Development of an instrument to measure mistreatment of women during childbirth through item response theory.

The objective of this study was to structure a proposal for an instrument to measure the mistreatment level of women during childbirth, through item response theory, based on the birth experience of postpartum women. A cross-sectional study was conducted, with the inclusion of 287 women who did not suffer complications during childbirth, randomly selected from two maternity hospitals in the capital of Rio Grande do Sul-Brazil, in 2016. Approximately 30 days after delivery, the women answered questions in a face-to-face interview about their birth experience (practices and interventions applied) and were inquired about their perception of having suffered disrespect, mistreatment or humiliation by health professionals. The set of practices was included in the item response theory model to design the instrument. Of the 36 items included in the model, 21 dealt with practices applied exclusively to women who went into labor, therefore two instruments were developed. The instrument including all women, containing 09 items, identified 23.7% prevalence of mistreatment to women during childbirth, while the instrument for women going into labor included 11 items and identified 22% prevalence. The items with the highest discrimination were: not having had a companion during labor (2.05; and 1.26), not feeling welcome (1.81; and 1.58), and not feeling safe (1.59; and 1.70), for all women and for those who went into labor, respectively. For those who went into labor, the items, did not have a companion during labor (1.22; PE 0.88) and did not feel comfortable asking questions and participating in decisions (1.20; PE 0.43) also showed greater discrimination. In contrast, when directly questioned, only 12.5% of women said they had experienced disrespect or mistreatment, suggesting that harmful practices are often not recognized as violent. Standardizing the measurement of mistreatment of women during childbirth can create more accurate estimates of its prevalence and contribute to the proposal of strategies to eliminate obstetric violence.

Introduction

Childbirth, an essentially physiological event, has over the years become a hospital procedure, with frequent use of routine interventions, often unnecessary and not based on scientific evidence supporting its benefits [1]. Expansion of access to health services, at population level, has provided greater safety to childbirth, which is supervised by qualified health professionals in an environment with technological resources and with the possibility of timely referral, if necessary, to higher complexity services [2, 3].

In 2019, in Brazil, 99.1% of births took place in health facilities and only 43.6% were vaginal [4]. The safety attributed to hospital childbirth care often contrasts with the risks of rigid routines and overburdened professionals in this setting, resulting in increased rates of induced labor, cesarean sections (c-sections), episiotomy, and other interventions that are frequently unnecessary [1, 5]. The combination of these factors impacts the quality of care, making it less compassionate, and hinders the women’s experience, often generating the legitimate feeling of mistreatment during childbirth.

A positive birth experience considers and incorporates women’s sociocultural beliefs and preferences. It includes the birth of a healthy baby in a safe environment with regard to clinical and psychological issues, with continuity of physical and emotional care, by a qualified technical team [5]. On the other hand, mistreatment refers to the different forms of violence that occur, whether physical, psychological, or verbal, or by unnecessary and harmful procedures, caused to women in prenatal care, childbirth, postpartum, and abortion [1, 6, 7]. The term obstetric violence is more commonly used in Latin America to refer to mistreatment during childbirth, and we agree with its use in the sense it poses a political stance, as it implies the recognition of structural problems that negatively affect women’s experiences, jeopardizing the realization of their rights [8]. However, in this paper we opted to use the term mistreatment because of its prevailing use in international scientific literature and also because the actual question made to the women participating in this study (see Materials and methods) contains the term mistreatment, and not the term obstetric violence.

