Are Mexican women dying from the effects of unsafe, illegal abortion because of restrictive laws in most of their country? Researchers from the Alan Guttmacher Institute, the US reproductive health and abortion advocacy organization, have estimated figures in excess of 1 million. Other researchers have claimed that Mexico has made no progress in reducing maternal mortality over the past 20 years, largely because of illegal abortions.
But new research by an international team led by Chilean epidemiologist Elard Koch shows that these findings are wildly misleading. Actual official data from Mexico and use of recognized research methods shows that maternal mortality has decreased by over 30 percent during the last two decades (82 percent over the last 50 years), and suggests that abortion figures touted in leading magazines may be inflated by as much as 1000 percent (10-fold).
In the following email interview with MercatorNet Dr. Koch explains how his team arrived at their conclusions on this topic, which is critical to health spending, the health of mothers and the lives of unborn children in developing countries.
Thinking of Mexico, an ordinary American or Australian might conjure up visions of teeming population, desperate poverty and rampant violence related to drug wars and a macho male culture — a place where life is terrible for most women, hundreds of thousands of whom want to abort the babies they had no real choice in conceiving in the first place. But because abortion is restricted by law in most of the country, women tend to have unsafe abortions from which many die. — Is this the picture of abortion in Mexico that experts tend to promote?
Some ideological advocates have portrayed Latin America as a very dangerous place for women since abortion is restricted in most countries of this territory. This view promotes an erroneous and pessimistic image of Latin American women and their motherhood. For instance, Chile exhibits the second lowest maternal mortality ratio in the American continent, second only to Canada. Our study published in PLoS ONE last May showed that the main factors positively influencing maternal health in Chile are women’s education, universal access to specialized prenatal care and childbirth attended by skilled professionals, access to nutrition programs for pregnant women, and availability of clean water and sanitation in most of the territory. None of these interventions are related with violence or gender issues.
Again, most people in the world fail to notice that the current Chilean government recently passed a policy allowing all the working female population to take six months paid leave after childbirth. Far from a macho culture, these facts show how paramount is the protection of women and motherhood for the country. Many developed countries do not have such full and cost-free coverage of prenatal care, childbirth and postnatal care, making motherhood an extremely difficult task to harmonize with professional and individual aspirations of women.
Mexico is following the same route of Chile to promote pre-, peri- and postnatal care in the public healthcare system. Clearly, legal abortion is not compatible with this picture of promoting and protecting motherhood as a fundamental value of society, simply because there is not a cause-effect relationship between the legal status of abortion and overall maternal mortality rates.
Why were you and your colleagues skeptical about the abortion and maternal mortality data for Mexico?
One of our lines of research deals with studying the factors influencing maternal health in Latin America. The Chilean natural experiment published last May is the first of a series of studies currently being conducted for Chile, Mexico, and other countries of the region. So far, our data show that Chile and Mexico displayed significant reductions of maternal mortality during the last half century. While we were conducting studies of Mexico Federal District recently, researchers from IPAS-Mexico published an article in International Journal of Gynecology and Obstetrics claiming an apparent lack of progress in Mexican maternal health during the last two decades. Since our analysis showed completely the opposite, we decided to re-evaluate the data obtained from official sources and compared them with those published by researchers of IPAS-Mexico.
Secondly, our study of Chile had already shown that there was no increase in hospitalizations due to abortion after therapeutic abortion was banned there in 1989. On the contrary, hospitalization due to any kind of abortion (whether spontaneous or induced) continued to decrease. Furthermore, if banning abortion resulted in more deaths from illegal abortion, then an increase in the abortion mortality ratio should have been observed. In fact, the abortion mortality ratio not only did not increase, but also decreased over 96% after the abortion ban in 1989, from 10.7 to 0.39 per 100,000 live births. This decrease in abortion-related mortality and especially abortion hospitalizations after the ban in Chile strongly suggests a parallel reduction in induced abortion rates.
Thirdly, in 1990, researchers of the Alan Guttmacher Institute (AGI) estimated approximately 160,000 illegal induced abortions in Chile. Taking into account the figures from our study in Chile, a sort of “paradox of illegal abortion” seems evident: if the figure of 160,000 illegal abortions was accurate and illegal abortion was equivalent to unsafe abortion as researchers of the AGI persistently claim, then a high number of deaths due to abortion should be expected. However, the opposite was observed.
To explain this apparent paradox, we hypothesized that something was not working well with the methodology estimating the figure of 160,000 abortions for Chile, and conducted a thorough review of the methodology applied in eight Latin American countries, including Mexico. This review was published in May in the journal Ginecología y Obstetricia de México. This triggered an intense correspondence debate regarding methods for estimating abortion figures, which is publicly available in this reputable Mexican journal.
Finally, results of the legalization of abortion in the Federal District of Mexico (Mexico DF) during 2007 were available for analysis. This offered, for the first time in a Latin American territory, the unique opportunity of a natural experiment for comparing estimates of abortion figures based on opinion surveys with actual figures of elective abortions after five years of legalization — a period of time sufficient to replace most of illegal abortions with legal procedures.
