The biodefender that cries wolf
DENVER — As Chris Lindley drove to work that morning in August 2008, a call set his heart pounding.
The Democratic National Convention was being held in Denver, and Barack Obama was to accept his party’s presidential nomination before a crowd of 80,000 people that night.
The phone call was from one of Lindley’s colleagues at Colorado’s emergency preparedness agency. The deadly bacterium that causes tularemia — long feared as a possible biological weapon — had been detected at the convention site.
Should they order an evacuation, the state officials wondered? Send inspectors in moon suits? Distribute antibiotics? Delay or move Obama’s speech?
Another question loomed: Could they trust the source of the alert, a billion-dollar government system for detecting biological attacks known as BioWatch?
Six tense hours later, Lindley and his colleagues had reached a verdict: false alarm.
BioWatch had failed — again.
President George W. Bush announced the system’s deployment in his 2003 State of the Union address, saying it would “protect our people and our homeland.” Since then, BioWatch air samplers have been installed inconspicuously at street level and atop buildings in cities across the country — ready, in theory, to detect pathogens that cause anthrax, tularemia, smallpox, plague and other deadly diseases.
But the system has not lived up to its billing. It has repeatedly cried wolf, producing dozens of false alarms in Los Angeles, Detroit, St. Louis, Phoenix, San Diego, the San Francisco Bay Area and elsewhere, a Los Angeles Times investigation found.
Worse, BioWatch cannot be counted on to detect a real attack, according to confidential government test results and computer modeling.
The false alarms have threatened to disrupt not only the 2008 Democratic convention, but also the 2004 and 2008 Super Bowls and the 2006 National League baseball playoffs. In 2005, a false alarm in Washington prompted officials to consider closing the National Mall.
Federal agencies documented 56 BioWatch false alarms — most of them never disclosed to the public — through 2008. More followed.
The ultimate verdict on BioWatch is that state and local health officials have shown no confidence in it. Not once have they ordered evacuations or distributed emergency medicines in response to a positive reading.
Federal officials have not established the cause of the false alarms, but scientists familiar with BioWatch say they appear to stem from its inability to distinguish between dangerous pathogens and closely related but nonlethal germs.
BioWatch has yet to face an actual biological attack. Field tests and computer modeling, however, suggest it would have difficulty detecting one.
In an attack by terrorists or a rogue state, disease organisms could well be widely dispersed, at concentrations too low to trigger BioWatch but high enough to infect thousands of people, according to scientists with knowledge of the test data who spoke on condition of anonymity.
Even in a massive release, air currents would scatter the germs in unpredictable ways. Huge numbers of air samplers would have to be deployed to reliably detect an attack in a given area, the scientists said.
Many who have worked with BioWatch — from the Army general who oversaw its initial deployment to state and local health officials who have seen its repeated failures up close — call it ill-conceived or unworkable.
“I can’t find anyone in my peer group who believes in BioWatch,” said Dr. Ned Calonge, chief medical officer for the Colorado Department of Public Health and Environment from 2002 to 2010.
“The only times it goes off, it’s wrong. I just think it’s a colossal waste of money. It’s a stupid program.”
Officials at the Centers for Disease Control and Prevention, the federal agency that would be chiefly responsible for rushing medications to the site of an attack, told White House aides at a meeting Nov. 21 that they would not do so unless a BioWatch warning was confirmed by follow-up sampling and analysis, several attendees said in interviews.
Those extra steps would undercut BioWatch’s rationale: to enable swift treatment of those exposed.
Federal officials also have shelved long-standing plans to expand the system to the nation’s airports for fear that false alarms could trigger evacuations of terminals, grounding of flights and needless panic.
BioWatch was developed by U.S. national laboratories and government contractors and is overseen by the Department of Homeland Security. Department officials insist that the system’s many alerts were not false alarms. Each time, BioWatch accurately detected some organism in the environment, even if it was not the result of an attack and posed no threat to the public, officials said.
At the same time, department officials have assured Congress that newer technology will make BioWatch more reliable and cheaper to operate.
The current samplers are vacuum-powered collection devices, about the size of an office printer, that pull air through filters that trap any airborne materials. In more than 30 cities each day, technicians collect the filters and deliver them to state or local health labs for genetic analysis. Lab personnel look for DNA matches with at least half a dozen targeted pathogens.
