Implementation of the presence of companions during hospital admission for childbirth : data from the Birth in Brazil national survey

Robust evidence of the benefits of continuous support during childbirth led to the recommendation that it should be offered for all women. In Brazil, it has been guaranteed by law since 2005, but scarce data on implementation is available. We aimed to estimate the frequency and associated socio-demographic, obstetric and institutional predictors of women having companionship during childbirth in the Birth in Brazil survey. Descriptive statistical analysis was done for the characterization of companions (at different moments of hospital stay), maternal and institutional factors; associations were investigated in bivariate and multivariate models. We found that 24.5% of women had no companion at all, 18.8% had continuous companionship and 56.7% had partial companionship. Independent predictors of having no or partial companionship at birth were: lower income and education, brown color of skin, using the public sector, multiparity, and vaginal delivery. Implementation of companionship was associated with having an appropriate environment, and clear institution al rules about women’s rights to companionship. Medical Chaperones; Midwifery; Maternal and Child Health; Parturition Resumo As evidências sobre os benefícios do apoio contínuo durante o parto levou à recomendação de que este apoio deve ser oferecido a todas as mulheres. No Brasil, ele é garantido por lei desde 2005, mas os dados sobre a sua implementação são escassos. Nosso objetivo foi estimar a frequência e fatores sociodemográficos, obstétricos e institucionais associados à presença de acompanhantes durante o parto na pesquisa Nascer no Brasil. Foi feita análise estatística descritiva para a caracterização dos acompanhantes (em diferentes momentos do tempo da internação), fatores maternos e institucionais; as associações foram investigadas em modelos bi e multivariada. Vimos que 24,5% das mulheres não tiveram acompanhante algum, 18,8% tinham companhia contínua, 56,7% tiveram acompanhamento parcial. Preditores independentes de não ter algum, ou parcial, foram: menor renda e escolaridade, cor parda da pele, usar o setor público, multiparidade e parto vaginal. A implementação do acompanhante foi associada com ambiência adequada e regras institucionais claras sobre os direitos das mulheres ao acompanhante. Acompanhantes de Pacientes; Tocologia; Saúde Materno-Infantil; Parto http://dx.doi.org/10.1590/0102-311X00127013 Cad. Saúde Pública, Rio de Janeiro, 30 Sup:S140-S153, 2014 ARTIGO ARTICLE S140 PRESENÇA DE ACOMPANHANTES DURANTE INTERNAÇÃO PARA PARTO S141 Cad. Saúde Pública, Rio de Janeiro, 30 Sup:S140-S153, 2014 Introdução Existem poucas experiências humanas comparáveis com o parto e o nascimento, em termos de intensidade física, psicológica e social, ou que sejam ao mesmo tempo tão desafiadoras e transformadoras1,2. Muito além de seus aspectos fisiológicos, o parto é um evento histórico e socialmente construído, de ampla variabilidade cultural e geográfica 3. O cuidado que uma mulher recebe durante o parto, assim como a qualidade da experiência para ela, são marcadores da posição que a mulher e o bebê ocupam nestas hierarquias sociais, incluindo dimensões como classe social, raça/ etnia, estado de saúde, estado civil, capacidade física, respeitabilidade sexual e outros 4. A assistência durante o parto depende também da disponibilidade e acessibilidade dos serviços de saúde, do acesso das mulheres à informação e a redes de apoio 5, de que tipos de profissionais de saúde predominam no mercado 6, e também do entendimento por parte dos serviços do que seria a assistência adequada para aquela população específica de mulheres que atendem 3,4. Em muitos países, a influência religiosa sobre a assistência à saúde e a noção de que o sofrimento no parto é um castigo que a mulher deve suportar por ter tido relações sexuais, contribuem para uma cultura de negligência dos direitos e do bem-estar maternos, para o uso liberal de intervenções invasivas, para a ausência de privacidade e de medidas de conforto durante o trabalho e parto 4,7 e, não raro, para a hostilidade e abusos explícitos, especialmente para as mulheres situadas na parte inferior das hierarquias sociais 4,7,8,9,10. O modo como a assistência é organizada, em termos de tipo de parto, pode ser influenciado ainda pelas crenças de gênero dos profissionais sobre o que se espera das mulheres na hora do parto. Essas crenças podem contribuir para o crescente número de cesarianas, especialmente com a associação de parto vaginal com o comportamento “não-feminino”, como a perda de controle, incontinência de secreções, indecência, indignidade e vulgaridade 11; e também com a crença de que o parto vaginal implica uma deformação genital, estando em conflito com a função adequada da vagina, que seria o fornecimento do prazer heterossexual 11,12. Essas crenças negativas contribuem para a ideia de que o parto é um evento repulsivo e constrangedor, que não deveria ser testemunhado por outra pessoa que não aqueles estritamente necessários no momento 7,12. Quando o parto hospitalar foi estabelecido como a regra pelos programas