Mistreatment of women during childbirth is considered a global public health problem, and its elimination appears as a priority in women’s health policies, drawn up by regulatory bodies, such as the World Health Organization (WHO)—Intrapartum Care for a Positive Childbirth Experience [5]—and the International Federation of Gynecology and Obstetrics (FIGO)—International Childbirth Initiative [9]. In Brazil, over the last 20 years there have been important governmental efforts to improve prenatal, childbirth and postpartum care, such as Programa de Humanização no Pré-Natal e Nascimento (Prenatal and Childbirth Humanization Program), instituted in 2000, and Rede Cegonha (Stork Network), in 2011. The private sector, with impressively high c-section rates (reaching around 80%) [10], received special attention in 2014, when the Projeto Parto Adequado (Appropriate Delivery Project), also called Nascer Saudável (Healthy Birth), was launched [11]. These initiatives, led by the Ministry of Health and the National Supplementary Health Agency–ANS, have shown its positive impact on childbirth care in Brazil [6, 11], however they have been weakened in recent years, especially since 2016, due to political changes in the country that have resulted in disinvestment as well as in clear disinterest in the continuity of the mentioned strategies. In this sense, the Rede Cegonha has recently lost its name and shape with the official institution of the new Rede de Atenção Materna e Infantil (RAMI)–Maternal and Child Care Network [12]. Unlike the first, this new governmental initiative puts marked emphasis on specialized care and high risk situations, while not including traditional birth attendants in any context of childbirth care. In terms of context, it’s important to note that the mentioned Childbirth Humanization Program was a governmental response not only to the epidemiological needs linked to avoidable maternal deaths but also to the demands of social and women’s movements claiming for their sexual and reproductive rights. Currently, the new RAMI likely entails risks to the progress achieved in more than 20 years.

A study developed in Brazil, from 2010 to 2013, identified a 25% prevalence in mistreatment of women during childbirth. This prevalence was 50% for women who had abortions [13]. While the research Nascer no Brasil (Birth in Brazil), conducted in the same period (2011–2013) with national representation, involving more than 16,000 postpartum women, identified a lower prevalence (5.1%) of mistreatment, abuse or violence by health professionals, as perceived by the women [14].

More recent surveys, such as one conducted in 2019 involving postpartum women from the Brazilian public and private networks, showed reduction in the rates of several interventions considered harmful when performed without precise indication: elective c-sections, episiotomy, application of manual pressure on the uterine fundus, and lithotomy position in labor. At the same time, it highlighted the increase in the prevalence of good practices, such as having a companion chosen by the woman, receiving food during labor, and the use of non-pharmacological methods for pain relief [11].

Even with current awareness of the impacts of mistreatment on quality and satisfaction with care, damaging the health of women and children [15-19], there are gaps in the understanding of this problem. These gaps are due, in part, to the absence of an indicator (or instrument) that measures mistreatment of women during childbirth, built from the perception of those who experience it, as well as the variability in how it is measured, which interferes with the estimation of its prevalence and comparability between studies [14, 18, 20]. Therefore, the aim of this study was to structure a proposal for an instrument to measure the Mistreatment Level of Women during Childbirth (MLWC), through the Item Response Theory (IRT), based on the childbirth experience reported from a sample of postpartum women. The secondary objective was to compare the prevalence of mistreatment achieved through the proposed instrument with the one measured through a specific and direct question applied to this sample of women.

Materials and methods

Study design and population

We conducted a cross-sectional study, including postpartum women who gave birth in two large maternity hospitals of Porto Alegre, Rio Grande do Sul (RS). One is a general university hospital, predominantly public, and the other is a private general hospital. Both are centers of excellence and referral institutions for both low and high-risk pregnancies, having performed, in the year 2016, 3725 and 4182 deliveries (26% of the city’s total), respectively. Reflecting the national trend, the private hospital stands for a much higher c-section rate (81%, compared with 33% in the public hospital) [4, 21]. In both facilities, deliveries are mainly assisted by obstetricians; the inclusion of midwives in childbirth care is recent and still incipient. Only the public facility, as a teaching hospital, includes medical and nursing students and residents in maternity care.

All women living in Porto Alegre who gave birth to full-term newborns in the two participating maternity hospitals without unfavorable outcomes at delivery (death or admission to intensive care) and with no formal contraindication for breastfeeding were considered eligible. These eligibility criteria were defined in order to avoid biases in the measure of the main outcomes of interest (satisfaction with childbirth and breastfeeding prevalence) in the research that originated this study [18, 22]. Women living in areas considered dangerous for home visits had to be excluded to preserve the safety of the research team.

