When Colorado environmental regulators received a federal report that warned of potential cancer risks in communities near Denver in August 2018, they alerted the public and the media and vowed to find out what residents were breathing. Two days after the information became public, the state started air testing for ethylene oxide, the cancer-causing compound used by a medical sterilization company that the federal Environmental Protection Agency identified as the culprit of the possible health threat. Within a month, Colorado officials had persuaded the company to enact new pollution controls, and a second round of air tests in October, after the upgrades, showed a dramatic reduction in concentrations of the toxic gas. By October, the Illinois attorney general had sued a suburban Chicago company to curb emissions, and a month later Michigan started testing the air around one of its ethylene oxide emitters. When confronted with the EPA’s National Air Toxics Assessment (NATA) at exactly the same time as Colorado, Illinois and Michigan, Georgia regulators decided not to publicize the EPA’s report and rejected as unreliable the idea of air testing, according to an Atlanta Journal-Constitution review of more than 22,000 emails and other documents. The Georgia Environmental Protection Division also kept the NATA report from then-Gov. Nathan Deal and incoming Gov. Brian Kemp, according to interviews with state officials, despite state law that says the agency has a duty to counsel the governor on environmental concerns affecting the state. By the time Colorado had convinced its medical sterilization plant to enact pollution controls, records show, Georgia was still three months away from even meeting with the operators of two metro Atlanta sterilizers that were also emitting potentially dangerous levels of the cancer-causing gas according to the EPA’s 2018 assessment. After site visits at the plants at the end of 2018, Georgia EPD waited for months for the two plants to produce data so that the state could conduct its own analysis using mathematical modeling, which Georgia only completed in June. The AJC’s review of Georgia’s response to the ethylene oxide crisis also found that state regulators appeared to miss the big picture — that the federal EPA now considered the risk of ethylene oxide exposure to be many times greater than before and that the public deserved to know. The agency’s narrow thinking and slow response contributed to a public panic in July when online health care publications broke the news about the potential health threat, the AJC found. Once the risk became public, Georgia reversed course and committed to air testing around the plants. Gov. Kemp quickly secured pledges of enhanced pollution controls at both of the ethylene oxide emitters in metro Atlanta — Sterigenics near Smyrna in Cobb County and Becton Dickinson (BD) in Covington. But those actions came a year after those taken in Colorado and Illinois, and eight months after proactive measures in Michigan, all states that EPA flagged for ethylene oxide risks. “I feel betrayed,” said Smyrna homeowner Jenni Shover, who took part in a protest this summer holding up a sign that read, “Hell hath no fury like a woman poisoned.” “They made a conscious decision not to tell the public,” Shover said. “There is no trust with them.” At town halls in Marietta and Covington in August, Georgia regulators presented new modeling that showed risks from ethylene oxide, or EtO, were less acute, confined to the immediate vicinity of the plants and not in surrounding neighborhoods. They also announced that both plants would invest millions to upgrade their emissions controls. But many residents aren’t buying it. “We need independent testing to get the facts,” Jason McCarthy, who leads a Covington-area chapter of the group called Stop EtO, said at one town hall. “We need independent testing to repair the trust that was broken.” Georgia EPD Director Richard Dunn and other officials stand by their handling of the ethylene oxide controversy. They said they thought from the outset that the 2018 report from EPA grossly overstated ethylene oxide dangers, a view shared by many EtO emitters, and required a deeper ground-level assessment. “The driver of our thinking was this is a screening tool, and we need further study,” Dunn said. “That was the sole driver.” Kemp spokeswoman Candice Broce said EPD will be more forthcoming with the public in the future. The NATA report has been online since 2018, but with EPD being “a very scientific branch of government,” they didn’t recognize a need for public relations, she said. “It wasn’t that they didn’t take any action, it’s that they didn’t publicize it in a way people would have expected,” Broce said. “And they’ve changed that practice. They’re moving forward, trying to be as clear as possible with press releases and public hearings, to be transparent and get tough questions from the community.” Failure to communicate The