Can government spy a terrorist?

Hooper, Charles L.

Now that Congress has apparently reined in the National Security Agency's spying powers by phasing out the NSA's bulk collection of phone records, those who support the U.S. Constitution and Fourth Amendment are cheering while those who fear another terrorist attack are grumbling.

The two sides will continue battling it out, each saying the other is wrong. However, we can shed some significant light on this issue by borrowing an insightful technique from medicine. By doing this, we will see that constitutional government is compatible with national security. We can have both.

Medicine and national security/ It may seem counterintuitive that medicine can provide insight into national security. But as University of Pittsburgh Medical Center professors Stuart Mendenhall and Mark Schmidhofer demonstrated in these pages a few years ago, medicine and national security deal with similar problems. (See "Screening Tests for Terrorism," Winter 2012-2013.)

Physicians face a quandary. Should they "spy" on a person's body to root out evil diseases? Or does that put the patient at more risk? Physicians aren't constrained by the Constitution, but they are constrained by fiscal prudence and the Hippocratic oath, which prohibits them from harming patients.

We might feel healthy today, but many of us are harboring dangerous diseases such as cancer. Doctors can use a number of diagnostic tests that might uncover such diseases. The question is, which tests should be used and when? Answering that requires a feat of mathematics that takes into account a number of factors:

* What percentage of the population harbors the disease?

* How accurate is the diagnostic test?

* What if the test yields a positive signal and the patient has the disease?

* What if the test yields a positive signal and the patient does not have the disease?

To help resolve this, doctors employ a calculation known as positive predictive value (PPV). It is a measure of the ratio of the true positive and false positive rates, and it equals the proportion of patients who test positive who really do have the disease. A PPV of 10 percent means that one in 10 people who test positive really does have the condition and that nine out of 10 people who test positive don't have the condition. For reference, the PPV for mammography for women 50 and older is 14 percent while the prostate cancer PSA test has a better PPV of 20-50 percent.

PPV and the TSA / Shift your attention from the NSA to that other homeland security agency, the Transportation Security Administration. The TSA has been in the news a lot lately because of dreadfully porous airport security. The two agencies are really peas in a pod: both regularly violate the Fourth Amendment to search innocent Americans with the hope of ultimately protecting those same innocent citizens. Both essentially have the same purpose, yet the TSA is easier to contemplate because of our familiarity with its agents and the straightforward purpose of its actions: the TSA searches our persons and our baggage to prevent weapons from entering commercial aircraft.

We can employ the PPV calculation to evaluate the actions of the TSA. Consider:

* How many Americans fly? There are over 800 million passengers annually on all scheduled flights.

* How many among this flying population are active terrorists? The data show the number is close to zero, but let's assume a much larger number: 100 terrorists and they each fly twice a year.

* How accurate are the TSA's tests? According to ABC News, which acquired a Department of Homeland Security document, "Undercover investigators were able to smuggle mock explosives or banned weapons through checkpoints in 95 percent of trials." So the answer is a measly 4.3 percent accuracy (three out of 70 correct) for flagging weapons when they are present. Let's also assume that if no weapons are present, the TSA's accuracy--not flagging innocent people--is much higher, say, 95 percent.

What is the TSA's PPV? It is a microscopic 0.0000214 percent. Consequently, for every true terrorist that the TSA flags as a potential terrorist, 4.7 million innocent passengers are also flagged as potential terrorists. Not only does this make a mockery of the whole "security" idea, but it doesn't provide much assistance in finding real terrorists, and these false positives cost the TSA time and money. For those of us incorrectly flagged as potential terrorists, it costs us time, personal privacy, potentially missed flights, black marks on our records, the ability to carry liquids and pocketknives, and huge amounts of stress.

More troubling, the real TSA PPV must be far lower than 0.0000214 percent. If our assumption is correct and there are 100 terrorists who take two flights each year, then, given the TSA's poor accuracy in finding weapons, 191 of those terrorists' attempts would have resulted in successfully boarding airplanes. With a 50 percent success rate, we would be seeing over 95 airline hijackings per year, year after year.

Thankfully, the record on foiling real terrorists is much better. The shoe bomber and the underwear bomber have been the only two post-9/11 attempts and neither succeeded, thanks largely to passenger activism; flight crews and motivated passengers are an important bulwark against nefarious in-air activities. If two terrorists succeeded in getting through the TSA's security system and the TSA has a 4 percent accuracy rate, then we can be confident that only two tried. Had three tried, three likely would have gotten through. We can conclude, therefore, that the TSA's actions and all those security lines we have endured over the years have likely prevented zero hijackings.

Having liberty and security/ Is the NSA any different than the TSA? Not likely. Even the FBI admitted that no major terrorist cases were cracked as a result of the Patriot Act's massive snooping powers. Based on the available evidence, both government agencies have spent massive amounts of taxpayer money and searched many innocent people, but have not protected us in the slightest, except perhaps through a deterrent effect. As the PPV calculation shows, mathematics thwarts both government agencies. Their security measures can never be accurate enough to overcome the extremely small prevalence of motivated terrorists.

Even without the NSA's bulk data collection and the TSA's ubiquitous airport security lines, the government has many tools available to apprehend terrorists when there is probable cause. Wholesale spying is not the answer. We should trim these two government agencies and realize that the choice is not between the U.S. Constitution and security; we can enjoy the blessings of both.

CHARLES L. HOOPER is president of Objective Insights, a firm that consults for pharmaceutical and biotech companies. He coauthored Making Great Decisions in Business and Life, Chicago Park Press, 2006.