Mathematical models and national security

Following an anthrax attack, who should receive the first antibiotics? If food aid is limited should 2,000 children each receive a little or should the aid be distributed to a few?

Tackling questions such as these has earned Lawrence Wein the nickname Dr Doom. The professor of management science at Stanford University’s Graduate School of Business teaches courses on operations, mathematical analysis and inventory and queueing theory. Such theories form the backbone of business operations for companies like McDonald's, so that it knows how many cashiers and people flipping burgers are needed to serve the lunchtime rush at any given franchise. But the main focus of his research applies these theories to questions of public health and national security.

How can an operations model be applied to public health?

One problem is blood transfusions. Typically the distribution is first in, first out, so patients are given the oldest blood that has been sitting on the shelf. This is what is called a queueing model, or a waiting in line model. But researchers suggest that for certain patients, such as those having cardiac surgery, the younger the blood that is transfused the better the outcome. So we looked at giving people the youngest blood first and asked if bad things would happen in the supply chain, like having to throw away blood when it gets too old, or running out of blood. We found it does not wreak much havoc on the supply chain. And we can predict that the one-year death rate after the transfusions could go down by about 100,000 people in the US if they switch systems.

After 9/11 you began to focus on catastrophic homeland security problems. What was your solution to distributing treatment in the event of an anthrax attack?

Anthrax is not a contagious disease but it can spread widely and is deadly. We used a mathematical model that assumed two pounds of anthrax were released from the window of a building in New York City. We looked at the winds and figured who would inhale how many anthrax spores and their probability of developing sickness.

We recommended that antibiotics be distributed through the postal service in different zones of New York City. You call a snow day – where people are asked to stay at home – then police-escorted postal workers go round giving antibiotics. Finally President Obama signed a directive for the 72 largest cities in the US. If there is a large-scale anthrax attack, bring in the postal workers.

What have you looked at in your research into food aid in Africa?

There is this stuff called Plumpy’Nut, a peanut butter paste that has really revolutionised treatment of severe acute malnutrition. But there’s only one company that owns the patent to it and there’s not enough of it. So if you have a scarce amount of food to give, which kids do you give it to? It’s a painful question to ask.

Current policy focuses essentially on how skinny you are. Those kids get the food. We looked at the data and found you shouldn’t base it just on how skinny kids are, you should base it on a combination of how skinny and how short they are.

The second insight was dosing. Suppose we have enough food for the full dose for 1,000 kids, do we give it out that way or do we scrimp it out to 2,000 kids at half the dose? We found you’re better off giving the full dose to a smaller number of kids rather than try to spread this out.

People in the field like this paper because no one’s really brought math to bear on this problem area. It is a data-poor field.

Your research on botulinum toxin was very controversial. The US government asked you not to publish your research on botulinum toxin. Why?

They said it was a road map to terrorism. Botulinum toxin is the deadliest substance known to man. It takes about a millionth of a gram to kill somebody. The proposal is that someone takes 10 grams of botulinum toxin and releases it somewhere upstage in a dairy supply chain. Tank trucks go from farms to dairy processing facility and dump milk into 50,000 gallon silos. Through the supply chain about 125,000 gallons of milk get contaminated and roughly 200,000 people die from this.

We made three recommendations. One is invest in prevention, lock the trucks and lock the milk tank. Another is heat pasteurisation, which actually inactivates much of the toxin without sacrificing nutrition and taste. And third was to develop a rapid highly specific test. The bottom line, the government actually intensified the pasteurisation formula for milk.

Academics complain there is insufficient crossover between the academic world and government. What have you done to see your research turned into policy?

It helps getting into a good journal. And I spend time writing op-eds for newspapers. If it goes to a wide readership, it can put pressure on the government to do something. I also go to Washington and talk to people in the Department of Homeland Security.

I’m at the age where I’m tenured and I don’t have to worry. I can afford to do that and I’m at the stage in my career that it is important to try.