de saúde em meados do Século 20 nos países industrializados, pela primeira vez na história, a maioria das mulheres começou a parir sem a presença de qualquer pessoa que lhe fosse familiar 2,7. Essa tem sido a situação em muitos países nas últimas décadas, uma vez que a presença de familiares era geralmente proibida no parto institucionalizado 13. No Brasil, essa foi também a regra, e fez parte da formação médica e de enfermagem. Em seu influente livro, Jorge de Rezende 14 (p. 233) recomenda: “Os familiares nervosos e as pessoas que, em grande número, vão às Maternidades como acompanhantes, devem ser afastados”. Os profissionais de saúde tendem a subestimar a importância da experiência do parto e se concentram principalmente nos seus resultados, como morbidade e mortalidade. Mas mesmo para mulheres saudáveis dando à luz bebês saudáveis, a experiência do parto é poderosa e duradoura, com consequências para sua saúde física e mental, autoimagem, relacionamento com o bebê, com o companheiro(a) e com a família 1,2,5,15. Embora algumas mulheres relatem a experiência do parto e nascimento como um evento empoderador 2,16, de êxtase ou até mesmo orgástico, para a maioria das mulheres a experiência do parto pode ser dolorosa e muito estressante. As causas de estresse e ansiedade incluem: a falta de familiaridade com os profissionais de saúde, com o ambiente hospitalar e suas rotinas, a dor e o desconforto (do próprio nascimento ou das intervenções), a preocupação e a incerteza sobre o que está acontecendo, as dificuldades de comunicação, a falta de empatia, o tratamento hostil, e a solidão 2,10,15. Na década de 1990, como resultado do ativismo das mulheres e da emergência da medicina baseada em evidências, um movimento mundial começou a documentar os benefícios emocionais e de saúde e alta satisfação materna, com a presença e apoio contínuo durante o parto. Grupos de pesquisadores e ativistas, em vários países, organizaram ensaios clínicos para randomizar mulheres com e sem acompanhantes. Esses estudos e as revisões sistemáticas decorrentes documentaram os muitos resultados positivos para a saúde materna e neonatal dessa intervenção simples. Isso levou à recomendação internacional pela Organização Mundial da Saúde (OMS) na década de 1990, de que “o apoio contínuo durante o trabalho tem benefícios clinicamente significativos para as mulheres e crianças e nenhum prejuízo conhecido, e que todas as mulheres devem ter apoio durante o parto e nascimento” 17 (p. 2). O acompanhamento, por um membro da família ou por uma doula (pessoa treinada para este papel), pode incluir o apoio emocional Diniz CSG et al. S142 Cad. Saúde Pública, Rio de Janeiro, 30 Sup:S140-S153, 2014 (presença contínua, encorajamento), informações sobre o progresso no trabalho e sobre técnicas para lidar com o momento, medidas de conforto (como toque, massagem, banho quente de aspersão ou imersão, ajuda para manter-se hidratada ou ir ao banheiro) e advocacy (ajudar a mulher e o(a) companheiro(a) a expressar seus desejos e necessidades para os demais) 13. As teorias que tentam explicar os efeitos do apoio contínuo sobre os resultados do parto, utilizam a hipótese de que este apoio otimiza a fisiologia do parto e os sentimentos de controle e competência das mulheres, reduzindo a dependência de intervenções médicas. A otimização das relações fetopélvicas pode ser promovida por meio do incentivo à mobilidade da mãe e do uso eficaz da gravidade e do suporte físico para adotar suas posições preferidas, além da recomendação de posições para situações específicas 13. A revisão sistemática mais recente sobre o apoio contínuo no parto (2011), mostra que as mulheres que tiveram esta intervenção eram mais propensas a ter parto vaginal espontâneo e menos propensas a utilizar analgesia intraparto ou para relatar insatisfação; o trabalho de parto era mais curto, estas mulheres eram também menos propensas a ter parto por cesariana ou parto vaginal instrumental, ter analgesia regional, ou o bebê ter score de Apgar baixo no quinto minuto de vida 13. Além disso, as análises de subgrupos mostraram que o apoio contínuo foi mais eficaz quando feito por alguém que não fazia parte da equipe do hospital, nem da rede social da mulher, e seus efeitos foram ainda mais intensos em ambientes em que a analgesia peridural não estava rotineiramente disponível 13. A presença de acompanhante melhora a segurança da mulher no parto, uma vez que parece aumentar a sua capacidade para se expressar, o que pode ser especialmente valioso em casos de complicações graves, ajudando as mulheres a serem ouvidas quando o atendimento de urgência é necessário 18. A presença de acompanhante pode ser considerada um indicador de segurança, de qualidade do atendimento e de respeito pelos direitos das mulheres na assistência 19. No Brasil, as evidências da primeira revisão sistemática sobre apoio no traba


Introduction
There are few human experiences that compare with the physical, psychological and social intensity of labor and birth, and that are the same time so challenging and transformative 1,2 .Far beyond its physiological aspects, childbirth is a socially and historically shaped event, with wide cultural and geographic variability 3 .