Sample and data collection

The data used in this study derives from former research designed to identify the factors associated with women’s satisfaction with childbirth. For this objective, the sample size calculation reached 276 women. More details on the methods and results of the former research have been published elsewhere [18].

Data collection occurred between January and August 2016. Every day, all eligible women who had given birth in the previous 24 hours received a number that was used for a draw. Each day, two women from the public maternity hospital and one from the private hospital were randomly selected and included in the study until the intended sample was reached. This proportion aimed to ensure a reasonable representation in relation to the use of public and private services, described in the literature as being around 70% and 30%, respectively, at national level [23, 24].

In the period from 31 to 37 days after delivery, an interview was conducted at the home or, rarely, in another place at the woman’s preference, to apply a structured questionnaire, which was specifically designed for this study, based on the previous experience of the researchers and the guiding documents of childbirth care in Brazil [6, 25]. The questionnaire had an average application time of 55 minutes and included items related to sociodemographic characteristics, women’s health and obstetric history, prenatal and childbirth care received. The socioeconomic level was assessed according to the Brazilian Research Enterprises Association, based on the possession of a series of domestic items and on the householder’s education level [26]. The grouping of categories from A to E corresponds to a range from better off (A) to worse off (E). Name, age, education and skin color were drawn from hospital records, while all other information considered in this study was collected during face-to-face interviews and the responses were referred by the participants. Women who were not found for the interview, after at least three attempts of contact by telephone and one in person, were considered a loss.

The interviews were carried out after a pilot project had been conducted and adjustments to the questionnaire had been made. The field team was composed of 12 interviewers trained for the job. Weekly meetings were held with the field team, seeking greater uniformity in data collection.

Some potential sources of bias were minimized with randomization (selection bias), face-to-face interviews (data quality), team training for data collection and weekly monitoring (measurement bias), and timing of interviews (1 month after delivery, not too short nor too long, avoiding at the same time gratitude bias and memory bias, respectively). Data was collected manually using paper copies of the questionnaires and then entered by two independent researchers, enabling reliable verification of the database. Moreover, responses to some key questions of the questionnaire were double-checked by telephone with 5% of the participants, selected by lot.

Statistical aspects

In this study, two different ways were used to measure mistreatment. First, through a direct question, to assess the woman’s perception of having suffered disrespect, mistreatment or humiliation, using the following question: Have you ever (during labor and childbirth care) felt disrespected, humiliated or mistreated by health professionals? having as response alternatives: yes, no and don’t know/do not remember. The second consisted in creating a measurement instrument for the latent trait mistreatment level of women during childbirth (MLWC) from a survey on childbirth experience, practices and interventions applied to the sample of women included in the study.

To create the measurement instrument, the IRT model known as the Two-Parameter Logistic Model was adapted [27, 28]. In the context of this study, the model predicts probability that a woman j, with MLWC θj, will respond to category 1 of item i (items presented in Table 1), i.e., P(Xij = 1 | θj), as follows: being i = 1,2,⋯,11 and j = 1,2,⋯,n,

Table 1

Characteristics of the sample of postpartum women regarding sociodemographic aspects, lifestyle, reproductive history, prenatal and childbirth care, according to the perception of mistreatment through the question: Have you ever felt disrespected, humiliated or mistreated by health professionals? Porto Alegre, 2016.

