When it comes to lead in drinking water, the CDC first failed DC's children and then children in communities all across the country and the world.
What went wrong and how can the CDC begin to fix it?
It starts with the influential CDC report, published in March 2004, and titled "Blood Lead Levels in Residents of Homes with Elevated Lead in Tap Water -- District of Columbia, 2004."
In February 2004, two weeks after the Washington Post informed District residents that for two and a half years the drinking water in thousands of DC homes had been contaminated with astronomical concentrations of lead, the CDC teamed up with the DC Department of Health (DOH) purportedly to assess health effects from the contamination. DOH was about to be exposed for knowing about the lead problem since 2002 and helping WASA cover up its serious nature. In less than six weeks, the CDC published a study that contradicted decades of prior scientific knowledge about child retention of lead from lead-contaminated drinking water (see also our 2/25/09 and 3/22/09 blog entries).
The CDC's conclusion was that long-term exposure to astronomical concentrations of lead in water "might have contributed to a small increase in blood lead levels (BLLs)" (emphasis added). No children, according to the CDC report, were identified with BLLs equal to or above the CDC's "level of concern" (10 micrograms per deciliter) whose exposure was attributable to the water, even in homes with the highest lead-at-the-tap concentrations (i.e., greater than 300 parts per billion).
CDC's supposedly authoritative report -- which was written, reviewed, edited, and published in less than 12 days -- utterly contradicted the EPA's predictive models of lead absorption by infants and young children from water. Aware that lead concentrations in the District's drinking water were exceeded only at Superfund sites, EPA calculated that a high percentage of DC infants would experience lead poisoning from the tap. Despite its highly dubious conclusion, the new CDC report was cited by scientists, public health experts, and communities with lead-in-water problems across the United States (and internationally) as evidence that high lead at the tap does not pose a significant health threat, even to the most vulnerable age groups.
Subsequently, Washington state, New York, and Montreal have all downplayed local problems with lead in drinking water by prominently mentioning the CDC's conclusions based on the agency's "study" in Washington DC. Other communities, like Los Angeles, responded to parent concerns about contaminated school water with egregious misstatements that could have been easily refuted if the CDC had ever delivered clear and unambiguous information about the dangers from elevated lead levels at the tap (see video below).
In DC, the CDC report brought a collective sigh of relief. Officials from EPA, DOH, and WASA claimed that the city's two-and-a-half year failure to properly notify residents about the unprecedented lead-in-water levels had no significant public health impact. The city's childhood lead poisoning prevention community openly labeled drinking water a "minor" source of lead and a "distraction" from other, "more important" sources. For four years, DOH left out water testing during environmental risk assessments at the homes of children with elevated BLLs. DC public schools failed to publicize elevated levels of lead in fountains and coolers. And just last year, DC's Mayor and key members of the DC City Council insisted on removing from a supposedly "comprehensive" lead poisoning prevention ordinance several sensible provisions aimed at addressing lead in drinking water over the strong objections of local and national lead poisoning prevention advocacy organizations.
All the while, the CDC sat back and watched. CDC officials participated in regular meetings with WASA and EPA, monitored and guided the lead program at the DOH, attended the task force meetings of the District's lead advocacy community, but they never made an effort to change DC's interpretation of the CDC report or to put the lead-in-water issue back on the city's radar screen.
Yet when crucial facts about the CDC report began to surface, uncovering major scientific shortcomings and misleading conclusions, the CDC's reaction revealed institutional resistance to transparency and accountability, blatant disregard for scientific integrity, and a complete inability to provide clear, cogent, and consistent answers to simple questions from the public. In response to evidence that the CDC report was being used widely to downplay the risks of lead in water, the CDC denied responsibility for the "no significant harm" claim and instead blamed readers for "misinterpreting" its findings.
In 2005, Virginia Tech Professor of Civil and Environmental Engineering and 2007 MacArthur Fellow Marc Edwards began an investigation into the CDC report and raised multiple and serious questions about the data behind it. For over three years Dr. Edwards has attempted to obtain answers from the CDC, with little success.
The highlights below from Dr. Edwards' interactions with the CDC illustrate the federal agency's unresponsiveness to legitimate questions about the validity and application of its report.
As background, the CDC report comprises two studies: the "citywide population study," which found no significant increase in elevated BLLs among DC residents in 2001-2004, and the "300-parts-per-billion study," which found that none of the 201 residents in homes with water lead levels above 300 parts per billion experienced elevated BLLs.
