Saturday, August 29, 2009

It Is Time for the CDC to Stop the Spin, Retract Its 2004 Report, and Apologize to DC

Remember back in April when Salon alleged that the Centers for Disease Control and Prevention (CDC) covered up harm from DC's 2001-2004 lead-in-water crisis through a highly influential, but also scientifically questionable, report? This report, coauthored by the DC Department of Health (DOH), claimed no significant public health impact from two and a half years of excessively high levels of lead at the tap, even among young children. The reassuring conclusion contradicted numerous prior research studies about the health effects of lead in drinking water as well as common sense.

Remember also that only hours after the online appearance of the Salon article, CDC issued a media statement characterizing Salon's claims as "inaccurate" and offering supposed rebuttals to them (for more information, see our 5/3/09 blog entry)?

We now have confirmation that, indeed, the CDC misled the public both in 2004 -- when it reported that the worst lead-in-water contamination in modern US history "might" have contributed only to a "small increase in blood lead levels," and in 2009 -- when it assured us publicly that Salon's criticisms of its 2004 no-significant-harm conclusion were unfounded. Shockingly, it looks like the CDC was actually aware of gross deficiencies in its data and analysis prior to the publication of its 2004 report, but failed to take simple steps to correct them.

The new revelations, covered in the Washington Post and the Public Radio International show Living on Earth and discussed in Congressional letters to the US Department of Health and Human Services and Mayor Fenty, came from the US House Subcommittee on Investigations and Oversight (of the Committee on Science and Technology). This Subcommittee has been looking closely at CDC's role in DC's 2001-2004 lead-in-water crisis (for more information, see our 3/22/09 blog entry). The Subcommittee's preliminary investigation, which sheds light on the data behind the CDC's 2004 report, supports Salon's allegations.

Here's some background:

To assess the public health impact of DC's 2001-2004 lead-in-water contamination, in early 2004 the CDC compared the percentage of DC residents (children and adults) with elevated blood lead levels (EBLLs) (i.e., equal to or above 10 micrograms per deciliter for children and 25 micrograms per deciliter for adults) for the years 1998-2003. They found that this percentage kept decreasing, despite the elevation of lead in the city's tap water, and concluded that if any harm had been done from the water crisis it was "small" and involved blood lead levels under the CDC's "level of concern" (i.e., 10 micrograms per deciliter).

But the CDC's dataset for children under 6 -- the age group most vulnerable to the effects of lead -- had an enormous hole. For 2003 -- the third full year of the city's lead-in-water contamination, when the public was still unaware of the problem and not taking systematic precautions to prevent exposure to high concentrations of lead at the tap -- the dataset included a significantly smaller number of blood lead test results as compared to the previous three years. The numbers of children tested per year as listed by the CDC are:

2000: 14,040 [175 children with EBLLs, or 1.2%]
2001: 16,042 [156 children with EBLLs, or 0.9%]
2002: 15,755 [122 children with EBLLs, or 0.8%]
2003: 9,229 [193 children with EBLLs, or 2%]

Since there was no apparent reason for a dramatic drop in blood lead testing in 2003, it looked like the CDC's database -- which came from the DOH -- was missing approximately 5,000 children's test results from a crucially important time period. By 2003, according to a peer-reviewed scientific paper in the journal Environmental Science & Technology (Edwards et al. 2009), some DC children were into their second or third year of exposure to high levels of lead in water. A thorough examination of blood lead levels from that time could have revealed health impacts that may not have been detectable at shorter exposure times and offered valuable insight into the true impact of DC's two and a half year lead-in-water contamination for the most vulnerable age group. If the missing test results included blood lead concentrations above 10 micrograms per deciliter -- which Salon alleged was the case -- their omission would have skewed the CDC's analysis by concealing harm.

In April 2009, the CDC admitted to Salon the absence of thousands of 2003 blood lead results from its study. However, it defended its no-significant-harm conclusion by claiming that all the missing results were below 10 micrograms per deciliter, and thus the missing data introduced no bias in the agency's analysis (on the contrary, the CDC suggested that the absence of these results probably exaggerated the percentage of children with EBLLs).

The CDC based its unsubstantiated assertion on information it allegedly obtained from DOH. The CDC's official story is that during the writing of its report, it asked DOH about the thousands of missing blood lead tests for 2003, and was told that the missing data had resulted from the failure of one commercial laboratory to report blood lead levels below 10 micrograms per deciliter (an omission that laboratories are not supposed to make).

To date there is no evidence that the CDC took any steps to confirm DOH's explanation before publishing its 2004 report or issuing its 2009 misguided media statement in response to the Salon article.

The CDC's lackadaisical approach to the missing data is especially troubling in the face of the following facts:
  1. The CDC was aware of serious problems with DOH's management of blood lead test results prior to the publication of its 2004 report. According to the Subcommittee, this knowledge "should have set off warning bells that CDC could not rely on the numbers being provided [by DOH] for public health statements."
  2. The DOH had known about the excessive levels of lead in DC's water since 2002, but had failed to take measures to protect the public until the Washington Post broke the story in 2004. Because of this history, the conclusion of the 2004 report -- which was supposed to determine whether the two-and-a-half year contamination had harmed DC residents -- would either incriminate or exonerate the DOH. By definition then, the DOH had walked into its partnership with the CDC with a clear conflict of interest, which may have compromised the integrity of the data it provided for the 2004 report.
Following the issuance of the CDC's 2009 media statement, WASAwatch posted a series of questions about the CDC's supposed rebuttal to Salon's allegations. Thanks to the US House Subcommittee on Investigations and Oversight, we now have answers to some of these questions:

Contrary to the CDC's claims, it looks like there was no one negligent lab that failed to report 5,000 blood lead levels under 10 micrograms per deciliter for 2003. After obtaining from the CDC the name of this alleged lab, the Subcommittee contacted it and learned that it "had continuously reported only elevated BLLs from 1999 until April of 2004," so there was no sudden change in its reporting practice for 2003.

