CDC Offers Weak Numbers on West Nile Spraying
Wait. Take five. That's all I'm asking. As we approach mosquito season in Dallas, the conversation about aerial spraying for West Nile once again begins to sound like free beer day at the ballpark.
Last week The Dallas Morning News published a story with a triumphal tone saying, "Aerial spraying of insecticide last summer was effective in reducing severe illness from the West Nile virus in North Texas counties" according to "a federal report released Tuesday."
But knowledgeable people are reading the same report and coming to very different conclusions. The report was released last week by the U.S. Centers for Disease Control and Prevention, an arm of the Department of Health and Human Services.
First: Why do we care? We care because the risks may be enormous. Science is coming to a finding that the kind of spray disseminated over Dallas last summer brings with it serious health hazards. Earlier this year Dr. David Bellinger of the Harvard School of Public Health and Boston Children's Hospital spoke to a Dallas teleconference on West Nile to warn Dallas officials about the substances sprayed here.
Bellinger said pesticides can have stealth effects on brain development in young children that may not show up for decades. He said these effects, "do not necessarily bring a child to medical attention, but can nevertheless limit their futures by causing intellectual deficits and abnormal behavior."
Last year I spoke with Sheldon Krimsky, a professor of urban and environmental policy planning at Tufts University in Massachusetts and author of several books dealing with public policy and environmental risk. He told me the most dangerous elements in what was sprayed here last year are not the active mosquito-killing chemicals but additives designed to shut down a mosquito's endocrine system.
The danger is that these substances do the same thing to our own endocrine systems, at least temporarily, and it's the endocrine system that minutely controls brain development in early life. For that reason, any decision to spray these chemicals from airplanes on residential areas must be carefully balanced against the risk. He said, "It's not a free ride in other words."
The CDC report released last week found a "measurable impact in preventing West Nile neuroinvasive disease." But the question asked by the critics is how measurable.
Neuroinvasive is the kind of West Nile that causes paralysis and often death. Non-neuroinvasive is another type of the disease that causes flu-like symptoms but not paralysis, lasting for a day to several weeks.
Some of the experts I spoke with last week said that on close examination the only measured impact they could find in the report was so slight as to seem almost like statistical sophistry. Gene Helmick-Richardson, a Ph.D. entomologist and pest control consultant in Dallas who has been a skeptic of the spray campaign here, said he did some work to dig out hard numbers on effectiveness in terms of human infections, something the report did not explicitly provide.
By drawing from one table in the report, Helmick-Richardson says he found that "you had a reduction of two tenths of 1 percent per hundred thousand [population], which means you might have five cases of neuroinvasive disease that were prevented by spraying three million dollars worth of pesticides."
Even this scant finding of a positive effect on human health is in doubt, some experts say, because the CDC tossed out one entire type of West Nile disease, the non-neuroinvasive form, from its computations. Because some numbers in the report seem to show the non-neuroinvasive form may have been more resistant to spraying than the other kind, critics told me they think tossing it out raises the possibility that including it would have wiped out the spray campaign's overall very narrow margin of success.
I took that question to Susan Hills, one of the authors of the report. She said by email: "We did not conduct an analysis that included neuroinvasive and non-neuroinvasive disease cases so I can't tell you what the results would be."
Dr. Laurence Boyd, a Dallas entomologist who has expressed skepticism about the spray campaign in the past, told me that the CDC's stated reason for tossing out non-neuroinvasive cases doesn't add up.
The report cites as its reason a "possible increase in the diagnosis of non-neuroinvasive disease in the treated areas after aerial spraying." It's a simple idea. Attendant on the spray campaign is a whole raft of publicity. More doctors are looking for West Nile, so more cases are found instead of being misdiagnosed as cold or flu.
But Boyd said that doesn't fly as a reason for not including non-neuroinvasive cases. The report is based on a comparison of last year's West Nile infection rates in sprayed areas with rates in areas that were not sprayed. Boyd points out the effect of publicity would have been the same for sprayed and non-sprayed areas.
