Recently, as part of the health care debate, and in response to the Stupak-Pitts Amendment attempting to restrict abortions in the pending health care reform bill, arguments have recently been made that universal access to health care actually has reduced the abortion rate when implemented, at least in developed countries, and in Massachusetts.
The Massachusetts example seems better at first glance, because it is a comparison of rates pre- and post- universal coverage:
The abortion rate… declined from 3.8 per 1000 population in 2006 to 3.6 per 1000 in 2008.
Of course, the main problem with this is that it’s not clear if this is a statistically significant difference, because it is a sample size of one. But even if the data were available for multiple states/countries before and after universal coverage, the trend has to be placed in the greater context of a longer time period. As it turns out,
The Massachusetts abortion rate has…dropped by a third, from 30 per 1000 women 15 to 44 years of age in 1991 to about 20 per 1000 in 2005, with most of the decrease occurring during the late 1990s.
So the trend was a decline in Massachusetts anyway, so there’s no a priori reason to believe that the cause of the decline was the implementation of universal coverage.
But the author, Dr. Patrick Whelan, successfully makes his point: the belief motivating the Stupak Amendment is that abortions will increase as a result of expanded coverage, but not if the government refuses to pay for them, or allow subsidized plans to pay for them. The Massachusetts data at least suggest that there is no apparent basis for this belief.
The world-wide abortion-rate data from the U.N. cited by T.R. Reid in the Washington Post has the advantage of being multiple data points, but the disadvantage that there is no before-and-after data available. Reid reasonably compares the U.S. abortion rate to that of the 32 other most developed countries which all have universal coverage in at least some form. It is a compelling notion that access to care corresponds to better access to pregnancy prevention, not to mention pre- and post-natal care, and therefore, results in fewer abortions.
Alas, although he makes a convincing argument that will provide talking points for progressives, his data were cherry-picked with no objective statistical analysis. He does point out what seems to be a smoking gun that out of developed countries, that the U.S. abortion rate is the highest (at 20.8 per 1000 women), but there is nothing inherent in that observation showing that universal health coverage reduces abortions.
Thus, it seemed it would be interesting to do some statistics on the data, to see if any trend pops out. In the table below (click for larger view), the abortion rate is listed for all 61 countries for which the U.N. provided data:
The most valid way to statistically compare the abortion rates between countries with and without universal health coverage is with a non-parametric test, because the sample sizes are so unbalanced – 48 countries have some form of universal health coverage, while 13 do not. A Mann-Whitney U test on the ranks of the data shows that there is no statistical difference (p>0.25) in abortion rates between countries with and without universal coverage.
The caveats in this analysis are first that it depends on the accuracy of the U.N. data, which presumably do not take into account illegal or under-the-table abortions (so to speak), whose rates will clearly vary greatly from country to country. (The zero values for Brazil, Panama, and Poland are a case in point.) Second, it also depends on the definition of universal coverage, which also varies greatly from country to country. For example, although Russia has universal coverage on paper, it is well known that their system is broken, and does not in practice provide universal access to health care.
But what if we focus on developed nations, as Reid did? We can’t compare those with universal coverage to those without, since they all have it but the U.S., but we can see if time since universal coverage was available has had any effect on the current abortion rate, and we can look more closely at the type of coverage provided. The next table summarizes these data.