Clearly, opinion surveys have overestimated the number of abortions, likely because these surveys are subjective in nature and strongly subjected to selection and recall bias.
What method did you use to gain a more accurate picture of abortion? What did you find?
It is important to note that any method used to estimate abortion figures will exhibit an error, leading to under- or over-estimation of such figures. This error, however, is not immediately evident and can only be identified when thoroughly evaluating the estimation procedure. In fact, statistical counting based on prospective records of elective abortions is substantially different from studies based on self-reported abortions.
In the first case, abortions are registered when the abortion procedure has been completed in a hospital or abortion facility. Thus, abortion figures will depend on the quality and integrity of the registry. The number of abortions may be underreported, especially when some facilities providing elective abortion services do not report them consistently. In the second case, that is, studies based on self-reported abortions of women or estimates based in the opinion of selected informants, abortion figures may be over- or underreported depending on the methodology utilized. When a sample of women of reproductive age is directly surveyed about induced abortion, it will generally be underestimated due to the strong stigmatization of abortion.
On the other hand, it is virtually impossible to establish figures of abortion using opinion surveys of a limited number of informants such as health workers. For example, asking a sample of individuals how many abortions they believe occur per each case of complicated abortion attended in a hospital or health facility is just a mere subjective opinion. This is basically the methodology used by researchers of the AGI. In the recent review of eight countries published in Ginecología y Obstetricia de México last May, it was established that this kind of survey is used to calculate an multiplying factor which is then applied to amplify by 5, 6, 7, or more the number of hospitalizations due to abortion obtained from in-hospital records.
In some cases, such as Colombia, the number of hospitalizations from abortion was estimated through additional surveys administered to directors of health facilities who were required to remember the number of hospitalizations in the “average month” and in the “past month”. These two estimates are averaged and multiplied by 12 to give an overall best guess as to the number of abortions carried out in a year-long period, assuming a constant monthly abortion incidence.
Clearly, the validity of these procedures is seriously questionable, mainly because the subjective nature of the surveys and because the methodology is strongly exposed to selection and recall biases per se.
Overestimation of abortion figures through the method used by researchers of the AGI becomes evident when comparing them side-by-side with actual figures of abortion in Mexico DF, where the Group for Information on Elective Reproduction (GIRE) carries out surveillance of abortion figures. As we showed in the International Journal of Women’s Health, the GIRE counted 12,221 elective abortions performed by women who lived in Mexico DF during 2009. The figure estimated by researchers from the AGI for the same year was 122,455, resulting in a 10-fold overestimation.
With regard to maternal and abortion mortality in Mexico, we employed official records from the National Institute of Statistics and Geography (INEGI), the General Directory of Health Information (DGIS) and the National System of Health Information of the Ministry of Health (SINAIS). Using these data, we were able to construct time series of maternal mortality ratio since 1957, and of the abortion mortality ratio since 1990. We could also analyze the relative contribution of different causes of death since 2000. To construct time series of mortality ratios, we divided the number of total or specific maternal deaths by the number of registered live births.
Maternal deaths were analyzed using the tenth (current) version of the International Classification of Diseases (ICD10), which contains several codes related to maternal deaths. Only nine (O00-O08) are associated with what is termed “pregnancy with abortive outcome”, to classify diseases or deaths related with pregnancies ending in abortion. However, the first four of these codes (O00-O03) classify ectopic pregnancy, hydatidiform mole, other abnormal products of conception, and spontaneous abortion. None of these conditions are related to induced abortion, whether legal, illegal, safe, unsafe, accidental, or incidental.
At this point, we detected two types of discrepancies in the study by researchers from IPAS-Mexico. The first was related to the numerator. They misused the coding in the ICD10 to analyze abortion mortality with specific emphasis on illegal induced abortion. In this case, researchers from IPAS-Mexico included all nine types of abortive outcomes to reach their estimates for illegally induced abortion.
The second discrepancy was due to the use of projected live births instead of official (and readily available) registered live births as the denominator of mortality ratios. Specifically, there was a significant difference between the projected (fictitious) figure of live births and the observed (actual) number of live births. We found that projected figures of live births were between 300,000 and 700,000 less than the number of registered live births, in the period between 1990 and 2010. In this regard, it is important to note that projected live births can be used only when official records of live births are unavailable or incomplete, which is not the case for the timeframe studied in Mexico.
Therefore, it is very hard to understand why researchers from IPAS-Mexico used fictitious denominators when actual data of live births are readily available. Clearly, dividing any number of deaths by a figure of projected live births which is substantially smaller than the figure of officially registered live births will unavoidably lead to a large error, thus overestimating maternal mortality ratios. Along with the numerator issues, these discrepancies led to a significant overestimation of mortality ratios in Mexico (up to 35%).
How many maternal deaths, exactly, can be attributed to induced abortion?
For 2009, out of the 1207 total maternal deaths registered in Mexico, only 25 can be attributed to illegal induced abortion.
Seventy-four deaths were classified under ICD10 codes O00-O08 — that is, deaths associated with “pregnancy with abortive outcome”. Out of these 74, only 25 were classified under codes O05-O07, which are the only codes that can be associated to illegal induced abortion.