	Sample n (%)	Felt disrespected, mistreated, or humiliated–n (%)
Predictor variables	n = 287	Yes n = 36–12.5%	No n = 251–87.5%
Sociodemographic
Age (year)
≤ 19 years	23 (8.0)	2 (8.7)	21 (91.3)
20–34 years	199 (69.3)	32 (16.1)	167 (83.9)
≥35 years	65 (22.6)	2 (3.1)	63 (96.9)
Color of skin
White	216 (75.3)	28 (13.0)	188 (87.0)
Black or brown	71 (24.7)	8 (11.3)	63 (88.7)
Socioeconomic level (n = 285) *
A–B	163 (57.2)	18 (11.0)	145 (89.0)
C–D–E	122 (42.8)	18 (14.8)	104 (85.2)
Education
College	124 (43.2)	15 (12,1)	109 (87.9)
Elementary and high school	163 (56.8)	21 (12.9)	142 (87.1)
Lives with a partner
Yes	248 (86.4)	33 (13.3)	215 (86.7)
Lifestyle and Health Status
Smoker
Current or past	92 (32.1)	11 (12.0)	81 (88.0)
Mental health condition
Current or past	38 (13.2)	6 (15.8)	32 (84.2)
Use of psychoactive medication
Current or past	55 (19.2)	9 (16.4)	46 (83.6)
Reproductive history
Previous births
One	142 (49.5)	21 (14.8)	121 (85.2)
Two	98 (34.1)	7 (7.1)	91 (92.9)
Three or more	47 (16.4)	8 (17.0)	39 (83.0)
Miscarriage history (n = 163) **
Yes	47 (28.8)	5 (10.6)	42 (89.4)
Last planned pregnancy
Yes	154 (53.7)	19 (12.3)	135 (87.7)
Prenatal Care
Number of visits (n = 282) *
≤ 7 appointments	46 (16.3)	4 (8.7)	42 (91.3)
8 or more	236 (83.7)	31 (13.1)	205 (86.9)
Companion of her choice
Yes, at least one appointment	218 (76.0)	31 (14.2)	187 (85.8)
Informed about her rights (n = 282) *
Yes, totally	152 (53.9)	15 (9.9)	137 (90.1)
Had a birth plan
Yes	15 (5.2)	1 (6.7)	14 (93.3)
Felt free to ask questions
Yes, totally	241 (84.0)	26 (10.8)	215 (89.2)
Childbirth care
Hospital status
Public	188 (65.5)	26 (13.8)	162 (86.2)
Private	99 (34.5)	10 (10.1)	89 (89.9)
Had to go to more than one maternity hospital
Yes	29 (10.1)	4 (13.8)	25 (86.2)
Had a companion
During labor	275 (95.8)	34 (12.4)	241 (87.6)
Delivery	283 (98.6)	36 (12.7)	247 (87.3)
Postpartum	275 (95.8)	34 (12.4)	241 (87.6)
Felt comfortable asking questions (n = 284) *
Yes	241 (84.9)	24 (10.0)	217 (90.0)
Understood information received
Yes	251 (87.5)	28 (11.2)	223 (88.8)
Went into labor
Yes	205 (71.4)	30 (14.6)	175 (85.4)
Had skin-to-skin contact with the baby (n = 281) *
Yes, immediately after delivery	167 (59.4)	23 (13.8)	144 (86.2)
Yes, after performance of procedures on baby	24 (8.5)	3 (12.5)	21 (87.5)
Felt welcomed in the environment (n = 281) *
Yes	220 (78.3)	23 (10.5)	197 (89.5)
Felt safe in the environment (n = 282) *
Yes	209 (74.1)	24 (11.5)	185 (88.5)
Had privacy (n = 280) *
Yes	235 (83.9)	22 (9.4)	213 (90.6)
Care provided to women who went into labor	n (%)	Felt disrespected, mistreated, or humiliated–n (%)
	205 (71.4)	Yes 30 (14.6)	No 175 (85.4)
Used pain relief methods
Yes	168 (82.0)	28 (16.7)	140 (83.3)
No	37 (18.0)	2 (5.4)	35 (95.6)
Analgesia requested but not received (n = 202) *
Yes	31 (15.3)	9 (29.0)	22 (71.0)
No	171 (84.7)	20 (11.7)	151 (88.3)
Offered liquids and light foods (n = 204) *
Yes	113 (55.4)	19 (16.8)	94 (83.2)
No	91 (44.6)	10 (11.0)	81 (89.0)
Encouraged to walk (n = 204) *
Yes	88 (43.1)	14 (15.9)	74 (84.1)
No	116 (56.9)	16 (13.8)	100 (86.2)
Trichotomy (n = 203) *
Yes	12 (5.9)	0 (0.0)	12 (100)
No	191 (94.1)	29 (15.2)	162 (84.8)
Enema (n = 203) *
Yes	7 (3.4)	0 (0.0)	7 (100)
No	196 (96.6)	29 (14.8)	167 (85.2)
Induction with oxytocin (n = 193) *
Yes, with or without consent	108 (56.0)	21 (19.4)	87 (80.6)
Yes, without consent	16 (8.3)	12 (75.0)	4 (25.0)
Amniotomy (n = 201) *
Yes, with or without consent	98 (48.8)	14 (14.3)	84 (85.7)
Yes, without consent	19 (9.5)	7 (36.8)	12 (63.2)
Fundal pressure maneuver (n = 200) *
Yes	22 (11.0)	4 (18.2)	18 (81.8)
No	178 (89.0)	24 (13.5)	154 (86.5)
Episiotomy (n = 204) *
Yes, with or without consent	72 (35.3)	11 (15.3)	61 (84.7)
Yes, without consent	21 (10.3)	3 (10.3)	18 (85.7)
Forceps (n = 200) *
Yes, with or without consent	11 (5.5)	1 (9.1)	10 (90.9)
Yes, without consent	5 (2.5)	1 (20.0)	4 (80.0)
Would like another position in labor (n = 151) *
Yes	11 (7.3)	5 (45.5)	6 (54.5)
No	140 (92.7)	20 (14.3)	120 (85.7)

