Introduction

Latin America has a high frequency of unintended pregnancies and pregnancies that end in abortion, resulting in a high abortion rate, which did not show a significant reduction when comparing the periods 1990–1994 and 2015–20191. Most countries in the region have restrictive abortion legislation2, including Brazil, where legal terminations of pregnancies are only permitted in cases of rape, risk to the woman’s life, and in pregnancies with anencephalic fetuses2.

Legal restrictions do not prevent abortions, but they do increase the possibility of unsafe abortions3. Global data show that, from 2015 to 2019, abortion rates were higher in countries with restrictive legislation, particularly in low- and high-income countries1.

In Brazil, there was a reduction in the proportion of women reporting having terminated a pregnancy between 2010 and 20214, as well as a reduction in the number of hospitalizations due to abortion complications since 19955,6. However, abortion remains an important cause of maternal morbidity5,7ranking among the five leading causes of maternal mortality in the country8. Furthermore, the illegality of abortion affects the quality of care, even for women with spontaneous abortions9,10,11.

The availability of legal abortion services in the country (services where only doctors terminate pregnancies in cases permitted by law) is limited, concentrated in 3·6% of the Brazilian municipalities, mostly those with more than 100,000 inhabitants, with a higher Human Development Index and located primarily in the Southeast region. These services are provided by public institutions or those affiliated with the Unified Health System (SUS)12. The concentration of services, the lack of intermunicipal public transportation, and the cost of travel to access legal abortion services are barriers to users who require these services13, resulting in a 4·8% lower legal abortion rate in municipalities without these services12. It is estimated that in Brazil, only one in 11 pregnancies resulting from sexual violence had access to legal abortion13.

The use of medications recommended by the World Health Organization3for medical termination of pregnancy is limited: mifepristone is not authorized in the country, while misoprostol is restricted to hospital settings, contradicting scientific evidence demonstrating its safety in primary care for pregnancies less than 12 weeks gestation3,14. This scenario is compounded by legislative initiatives attempting to make access to legal abortion even more restrictive in the country, either by creating additional obstacles to access in legally permitted situations or by proposing to suspend existing permissions15.

Currently, all clinical abortion care in Brazil is performed exclusively by medical professionals. In a literature review that aimed to identify studies that assessed physicians’ knowledge, attitudes, and practices regarding abortion care in Brazil we identified 35 studies, 18 related to the topic, with 10 published more than 10 years ago and only one national study conducted in 2001/2002. This study16 reported inaccurate knowledge of abortion legislation, resistance to expanding legal permissions for termination of pregnancy beyond those in place, unfamiliarity with clinical guidelines, and lack of experience with manual vacuum aspiration (MVA) or misoprostol with mifepristone or methotrexate. Therefore, the objective of this study is to update the evidence on the knowledge, attitudes, and practices of physicians working in Brazilian public and private maternity regarding abortion care.

Methods

Nationwide, cross-sectional, hospital-based study, carried out between 2021 and 2025, integrated with the study “Birth in Brazil II: national research on abortion, labor and childbirth (NBII)”17. We used the STROBE Statement—Checklist to guide our report.

Study population

Physicians who provided clinical care to women hospitalized for childbirth or abortion care were considered eligible. Professionals working exclusively in management or outpatient settings were not considered eligible.

Sampling plan

For the NBII survey, a two-stage probability sample (hospitals and women) was selected, with hospitals stratified by national macro-region, location (metropolitan or non-metropolitan region), type (public, mixed, private), and size (100–499, ≥ 500 births/year), resulting in 60 sampling strata17. The hospitals included are responsible for 97,7% of births in the country.

All hospitals selected for the NBII were included in this KAP study. We used a convenience sample, due to difficulties in obtaining complete lists of obstetricians from each hospital, particularly in services with open staff models. We included two obstetricians in hospitals with 100–499 births/year and four obstetricians in hospitals with ≥ 500 births/year. Obstetricians who provided clinical care to women hospitalized for childbirth or abortion care were considered eligible. Professionals working exclusively in management or outpatient care were excluded17.