The care that a woman receives during labor and delivery, like the quality of the experience of birth itself, are markers of the woman's and baby's place in social hierarchies, including dimensions such as social class, race/ethnicity, health status, marital status, physical ability, sexual respectability, and others 4 .
Care in childbirth depends on the availability and affordability of health services, on women's access to information and support from their social networks 5 , on what kind of health providers predominate the birth market 6 , and also on the services' understanding of what is appropriate care for the particular population of women that they assist 3,4 .In many countries, religious influences on health care and the notion that suffering in childbirth is a legitimate penalty that women should endure for having sex, contribute to a culture of neglect of maternal rights and wellbeing, to the absence of privacy or comfort measures during labor and birth, to the liberal use of invasive interventions 4,7 , and sometimes to overt hostility and abuse, especially for women at the bottom end of social hierarchies 4,7,8,9,10 .
The way care is organized, in terms of mode of delivery, can be shaped by providers' and women's gendered beliefs of what is expected from women in relation to birth.These beliefs may contribute to the growing number of c-sections, particularly with the association of vaginal birth with "nonfeminine" behavior, such as loss of control, messiness, leakiness, indignity and vulgarity 11 ; and also with the belief that vaginal birth implies a form of genital deformation, conflicting with the appropriate function of the vagina, which would be to provide heterosexual pleasure 11,12 .These negative beliefs contribute to the idea that childbirth is a disgusting and shameful event that should not be witnessed by anyone other than those strictly necessary at that moment 7,12 .
When hospital birth was established as the rule by health programs in industrialized countries in the mid-20 th Century, for the first time in history most women started to deliver without the presence of a person familiar to them 2,7 .This has been the usual situation in many countries for decades, as the presence of relatives was usually prohibited in institutional births 13 .In Brazil this was also the rule, and a part of medical and nurs-ing training.In his influential book, De Rezende et al. 14 (p.233) recommended: "The family members of the laboring woman, who anxiously go to the maternity wards in great numbers, should be kept distant".
Providers tend to underestimate the importance of the experience of childbirth and focus mainly on its outcomes, such as morbidity and mortality, but even for healthy women giving birth to healthy babies, the experience of childbirth is powerful and long-lasting, with consequences for women's physical and mental health, self-image, relationship with the baby, partner and family 1,2,5,15 .Although some women report the experience of labor and birth as an empowering, ecstatic or even orgasmic event 16 , for most women, the experience of childbirth can be stressful and painful.Causes of stress and anxiety include lack of familiarity with providers, hospital environment and routines, pain and discomfort (from birth itself or from interventions), worry and uncertainty about what is happening to them, problems of communication, lack of empathy, rude treatment, and loneliness 2,10,15 .