*Missing data correspond to responses “I don’t know” or “I don’t remember”.

**Missing data corresponds to women that were not applicable for this response.

where X is the dichotomous item that takes the values 1 or 0;

θj represents the MLWC of the jth woman, estimated on the scale with mean zero and standard deviation 1;

b is the position parameter of item i, estimated on the same scale as θ, which represents the severity of the content measured in item i, that is, the level of mistreatment necessary for the woman to answer category 1 of item i with probability equal to 0.5;

a is the discrimination (or slope) parameter of item i—low values of this parameter indicate that women with different levels of mistreatment are about equally likely to answer category 1 of item i, and very high values of this parameter discriminate women basically into two groups: the group that has a level of mistreatment below the value of the parameter bi and the group that has a level of mistreatment above the value of the parameter bi;

D scale factor (constant, equal to 1 or 1.7).

One of the products of the IRT models are the so-called item characteristic curves (ICC). The ICCs of the two-parameter logistic model used in this study describe how changes in MLWC are related to changes in the probability of response in category "1" of each item of the measurement instrument. Items with high discrimination values (it is desirable for items to have estimates for the parameter a greater than or equal to 1) have the steepest ICCs [27].

Another product of the IRT models are the item information curves (IIC) and test information curves (TIC). The IICs describe the amount of psychometric information that each item adds to the estimate of MLWC for different levels of that latent trait. The TIC is a sum of the IICs and shows for which MLWC the created measurement instrument estimates the most accurate latent trait.

The two-parameter logistic model must satisfy two assumptions: local independence and one-dimensionality (a single latent trait is determining the responses to the items). These assumptions are related, so that if the measurement instrument is unidimensional, local independence is met [27, 29]. The assumption of one-dimensionality can be made flexible, which is known as sufficient one-dimensionality, that is, it is enough that there is a preponderant dimension (explanation proportion of the first dimension at least equal to 20%) so that the model can be used [30-32]. Verification of sufficient one-dimensionality of the proposed measurement instrument was performed by means of exploratory factor analysis, using the tetrachoric correlation matrix, since the items are dichotomous responses.

The cut-off point of the IRT score for defining mistreatment was 0.5 standard deviation above the mean. The binomial test was used to compare the prevalence of MLWC with the prevalence obtained by the perception of violence, measured through a direct question.

To create the MLWC measure, the ltm, version 1.1–1, and psych, version 2.1.6 packages were used. The remaining analyses were performed using the SPSS software, version 18.