Data collection

In each hospital, the coordinator of the obstetrics service and first available professionals were approached and invited to participate in the study by the NB II research team. Participation was voluntary, during working hours, through completion of a self-administered, anonymous questionnaire.

The instrument was developed by the research coordination team and addressed issues considered a priority in the Brazilian context, particularly regarding legal permissions for termination of pregnancy in the country, care practices for women undergoing abortion, and participation in legal abortion care.

Professionals who declined, did not return the questionnaire, or returned it blank were considered refusals. Replacements were used for refusals until the sample size was reached in each hospital or when no more professionals were available.

Data analysis

Knowledge, attitudes, and practices were analyzed for the total sample and categorized by regions of the country (North, Northeast, Southeast, South, and Midwest) considering the differences in access to legal abortion in these regions. The WHO18 and the Brazilian Ministry of Health care guidelines19 available at the time of the study were used as the reference standard.

We used weighting and calibration procedures. For weighting, because information on the total population of obstetricians involved in childbirth and abortion care was unavailable, selection probabilities could not be directly estimated. As a proxy, we used the number of live births, based on the assumption that hospitals with more births would proportionally have more obstetricians. Basic weights were calculated as the inverse of the selection probability, defined as the ratio between the total number of annual live births within each of the 60 strata and the number of sampled obstetricians. Finally, weights were calibrated using data from all hospitals with ≥ 100 live births in the 2022 Live Birth Information System, based on the joint distribution of region and hospital type. Differences in proportions were assessed using the chi-square test (α = 0.05). All analyses accounted for the complex sampling design by using the Complex Samples module of IBM SPSS Statistics, version 26.0 (IBM Corp., Armonk, NY, USA).

Ethical aspects

The NBII study was approved by the National Research Ethics Committee (Comissão Nacional de Ética em Pesquisa - CONEP), on June 22, 2020, process number 21633519.5.0000.5240. All methods were performed in accordance with the relevant guidelines and regulations.

Abortion is a sensitive topic in Brazil, and we wanted to ensure that professionals felt comfortable expressing their opinions and practices. Signing the informed consent form could raise concerns among professionals that their identity could be identified without guaranteeing their anonymity. Therefore, for this study with physicians, we opted to use an invitation letter, replacing the free and informed consent form, to guarantee the anonymity of the respondents. The invitation letter contained information about the research objectives and methods, the voluntary nature of participation, the risks and benefits of participating in the research, the guarantee of confidentiality and anonymity, as well as the contact details of the study coordination team. The questionnaire did not contain any information that could identify the professional or the hospital. We only obtained information on demographic characteristics, type of hospital (public, mixed or private) and region of location. The return of the completed questionnaire was considered as an agreement to participate. The use of the invitation letter was approved by CONEP.

Results

Data from 1,267 physicians were analyzed, corresponding to 86% of the planned sample. Refusals were the main reason for not reaching the planned sample. Most professionals worked in the Southeast and Northeast regions of the country, in public hospitals, with 500 or more births per year, and with teaching activities. Most respondents self-reported as female, had a mean age of 43·2 years, self-identified as white, were married/cohabiting, and with children. Nearly 90% reported having a religion, with 90·9% considering religion important or very important to their lives. Approximately 40% had graduated less than 10 years ago, 87% had completed a residency or specialization in obstetrics and gynecology, and 65·2% had been practicing for less than 10 years. Almost all professionals had formal employment contracts and 85% worked on shifts. Significant differences by region were observed for all physician characteristics, except for having a specialization and working in routine services (Table 1).

Table 1 Characteristics of physicians by region. Brazil, 2021–2025.
Full size table

Almost all professionals reported receiving training on misoprostol use and uterine curettage during medical training. For MVA and legal abortion care, this figure was close to 75% during medical training and 30% during in-service training. One-third of the physicians interviewed were unfamiliar with the Brazilian Ministry of Health’s standards of care for abortion. Half of the professionals had inadequate knowledge about the use of conscientious objection to deny legal abortion care. Knowledge about specific aspects of misoprostol use and MVA was high, being lower for non-hospital misoprostol use and MVA for infected abortions. Significant differences by region were observed for training in MVA, knowledge of technical standards for humanized abortion care and anencephalic pregnancies, and specific aspects of MVA (Table 2).