In the 1990s, as a result of women's activism and of the emergence of evidence-based healthcare, a worldwide movement started to document the emotional and health benefits, and high maternal satisfaction, with continuous presence and support during childbirth.Groups of researchers and activists in favor of companionship in several countries organized clinical trials to randomize women with and without companions.These trials and the subsequent systematic reviews of them, built the gold-standard evidence documenting the many maternal and neonatal positive health outcomes of this simple intervention.That led to the international recommendation by the World Health Organization (WHO) in the 1990s, that "continuous support during labour has clinically meaningful benefits for women and infants and no known harm, and that all women should have support throughout labor and birth" 17 Companionship, by a family member or by a doula (a trained labor support person) can include emotional support (continuous presence, reassurance and praise), information about progress in labor, advice regarding coping techniques, comfort measures (such as touch, massage, warm baths/showers, promoting adequate fluid intake and output) and advocacy (helping the woman and partner to articulate their wishes and needs to others) 13 .Theories that try to explain the effects of labor support on childbirth outcomes hypothesize that labor support enhances labor physiology and mothers' feelings of control and competence, reducing reliance on medical interventions.Enhanced fetopelvic relationships may be accom-plished by encouraging mobility and effective use of gravity, supporting women to adopt their preferred positions and recommending specific positions for specific situations 13 .
The latest systematic review of continuous support on childbirth (2011) shows that women allocated to this intervention were more likely to have a spontaneous vaginal birth and less likely to have intrapartum analgesia or to report dissatisfaction; their labors were shorter, they were less likely to have a caesarean or instrumental vaginal birth, regional analgesia or a baby with a low 5-minute Apgar score 13 .In addition, subgroup analyses suggested that continuous support was most effective when provided by someone who was neither part of the hospital staff nor the woman's social network, and its effects were even stronger in settings in which epidural analgesia was not routinely available 13 .Having a companion improves women's safety in childbirth, since it promotes women's ability to speak up about their needs, which can be particularly valuable in cases of serious complications, helping women to be heard when urgent care is needed 18 .Having a companion can be considered an indicator of safety, quality of care, and respect for women's rights in maternal care 19 .
In Brazil, the evidence of the first systematic review on support in labor (1996) was used to propose the first laws about the right to companions at birth in the 1990s, such as those adopted in São Paulo, Rio de Janeiro, and Santa Catarina 20,21,22 .In 2005, a national law which applies to all women in public and private services (n.11,108 23 ) affirmed the right of all women to have a companion of their choice during antenatal care, and during all stages of hospital stay for birth, including labour, delivery and post-partum periods.Social movements promoting the humanization of childbirth and male enhanced participation in fatherhood led to new public policies involving men, including the fathers' antenatal care programs and campaigns to promote their presence in childbirth 24,25 .
Although companionship in birth is an official maternal health policy of the Brazilian Ministry of Health, known as the Rede Cegonha, the pace and extent of implementation is unknown, since until recently no national data on companionship were available in the Brazilian Health Informatics Department (DATASUS).Our hypothesis is that there is an increase in the incorporation of the presence of companions during hospital stay, but this is unequal and related to ability to pay; and women who are at the bottom of the social hierarchy -poorer, less educated, non-white, having vaginal births, and delivering in services that refuse to implement the law, are more frequently deprived of the benefits of companionship during birth.
The aim of this paper is to analyze the implementation of accompaniment during hospital admission, labor, delivery and postpartum, in the Brazilian public, mixed and private sectors; to identify who is the companion; and the maternal, institutional and environmental factors (privacy, type of beds, accommodations), associated with the presence of companions in those sectors.

Methods
Birth in Brazil is a national hospital-based study of postnatal women and their newborn, conducted from February 2011 to October 2012.
The study sample was selected in three stages.The first was composed of hospitals with 500 or more births/year stratified by the five country regions, location (state capital or not), and type of hospital (private, public or both).The second was composed of days (minimum of seven days in each hospital) and the third composed of postnatal women.In each of the 266 hospitals sampled, 90 postnatal women were interviewed, totaling 23,940 subjects.More information about the sample design is detailed in Vasconcellos et al. 26 .
In the first wave of the study, interviews were conducted with postnatal women during hospitalization and data extracted from the medical records of the mother and newborn and antenatal notes photographed.Follow-up telephone interviews were conducted before six months and at twelve months after birth to address maternal and newborn outcomes.Detailed information about data collection is reported in do Carmo Leal et al. 27 .