Ethical aspects

This study complies with the standards governing research with human subjects [33] and was approved by the research ethics committees of the institutions involved (CAAE 49938015.3.0000.5327 and 46775115.0.3002.5330). All women who agreed to participate in the study signed an informed consent form.

Results

Among the postpartum women selected for this study, 379 were eligible. Of these, 287 were effectively interviewed. There were 25 (6.6%) refusals, and 67 women (17.7%) were lost due to contact failure to schedule the interviews (data in S1 Fig). The women not interviewed due to losses and refusals differed in terms of education and skin color, showing less education (p<0.01) and a higher prevalence of white skin color compared to those interviewed (p = 0.032).

To develop an instrument capable of measuring the MLWC, 36 items (variables) were initially included, considered as observable facets of the latent trait. As 21 items referred only to women who had gone into labor, including those who ended up having a c-section, we opted for the development of two measures, with the purpose of exploring the particularities related to mistreatment in each of the groups: for all women and for women who went into labor.

Due to the occurrence of missing data, lack of variability in responses and internal consistency measures (Crombach’s Alpha), several items were eliminated in the exploratory phase for the development of the instruments for measuring the two MLWC (data in S1 Table). For instance, episiotomy and amniotomy, which were highly prevalent in our sample (35.3% and 48.8% amongst the 205 women who went into labor, respectively–Table 1), were eliminated because they did not contribute psychometric information in the estimation of the latent trait.

Due to the sample size and the low frequency of response in specific categories, the items (variables) were dichotomized (data in S2 Table). In the final model, the measurement instrument covering all postpartum women consisted of nine items (MLWC1), while the measurement instrument covering only women who went into labor consisted of 11 items (MLWC2)—data in S2 Table. The assumption of sufficient one-dimensionality was met in both measures. For MLWC1, the first factor explained 32% of the total variance among the nine items of the measure, and for MWC2, the first factor explained 31% of the variance among the 11 items of the measure.

The prevalence of mistreatment was 12.5% and 23.7% (p<0.001), for perceived disrespect, mistreatment or humiliation, measured through the direct question; and for the MLWC, constituted through a set of variables, respectively. When only the women who went into labor were considered, these proportions were 14.6% and 22.0% (p = 0.002), respectively. Table 1 shows the characteristics of the sample, which was composed mostly of women aged 20–34 years, white, with a partner and with frequent prenatal care (more than eight visits). Regarding the place of delivery, 21.7% of women did not feel welcome in this environment, 25.9% did not feel safe, and 16.1% reported having no privacy.

Of the 287 women studied, 71.4% went into labor. The induction or acceleration of labor with oxytocin occurred in more than half of the women, similar prevalence to the occurrence of amniotomy, and about 10% of these women did not consent to these procedures. As for the delivery position, 99.3% gave birth in the lithotomy position, and only 7.3% of the women interviewed said they would like other positions. Exposure of women to fundal pressure maneuvers occurred in 11.0% of deliveries (Table 1).

Forty-eight percent of the women in our sample went through a c-section and almost 29% did not initiate labor (were submitted to an elective c-section). Table 2 presents c-section rates according to key characteristics: hospital status (public x private), women’s age, skin color, education level, socioeconomic level, parity, and feeling mistreated, disrespected or humiliated. Higher c-section rates were associated with private hospital, as well as with older, white and highly educated women, with a less vulnerable socioeconomic level. The opposite associations were found for women who went into labor. C-section rates were lower amongst women who did not feel mistreated.

Table 2

C-section rates and proportion of women who went into labor according to key characteristics (n = 287).

Porto Alegre, 2016.