While researching the 300-parts-per-billion study, Dr. Edwards discovered that there was an unreported delay of months to a year between the time residents were informed that their water was contaminated with lead (and took measures to protect themselves from exposure) and the time their blood was tested for lead. Because the half-life of lead in blood is 28-36 days, BLLs can drop relatively quickly when the exposure source is eliminated. For this and other reasons, Dr. Edwards concluded that the blood lead samples for the CDC study were drawn from a subset of the DC population that had very low likelihood of elevated BLLs from drinking water. Thus, contrary to the public presentation of that data as "worst case," the CDC study had by design almost no chance of detecting any link between excessively high lead-in-water levels and elevated lead in blood. In July 2006, when confronted by reporters, the CDC admitted the undisclosed water/blood sampling gap, but to date it has failed to add a clear disclosure about this fatal limitation in its report.
Troubled by the haste with which the CDC report was written, and aware that the DOH had featured the report prominently two days after its publication in sworn testimony at a City Council hearing about the agency's own role in the 2001-2004 crisis, Dr. Edwards sent a letter to the CDC asking whether the preparation and presentation of the report followed the journal's written publication guidelines (regarding clearance, review, and production). He followed up with dozens of e-mails and phone calls, as well as a Freedom of Information Act request. To date, the CDC has not answered Dr. Edwards' questions.
Dr. Edwards sent a second letter to the CDC requesting copies of all e-mails between the CDC and DOH concerning the CDC report. Four months later, with no e-mails in hand, Dr. Edwards wrote back to the CDC stating,
"By refusing to fulfill my [Freedom of Information Act] request and refusing to answer the simplest questions I have posed, CDC is helping to perpetuate flawed (or even fraudulent) science that directly impacts public health. [...] CDC's refusal to produce the documents in a timely manner, failure to produce the data behind this study, and complete lack of response to my earlier questions illustrates a callous disregard for principles that govern scientific inquiry. I am extremely disappointed that CDC is willingly sacrificing public health to hide data behind an obviously flawed study."To date, the CDC has not released the requested e-mails, or answered questions about whether written CDC policies were violated in publishing the 2004 report.
Dr. Edwards sent to the CDC questions related to his discovery that 2 of the 17 children in the 300-parts-per billion study had not consumed tap water for two years prior to collection of their blood. Later, documents obtained from DOH through Freedom of Information Act requests revealed that none of the 17 children tested in the CDC study were drinking unfiltered tap water. To date, the CDC has not answered Dr. Edwards' questions about this issue, nor has it added any disclosure to its report about this important fact.
Dr. Edwards submitted to the CDC a request for all the scientific data behind the 300-parts-per-billion study. To date, the CDC has not sent him this set of data.
After several cases of lead poisoning in North Carolina were tied to lead-in-water levels far below 300 parts per billion, the CDC contended that its 2004 report had never suggested that consuming water with 300 parts per billion lead is safe. However, the agency again did not disclose the errors behind the misleading conclusions of its 300-parts-per-billion study. To date, the CDC has still not explained what message readers ought to draw from this study, when nearly all of the residents who participated were drinking bottled and filtered water for months to a year before having their blood tested.
Contrary to sworn testimony under oath by the CDC study's DOH co-authors, Dr. Edwards discovered that numerous 2004 environmental risk assessments at the homes of DC children with elevated BLLs found drinking water to be the sole identified lead exposure source or a contributing source of lead exposure. In response to this finding, the CDC told WAMU radio that it planned to conduct a follow up study to determine whether its 2004 conclusions were correct. As part of this new study, the CDC stated that it would review DC's 2004 often cited environmental risk assessments. The CDC estimated that a definitive follow up report would be published several months later. To date, the CDC has not published such a report (CDC recently announced that a paper will be published later in 2009).
Directly contradicting its comments from two months earlier, CDC told Salon that it had decided against reviewing DC's 2004 environmental risk assessments.