The new, larger units would be automated labs in a box. Samples could be analyzed far more quickly and with no need for manual collection.
Buying and operating the new technology, known as Generation 3, would cost about $3.1 billion over the next five years, on top of the roughly $1 billion that BioWatch already has cost taxpayers. The Obama administration is weighing whether to award a multiyear contract.
Generation 3 “is imperative to saving thousands of lives,” Dr. Alexander Garza, Homeland Security’s chief medical officer, told a House subcommittee on March 29.
But field and lab tests of automated units have raised doubts about their effectiveness. A prototype installed in the New York subway system in 2007 and 2008 produced multiple false readings, according to interviews with scientists. Field tests last year in Chicago found that a second prototype could not operate independently for more than a week at a time.
Most worrisome, testing at the Pacific Northwest National Laboratory in Washington state and at the Army’s Dugway Proving Ground in Utah found that Generation 3 units could detect a biological agent only if exposed to extremely high concentrations: hundreds of thousands of organisms per cubic meter of air over a six-hour period.
Most of the pathogens targeted by BioWatch, scientists said, can cause sickness or death at much lower levels.
A confidential Homeland Security analysis prepared in January said these “failures were so significant” that the department had proposed that Northrop Grumman Corp., the leading competitor for the Generation 3 contract, make “major engineering modifications.”
A spokesman for the department, Peter Boogaard, defended the performance of BioWatch. Responding to written questions, he said the department “takes all precautions necessary to minimize the occurrence of both false positive and false negative results.”
“Rigorous testing and evaluation” will guide the department’s decisions about whether to buy the Generation 3 technology, he said.
Representatives of Northrop Grumman said in interviews that some test results had prompted efforts to improve the automated units’ sensitivity and overall performance.
“We had an issue that affected the consistency of the performance of the system,” said Dave Tilles, the company’s project director. “We resolved it. We fixed it.... We feel like we’re ready for the next phase of the program.”
In congressional testimony, officials responsible for BioWatch in both the Bush and Obama administrations have made only fleeting references to the system’s documented failures.
“BioWatch, as you know, has been an enormous success story,” Jay M. Cohen, a Homeland Security undersecretary, told a House subcommittee in 2007.
In June 2009, Homeland Security’s then-chief medical officer, Dr. Jon Krohmer, told a House panel: “Without these detectors, the nation has no ability to detect biological attacks until individuals start to show clinical symptoms.” Without BioWatch, “needless deaths” could result, he said.
Garza, the current chief medical officer, was asked during his March 29 testimony whether Generation 3 was on track. “My professional opinion is, it’s right where it needs to be,” he said.
After hearing such assurances, bipartisan majorities of Congress have unfailingly supported additional spending for BioWatch.
The problems inherent in what would become BioWatch appeared early.
In February 2002, scientists and technicians from Lawrence Livermore National Laboratory deployed a prototype in and around Salt Lake City in preparation for the Winter Olympics. The scientists were aware that false alarms could “cause immense disruptions and panic” and were determined to prevent them, they later wrote in the lab’s quarterly magazine.
Sixteen air samplers were positioned at Olympic venues, as well as in downtown Salt Lake City and at the airport. About 5:30 p.m. on Feb. 12, a sample from the airport’s C concourse tested positive for anthrax.
Utah Gov. Mike Leavitt was at an Olympic figure skating competition when the state’s public safety director, Bob Flowers, called with the news.
“He told me that they had a positive lead on anthrax at the airport,” Leavitt recalled. “I asked if they’d retested it. He said they had — not just once, but four times. And each time it tested positive.”
The Olympics marked the first major international gathering since the Sept. 11, 2001, airliner hijackings and the deadly anthrax mailings that fall.
“It didn’t take a lot of imagination to say, ‘This could be the real thing,’” Leavitt said.
But sealing off the airport would disrupt the Olympics. And “the federal government would have stopped transportation all over the country,” as it had after Sept. 11, Leavitt said.
Leavitt ordered hazardous-materials crews to stand by at the airport, though without lights and sirens or conspicuous protective gear.