Table 2 Physicians’ knowledge of abortion care. Brazil, 2021–2025.
Full size table

Approximately 90% of professionals agreed with legal permissions for termination of pregnancy in cases of risk to the woman’s life and anencephalic pregnancies, with a lower proportion, 80%, in cases of sexual violence. Agreement with expanding permissions was less than 20% for several conditions (contraceptive method failure, financial situation, studies/work, any circumstance), reaching 83·2% for other malformations incompatible with life and 43·7% for harm to the woman’s physical and/or psychological health. Approximately 80% of professionals agreed with the advantages of MVA, with lower agreement regarding misoprostol, particularly for non-hospital use (32·3%). Three-quarters of physicians said that professionals should maintain medical confidentiality when providing care to women who likely terminated their pregnancies, and less than half state that they would only claim conscientious objection when another professional was available to provide legal abortion care. Significant differences by region were observed for all aspects investigated, except for some specific aspects of misoprostol and MVA use, medical confidentiality and recourse to conscientious objection (Table 3).

Table 3 Physicians’ attitudes toward abortion care. Brazil, 2021–2025.
Full size table

Approximately 60% reported the existence of clinical guidelines for abortion care in the service where they are currently working, which were followed routinely or occasionally by 57·4% and 38·4% of professionals, respectively. Only 40% reported that MVA was always available in the service, with 21·9% reporting occasional availability. Over 90% reported using MVA when available, and less than 10% reported using electric aspiration for uterine evacuation when MVA was unavailable. Over 80% of professionals reported having provided care to women who likely terminated their pregnancies, with less than half reporting empathy for the woman’s situation, with 10% reporting the women to the legal authorities. One-third of professionals reported that women hospitalized for abortion care experience prejudice or are unaware of its occurrence. Significant differences between regions were observed in all aspects of care investigated, except for having provided care to women who likely terminated their pregnancies and for women experiencing prejudice in services (Table 4).

Table 4 Physicians’ practices in abortion care. Brazil, 2021–2025.
Full size table

Approximately 60% of professionals reported having provided legal abortion care in their professional lives. Half currently worked in services that provided this care, with 80% being part of the team, either in all situations (41·1%) or in specific situations (38·9%), with anencephaly (83·4%) and risk to the woman’s life (69·0%) being the most frequent. Approximately 40% reported that care is always or occasionally provided by a multidisciplinary team, 20% that the service requires a police report or court order to perform the legal abortion, and 50% that service professionals claim conscientious objection to not providing legal abortion care. However, a high proportion of professionals were unable to provide information about these aspects of the service’s operation. The main reasons for not participating in the legal abortion care team were religion and conscientious objection. No significant differences by region were observed for legal abortion care in situations of anencephaly and risk to the woman’s life; requirement of a police report or court order; recourse to conscientious objection by professionals; and reasons for not joining the legal abortion care team. However, significant differences by region were observed regarding legal abortion care during professional life and current work in a service that provides legal abortion care, with lower values observed in the Midwest region. A lower proportion of professionals who reported never providing legal abortion care was reported in the South and Southeast regions, while greater care for the specific situation of sexual violence was observed in the South region. Fewer professionals in the North region reported that care was provided by multidisciplinary teams (Table 5).

Table 5 Physicians’ practices in legal abortion care. Brazil, 2021–2025.
Full size table

Discussion

The results of this study reveal regional differences in the characteristics, knowledge, attitudes, and practices of medical professionals providing abortion care in Brazil. Common characteristics include: (i) high agreement with legal permissions related to congenital anomalies incompatible with life and situations that endanger the woman’s life, with very low agreement with permissions that would extend the legal termination of pregnancy; (ii) the existence of prejudice and a lack of empathy among professionals toward women who have likely terminated their pregnancies; and (iii) limits on the provision of legal abortion, with limited availability of a multidisciplinary team, the requirement for a police report or court order to provide care, and a high reliance on conscientious objection to not provide care. Almost 40% of professionals were unfamiliar with the routines of the service where they work, being unable to provide information on specific aspects of legal abortion care which made it difficult to interpret some significant differences between regions.