On 1/11/09, Dr. Edwards made yet another plea to the CDC for the release of information:
"For more than a year now, I have been trying to get simple answers to questions I have about the [CDC report]. [...] Unless I hear from you before next Tuesday at 5:00 pm (January 16), I have decided to document my concerns to the CDC Associate Director of Science. At present, lacking the innocent explanation that I have been trying to find for more than a year now, my report will document suspicions of scientific misconduct [...]. I regret that it has come to this. [...] I have been given a run-around for more than a year while the findings of this study have spread throughout the public health community and caused undeniable harm."After a week of silence, Dr. Edwards submitted to the CDC a 27-page letter (available upon request) alleging possible scientific misconduct -- fabrication and falsification of data, based largely on information he had pieced together from the DOH, parents of young children in DC, newspaper reports, and official statements by government representatives. This letter reiterated very specific and highly credible problems regarding data inconsistencies and contradictions, the absence of thousands of critical BLL data points (including BLLs above the CDC's "level of concern"), biased inclusion and exclusion of residents into the 300-parts-per-billion study, and additional evidence that the sub-population selected for this study was by no means the "worst case," as had been stated repeatedly by authors of the study under oath.
On 1/23/07, the CDC sent Dr. Edwards the data he had requested for the first part of the report (the citywide population study), with critical information deleted that was necessary to reproduce the analysis (as well as conduct new analyses). It informed Dr. Edwards that if he wanted this information, he would have to pay an additional fee of $166.14. Dr. Edwards sent CDC a check for this amount, which was cashed in April 2007. To date, CDC has still not sent Dr. Edwards the data it promised him. As for the 300-parts-per-billion study, the agency stated that it does not have the data for it and referred Dr. Edwards to the Food and Drug Administration (several co-authors on the CDC report were affiliated with the FDA).
Dr. Edwards sent the CDC two Canadian newspaper articles quoting Joe Schwarcz, PhD, a McGill University chemistry professor who cited the CDC study to reassure the public that widespread lead-in-water problems in Montreal were no cause for serious concern. The first article quoted Dr. Schwarcz as saying that DC:
"...got thousands of people to actually give blood and they found that although the water level was sometimes as high as 300 parts per billion, which is astounding, it didn't influence the blood levels."The second article, again citing the CDC report, made light of Montrealers' rush to buy water filters:
"Joe Schwarcz, director of the McGill University Office for Science and Society, noted there must be 'glee in the Brita boardroom' this week. But he pointed out a 2004 study in Washington, DC, on residents who were exposed to lead levels in their drinking water that were 300 times Quebec's norms, showed no signs of lead poisoning in the blood, indicating that poisoning rarely results from drinking water."In his introductory note to the CDC, Dr. Edwards wrote:
"...after reading this story, how many people in Montreal are going to take seriously the threat from lead in the water? How on earth can CDC sit by and do nothing to correct this horrific misconception? KIDS IN SCHOOLS ALL OVER THE US ARE BEING ENDANGERED, AS WE SPEAK, BY THAT CDC [REPORT]. NOT TO MENTION IN CITIES, LIKE MONTREAL, AND ELSEWHERE IN THE US, WHERE AUTHORITIES FEEL JUSTIFIED IN NOT EVEN TELLING PEOPLE ABOUT HIGH LEAD IN WATER. CDC's inaction on this issue over the last 14 months is completely inexcusable. It will take years to even get back to where we were on this issue before the DC blood lead fraud took place."In response to these concerns, the Office of Science at the CDC's National Center for Environmental Health contacted Dr. Edwards to tell him that no one at CDC disputed that the residents in the 300-parts-per-billion study were probably not consuming tap water for months to a year before their blood was drawn. Stating that additional delays in clarifying the 300-parts-per-billion study would be "unconscionable," the Office of Science led Dr. Edwards to believe that the scientific record would be corrected "in no uncertain terms" within one week. The Office of Science also suggested that it could not investigate Dr. Edwards' scientific misconduct allegations, unless Dr. Edwards attributed the allegations to specific CDC employees.
