Press Box

Body Count

Doing the math on the Washington Post’s momicide series.

How many panicked pregnant women and postpartum mothers ordered body armor after reading Donna St. George’s three-part series about “maternal homicide” in the Washington Post earlier this week? You could hardly blame them. St. George’s blood-splattered narrative stacks new and future moms in the morgue like so much cordwood.

Again and again, St. George turns to researchers who, while they can’t supply numbers, encourage her to believe that maternal homicide is frequent and going unnoticed. “People have this misconception that pregnancy is a safe haven,” one tells her. When St. George tells a Texas professor that the Post has documented 1,367 maternal homicides in the United States over 14 years, the professor responds, “That’s a formidable number—and that’s just the tip.” Another researcher laments the lack of good government statistics on maternal murder: “The system is flawed,” she says. Another says of maternal homicide, “I thought it was a tragedy. I didn’t think it was a trend.”

Is maternal homicide on the rise? Are pregnant women and new mothers in any special danger of murder? And what the hell is maternal homicide? On Dec. 20 (“The Muddled Maternal Murder Series“), I took St. George to task for the alarmism of her series and chided her for not only failing to prove anything with her series but failing to promise to prove anything.

The pivotal research in her piece is “Enhanced Surveillance for Pregnancy-Associated Mortality–Maryland, 1993-1998,” a 2001 study of 247 “pregnancy-associated” deaths in Maryland between 1993 and 1998 published in the Journal of the American Medical Association. It found that “a pregnant or recently pregnant woman is more likely to be a victim of homicide than to die of any other cause,” which St. George quotes favorably.

But the horror of this JAMA study recedes as you read it. We all know what a pregnant woman is: someone who’s carrying a baby. But what is a “recently pregnant” woman? The JAMA study defines the phrase very broadly. By its definition, mothers who give birth are recently pregnant for the 365 days following delivery. Women whose pregnancies end for any reason are also recently pregnant for 365 days after termination. So, a woman who had an abortion, miscarried, or gave birth to a baby would qualify for inclusion in this mortality study if she died with a year of that event.

How many Post readers were aware of this definition when they encountered JAMA’sfinding that homicide was the likeliest case of death for pregnant or recently pregnant women? Few, I’ll bet. When I consulted the fine print of the JAMA study I learned that of the 247 “pregnancy-associated” deaths collected for examination, 50 were homicides. But only 23 of the homicides were committed during pregnancy, or 9 percent of the total deaths recorded. The other 27 killings came within one year of delivery or termination of the pregnancy, and only three came within 42 days of delivery or termination.

Elsewhere in the first installment of her series, St. George tells readers that in Maryland, “slightly more than 10 percent of all homicides among women ages 14 to 44 happened to a pregnant or postpartum woman in the past decade. If that held true nationally, it would suggest about 295 maternal homicides nationwide a year.” (Emphasis added.)

Outside of context, 10 percent sounds appalling. But my back-of-the-envelope calculations reveal that about 10 percent of all American women between the ages of 14 and 44 were pregnant in 2000. So, if you were to murder women in this age bracket at random, 10 percent of your victims would be pregnant. (See my math here.)

Because St. George’s group of Maryland murder victims includes postpartumwomen, this means the percentage of pregnant women murdered has got to be less than 10 percent. An unstated premise in St. George’s series is that the new findings refute the intuitive sense that pregnancy provides some protection from murder. But, examined closely, St. George’s reporting seems to support the idea that pregnant women are murdered less often than non-pregnant women.

You’d never know from reading the series that all American women—pregnant, non-pregnant, pre-pubescent, post-menopausal, geriatric—are safer from murder today than they were in 1995. According to FBI statistics, 3,147 females between the ages of 13 and 44 were murdered that year. The death harvest has hovered around the 2,000 figure since 1999.

If you ordered that body armor, keep the receipt.

The Back-of-the-Envelope Calculations
The authors of the JAMA study, Isabelle L. Horon and Diana Cheng, recoil from the question of whether the homicide rate is higher among pregnantand postpartum women than others because of the “complexities” of figuring the number of pregnant women in the general population.

I am not so daunted.

The National Center for Health Statistics recorded approximately 4.04 million live births in 2000 for women between 14 and 44, so all of these women were pregnant during 2000. Approximately 1 million women between the ages 14 and 44 who were pregnant in 2000 didn’t give birth until the first three months of 2001. This brings my total to 5.04 million pregnant women in 2000.

(My count is incredibly conservative because I don’t include the 1 million fetal losses in 2000—miscarriages from eight weeks forward—or the 1.3 million abortions estimated by the center. Some of these women could have aborted, miscarried, and given birth—or any combination of two—in the same year and ended up being counted a couple of times. For the purposes of streamlining my argument, I deliberately ignored the minuscule number of women who had two successful pregnancies in 15 months in 2000.)

Based on Census Bureau figures, I estimate that there were about 52.3 million women between the ages of 14 and 44 in the U.S. in 2000. So, in the year 2000, 9.6 percent (5.04 million divided by 52.3 million) of all women between the ages of 14 and 44 were pregnant and produced a live birth.