“He was ready to close the airport and call the National Guard,” recalled Richard Meyer, then a federal scientist assisting with the detection technology at the Olympics.
After consulting Meyer and other officials, Leavitt decided to wait until a final round of testing was completed. By 9 p.m., when the results were negative, the governor decided not to order any further response.
“It was a false positive,” Leavitt said. “But it was a live-fire exercise, I’ll tell you that.”
The implication — that BioWatch could deliver a highly disruptive false alarm — went unheeded.
After the Olympics, Meyer and others who had worked with the air samplers attended meetings at the Pentagon, where Deputy Defense Secretary Paul D. Wolfowitz was building a case for rapidly deploying the technology nationwide.
On Jan. 28, 2003, Bush unveiled BioWatch in his State of the Union address, calling it “the nation’s first early-warning network of sensors to detect biological attack.”
The next month, a group of science and technology advisors to the Defense Department, including Sidney Drell, the noted Stanford University physicist, expressed surprise that “no formal study has been undertaken” of the Salt Lake City incident. The cause of that false alarm has never been identified.
“It is not realistic to undertake a nationwide, blanket deployment of biosensors,” the advisory panel, named the JASON group, concluded.
The warning was ignored in the rush to deploy BioWatch. Administration officials also disbanded a separate working group of prominent scientists with expertise in the pathogens.
That group, established by the Pentagon, had been working to determine how often certain germs appear in nature, members of the panel said in interviews. The answer would be key to avoiding false alarms. The idea was to establish a baseline to distinguish between the natural presence of disease organisms and an attack.
The failure to conduct that work has hobbled the system ever since, particularly in regard to tularemia, which has been involved in nearly all of BioWatch’s false alarms.
The bacterium that causes tularemia, or rabbit fever, got its formal name, Francisella tularensis, after being found in squirrels in the early 20th century in Central California’s Tulare County. About 200 naturally occurring infections in humans are reported every year in the U.S. The disease can be deadly but is readily curable when treated promptly with antibiotics.
Before BioWatch, scientists knew that the tularemia bacterium existed in soil and water. What the scientists who designed BioWatch did not know — because the fieldwork wasn’t done — was that nature is rife with close cousins to it.
The false alarms for tularemia appear to have been triggered by those nonlethal cousins, according to scientists with knowledge of the system.
That BioWatch is sensitive enough to register repeated false alarms but not sensitive enough to reliably detect an attack may seem contradictory. But the two tasks involve different challenges.
Any detection system is likely to encounter naturally occurring organisms like the tularemia bacterium and its cousins. Those encounters have the potential to trigger alerts unless the system can distinguish between benign organisms and harmful ones.
Detecting an attack requires a system that is not only discriminating but also highly sensitive — to guarantee that it won’t miss traces of deadly germs that might have been dispersed over a large area.
BioWatch is neither discriminating enough for the one task nor sensitive enough for the other.
The system’s inherent flaws and the missing scientific work did not slow its deployment. After Bush’s speech, the White House assigned Army Maj. Gen. Stephen Reeves, whose office was responsible for developing defenses against chemical and biological attacks, to get BioWatch up and running.
Over the previous year, Reeves had overseen placement of units similar to the BioWatch samplers throughout the Washington area, including the Pentagon, where several false alarms for anthrax and plague later occurred.
Based on that work and computer modeling of the technology’s capabilities, Reeves did not see how BioWatch could reliably detect attacks smaller than, for example, a mass-volume spraying from a crop duster.
Nevertheless, the priority was to carry out Bush’s directive, swiftly.
“In the senior-level discussions, the issue of efficacy really wasn’t on the table,” recalled Reeves, who has since retired from the Army. “It was get it done, tell the president we did good, tell the nation that they’re protected.… I thought at the time this was good PR, to calm the nation down. But an effective system? Not a chance.”
Why no illness?
It wasn’t long before there was a false alarm. Over a three-day period in October 2003, three BioWatch units detected the tularemia bacterium in Houston.
Public health officials were puzzled: The region’s hospitals were not reporting anyone sick with the disease.