Data on income level/social class were retrieved using the ABIPEME score (http://www.abipeme.org.br), which divides the population based on family level of consumption; it is presented in 5 classes (from A, the highest to E the lowest).The covariates included: age (12-19 years, 20 to 34 years and 35 years or more), skin color (self-reported by the woman: white, black, brown), education (incomplete or complete elementary school, high school and college), socioeconomic score ABIPEME (class A/B, C and D/E), mode of delivery (vaginal/cesarean), region (North, Northeast, Southeast, South, Central), and source of payment.Women who delivered in public health care facilities and women who delivered in mixed health care facilities that were not paid by health insurance policies were classified as "public source of payment".Women whose delivery was paid through a health insurance policy, and the delivery occurred in mixed or private hospitals, and women who delivered in private facilities, regardless of whether the delivery had been paid or not by the health insur-ance policy, were classified as "private source of payment".
Covariates of the health service structure included: if the service had any accommodation for the companion (at least a chair), the place that the woman stayed during labor (surgical theater, obstetric center, room, infirmary), if it was a training hospital, if it was accredited as a Babyfriendly Hospital, if it was accredited as a Galba de Araújo (best practices award) hospital, if the hospital had a clear policy of respecting the law (companions allowed during all periods of childbirth care), if privacy during care was present at all stages of care.
The presence of companions was assessed using two instruments: medical records and face-toface interviews.In the medical records, information on companionship was missing from 71.2% of the records, reflecting the limited importance given to this intervention, and the lack of service accountability to report whether or not it complied with the law.On the other hand, reporting on faceto-face interviews was very complete and reliable (0.2% missing), so this was the instrument used in this analysis.Women were asked (1) if they had a companion during their hospital stay; if not, why not; (2) who the companion was; (3) if the companion was the one chosen by them; (4) if they considered that having a companion during birth was helpful, and to what extent, (5) if the companion was present during all stages of their hospital stay, including (a) admission, (b) labor, (c) delivery, (d) immediate post-partum/recovery, and e) rooming-in.We aimed to distinguish between "visit" (present in regular visiting hours) and companion (during specific moments).Data on women´s socio-demographic and obstetric history were retrieved from medical records and face-to-face interviews.Data on institutional ambience and routines were retrieved from the hospital structure and process form.
The outcome variable in this study ( 5) "if the companion was present during all stages of her hospital stay" was later recoded with the alternatives: "no companionship at all", for those who had a total absence of companions in every stage of hospital stay; "partial companionship" including women who had any companion in some stage of stay (admission or labor or delivery or immediate post-partum/recovery, or rooming-in, (this alternative can also be called "partial absence"); and "continuous presence/companionship" for those who had companions according to what is defined in the Brazilian legislation (during all stages of hospital stay).To recode these variables, women who did not go into labor (elective cesareans) were excluded from the analysis of having "companionship during labor".
A multinomial regression model 28 was fitted to estimate the odds ratio (OR) of each covariate on "no companion" and "partial companion" as compared with "continuous companionship".In all analyses, we adopted a significance level of 5% considering the characteristics of the complex plane sampling: strata, conglomerate and weighting.The statistical program used was IBM SPSS, version 19.0 (IBM Corp., Armonk, USA).
This research was guided by Resolution n. 196/96 of the Brazilian National Health Council, which provides guidelines and standards of human research under the research protocol CEP/ ENSP (n.92/10).All directors of institutions and postpartum women signed an informed consent form.

Results
Table 1 shows the descriptive statistics for this survey.All women were eligible for the questions on companionship during childbirth.Of them, 24.5% had no companion at all and 75.5% had some form of companionship during their hospital stay: 18.8% had continuous companionship and 56.7% had partial companionship.
Considering only women who had a companion at any stage, this was most frequently the woman's partner (35.4%), followed by the mother (26.3%), sibling or friends.Doulas were present in 0.1% of cases.Of those women who had a companion, 92.8% of them were reported as being the one of the women's choice.
Women were more frequently accompanied during admission (70.1%), but more rarely during labor (42.1%, considering only those women who went into labor).During the moment of delivery itself, 32.7% had a companion.In the immediate post-partum stage, 36.9% had a companion, and 61.3% in the post-partum apartment or roomingin ward.
The vast majority of women considered that having a companion during labor and birth "was very helpful" or "helpful" for women in order to have a better and calmer birth experience (91.2% of valid responses); only 2.7% of the total sample thought it was not helpful and made the woman more anxious.
