Thursday, March 23, 2006

Homework time again

Among the many vexing issues we will face if a pandemic materializes is what to do about schools -- daycare, elementary, highschools, colleges. Previous data from Boston's Children's Hospital indicated school age children might be the drivers of seasonal influenza epidemics. Their syndromic surveillance system saw the outbreak signal in that age group about a month ahead of the general population. Other data had suggested that closing schools might have ameliorated flu risks in communities where it had been tried. But now computer modelers have entered the fray with a prediction that school closing wouldn't work after all.

The paper has only been given orally at a meeting so I have only seen news reports, but it sounds like what is called agent-based modeling where individual computer "agents" are given attributes of the general population (age, sex, how many contacts they have a day and with whom, how far they travel, where they work or go to school, etc.) and other relationships governing the rate at which the disease is transmitted is put on top of that. Then scenarios are run many times and some idea of the probability of various outcomes is determined. It takes a lot of computing power and it is hard to know whether it produces results that are at all accurate. It still depends on a lot of assumptions. But it is a useful technique to get a feel for the range of things that might happen and to detect unusual or counter-intuitive possibilities.

This week the result of one of these computer simulations was presented at the International Conference on Emerging Infectious Diseases in Atlanta. Here is what the Emory University researchers found:
Using statistics and computer models, U.S. researchers simulated an influenza outbreak in a small urban community of several thousand people, where people made contact in a variety of places where disease could be transmitted, including schools, homes, day care centers, work places and long-term care facilities.

The results suggested that closing schools might simply send children to other places where they could encounter a virus.

"When we assume school closing, it doesn't mean the children are sent home. They will meet at the movies or other places in the community," said Michael Haber, the study's author and a professor in the Emory University Department of Biostatistics. "Children who are not ill may become sick outside the school."

Much more likely to help slow the spread of bird flu and other viruses is home confinement of anyone who is ill or exposed. The study found that under certain circumstances, infection rates could be reduced up to 52 percent and death rates up to 60 percent by home confinement. (Reuters)
This is not a counterintuitive result. It says that if you close the schools you better be sure the kids don't just congregate elsewhere, especially mixing with a random assortment of adults not related to them and it points up the extremely difficult nature of the school problem. If "social distancing" (a euphemism for artificially making it difficult for people to be near each other) is going to be a prominent tactic during a pandemic, school closings will be one application. Federal and state officials are already discussing plans for this. And they should. But schools serve purposes besides education: childcare for working parents, nutrition via school lunches for poor children, ways to disseminate information to parents and much more. If you close them you also have to make provisions for the consequences, as the modeling suggests. Closing wouldn't be for a day or a week, but probably for weeks or a month or more. The consequences would ripple throughout the community. There is not good time for a pandemic, but this is an especially bad time in the US. The tax cut craze at home and the export of hundreds of billions abroad for the Iraq War has left us defenseless. If irony weren't already dead . . . .

So what to do about schools? Parent Teacher groups and local school officials should designate a committee to start meeting on a regular basis to review developments, hold public meetings of the school community and consult with other agencies and voluntary organizations. There are a number of immediate problems to solve. Who is going to make the decision to close and on what basis? When a closure occurs, what will be the various consequences? Taking them one at a time and dividing the labor, what are some ways we can think of ahead of time to work around those consequences?

No one has the answer to these tough questions. But thinking them through ahead of time may come up with some good approaches and may avoid some disastrous mistakes made in a moment of panic or disarray. The more advance work that can be done in a calm atmosphere the better everyone will react in case action is needed. If a pandemic occurs, all of these problems will get solved because they will have to be solved. But they might be solved badly or destructively. An investment in foresight can pay huge dividends later.

Once again. Homework time.