Still without the raw data he had requested about the 300-parts-per-billion study, Dr. Edwards wrote to the Office of Science:
"I hope that you all agree, that when the raw data for the study cannot be found and produced 20 months after I first asked for it, it raises scientific questions that go beyond a 'misunderstanding' that readers took from reading the work. I therefore request that when the 'clarification' to the [CDC report] is published, that it also explicitly state that either 1) CDC cannot find and produce the data used for the [report's] '300 ppb' study, or 2) CDC had nothing to do with the data, analysis or interpretation related to the '300 ppb' study. At least then I can stop people who refer to this horrible excuse for a scientific publication from invoking CDC's reputation to support their flawed conclusions."August 2007
Over four months after stating that CDC would release a clarification about its 300-parts-per-billion study within days, the Office of Science informed Dr. Edwards that a correction had been posted online. At the bottom of CDC's Q&A page on lead in water, CDC wrote the following addendum:
"Following the release of the [CDC report] some reports have suggested erroneously that the Centers for Disease Control and Prevention has determined that lead in residential tap water at concentrations as high as 300 parts per billion is 'safe.' CDC would like to reiterate the key message from the 2004 article that because no threshold for adverse health effects in young children has been demonstrated (no safe blood lead has been identified), all sources of lead exposure for children should be controlled or eliminated. Lead concentrations in drinking water should be below the US Environmental Protection Agency's action level of 15 parts per billion."The addendum, which is not posted on the actual report, offered no explanation about why the public's interpretation of the 300-parts-per-billion study was "erroneous," or what the CDC report actually concluded about the health effects of water with excessively high lead. The addendum also made no disclosure about the time gap between water and blood tests, or the fact that all the children in the sample group were reported as drinking bottled or filtered water. If the "key message" of the CDC report was the obvious truth that all sources of lead must be controlled and that water lead must be below 15 parts per billion -- facts that were known long before 2004 -- then the CDC report offered nothing new, especially about the health impact of astronomical levels of lead in the District's drinking water in 2001-2004.
By placing blame on the public and various officials for "misinterpreting" its report and by diverting attention away from the horrific flaws of its two studies (the citywide population study and the 300-parts-per-billion study), the CDC tried to evade any responsibility for the actual message of its work, which was that, contrary to all reasonable expectations, long-term consumption of water with lead levels even above 300 parts per billion had no significant health impact, even in young children (and that no one in DC showed evidence of harm from the years of high lead in water). CDC and DOH purposefully disseminated that very message in no uncertain terms (see for example, the video clip of DOH's 9/04 testimony to City Council).
Dr. Edwards submitted to the CDC a second letter (available upon request) alleging possible scientific misconduct, this time by specific CDC employees. This letter, a continuation of the one he had submitted nine months before, focused on a) unusual circumstances related to the publication of the CDC report; b) purposeful omission of facts that confound the scientific analysis in the report; and c) CDC's refusal to properly correct the scientific record, even after acknowledging fatal flaws in the study.
The CDC sent Dr. Edwards a letter in response to his first allegations, stating that it had "examined CDC's role in the study and [had] found no evidence of misconduct." The letter offered no refutation or explanation in response to Dr. Edwards' specific allegations. Instead, it referred Dr. Edwards to the DC Office of the Inspector General (OIG). Dr. Edwards told the CDC that he had already contacted OIG, which had referred him to CDC, because the 2004 report was a CDC publication.
A Salon article has just revealed that in 2007 the CDC presented data at the annual meeting of the American Public Health Association (APHA) which showed that many children in Washington DC experienced elevated BLLs from the 2001-2004 lead-in-water crisis. The CDC chose not to publicize these findings or post its PowerPoint presentation on its website (the presentation is available on the APHA website for a fee). Nor did CDC inform EPA, DOH, the DC Department of the Environment, or any of the residents in Washington DC about its finding.
The CDC sent Dr. Edwards a response to his second scientific misconduct letter stating that it had "examined CDC's role in the study and [had] found no evidence of misconduct." CDC offered no explanation for, or refutation of, any of Dr. Edwards' allegations.
When a new study by Edwards et al. showed that in 2001-2004 hundreds, if not thousands, of DC infants and toddlers experienced elevated BLLs, the CDC kept quiet the fact that it had already presented similar findings at the APHA annual meeting in 2007. CDC then sat and watched as WASA questioned the scientific merit of the Edwards et al. study and called for an independent investigation into the obvious disparity between the Edwards' et al. conclusions and those from the 2004 CDC report. CDC never disputed that there was an obvious contradiction between the Edwards et al. study results and their 2004 report. To the press, the CDC again affirmed that it found "no direct link" between the unprecedented lead-in-water contamination and DC children's blood lead levels and no "evidence of a public health crisis."
Today, the CDC continues to stand by its 2004 findings, without producing evidence to rebut Dr. Edwards' specific allegations of fabrication and falsification of data. Moreover, it refuses to straightforwardly acknowledge the "no significant harm" message of its report that was sent and received. Instead, CDC points to vague language in the paper about a possible "small increase in blood lead levels" from DC's lead-contaminated water and to a broad, standard warning that "public health interventions should focus on eliminating all lead exposures in children."