For those women who did not have a companion (data not shown in table), reasons given for not being accompanied were chiefly institutional noncompliance with the national law, mainly "no kind of companions admitted at the hospital" (52%), and other forms of restrictions (only for cesareans, or for adolescents, for female companions, for those who attended a course, for those who paid extra, etc.).Another cause was "she did not have someone to stay with her" (18%), either because her partner had to stay with the other children, or for lack of a social support network, or because of "unexpected admission for birth", in situations when the patient came by herself just for an antenatal check and had an immediate indication for a cesarean.Some women who had been transferred from other services reported access obstacles such as "the ambulance did not allow bringing in the companion".
There were variations in the lack of information about companionship received by women before labor, as women reported that they "did not know it was allowed", "did not know it was allowed in vaginal births", "did not know it was allowed for non-adolescents".Only 5.7% of unaccompanied women (1.4% of the total number of women) said that they were alone because they did not want to have any companion (data not shown in table ).
There was a large variation in terms of the implementation of accompaniment among regions: 23.1% and 22.6% of women had companionship according to the law in the Southeast and South regions (the best results), and only 11.7% in the North.The Central region had the worst results, with over 38.9% of women with no companionship at all (Figure 1).
Table 2 shows the bivariate analysis for the presence of companions according to maternal socio-demographic and obstetric characteristics."Partial companionship" showed relatively little variation from the average of around 55%, with the relevant differences concentrated in the extremes (continuous companionship and no companion at all).Women more frequently had no companionship if they had had a vaginal birth, lived in the Central region, had lower incomes, fewer years of education, were black or brown, multiparous and used the public system.Women's form of payment for childbirth care was highly associated with having or not having any companion: while in the public sector 29.5% had no companion, in the private sector it was 4.7%.Women's marital status and the kind of provider were not statistically different in the odds of companionship; maternal age was not significant in this bivariate analysis.
We investigated the role of hospital structure and processes in their association with the presence of companions in a bivariate analysis (Table 3).As expected, the services which allowed women to have companions throughout the hospital stay, which had minimal accommodation facilities (such as chairs for all companions), who provided minimal privacy for women during their stay, those accredited "Baby-Friendly Hospitals" and those services which received the Galba de Araújo award for humanized services, were less associated with having "no companionship" and more with "continuous companionship".Having a high risk unit, a Pregnant Woman Support House, if the service was a training hospital, the place where the birth occurred, and type of provider were not associated with companionship.We found that in services where the health manager says that all women have the right to have a companion during all her hospital stay, only 32.5% of women report having one.If a woman wants to have a companion according to the law, and had someone to stay with her, the safest places to go are the Galba de Araújo accredited hospitals, with 46.2% of women accompanied according to the interviews.
Table 4 presents the comparison between "continuous companionship", as the reference, and having "no companion" (total absence of companions) in the first block, and "partial companion" (partial absence of companions) in the second block.The maternal factors associated with having higher chances of total absence of companions were: vaginal delivery (OR = 1.6), living in the Central region (OR = 5.0), social class (OR = 2.8 for D + E and OR = 1.8 for class C), maternal education (OR = 1.8 for up to 11 years), ethnicity (OR = 1.3 for brown), multiparity (OR = 1.6) and use of the public sector (OR = 13.3).The factors of the hospital structure associated with higher chances of total absence of companions were: not having a chair for the companion (OR = 3.4) or only for some beds (OR = 2.0), not having a policy of allowing companions (OR = 4.1), not being a Baby-  Friendly Hospital (OR = 2.3), not having a Galba Araújo award (OR = 4.1) and having never applied for this award (OR = 3.9).The maternal factors related to having higher chances of having a companion only at some point (partial absence) were: vaginal delivery (OR = 2.5), living in the Central region (OR = 2.1), social class (OR = 2.2 for D + E and OR = 1.5 for class C), maternal education (OR = 1.2, between 8 and 11 years), ethnicity (OR = 1.2 for black and brown ), multiparity (OR = 1.2) and using the public sector (OR = 3.2).The factors related to the structure of the hospital associated with higher odds of partial absence of a companion were: not having a policy of allowing companions (OR = 2.3), not being a Baby Friendly Hospital (OR = 1.8), not having a Galba Araújo award (OR = 2.9) and having never applied for this award (OR = 2.7).

