In Brazil, women with higher levels of education are more likely to terminate a pregnancy, and less likely to experience complications after the termination, than those with less education, according to a cross-sectional study conducted among female and male civil servants in Campinas, Sao Pãulo.1 Despite the country’s strict abortion laws, more than half of respondents who had had (or whose partner had had) an unwanted pregnancy reported that the pregnancy ended in abortion. The only measured characteristic associated with opting to terminate an unwanted pregnancy was education level: Respondents with at least some college education were more likely than others to say that an unwanted pregnancy had resulted in abortion (prevalence ratio, 1.6). After an abortion, women with some college education were less likely than those with lower levels of education to need medical care (13% vs. 38%) and hospitalization (8% vs. 28%).

Although abortion is restricted to women who were raped or whose life is endangered by their pregnancy (and, since 2012, in cases of fetal anencephaly), an estimated one million induced abortions occur in Brazil each year; only a small proportion can be classified as legal. And, despite scant data, it is generally believed that women’s risk of complications from unsafe abortion increases as socioeconomic level decreases. Using education level as a proxy for socioeconomic status, the researchers sought to assess the association between education level and induced abortion, and to examine women’s access to physician-performed abortions and their risk of complications.

In January 2010, the researchers sent self-administered questionnaires to 15,800 male and female civil servants; the survey was resent after a month to increase the response rate. In all, 1,660 questionnaires were returned (response rate, 11%). Women were asked if they had ever had “an absolutely unwanted pregnancy,” whether they had felt the need to terminate the pregnancy and, if so, what they had done; men were asked the same questions about their female partners. The survey also collected social and demographic information that referred to the time of the unwanted pregnancy, including age, number of children, marital status, education level and contraceptive use. The researchers used chi-square tests to assess differences in reports of unwanted pregnancy and induced abortion by social and demographic characteristics; multivariate Poisson regression models were constructed to assess associations between demographic variables and induced abortion, the need for medical care and the need for hospitalization.

Nearly three-quarters of participants (73%) were women. Overall, 18% of respondents reported that they or a partner had had an unwanted pregnancy; the proportion was even higher (24%) among respondents who had not attended college. At the time of their unwanted pregnancy, almost half of participants were aged 18–24, nearly three-fourths had no children and a similar proportion were not in a stable union; 45% had at least some college education. Thirty-three percent of respondents had not been using contraceptives, while 23% had been using a hormonal or surgical method or an IUD, and 45% had been using a barrier method, a behavioral method (rhythm or withdrawal) or both.

Among respondents reporting an unwanted pregnancy, 56% had had (or their partner had had) an induced abortion. At the time of the pregnancy, 20% of those reporting an abortion had been in a stable union, and 54% had been using a barrier method, a behavioral method or some combination thereof. Fifty-six percent had some college education. The most common reason for having had an abortion—given by 37% of both men and women—was that the respondent had not wanted to be a single parent. In bivariate analyses, living in a stable union, having attended college, and using barrier or behavioral contraceptive methods were associated with an elevated likelihood of having had an abortion; college attendance was the only characteristic that remained statistically significant in the multivariate regression analysis (prevalence ratio, 1.6).

Sixty-two percent of induced abortions had been performed by physicians, 10% by midwives and 2% by providers with no formal training; 26% were self-induced. Misoprostol was used in 18% of abortions, and unspecified medicines in 12%; medication was provided by physicians or midwives in 5% of abortions, and 26% of respondents reported self-administering drugs. Three percent of respondents said catheters and herbs had been used, and some women had used more than one method. Seventy-six percent of abortions reported by respondents with some college education had been performed by a physician, compared with only 41% of those among respondents with less education.

Twenty-three percent of women who had had an abortion had needed medical care afterward; 17% had required hospitalization. Women whose abortion had been performed by a physician were less likely than those who had used another provider or self-induced to need medical care (12% vs. 40%) or require hospitalization (8% vs. 30%). Moreover, respondents with some college education were less likely than those with less education to report a need for medical care (13% vs. 38%) or hospitalization (8% vs. 28%) after the abortion. Finally, regression models revealed that having an abortion performed by a physician was associated with a lower likelihood of requiring medical care or hospitalization after an abortion (prevalence ratio, 0.3 for each).

The researchers note several limitations. The sample was not representative of the Brazilian population or even (because of the low response rate) civil servants. Furthermore, those who participated may have done so because they had relatively liberal views on abortion, and because most of the abortions had occurred years or even decades earlier, the hospital admission rates might not reflect current risk. Despite these limitations, the researchers suggest that their study provides objective evidence that less educated women face an elevated risk of having complications from unsafe abortion, underscoring “the social inequalities associated with abortion in Brazil.”
L. Melhado


1. Dias TZ et al., Association between education level and access to safe abortion in a Brazilian population, International Journal of Gynecology and Obstetrics, 2014, doi: 10.1016/j.ijgo.2014.09.031, accessed Jan. 22, 2015.