MAUI, HAWAII -- Maternal mortality figures in this country may be underreported by as much as a half, according to a study conducted in Michigan.
The study found that while figures in Michigan's vital statistics reports would have indicated a statewide maternal mortality of 7.6 deaths per 100,000 live births annually in recent years, enough additional pregnancy-related deaths could be found to boost the ratio to 18/100,000 annually.
If Michigan's system underestimates maternal mortality to such an extent, then it is likely that other states underestimate maternal mortality as well, Dr. James D. Gell said in an interview Dr. Gell reported the data in a poster presentation at the annual meeting of the Central Association of Obstetricians and Gynecologists.
This underestimation of maternal mortality gives people a false sense of security, not only about the health of mothers, but also about the state of women's health in general. Maternal mortality, like infant mortality, is often considered one of the basic barometers of a nation's health care delivery, he said at the meeting, which was held jointly with the annual meetings of Districts VI, VII, and IX of the American College of Obstetricians and Gynecologists.
"Even if you believe the ratio is under 10/100,000, there are still a lot of preventable deaths out there," said Dr. Gell of Wayne State University, Detroit.
An important reason for the underreporting is that ICD-9 coding does not necessarily require mention of whether a woman was pregnant or even in the postpartum period when she died, he said.
Coding may list a death as "hypovolemic shock secondary to hemorrhage." But it does not require mention of the fact that the hemorrhage may have been due to an ectopic pregnancy or delivery. Therefore, it does not get picked up by a state's vital statistics system as a maternal death.
The investigation compared the state's vital statistics reporting for 1986-1995 with reports of possible maternal deaths submitted directly to the Michigan Maternal Mortality Study.
In 1985, study investigators had sent a letter to hospital administrators, police departments, and emergency medical service providers, asking them to report possible maternal deaths.
The submitted reports were then matched against the vital statistics reports. When they did not match, the death was investigated by the Michigan Department of Community Health and reviewed by the state's Maternal Mortality Committee to determine if it qualified as pregnancy associated or related.
Between 1986 and 1995, Michigan vital statistics reported 1,445,299 live births with 111 maternal deaths, for an average annual ratio of 7.6/100,000, which varied from 5.7 to 10.7.
The study identified an additional 160 deaths, raising the total to 271 deaths, for an average ratio of 18.6/100,000, ranging from 15.8 to 23.1.
Michigan state officials already have plans to remedy this underreporting, Dr. Cell said. In 2002, the state is going to include a check box for pregnancy on its death certificates.
Other findings of the study included:
* Forty-eight percent of the maternal deaths occurred in unmarried women.
* African American women accounted for 20% of the live births between 1986 and 1995 but 47% of the maternal deaths.
* There has been a shift in the causes of mortality since 1950. From 1950 to 1985, the three leading causes were obstetrical hemorrhage (32%), sepsis (21%), and pregnancy-induced hypertension (18%).
From 1986 to 1995, the three leading causes were cardiovascular disease (15%), sepsis and pneumonia together (13%), and pulmonary embolism (12%).