How will the technologies of the future help protect us against terrorism? The new book Safe: The Race To Protect Ourselves in a Newly Dangerous World examines innovative techniques for sniffing out attacks before they happen and for limiting damage if a strike does occur. In today's excerpt, the second of a three-part series, Martha Baer, Katrina Heron, Oliver Morton, and Evan Ratliff look at a computer chip that could be the best weapon against bioterrorism. Yesterday's excerpt explained how to recognize potential terrorists with facial heat sensors and automated video cameras. Tomorrow's selection considers a new technology that allows the government to root through citizens' private data without behaving like a police state.
In October 2001, Sen. Tom Daschle's office received an envelope containing a few grams of anthrax. The extremely dangerous particles could drift in the air with ease and, once settled, loft themselves back up at the gentlest disturbance. There may have been enough anthrax in the envelope to deliver 100,000 lethal doses; spread through the building it could have killed hundreds of senators, staff members, and building staff. But it didn't, for one simple reason. Along with the fine white powder in the envelope was a letter that revealed it was anthrax.
The problem is what to do when the attacker is not so forthcoming. The stand-alone biological-agent detectors available today are able to distinguish possibly suspicious biological particles in the air with the use of antibody tests, but these tests are far from reliable enough for routine use. The best they can do is send a signal for someone to come and take a sample to a lab for full analysis if there is something in it that looks suspicious. But lab results can take hours. There may be some value in having these current sensors around a military base during a war, or around vital government buildings as a routine matter (and indeed such machines are already found in some sensitive parts of Washington). But they're a very long way from a system that can spot a biological attack whenever and wherever it takes place.
On Star Trek episodes in the 1960s, the crew of the Enterprisealready had handheld detectors called tricorders that identified any sign of life. But there are two essential barriers to developing a tricorder. One is that there is a bewildering range of living things in the environment to sense. A detector that just picks out anthrax or smallpox is not enough. You need one that can pick out anything threatening, which means you need a device that can run a wide range of tests in parallel. The other problem is sorting out bad bugs from normal ones. The E. coli bacteria that live harmlessly in our guts, for example, and the vicious 0157 strain of the same bug that can kill in a matter of days are vastly different at the level of their genes but almost identical from the outside.
At the moment, there's really only one feasible way to put together a sensor network for detecting biological attacks: assign the job not to technologies but to people. Today, as throughout the history of public health, most disease outbreaks are spotted when a clinician recognizes something unusual or out of place; the anthrax attacks were recognized because a doctor in Florida had recently been trained in recognizing the symptoms. The trouble is that clinicians in any given hospital or clinic won't see the big picture, just the patients they're treating.
A study carried out at Johns Hopkins made this point vividly. Researchers there considered a scenario in which an anthrax attack in Baltimore took place during a professional football game, infecting 16,000 fans without being detected. The head of infectious diseases at the medical school asked the staff how quickly an influx of affected patients might be noticed at that time of year. The physician in charge asserted that patients turning up with the symptoms of anthrax would be diagnosed as having flu and sent home. The ER's lead lab technician said he had not seen Bacillus anthracis isolated from a patient in all his 25 years on the job. If it turned up in a blood test, staff would consider it meaningless contamination caused by one of anthrax's benign relatives. Only if it turned up in three different tests would the staff conduct further investigations, and those tests would take 48 hours. Forty-eight hours after the first blood tests, a large number of the infected would be dead or dying. This scenario's final death toll was about 4,000.
None of this should be seen as a critique of Johns Hopkins. Most people with flulike symptoms in November really do have the flu or something similar and almost all Bacillus species in blood cultures are contaminants. George Poste, chair of the Defense Science Board's bioterrorism task force, explains this in terms of an old medical school adage designed to warn students away from recondite but unlikely diagnoses: "When you hear hoof beats it could be a zebra—but it's probably horses."
The problem is that in a bioweapon attack the hoof beats will indeed be zebras—quite possibly vast trampling herds of them. This is why Poste is pressing for new diagnostic technologies. Although we do not yet have the miniaturized versions of canaries, food-tasting mice, and Kuwaiti Fighting Chickens that will one day harness pure biology to detect and identify pathogens, we can combine elements of biology with electronic information processing. A combination of biotechnology and semiconductor manufacture techniques makes it possible to put hundreds of thousands of probes capable of recognizing DNA sequences and protein structures onto a single silicon chip.
Poste suggests that chips be made that would recognize the protein and gene signatures of 50 or so of the most plausible bioweapons. Such chips could then be used to test blood from patients on a routine basis. All those hypothetical fans feeling flu-y after an attack at the football game would be tested with these chips upon showing up at the hospital. Most of the time the tests would show nothing unusual, because the steed that had brought the patient to the clinic would be a horse. Occasionally, though, these novel blood tests would turn up a zebra—which is why Poste calls the technology the Z-chip.
The idea would be to make Z-chip systems easy to use and, by buying in bulk, very cheap. The system would use already developed techniques for handling small samples of fluids to get the blood to the relevant portions of the chip automatically. The patterns formed on the chip would be quickly interpreted—and at the same time as they were delivered back to the clinic, they'd also be put into a national system looking for greater patterns.
The Pentagon takes Poste seriously and is allocating millions of dollars to Z-chip research. In some ways, the Z-chip looks like the sort of biodetector you might want plastered all over towns, cities, and subway stations—something that can pick up the patterns associated with all sorts of pathogens. But it would never be able to work in such environments. The Z-chip needs the pathogens it detects to be processed by the human body before it can get to work. The human body puts a lot of effort into making sure that its bloodstream is very stable and clean, its acidity and temperature tightly controlled. It is a nicely noise-free environment in which to look for a signal.
Later versions of the technology, Poste imagines, could leverage their intimacy with the body's biology further still. Z-chips that looked at the genes expressed in white blood cells could listen in on the body's own surveillance system, a far more subtle and perceptive judge of what's going on than human technology can yet manage. The technology might end up as something that simply makes the information that you already have inside you accessible to your eye and mind. Making the way things work more visible, letting the whole system know itself so that it can talk to itself, is the high road to resilience.