Additionally, the Subcommittee did precisely the kind of data-checking that the CDC should have done before publishing its 2004 report. It requested all the 2003 data from all District labs that analyzed blood lead levels for DC children under 6. The Subcommittee found that the labs had reported to DOH several thousand more test results (i.e., at least 13,758 rather than 9,229) and several hundred more EBLL cases (i.e., at least 486 rather than 193) than the CDC included in its analysis. The Subcommittee's examination of this more comprehensive dataset revealed that at least 3.5% of DC children who were tested in 2003 had EBLLs, not 2%, as the CDC claims to this day. Shockingly, some of these children had blood lead concentrations more than 600% as high as the CDC's "level of concern."

The CDC, it thus seems, based its 2004 no-significant-harm conclusion on an analysis that omitted a majority of the children with EBLLs in 2003. In other words, the nation's premier public health agency allowed itself to compromise the integrity of its report and the soundness of its health message by embracing DOH's "negligent lab" theory unquestioningly. Moreover, it disclosed none of the problems with the missing data and DOH's conflict of interest in its report.

In its 8/3/09 letter to the US Department of Health and Human Services, the House Subcommittee stated that "CDC's inability or unwillingness to validate and verify the data it was being provided by [DOH] raises serious questions about the ability of the CDC's lead program to ensure the integrity of the data provided to it for other years [i.e., in addition to 2003] by DC as well as from other CDC cooperating states and cities."

More grave flaws in the CDC's 2004 no-significant-harm report that we outlined in our 3/22/09 blog entry raise additional questions about the ability of the CDC's lead program to ensure the trustworthiness of the public health information it provides to the public.

So what do the Subcommittee's new findings mean for DC?
  • During the third year of the city's unprecedented lead-in-water contamination, it is likely that more than twice as many cases of EBLLs among children under 6 were identified than the CDC and DOH claim. There is a distinct possibility that at least some of the unacknowledged cases did not receive necessary interventions to eliminate the lead exposure source(s) in the child's home and offer treatment when warranted.
  • In light of the fact that the number of children who were tested in 2003 (to the best of our knowledge, about 13,758) was far smaller than the number of children who lived in DC at the time (approximately 39,356), it is likely that many untested children also experienced EBLLs, but their condition went undetected.
  • The CDC's claim that there was no significant rise in EBLLs from DC's lead-in-water contamination is utterly unsubstantiated and, in the context of a 2009 peer-reviewed scientific study and an unpublished 2007 CDC analysis that did find harm, most likely false.
  • Had the CDC's 2004 analysis not excluded thousands of crucial blood lead test results, it may have revealed significant harm from DC's lead-in-water contamination. In that case, the CDC's 2004 report would have presented a very different conclusion than the one published, and DC may have taken measures it still has not taken to educate the public about the hazards of lead at the tap and protect children from future exposures. Other cities in the US and internationally that have used the CDC's report to downplay the health risks of lead-contaminated water may have also responded to their own lead-in-water problems in a more appropriate and health-protective way.
The CDC is supposedly the nation's premier public health agency whose statements have great influence over personal behaviors and public policy. Yet its 2004 report about DC's historic lead-in-water contamination reveals a severely botched response to a serious public health crisis that reflects gross governmental irresponsibility and negligence. Not only were DC residents exposed to two and a half years of astronomical levels of lead at the tap but, thanks to the CDC, we were also misled into believing that our children survived the contamination essentially unscathed and that elevated levels of lead in water pose practically no health risk.

Even worse, the CDC continues its no-significant-harm spin to this day, despite the fact that it has been unable to refute any of the serious allegations against the integrity of its report or say why a study using such a severely compromised dataset should be considered credible at all. In response to the 8/7/09 Public Radio International story, the CDC lamely re-asserted that its 2004 analysis identified no children "with BLL 10 ug/dL, even in homes with the highest water lead levels."

This statement is outright inaccurate because for every year between 1998 and 2003, the CDC did identify children with EBLLs -- it just didn't identify a dramatic rise in EBLL cases in 2003. Second, this statement fails to address the problem of the missing blood lead test results, which was the sole focus of the Public Radio International piece. Third, this statement does nothing to refute the apparent fact that the CDC's 4/10/09 assertions have now been shown to be as flawed as the agency's original report. Fourth, this statement does not disclose the disturbing detail that the CDC looked only at 17 children who resided in homes with the highest water lead levels, and these children had stopped drinking unfiltered tap water months to a year before their blood was tested for lead (for more information, see our 3/22/09 blog entry). Because blood lead levels tend to drop quickly after exposure is stopped, the chance of finding high lead in the blood of these children from high lead in water was minuscule to non-existent.

It is deeply worrisome that five years after the publication of its misleading report the CDC is still disseminating deceptive information about the health impact of lead-contaminated water, while at the same time refusing to answer simple questions from the public about its report's validity (for more information on CDC's stonewalling, see our 2/25/09 blog entry). It is equally disturbing that the CDC has failed to release all the information requested by the US House Subcommittee on Investigations and Oversight (see p. 5 of the Subcommittee's 8/3/09 letter to the US Department of Health and Human Services).

Congressional hearings about the CDC's involvement in DC's 2001-2004 lead-in-water crisis are likely to take place in September.

In the meantime, the CDC's Lead Poisoning Prevention branch would be well-advised to show a modicum of respect for sound science and public health by stopping its destructive spin, retracting its 2004 report, and apologizing to DC for downplaying a serious -- and in some cases ongoing -- public health risk.