A registry of vital statistics needs to fulfill four requirements in order to be considered as reliable by the World Health Organization (WHO): integrity, medical certification, location of residence, and civil registration. In this regard, civil registration in Mexico and Chile are characterized as virtually complete by the United Nations and the WHO, with good attribution of causes of death. In fact, both countries are classified in list A (complete vital statistic records) regarding the registry of maternal death causes, along with other 63 countries in the last report of global maternal mortality by WHO.
Not infrequently, however, it is thought that criminalization of abortion may lead to underreporting by misclassification of maternal deaths suspected to be the result of an illegally induced abortion. Nevertheless, there are codes in the ICD10 that allow safeguarding of both professional and patient confidentiality when physicians suspect an illegal abortion. Moreover, nowadays in most countries, physicians are subject to legal sanctions if they are found guilty of distorting or misclassifying actual causes of death. Thus, the reliability of mortality ratios associated with abortion is likely to depend on the registry quality of each territory studied. And, as explained before, civil registration in Mexico is classified in List A by WHO.
Regarding the problem of underreporting of elective abortion figures in the Federal District of Mexico (Mexico DF) after abortion legalization, we think it is probable that not all abortions carried out in private clinics are being reported. However is very unlikely that the level of underreport exceeds the extreme case of 100 percent, mainly because most Mexican women (over 70 percent) use public health facilities that are free. Part of the Mexican population prefers using private health facilities at their own monetary cost. Thus, probably the underreport rates would follow the same trend, mainly explained by the proportion of women seeking abortions in private clinics.
For example, if we imagine a underreport rate of 100 percent, approximately 50 percent of abortion should be conducted in public health facilities and 50 percent in private clinics. This means that if 12,221 abortions were conducted in public health facilities, then 12,221 abortions would have been performed in the private sector. Thus, considering that over 70 percent of the general population use the public healthcare system, it is very unlikely that in private clinics the number of abortions there surpasses the number of abortions registered in public health facilities. Even with a 100% underreport rate, the figures for induced abortion published for Mexico DF by researchers of the AGI would be significant overestimates.
Among the various causes of maternal death you put the spotlight on partner violence against women during pregnancy, which you say has increased “alarmingly” in recent years. Has anyone suggested that this could be due to lack of access to abortion? Do you have your own theory about causes of this trend?
We were very interested in the trend observed in the maternal mortality ratio due to “other abortions” (O05 of ICD10), which increased from 0.11 to 0.23 maternal deaths per 100,000 live births between 2002 and 2008. At the same time, national surveys of violence against women were conducted in 2003 and 2006, showing an increase of violence against women during pregnancy, from 5.3 percent to 9.4 percent between those years. This means that violence against pregnant women doubled in three years. I think that this can be regarded as “alarming”.
Given that code O05 is employed in the clinical setting to classify maternal deaths with abortive outcome due to accidental injury or violence from a third party, we became interested in the correlation between increase in violence and the increase in maternal deaths classified under code O05. However, this is just a correlation so far and should not be regarded as a cause-effect relationship. It requires further analysis before starting to make assumptions about its implications.
If not legalized abortion, what is it that women in Mexico and other countries of Latin America most urgently need if motherhood is to be made safer?
Nowadays, promoting changes in the legal status has been the main solution to maternal mortality in developing countries with restrictive abortion laws. In this context, as scientists we aim to provide objective evidence to help developing countries to better allocate the limited resources they have to continue their progress in maternal health.
Our studies in Chile and Mexico show that the legal status of abortion is not associated with overall maternal mortality rates. In other words, there is no cause-effect relationship between the legal status of abortion and maternal mortality.
In addition, since over 98 percent of causes of maternal death are unrelated to induced illegal abortion in Mexico, little or null improvement in overall maternal health should be expected from abortion legalization or abortion prohibition.
It is clear that decreasing abortion-related mortality is always a relevant factor that positively influences maternal health in a country. However, nowadays hundreds of Mexican women continue to die due to hemorrhage, eclampsia and indirect causes; this suggests very concrete strategies that clearly are unrelated to the legal status of abortion. Hence our recommendation is to provide universal access to prenatal and perinatal care and emergency obstetric units with specialized health professionals, especially in the most vulnerable regions. In addition, such strategies can also decrease abortion related mortality in countries with restrictive abortion laws, as has been scientifically well documented in Chile.
Elard Koch is Assistant Professor of Epidemiology in the Department of Family Medicine at the University of Chile, and Associate Professor and Director of Research for the Institute of Molecular Epidemiology, Centre of Embryonic Medicine and Maternal Health at the Universidad Católica de la Santísima Concepción.
The study discussed here is: Koch E, Aracena P, Gatica S, Bravo M, Huerta-Zepeda A, Calhoun B C (2012) Fundamental discrepancies in abortion estimates and abortion related mortality: A re-evaluation of recent studies in Mexico with special reference to the International Classification of Diseases. Int J Women Health 4:613-623. Available at http://www.dovepress.com/articles.php?article_id=11688