Five years after the CDC report's publication, CDC's position on the health risks from lead at the tap generates more confusion than clarity. Based on a compilation of CDC's rare comments to the press and DC lead poisoning prevention advocates, here's what the CDC is saying:
- The 2004 report does not say that high lead in DC water caused no significant harm, and if anyone makes such a claim their motives ought to be questioned. The report notes concerns about a "possible health impact," but it found no direct link between elevated water lead and elevated blood lead. The health effects from DC's 2001-2004 contamination are likely to be "very slight."
- The 2009 Edwards et al. publication, which found serious and widespread harm from DC's lead-in-water crisis, is essentially in agreement with CDC's 2004 report.
- The 2004 report is not scientific. It is missing thousands of blood lead data points from a critical time period, it fails to focus on the most vulnerable populations, it makes claims about the health effects of highly contaminated water based on blood tests of residents who didn't drink this water, and is it being widely "misinterpreted." (Yet this completely discredited report continues to be displayed on the agency's journal and website, where it continues to be cited as evidence that high levels of lead in drinking water do not have significant health consequences, and where it can mislead anyone who would think that the CDC would never publish a report that was "not scientific").
- The 2007 CDC presentation (which found highly significant proof of serious harm to DC residents from the high lead in water and agrees with conventional scientific knowledge about the health impact of lead at the tap) is not appropriate for public distribution or discussion because it does not meet CDC standards for "best quality science."
One thing we can probably all agree on is that we expect more from the nation's premier public health agency. In the last 5+ years, CDC's actions and public health messages about lead in drinking water have neither improved the health of the children of the United States, nor promoted health education that can help us prevent exposure to hazardous levels of lead at the tap. On the contrary, the CDC's 2004 report has done great disservice to our children. If the CDC does not retract it, it must immediately disclose its severe limitations. Here are our suggestions:
What the CDC must disclose about its 2004 report
Apart from acknowledging publicly, clearly, and fully all the flaws and limitations in its citywide population study and 300-parts-per-billion study as described by Edwards et al., the CDC must disclose:
I. Conflicts of interest
- Eight of the CDC report co-authors were employees of the DOH, an agency that had direct knowledge of the 2001-2004 lead-in-water crisis, but had taken little action to address it. Some DOH co-authors coordinated the 2004 environmental risk assessments at the homes of children with elevated BLLs, several of which showed drinking water to be the sole or a contributing source of lead, but they misrepresented these findings under oath. DOH was the source of all the data for the CDC report, and thousands of blood lead test results from 2003 have now been discovered to be missing. In July 2004, DOH was sued (along with WASA) for personal injury of a DC child with severe lead poisoning, and the lawsuit alleges that the child's exposure was at least partly from high lead in drinking water. Two DOH employees were fired for their role in the lead-in-water crisis.
- Lead service lines are not the only source of lead in drinking water. Lead solder and leaded brass are important sources as well, and the majority of homes in DC have them. In 2001-2004, contrary to WASA's frequent misstatements, over 15% of District homes with no known lead service line had lead-contaminated water. The CDC report's focus on lead service lines must not confuse readers into thinking that residents in homes without a lead service line are protected from lead-in-water problems. Indeed, a reasonable analysis indicates that in 2001-2004 more DC children were lead poisoned from high lead in water in apartments and homes without lead service lines (as in the case of a child where the likely lead source was contaminated water at Wilkinson Elementary School, which has no lead service lines), than in homes with lead service lines. In North Carolina lead poisoning from water occurred even though the city has no lead water lines.
- The report's characterization of lead paint and dust hazards as "high-dose lead sources" and the implication that lead-contaminated water is a "low-dose lead source" is inaccurate. Some lead paint, manufactured prior to about 1940, can contain 10-50% lead by weight, but lead pipes are 100% lead by weight. Lead solder, commonly used in homes prior to 1986, contains 50% lead by weight. According to Dr. Edwards, metallic lead is much more dense than lead compounds in paint, so a particle from lead solder or partly replaced lead service line will always have much more lead than an equivalent sized lead paint particle. Lead solder alone has been linked to childhood lead poisonings in North Carolina and Massachusetts.
- what happened in DC in 2001-2004 specifically, and
- the possible importance of drinking water as a potentially significant source of lead generally.
Video clip: 10/30/08 NBC Los Angeles segment about highly misleading risk communication messaging to concerned parents at an LA school with high levels of lead in water.