Thursday, February 05, 2009

By Their Own Hand



Suicides among the U.S. military may have killed more members of the American military forces in January than combat in Afghanistan and Iraq did:

The Army is investigating a stunning number of suicides in January — a count that could surpass all combat deaths on America's two warfronts last month.

According to figures obtained by The Associated Press, there were 24 suspected suicides in January, compared to only four in January of 2008, six in January of 2007 and 10 in January of 2006.

Yearly suicides have been rising steadily since 2004 amid increasing stress on the force from long and repeated tours of duty in Iraq and Afghanistan.

The service has rarely, if ever, released a month-by-month update on suicides, but officials said Thursday that they wanted to re-emphasize "the urgency and seriousness necessary for preventive action at all levels" of the force.

An alarmed Army leadership also took the unusual step of briefing congressional leaders on the information Thursday morning.

The big monthly count follows an annual report last week showing that soldiers killed themselves at the highest rate on record in 2008. The toll for all of last year — 128 confirmed and 15 pending investigation — was an increase for the fourth straight year and even surpassed the suicide rate among civilians.

What makes studying these numbers so hard statistically is that they should be compared to a suitably selected control group, and that control group depends on the questions we want the data to answer. Do we want to know if people in military in general commit more suicides? Or only during war times? Or only during the current wars in Iraq and Afghanistan?

All that may sound trifling and even callous as a response to the quote I gave above. But we can't really tell how unusual the rate is without knowing what we should compare it to.

For example, it's not really meaningful to compare military suicide rates to the overall suicide rate in the U.S., because the latter is based on the averages from all age groups. Because suicide rates rise with age, we'd expect the military rate to be lower for that reason alone.

On the other hand, suicide rates are higher for men than for women and the military has a skewed gender distribution with many more men than women. That, in turn, should pull up the expected suicide rate when compared to the general population in the same age groups.

Finally, the military may just have a higher suicide rate even during peace times, and the relevant control group for evaluating the most recent news could be the average peace-time military rate of suicide.

In any case, that the rates are rising so rapidly suggests that we have a real crisis and that fast action is necessary. That action should probably also evaluate whether all the deaths of female military declared as suicides in Iraq and Afghanistan really were suicides.