"The publicity in the areas that were sprayed is mostly television," he said. "The television stations also reach the areas that were not sprayed."
So if TV stories increased the vigilance of doctors in the sprayed areas, it would have increased their vigilance by the same amount in non-sprayed areas. The comparison would still have been valid. So why was non-neuroinvasive tossed out?
In the treatment of non-neuroinvasive disease in the report, Boyd sees troubling evidence of a "circularity" he thinks is evident in another important part of the report dealing with mosquito populations. This one is a major sticking point for critics of the report. Helmick-Richardson told me he was "appalled" it was not even mentioned in the Morning News story by Sherry Jacobson.
"Did she even read the report?" he wondered.
Because the spray campaign here was carried out at a time when West Nile infection rates were already plummeting as part of the natural cycle of the disease, all of the areas covered by the report, sprayed and non-sprayed, showed dramatic declines in infection. That made a tricky statistical business of attributing degrees of decline to the spray.
But if the spray was effectively killing mosquitoes, it should not have been difficult to measure declines in mosquito population. Indeed, right after the spray campaign here last year CDC experts endorsed a preliminary report claiming a 93 percent kill rate for mosquitoes.
The shocker in this new report was that after a closer look the CDC had to back away from the 93 percent kill-rate claim. Far far away. In fact the final report found that mosquito populations in sprayed areas actually increased slightly during the spraying.
The CDC report seeks to minimize this fact with a kind of off-the-cuff theory that the increase wasn't a threat to anyone. This is where Boyd finds his second "circularity."
The report states there were "small but statistically significant increases in mosquito abundance." But it suggests that new mosquitoes were not a problem, because they hadn't had time to bite any infected people or birds and become carriers.
But later the report states that blood-fed mature mosquitoes, the ones that could be carriers, are the ones most likely to escape the effect of the spray. Why, Boyd wonders, would that not make the blood-fed disease carriers more numerous than the newly emergent mosquitoes in the rising overall population of mosquitoes?
"It's a circular argument," he says. "Very, very contradictory."
But why would the CDC have any agenda here other than science? If the critics are right and they've got a thumb on the scale, what would be their motivation?
Some experts, even those far from us, have taken special note of the early appearance and ongoing role in Dallas last year of the CDC as aggressive champions of aerial spraying. David Pimentel, an emeritus professor of entomology at Cornell University and an expert on mosquito control, told me he was surprised to learn the CDC was in Dallas pushing aerial spraying because in the past it has always thrown its weight behind what he called safer and more effective methods of mosquito control.
Pimentel, who was on the staff of the CDC at one point in his career, said that before Dallas, the CDC always has recommended that communities hand-spray places where mosquitoes breed more than relying on airplanes. "They recommended in their publications anyway that you go after the larvae and pupae, and I agree with them that this is an effective means because you don't have to worry about the wind and so on. It's safer and more effective."
In looking for evidence of a biased advocacy for spraying in this new report, both Boyd and Helmick-Richardson point to another aspect of the report not even mentioned in the Morning News story — the margin of error. According to the report, the CDC actually is not sure whether infection rates decreased six times faster in sprayed areas than non-sprayed or decreased at the same rate. If it's the same rate, then the spraying had no effect at all on infection rates.
To be sure of this, I asked the authors of the report if the lower parameter in their margin of error meant infection rates in sprayed areas and in non-sprayed areas may have been the same.
They wrote back that I was correct that the low end of the "confidence interval" in their report "would indicate that the results were similar in treated and non-treated areas."
But that's not actually true. The number in question is a factor of one. Multiplying by one does not produce results that are "similar." It produces results that are the same.
Again, the stakes are high in this. Your daily newspaper, for whatever reason, has told you virtually nothing about the risks. All I am saying is that we all need to demand better answers to tougher questions before we do this again.
We all come to these emergencies with the best of intentions, but few of us leave our egos at home. This is serious. It ain't just baseball.
Get the This Week's Top Stories Newsletter
Every week we collect the latest news, music and arts stories — along with film and food reviews and the best things to do this week — so that you'll never miss Observer's biggest stories.