Published in the National Post
Seventy thousand. We've been told this is the number of women who die each year as a result of unsafe abortions in the developing world. Furthermore, we are told, this accounts for 13% of all maternal deaths.
It's a tragedy whether 700 or 70,000 women die from abortions. However, the issue is much bigger than the number. The assumptions used to calculate the 70,000 are debatable and the methodology isn't up to scientific standards. In the end, 70,000 is nothing more than a stab in the dark by the World Health Organization's own admission.
The source for the 70,000 lies in a series of reports published by WHO called Unsafe Abortion. There, researchers repeatedly clarify how hard it is to study the issue due to the lack of data. One citation (of many) reads: "As there are no feasible data collection methods that can reliably reflect the overall burden of unsafe abortion, one is left to work with incomplete information on incidence and mortality from community studies or hospitals ... This is then adjusted to correct for misreporting and under-reporting." It's no big surprise that some of the world's poorest or war-torn nations don't keep impeccable abortion statistics. So researchers rely on assumption after assumption.
Like this one: Abortions in the developing world, they say, are always under-reported. This assumption leads researchers to consistently inflate abortion numbers. Entirely lacking is any rigorous defence of this opinion. There's every reason to believe that local attitudes toward abortion, most of which are significantly less approving of abortion than those in the West, might lead women to have fewer abortions, not more.
Then there's the assumption that allows for local data to be applied nationally. "It was assumed that sub-national data could be extrapolated to country level with adjustments," write researchers. That's about as rigorous as assuming that since 2% of Saskatchewan is francophone, so too is 2% of Canada.
What about this one? Researchers assume that half of induced abortions would result in hospitalization for complications, using hospitalization rates to attempt to calculate unsafe abortion rates and then death from unsafe abortion. Yet that is entirely unreliable: Depending on the method used to abort, the prior health of the mother and access to basic antiseptic supplies, this ratio could be far too low or far too high.
In the developing world, reliable data about reproductive health is largely confined to the number of live births and hospital admissions. With only these two data sets that are reliable, researchers make a series of assumptions that simply don't hold up: They assume that the proportion of births in hospital is comparable to the proportion of miscarriages and abortions outside and try to link the deaths of women with their best guesses about "unsafe" abortion. In Canada, for example, births that occur in hospitals and miscarriages (or abortions that result in a hospital admission) do not correlate and vary significantly by region. This assumption is, by logic, more inaccurate in the developing world.
No one can confirm or check the country by country results for death from unsafe abortion. "Because of the level of uncertainty," write researchers, "country estimates were calculated solely for the purpose of aggregation and are not published."
Finally, researchers do not appear to consider the variation among regions. One of the more detailed studies cited in a WHO report published in 2004 is almost two decades old and based exclusively on 144 women living in the same town in Nigeria. This fails to meet the accepted standards for meaningful conclusions in social science or epidemiology.
So why the obfuscation?
Activists know very well that statistics play a role in directing policy dollars. The stakes just went up with the promise of big money at the G8/G20 meetings. If unsafe abortion is a major cause of maternal death in the developing world, appeals to increase funding for this gain credibility. But if the discussion turns to the much more prevalent risks of dying in childbirth, effective foreign aid might focus on making birth safer.
Abortion-rights activists are busy pointing fingers at those who exclude abortion from maternal health, saying that medical care should be based on hard science, not ideology. True enough. It's just that there appears to be more ideology than science to the 70,000 number so often cited.
-Andrea Mrozek is manager of research at the Institute of Marriage and Family Canada. Rebecca Walberg is president of the Wakefield Centre for Policy Research