state had plenty of time to craft a regulatory and communications response, the AJC found. In summer 2017, Georgia regulators got an early peek at the EPA’s findings, which suggested two Atlanta-area plants might be emitting harmful levels of ethylene oxide, an invisible, odorless gas that had been reclassified as a carcinogen in the final days of the Obama administration. Federal officials told states not to share its “draft results” with anyone outside of their agencies. A Georgia modeling and data manager at the time reviewed the preliminary findings and sent an email to an EPA air quality official, saying she didn’t understand “strange hot spots” in Newton and Cobb counties. The EPA official responded that the risk factor for ethylene oxide “went up by about a factor of 60.” The exchange triggered some internal discussion within EPD, but no deeper investigation. In the final NATA report, the areas of risk shifted slightly. But neighborhoods around the Sterigenics and BD plants remained on the list of ethylene oxide hot spots. EPA distributed talking points to state agencies, including Georgia EPD, but opted against issuing a press release. Karen Hays, the director of the Georgia EPD’s air protection branch, said she did not recall agency leaders having any discussions about informing the public. “As far as communication, trying to communicate something that is this complicated is really challenging for us because there is a lot of information that is quite frankly contradictory,” she said. “To take what’s in the NATA and spread it out to the public without doing a detailed evaluation of what’s really happening, to me, is just not appropriate.” Colorado: ‘reassure the public’ Colorado, Illinois and Michigan had no such qualms about releasing information, the AJC found. Colorado regulators issued a press release the same day the federal report was made public, Aug. 22, 2018, linking to EPA’s website and describing the state’s plans for dealing with the Denver-area plant, Terumo BCT. “Measuring actual exposures, rather than computer-modeled exposure as in the EPA analysis, will provide a more accurate picture of whether there is a potential health risk,” the press release from the Colorado Department of Public Health and Environment said. The announcement said that even though Terumo BCT was in compliance with state and federal pollution control requirements, and even though the state’s own cancer registry showed no elevation in cases near the plant, the first round of air sampling would be done within a month. Testing started two days later — lasting seven days with samples taken from four sites around the plant. The results backed up what the federal report’s modeling showed — air concentrations that well exceeded the EPA’s cancer danger threshold. “We knew that once (NATA) went out, people were going to see it, and we would get questions on it,” said Gordon Pierce, a program manager in the Colorado’s air pollution control division. “So if we didn’t have answers or have work ongoing to try and get answers, we would be called to task on it.” Pierce said his department could have relied on computer modeling alone to evaluate the cancer risks, the approach taken in Georgia, but that might not have satisfied people living nearby. Colorado tested the air again for seven days in October 2018, after Terumo BCT installed new pollution controls. This time samples were taken at eight points, including as far as a mile away. The tests found a 2- to 5-fold reduction in cancer risk, although it remained above the EPA’s acceptable range for ethylene oxide. Pierce said he’s waiting for guidance from EPA on what to do next. “It was to reassure the public,” Pierce said. “It was also to make sure that what we were being told by the company matched.” After an open house to explain the testing results to residents, publicity subsided and the state had the issue behind them by Christmas. Excluding staff time, the air sampling cost Colorado $23,000, according to a department spokeswoman. Illinois, Michigan also act In the Chicago suburb of Willowbrook, federal scientists were so alarmed by emissions modeling at a Sterigenics facility that the agency conducted its own air sampling, months before the NATA report’s publication. EPA also called in a specialized division of the Atlanta-based Centers for Disease Control and Prevention to assess the results. Based on preliminary modeling and air testing, the CDC specialists said ethylene oxide in the air posed “a public health hazard to these populations.” The NATA report and the Willowbrook air tests were released together in August 2018, triggering organized protests and a series of angry public meetings. By October, the state attorney general and DuPage County sued Sterigenics to curb emissions. After the Village of Willowbrook conducted its own air tests in February, which showed high levels of the gas, the state temporarily closed the plant until it could meet strict new