Dr. Mary desVignes-Kendrick, the city’s health director, wanted to question hospital officials in detail to make sure early symptoms of tularemia were not being missed or masked by a flu outbreak. But to desVignes-Kendrick’s dismay, Homeland Security officials told her not to tell the doctors and nurses what she was looking for.
“We were hampered by how much we could share on this quote-unquote secret initiative,” she said.
After a week, it was clear that the BioWatch alarm was false.
In early 2004, on the eve of the Super Bowl in Houston, BioWatch once again signaled tularemia, desVignes-Kendrick said. The sample was from a location two blocks from Reliant Stadium, where the game was to be played Feb. 1.
DesVignes-Kendrick was skeptical but she and other officials again checked with hospitals before dismissing the warning as another false alarm. The football game was played without interruption.
Nonetheless, three weeks later, Charles E. McQueary, then Homeland Security’s undersecretary for science and technology, told a House subcommittee that BioWatch was performing flawlessly.
“I am very pleased with the manner in which BioWatch has worked,” he said. “We’ve had well over half a million samples that have been taken by those sensors. We have yet to have our first false alarm.”
Asked in an interview about that statement, McQueary said his denial of any false alarm was based on his belief that the tularemia bacterium had been detected in Houston, albeit not from an attack.
“You can’t tell the machine, ‘I only want you to detect the one that comes from a terrorist,’” he said.
Whether the Houston alarms involved actual tularemia has never been determined, but researchers later reported the presence of benign relatives of the pathogen in the metropolitan area.
Fear in the capital
In late September 2005, nearly two years after the first cluster of false alarms in Houston, analysis of filters from BioWatch units on and near the National Mall in Washington indicated the presence of tularemia. Tens of thousands of people had visited the Mall that weekend for a book festival and a protest against the Iraq War. Anyone who had been infected would need antibiotics promptly.
For days, officials from the White House and Homeland Security and other federal agencies privately discussed whether to assume the signal was another false alarm and do nothing, or quarantine the Mall and urge those who had been there to get checked for tularemia.
As they waited for further tests, federal officials decided not to alert local healthcare providers to be on the lookout for symptoms, for fear of creating a panic. Homeland Security officials now say findings from lab analysis of the filters did not meet BioWatch standards for declaring an alert.
Six days after the first results, however, CDC scientists broke ranks and began alerting hospitals and clinics. That was little help to visitors who already had left town, however.
“There were 100 people on one conference call — scientists from all over, public health officials — trying to sort out what it meant,” recalled Dr. Gregg Pane, director of Washington’s health department at the time.
Discussing the incident soon thereafter, Jeffrey Stiefel, then chief BioWatch administrator for Homeland Security, said agency officials were keenly aware that false alarms could damage the system’s credibility.
“If I tell a city that they’ve got a biological event, and it’s not a biological event, you no longer trust that system, and the system is useless,” Stiefel said on videotape at a biodefense seminar at the National Institutes of Health on Oct. 6, 2005. “It has to have a high reliability.”
Ultimately, no one turned up sick with tularemia.
Culture of silence
Homeland Security officials have said little publicly about the false positives. And, citing national security and the classification of information, they have insisted that their local counterparts remain mum as well.
Dr. Jonathan Fielding, Los Angeles County’s public health director, whose department has presided over several BioWatch false positives, referred questions to Homeland Security officials.
Dr. Takashi Wada, health officer for Pasadena from 2003 to 2010, was guarded in discussing the BioWatch false positive that occurred on his watch. Wada confirmed that the detection was made, in February 2007, but would not say where in the 23-square-mile city.
“We’ve been told not to discuss it,” he said in an interview.
Dr. Karen Relucio, medical director for the San Mateo County Health Department, acknowledged there was a false positive there in 2008, but declined to elaborate. “I’m not sure it’s OK for me to talk about that,” said Relucio, who referred further questions to officials in Washington.
In Arizona, officials kept quiet when BioWatch air samplers detected the anthrax pathogen at Super Bowl XLII in February 2008.
Nothing had turned up when technicians checked the enclosed University of Phoenix Stadium before kickoff. But airborne material collected during the first half of the game tested positive for anthrax, said Lt. Col. Jack W. Beasley Jr., chief of the Arizona National Guard’s weapons of mass destruction unit.
