your credibility.”

But the threat of being sued as an individual weighed on her. “I was worried that my credibility as a scientist would be marred,” Osterweil said. A single mother, she also feared that she would lose her job. “Regardless of how good of a scientist you are, they can still ruin your credibility.”

The legal threat is just one of the forms of harassment endured by the whistleblowers, who described incidents of shouting, name-calling, and disparagement of their work in front of colleagues.

Irwin described one interaction in which a highly ranked colleague, who also happens to be tall and muscular, was trying to convince him to change the safety threshold of a chemical. “He came to my cube one time, towering over me while I was in my chair,” said Irwin. “He wanted to have a benchmark of two instead of 1,000. And he was shouting at me.” Irwin said he felt threatened and that the intimidation was a key part of the pressure to change scientific findings. “That’s the main mechanism, how they achieve their goal for us to delete hazards — they bully us to do it.”

In the case of the chemical that she believed presented a developmental hazard, Osterweil also faced escalating pressure to use a study done by the manufacturer, which had already been deemed scientifically unacceptable. Although she wasn’t personally sued, by August 2020 the case in which they were told to “lawyer up” was escalated to the level of “hair on fire.” In November, a consultant who worked for the company that submitted the chemical and had previously worked at the EPA arranged a meeting with one of Osterweil’s superiors, who was a personal friend. And on December 21, 2020, when Osterweil was out of the office, an updated assessment of the chemical was uploaded to the agency’s system.

Although she had resolutely refused to change her conclusion about the threat posed by the chemical, the new document dismissed the information Osterweil had inserted about the developmental hazards. Instead, it was deemed unlikely to present an unreasonable risk — making it a “not likely,” in the agency’s parlance. The industry had once again won the battle over a chemical assessment — getting the risks of its product minimized and speeding up the review process. In an email sent by Osterweil’s supervisor that day, she wrote, “Please move this Not Likely expeditiously, as the submitter is anxious.”

In June, after several of the whistleblowers had filed complaints with the EPA’s science adviser and the EPA inspector general, the assessment was again changed, this time to reflect that the chemical “may present an unreasonable risk of injury to health.” It was nevertheless allowed onto the market.