The highest proportion of mixed-race professionals in the North and Northeast regions, working in public services, on shifts, and as public servants probably reflects both the demographic composition and the lower availability of private services in these regions. Also noteworthy is the higher proportion of male professionals, those aged 50 or older, with children, and with longer training and experience.

Standards of care and clinical guidelines should guide the services’ performance based on scientific evidence. In this study, approximately one-third of professionals were unaware of the existence or content of the Brazilian Ministry of Health’s standards of care, and only 60% reported the existence of clinical guidelines in their services, with just over half of them using them routinely. The high proportion of professionals unaware of the characteristics of abortion care in their services reinforces their lack of familiarity with existing guidelines.

Limited training during medical school or in the services on topics related to the use of MVA and legal abortion care are noteworthy and have been previously reported16,20. This is reflected in inadequate knowledge about conscientious objection and specific aspects of MVA, such as its preferential indication for infected abortions19. Knowledge about misoprostol was high, but there were less knowledge and agreement about using misoprostol at home. This is likely due to the Brazilian context, where misoprostol use is restricted to hospital use, but may also reflect a lack of awareness of current guidelines recommending the use of misoprostol, with or without mifepristone, in primary care or at home, and with care provided by non-physician professionals3. Professionals’ limited knowledge of this type of care, as well as the political persecution of legal abortion services that try to implement outpatient abortions, hinders the expansion of access to legal abortion in the country or even medical harm reduction practices related to unsafe abortion21.

The low use of MVA had already been previously reported by obstetricians in a nationwide study conducted in 2001/200216but the lack of information on the reasons for non-use makes it difficult to compare the results. In this study, across all regions, the low routine availability of MVA, with high use when available, indicates that the lack of supplies is the main limiting factor for its expanded use in the country. Data from the SUS Hospital Information System (SIH/SUS), a system that records publicly funded hospitalizations in Brazil, indicate that over 90% of surgical uterine evacuation procedures in the country are performed through curettage22, a much higher rate than that observed in other Latin American countries7. However, in a study conducted in 20 hospitals located in three Brazilian states, the use of MVA was higher, observed in 32% of uterine evacuation cases, suggesting variable use across hospitals23. Considering the advantages of MVA, efforts should be made to ensure that supplies are available in Brazilian services, as the most recent WHO guideline3recommends against the practice of dilatation and sharp curettage for surgical abortion when gestational age is < 14 weeks. Also noteworthy is the low use of electric aspiration, when MVA is absent, as an alternative to uterine curettage, due to the lower risk of complications18.

In addition to limited healthcare resources, there is a lack of empathy among professionals, prejudice among staff, and the report of women who likely terminated their pregnancies to the police, which demonstrates the impact of legal restrictions to pregnancy terminations, as well as the social stigma surrounding abortion10. Studies in Brazil and other countries have highlighted negative experiences reported by women hospitalized for abortion complications7,23,24, with a high prevalence of anxiety and stress symptoms25.

Regarding legal abortion, the less favorable attitude towards expanding legal permissions for termination of pregnancy was like that observed in a study with obstetricians in 2001/200216, showing that the perception of obstetricians has not changed in the last 20 years. The professionals’ more favorable attitude toward legal termination in cases of anencephaly and risk to the mother’s life is reflected in their practice, with greater participation in these specific situations. Similarly, inadequate knowledge about the use of conscientious objection, considered correct in any situation, and their opinion on its use, with only half of the professionals claiming that it can only be used when another professional is available, is also reflected in the high proportion of professionals who use conscientious objection to avoid participating in legal abortion care. Previous studies in the country have identified similar results among medical students and medical and management professionals26,27. The WHO states that access to and continuity of comprehensive abortion care must be protected against the barriers created by conscientious objection. In contexts where it proves impossible to regulate conscientious objection without affecting the rights of people seeking abortion, conscientious objection to the provision of abortion services may become indefensible3.

