Out of 1,500 questionnaires mailed, we received 572 completed questionnaires (38% response rate). As shown in Table 1, the majority of the respondents were women (56%), between 26 and 45 years old (56%), and Catholic (72%). About 52% of the respondents lived in the Southeast Region of Brazil, which includes the states of Rio de Janeiro, Minas Gerais, São Paulo, and Espirito Santo. The geographical distribution of respondents was roughly equivalent to that of the FEBRASGO membership overall .
Two hundred seventy-six physicians (48%) correctly identified that abortion is legal in Brazil to save the life of a woman and in the case of rape. Also, the large majority (70%) reported awareness that the Brazilian judicial system can grant court authorization of an abortion in cases of severe fetal malformations (Table 2). Considerable confusion existed over the Ministry of Health guidelines on the gestational age limit for abortion. One quarter of physicians did not know of such a gestational age limit guideline, while 40% thought it was 12 weeks and 30% thought it was 20 weeks. In terms of abortion procedures, 86% and 90% of physicians reported knowledge of MVA and dilation and curettage (D&C), respectively. A large majority (86%) also reported knowledge of misoprostol or other prostaglandins for abortion, while few physicians knew of the mifepristone and misoprostol regimen (37%) or the methotrexate and misoprostol regimen (27%).
Table 3 contrasts the percentage of physicians who believed that abortion is currently legal in Brazil, according to different circumstances, with the percentage of physicians who believed abortion should be legal. In every circumstance except for rape, more physicians believed that abortion should be legal than believed abortion is currently legal. In the case of rape, almost all (93%) correctly identified that abortion is legal and a lesser percentage (85%) thought that it should be legal. A sizeable difference emerged in the case of severe fetal malformations, where 36% of physicians incorrectly believed that it was a legal case for abortion, while 89% thought that abortion should be legal in this case. We observed a similarly large difference when the woman's health is at risk; 7% incorrectly thought that the law currently allowed for an abortion in this case, whereas 30% of physicians thought that abortion should be legal in this case.
Overall, we found that only 3% of physicians agreed with the current law that allows for abortion only in cases of rape or to save the woman's life. Most physicians (77%) thought the law should be liberalized to allow for a legal abortion under more circumstances and 17% thought the law should be more restrictive. The overwhelming majority (95%) expressed their support for public funding of abortion services (Table 4).
In addition to sharing their knowledge and opinions regarding abortion law, physicians also reported on their abortion-related practices (Table 5). Nearly 70% had never received any training on abortion procedures. Among the 33% of respondents that had ever performed an abortion, the most common procedures were D&C (60%) and the use of misoprostol or other prostaglandins (68%) for medical abortion. While most of the respondents (73%) had performed an abortion within the first 12 gestational weeks, there was also a significant number (44%) that had performed an abortion between 13 to 20 gestational weeks. Fifty-three percent had performed an abortion in the case of severe fetal malformations, 25% had performed an abortion to save a woman's life, and 19% had performed one in a case of rape.
We encountered several statistically significant relationships between physicians' sociodemographic characteristics and abortion-related practices with their opinion of abortion law (Table 6). In a bivariate analysis, physicians who felt the law should be more liberal were more likely to have correct knowledge of abortion law, to be familiar with the abortion law regarding severe fetal malformations, and to support public funding for abortion services (p < 0.10). Catholic physicians and Evangelical physicians were not significantly different in their abortion opinions from those who reported that they were not religious; however, physicians who reported having "other" religious affiliation had lower odds of thinking the abortion laws should be more liberal. These variables were all included in an initial multivariate analysis of physician opinion, in addition to region of residence and whether or not they had ever performed an abortion. The latter variables were included due to their hypothesized association with physician opinion and their importance in the literature. In the final model, we combined Catholic and Evangelical physicians into a single group and compared them with non-religious respondents and respondents reporting "other" as their religion. We found that Catholic and Evangelical physicians were not significantly different in their abortion opinions from physicians who reported no religious affiliation. In contrast, physicians whose religious affiliation was "other" had a decreased odds of thinking abortion law should be more liberal compared to physicians with no religious affiliation (OR 0.27 (95% CI: 0.10, 0.74). In addition, physicians who had correct knowledge regarding the status of abortion law in Brazil had a 46% higher odds of favoring a more liberal abortion law compared to those who did not have correct knowledge (OR 1.46 (95% CI: 0.96, 2.22), but this finding was not statistically significant. Finally, physicians who were in favor of public funding for legal abortion services had six times the odds of favoring a more liberal abortion law compared to those who were opposed to public funding of legal abortions (OR 6.01 (95% CI: 2.53, 14.28); however, it must be taken into account that only a small number (24) had said that they opposed public funding for legal abortions. Region of residence, correct knowledge of the law regarding abortions in cases of severe fetal malformations, and experience performing an abortion were not significantly related to physician opinion in the multivariate analysis and thus not included in the final model.