Discussion
In this survey, approximately one quarter of women had no companion at all, less than one in five had continuous companionship, and 55.2% had partial companionship.The woman's partner was present in one in four births in the total sample, and in 35.4% of the births when women had any companions.Women with lower incomes and those who were less educated, black or brown, and who used the public sector had fewer companions at any moment of birth.The vast majority of women reported that having a companion was very helpful or helpful.Of those women who did not have a companion, this was mostly because of hospital prohibition; only 1.4% of the total number of women did not want to have a companion.Implementation was associated with change in institutional culture and rules, such as having * Reference: continuous presence of a companion at all moments of hospital stay for birth (as defined by the legislation).
a clear policy allowing companions for all women at all stages of care, environment and changes in facilities such as having chairs for all companions.The quality of information in medical records was poor, and asking women about their experience was a reliable source of information, which helped to compensate for these deficiencies.
In another analysis in the Birth in Brazil survey, focusing on women's satisfaction 29 , with regard to women's experience of abusive treatment during childbirth care, it was found that women who went into labor reported a higher incidence of physical, verbal or psychological violence, compared with those who did not go into labor.They found that the presence of a companion significantly inhibited the occurrence of all forms of violence for women attending publicly funded care.That is, if a woman had a companion present, she was less vulnerable to violence, even in a public service and going into labor.
Qualitative studies about public services show that women know that there is a right to companionship, but they do not trust that services will respect their right, and fear confrontation and retaliation during their hospital stay if they insist on having a companion, based on the reports of the many obstacles they hear about from the experiences of relatives and friends 30,31 .
Qualitative research about the experience of companions in the private sector shows that even respondents who said that they were "present in all stages of the hospital stay", also reported that they were present at "some moment of each stage", not necessarily during "all times in that stage".For example, companions were with the woman at admission, but had to separate from her when she was filling in papers; they then stayed for part of labor, and at the moment of birth, most partners were not allowed to stay with the woman until the very moment of delivery itself, usually left the room with the newborn, so women would stay without a companion during the rest of the surgery, and then most women stayed alone for most of the immediate postpartum period.In qualitative studies, the immediate postpartum period was considered by some women who had cesareans as the worst part of the experience of childbirth, as they find themselves alone and helpless in a separate room with no news about the baby, and not being able to move or search for help because they were under the effects of analgesia 30,32 .This can be the reality of those who had the best outcomes in this research, classified as "continuous companionship".
Qualitative studies indicate that the role of companions in Brazil can be very different from the active, helping-hand birth partner envisioned in other settings and clinical trials, since generally male companions have their potential contri-butions very restricted 31,32 .In vaginal birth, they are frequently only allowed to stay still beside the woman's head, are openly advised not to see the stretching of the vagina, the delivery itself or sutures, or the expelling of the afterbirth, because of the risk of losing sexual interest for the birthing woman; these commands are often made in front of the birthing woman 11,12,19,31 .This can partly explain why there are more companions in cesarean sections, when those constraints related to notions of women's decency and shame are prevented by an ordained, de-genitalized birth.
Even in such contexts, where little active participation is expected or admitted from companions -as from women themselves -this passive presence is highly valued, even if just as a witness of this life-changing moment.In our research, almost all women found this presence very helpful and thought that it contributed to making birth a better, calmer experience.In another analysis of the Birth in Brazil survey, women who had companions reported more satisfaction with care, received better information and felt more respected by providers, as well as feeling more protected from any form of violence during the experience 29 .
Fifteen years after the first official initiatives to promote companionship for all women during labor and birth in Brazil in the 1990s, there are advances to celebrate, but much to be improved, especially in terms of equity.There is a clear increase in companionship in both public and private sectors, if compared with the first national data available, the Brazilian National Survey of Demography and Health (PNDS 2006) 27 .With data from births between 2001 and 2005, this survey showed that only 16% of women had any companion, 9% in the public sector, 32% in the private.However the way the question was asked in the PNDS ("Did you have a companion during childbirth?") does not make it clear if this referred to any moment of the hospital stay for birth, or during delivery itself.The recent initiative of the Stork Network ombudsman will help to provide real time information on implementation of companionship and other policies to improve maternity care.