The North region had the highest proportion of reported services offering legal abortion care. This result is consistent with a previous study that showed that the North region had the highest proportion of municipalities offering legal abortion care in 2019, although the Southeast region had the largest number of municipalities offering this service in the country12. This result may also have been affected by the higher proportion of public hospitals in the North region, as the provision of legal abortion services in the country is almost exclusively publicly funded12. However, it is noteworthy that the North region had the highest proportion of professionals who never participate as members of the legal abortion care team and the highest proportion who reported not maintaining professional confidentiality and reporting women who likely terminated their pregnancies to the legal authorities. The higher proportion, in this region, of professionals who had a religion and who reported the importance of religion in their lives may be one explanation for these results, as religion is frequently cited as a reason for not providing abortion care16,20,28. A higher proportion of male and older professionals may also explain this finding, as a previous study in Brazil showed greater empathy among younger and female professionals with women hospitalized for legal abortion20. Previous studies indicate that the North region has the worst indicators in abortion care, with low rates of legal abortions13; high rates of hospitalizations due to abortion complications6,22; and a high abortion-specific maternal mortality rate among women aged 10 to 49, being the only region without a declining trend between 1996 and 2022 in the country29.

The main limitation of this study is the use of a convenience sample. The characteristics of the professionals included in the study largely reflect those observed in the most recent medical demographic survey conducted in the country30. However, the sample overrepresented formally employed obstetricians, observed in only one-third of the physicians working in the country. This may limit generalizability - especially for private services, where many physicians work without formal contracts and may be less bound by institutional protocols. It is possible that the proportion of professionals who are unable to provide information about the characteristics of the service could be even higher if these professionals without formal contracts were included. Hospitals with fewer than 100 births per year were not included in this study and the results are not applicable to professionals working in these hospitals. Our study also did not assess the KAP of professionals who provide abortion care in non-hospital settings. Although current recommendations for abortion care include the use of medical abortifacients and the performance of MVA in primary care3, available data suggest high hospital use for abortion completion in Brazil, reported by nearly half of women in the three editions of the Abortion National Survey (Pesquisa Nacional do Aborto)4.

Among the strengths, we highlight the scope of the research, which is the largest study to investigate physicians’ knowledge, attitudes, and practices regarding abortion care in Brazil; the analysis by region of the country, allowing the identification of differences that can enable the development of differentiated strategies, considering Brazil’s continental size and regional diversity; and the use of a self-completed and anonymous form, aiming to avoid responses considered desirable by ensuring confidentiality and anonymity. However, the potential for social desirability bias cannot be ruled out, and self-reported knowledge, attitudes, and practices should be viewed as conservative outcomes, suggesting that abortion assistance in the country may be even more limited.

In conclusion, access to legal abortion in Brazil is limited, and legislative proposals seek to further restrict it. Medical professionals responsible for providing care tend to hold a conservative view on termination of pregnancy, despite evidence that the ban increases unsafe abortions and, consequently, maternal morbidity and mortality. Low availability of MVA was observed in all regions, and ensuring supplies is essential for expanding MVA use in the country, replacing uterine curettage in pregnancies up to 13 weeks, in accordance with the best scientific evidence3. Less knowledge and agreement about medical abortion outside the hospital setting were also observed, which may be a barrier to expanding its use in primary care and by other health professionals, as recommended by the WHO3.

Problems in existing legal abortion care services have been identified, including low availability of a multidisciplinary team, which is recommended by Brazilian standards of care19; the service requirement for a police report or court order; and the frequent reliance on conscientious objection to avoid providing care in these cases. Prejudice, a lack of empathy, and reporting women who likely terminated their pregnancies to authorities threaten the quality of care for women hospitalized for abortion, even those hospitalized with spontaneous miscarriages, due to the suspicion surrounding these women in a context of legal restrictions20.

To expand access to safe abortion, it is necessary to review the authorizations for the use and marketing of medications for medical abortion; allow other professionals, such as primary care physicians and nurses, to provide care; promote training that addresses women’s sexual and reproductive rights beyond biological and clinical management aspects of abortion care; as well as the need to implement clinical guidelines that ensure care based on scientific evidence and on women´s needs.