The Guard rushed some of the genetic material to the state’s central BioWatch lab in Phoenix for further testing. Federal and state officials convened a 2 a.m. conference call, only to be told that it was another false alarm.
Although it never made the news, the incident “caused quite a stir,” Beasley said.
The director of the state lab, Victor Waddell, said he had been instructed by Homeland Security officials not to discuss the test results. “That’s considered national security,” he said.
The dreaded call
In the months before the 2008 Democratic National Convention, local, state and federal officials planned for a worst-case event in Denver, including a biological attack.
Shortly before 9 a.m. on Aug. 28, the convention’s final day, that frightening scenario seemed to have come true. That’s when Chris Lindley, of the Colorado health department, got the phone call from a colleague, saying BioWatch had detected the tularemia pathogen at the convention site.
Lindley, an epidemiologist who had led a team of Army preventive-medicine specialists in Iraq, had faced crises, but nothing like a bioterrorism attack. Within minutes, chief medical officer Ned Calonge arrived.
Calonge had little faith in BioWatch. A couple of years earlier, the health department had been turned upside down responding to what turned out to be a false alarm for Brucella, a bacterium that primarily affects cattle, on Denver’s western outskirts.
“The idea behind BioWatch — that you could put out these ambient air filters and they would provide you with the information to save people exposed to a biological attack — it’s a concept that you could only put together in theory,” Calonge said in an interview. “It’s a poorly conceived strategy for doing early detection that is inherently going to pick up false positives.”
Lindley and his team arranged a conference call with scores of officials, including representatives from Homeland Security, the Environmental Protection Agency, the Department of Health and Human Services, the Secret Service and the White House.
None of the BioWatch samplers operated by the state had registered a positive, and no unusual cases of infection appeared to have been diagnosed at area hospitals, Lindley said.
The alert had come from a Secret Service-installed sampler on the grounds of the arena where the convention was taking place. The unit was next to an area filled with satellite trucks broadcasting live news reports on the Democratic gathering. Soon, thousands of conventioneers would be walking from Pepsi Center to nearby Invesco Field to hear Obama’s acceptance speech.
Had Lindley and Calonge been asked, they said in interviews, they wouldn’t have put the BioWatch unit at this spot, where foot and vehicle traffic could stir up dust and contaminants that might set off a false alarm. As it turned out, a shade tree 12 yards from the sampler had attracted squirrels, potential carriers of tularemia.
The location near the media trailers posed another problem: how to conduct additional tests without setting off a panic.
EPA officials “said on the phone, ‘We have a team standing by, ready to go,’” Lindley recalled. But the technicians would have to wear elaborate protective gear.
The sight of emergency responders in moon suits “would have derailed the convention,” Calonge said.
On the other hand, sending personnel in street clothes would risk exposing them to the pathogen.
“This was the biggest decision we ever had to make,” Lindley said.
When the conference call resumed, Lindley said the state would collect its own samples, without using conspicuous safety gear. “No one was willing to say, ‘That’s the right response, Colorado,’” Lindley recalled. “Everybody was frozen. We were on our own.”
State workers discreetly gathered samples of soil, water and other items for immediate DNA analysis. No pathogen was found.
At 3 p.m., Lindley told participants in another national conference call that his agency was satisfied there was no threat. “I said: ‘We are doing no more sampling. We are closing up this issue,’” Lindley recalled.
Lindley and Calonge, having staked their reputations on not believing BioWatch, were vindicated: Barack Obama gave his acceptance speech on schedule. No one turned up sick with tularemia. And, to their surprise, news of the false alarm never became public.
Officials responsible for BioWatch insist that the false alarms, which they refer to as “BioWatch actionable results,” or BARs, have been beneficial.
Each incident “has provided local, state and federal government personnel an opportunity to exercise its preparedness plans and coordination activities,” three senior Homeland Security BioWatch administrators told a House subcommittee in a statement in July 2008. “These real-world events have been a catalyst for collaboration.”
Biologist David M. Engelthaler, who led responses to several BioWatch false positives while serving as Arizona’s bioterrorism coordinator, is one of the many public health officials who see it differently.
“A Homeland Security or national security pipe dream,” he said, “became our nightmare.”