Unfortunately, the ambivalent wording of Law n. 11,108/05 in practice is interpreted as if doctors and services have the authority to violate the law.There are no sanctions if health authorities or providers refuse to allow a companion, justifying their disregard of these rights based on their "medical autonomy", or on the lack of privacy on the wards, or that they did not have time to adjust to the new legislation 33 .This can hardly be considered justifiable after nine years of adjusting to the national law (2005).The absence of any mandatory reporting, or sanctions for ignoring women´s rights contribute to the low and incomplete implementation rates.Some local experiences in Brazil show that it is possible to ensure women's rights during childbirth as a public policy.In the city of Belo Horizonte, the capital of Minas Gerais State, this was guaranteed -before and differently from the National Law for all women -in the public (Brazilian Unified National Health System -SUS) and private sectors.This is an example of the power of the political will of local policy-makers: differently from the rest of the country, in Belo Horizonte, women in the public sector had more access to companionship during all the period of hospital stay for birth than those in the private sector, ranging from 75% (at the university hospital) to 98.3% (at a Birth Center) while in the private sector, these levels of implementation have not been achieved 34 .
One limitation of this study is that as the instrument used did not allow the identification of women who gave birth paid by direct disbursement, it is possible that some women had their delivery assisted in mixed health care facilities and were classified as having a public source of payment, despite having paid for their delivery care.However, as these women had very similar socioeconomic characteristics to women attending public hospitals, it is likely that misclassification occurred in a few cases.As it is a nondifferential misclassification with respect to the outcomes studied, it is expected that there has been attenuation of the magnitude of the observed associations.

Conclusions
Continuous support in labor and birth is a safe, highly effective intervention in improving maternal and neonatal outcomes, with high maternal satisfaction rates, very low cost, and it is a formal right of Brazilian childbearing women, as established by Law n. 11,108/05.Although there are advances to be celebrated in implementing this policy, such as a slow but consistent improvement in the presence of companions since the latest national data were taken, most women in Brazil cannot count on having a companion during labor or delivery.
This study shows that having the continuous presence of a companion during delivery in most settings is still a privilege for women with higher income and education, who are white, pay a private provider and have a cesarean.The resistance to the implementation of companions for all women during all moments of childbirth, and to the promotion of a woman-friendly birth environment, confirms the social movements' complaint that there is a conflict of interest in the organization of childbirth care, aimed at "worsening birth to sell c-sections" 35 : if a better experience of childbirth were possible, less women would comply to having a C-section just to feel protected from violence, loneliness and indignity.At the same time, implementation depends on the political will of health authorities and service managers, in the public and private sectors, to change the traditional culture of neglect with women's wellbeing, safety and comfort during childbirth.This includes the authority and the negotiation skills to promote a culture of respect of women's rights, overcoming providers' resistance to change discriminatory culture and routines.
Companionship can be considered a marker of safety and quality of care, and also an indicator of the incorporation of several of the SUS principles, such as integrality (comprehensiveness) of health care, universality, equity, and humanization.If companionship in childbirth is to be implemented as a universal, equitative health policy, regular information on the presence of companionship should be part of service protocols and provider training, including the accountability for those services who refuse to protect and fulfill women's rights.
Positive experiences should be analysed and used to inspire action, such as those of municipalities that were able to make companionship universal by conditioning SUS reimbursement to the presence of companions, training of providers and service doulas, and informing women and their potential companions about their rights in childbirth, during antenatal care, and hospital admission.

Figure 1 Frequency
Figure 1 Frequency distribution by regions of companionship during hospital stay for childbirth.Brazil, 2011-2012.
Chaperones Médicos; Tocología; Salud Materno-Infantil; Parto Contributors C. S. G. Diniz collaborated in the study design and study coordination, proposed the data analysis and wrote the draft of the paper.E. d'Orsi, R. M. S. M. Domingues, J. A. Torres, M. A. B. Dias, C. A. Schneck, S. Lansky, N. Z. F. Teixeira, S. Rance and J. Sandall collaborated in the study design and study coordination, in data analysis and reviewed the final draft of the paper.

Table 1
Frequency distribution of companionship during hospital stay for childbirth.Brazil, 2011-2012.
* All values were corrected by the sample weights; ** Only for women who had a companion; *** Only for women who went into labor; # Only for those who had a companion during childbirth.

Table 2
Companionship during different moments of hospital stay for childbirth according to maternal factors.Brazil, 2011-2012.

Table 3
Companionship during different moments of hospital stay for childbirth according to hospital structure and processes.Brazil, 2011-2012.

Table 4
Multinomial regression model for companionship during admission for childbirth *, according to maternal and hospital characteristics.