standards. Last month, Sterigenics announced the plant will not reopen. “We’re going to be done up here by the time you guys get started (air testing),” Willowbrook Mayor Frank Trilla told the AJC last month. “By the time you get done, they’re going to have shutters on their doors.” Georgia’s air quality regulators watched the Willowbrook controversy unfold, swapping news stories about heated public meetings and sharing the federal air testing report, emails show. That still didn’t spur anyone to craft a communications strategy. The records show EPD officials taking a skeptical view of Illinois’ enforcement efforts. A risk assessment program manager downplayed the air testing, writing in an email that “my concern … is the assumption that a casual relationship can be easily drawn between chronic exposure to EtO air emissions and elevated cancers in the population surrounding a facility under routine monitoring.” Observing the Willowbrook controversy from across Lake Michigan, however, the Michigan Department of Environment, Great Lakes and Energy wanted no such firestorm on its watch. Three months after NATA posted, Michigan conducted air tests around its most problematic plant, Viant Medical in Grand Rapids. The neighborhood had already been on the department’s radar because of a possible cancer cluster believed to be linked to an old city dump. But with outrage building in Willowbrook, Michigan went forward with a press release in January, announcing Michigan had issued Viant Medical a violation notice for failing to control its ethylene oxide emissions, according to Chris Ethridge, a field operations manager for the Michigan’s air quality division. “We had measurements that were of concern, so we felt compelled to notify the community,” Ethridge said. “I think it’s kind of a matter of ethics.” A second round of air testing in March found emissions levels still above state standards. The company has announced that it will halt its sterilizing operations at the plant by the end of the year, characterizing it as a business decision. Under a proposed consent order, the company faces a $110,000 fine for its emissions. Air testing dismissed EPD defended computer modeling as being more scientifically sound than air testing, but experts told the AJC modeling should be buttressed by testing data around emission sources. Modeling analyzes companies’ self-reported emissions, building and smokestack dimensions, and years of meteorological data to estimate where gases go and plot areas of potential exposure. Monitoring samples the air to determine what compounds are present and in what amounts. There are drawbacks: it can take months and may not pinpoint the exact sources of a chemical. But air testing can be used to check whether the modeling and self-reported company data match the physical conditions in the air. “It’s good to verify your model with observations,” said Jennifer Kaiser, a Georgia Tech assistant professor in environmental engineering and atmospheric sciences. “Testing is sort of ground truth.” The federal EPA says air testing has discovered ethylene oxide in higher concentrations than expected and in places where scientists hadn’t forecasted. Industry groups contend air sampling fails to take into account background levels of ethylene oxide produced by vehicle emissions and other natural sources. They also say the EPA’s risk level for the compound is too strict, lower in fact than what is naturally made within the human body. Georgia EPD echoed some of those arguments, saying in an interview that early test results in Illinois were invalidated because other chemicals collected by sampling canisters skewed the data. Dika Kuoh, Georgia’s assistant air protection branch chief, said the science of ethylene oxide air testing hadn’t been perfected when EPA published the NATA report in August 2018. Federal regulators released air testing guidance only a few months ago. “Modeling is always the first tool that’s used in these situations because it is the one that gives you the quickest result,” Kuoh said. “You go monitoring, you are just measuring the air, and you could be measuring all kinds of things that are not the facility or related to the facility.” Public outrage when the health risks became known, however, forced EPD to rethink this position two months ago, agreeing to conduct long-term testing near both plants. “I will say this, from my perspective, the monitoring is due to demands from the community,” EPD Director Dunn said. “It’s less so from a scientific perspective.” Shover, who protested Sterigenics in August, has been living near the plant for nearly two decades. She says Georgia already squandered its chance to find out what the two plants were pumping out before they knew they were under scrutiny. With every ache, pain and itch, she said she worries about cancer. “We will never have that data — never,” Shover said. “And that is unforgivable.” Staff Writer Meris Lutz contributed to this story.