	C-section	p-value*	Went into labor	p-value*
	n = 138 (48.1%)		n = 205 (71.4%)
	n (%)		n (%)
Hospital status		<0.001		<0.001
Public	60 (31.9)		164 (87.2)
Private	78 (78.8)		41 (41.4)
Age (year)		<0.001		<0.001
< 35 years	89 (40.1)		172 (77.5)
≥35 years	49 (75.4)		33 (50.8)
Color of skin		0.056		0.015
White	111 (51.4)		146 (67.6)
Black or brown	27 (38.0)		59 (83.1)
Education level		<0.001		<0.001
College	83 (66.9)		66 (53.2)
Elementary and high school	55 (33.7)		139 (85.3)
Socioeconomic level (n:285) **		<0.001		<0.001
A–B	97 (59.5)		96 (58.9)
C–D–E	41 (33.6)		107 (87.7)
Parity		0.409		0.191
Primiparous	72 (50.7)		96 (67.6)
Multiparous	66 (45.5)		109 (75.2)
Feeling of mistreatment		0.032		0.114
Yes	11 (30.6)		30 (83.3)
No	127 (50.6)		175 (69.7)

*Fischer’s Exact test.

**Missing data correspond to responses “I don’t know” or “I don’t remember”.

Table 3 describes the severity and discrimination measures for each item analyzed by IRT. For MLWC1, the items with the highest severity were the following: not having had a companion during labor (5.18; standard error—SE 2.95), delivery (3.01; SE 0.79), postpartum (4.47; SE 2.13), and not having understood information given by health professionals (3.11; SE 1.15).

Table 3

Parameter estimates of items of the measuring instrument for Mistreatment Level of Women during Childbirth (MLWC1) in postpartum women in public and private hospitals (n = 287).

Items	Items (variables)	Severity*	Discrimination*
	Did not have a companion
1	During labor	5.18 (2.95)	0.64 (0.41)
2	In postpartum	4.47 (2.13)	0.76 (0.42)
	Understood information given
3	No or not all	3.11 (1.15)	0.68 (0.28)
	Did not have a companion
4	At delivery	3.01 (0.79)	2.05 (1.08)
	Had privacy during birth
5	Little or not at all	2.48 (0.81)	0.73 (0.27)
	Felt comfortable to ask questions and participate in decisions
6	No or sort of	2.06 (0.52)	0.99 (0.32)
	Felt welcomed in the environment
7	More or less, little or not at all	1.07 (0.20)	1.81 (0.56)
	Felt safe in the environment
8	More or less or not at all	0.94 (0.19)	1.59 (0.49)
	Had skin-to-skin contact in the delivery room
9	No or only after procedures	0.93 (0.19)	0.43 (0.19)

* Point Estimation (Standard Error)

For the MLWC1, the characteristic curves (ICC) and the information curves (IIC) of the items ’No companion at delivery’ and ’No privacy at delivery’ show that the higher the MLCW, the greater the likelihood of the woman answering that she had no companion at delivery and no privacy at delivery (Fig 1). The difference is in the slope of the characteristic curves of the two items, with the item ’No privacy at delivery’ having the less steep curve, because this item has a much lower discrimination ability (0.73—Table 3) than the other item (2.05—Table 2). The two IICs reflect this difference. They present the amount of information that each item provides to the estimation of the latent trait MLWC1. The item with the higher discrimination ability has its peak at a higher amount of information (approximately at 0.9) than the other item, which has its peak at a lower amount of information (approximately 0.13).

Fig 1

Characteristic and information curves of the items "no companion" and "no privacy" at delivery for the MLWC1 model (n: 287).

Fig 2 presents the standard error of the MLWC1 measure, i.e., the precision of the estimate of the latent trait MLWC1 for any level of this trait. It can be seen that postpartum women whose MLWC1 value is in the range of approximately 0.5 standard deviation above the mean will have more accurate estimates of this trait than those whose MLWC1 is below this value. This means that the proposed measurement instrument performs better in postpartum women with higher mistreatment levels.

Fig 2

Standard error for the MLWC1 model (n: 287).

For MLWC2 (women who went into labor, Table 4), the items with the highest severity were as follows: not having had a companion in the postpartum period (4.98; PE 3.18), having undergone the fundal pressure maneuver (4.84; PE 3.19), and not having had a companion during labor (4.40; PE 2.37) (Table 4).

Table 4

Parameter estimates of the items of the measuring instrument for Mistreatment Level of Women during Childbirth (MLWC2) in postpartum women who went into labor (n: 205).

Items	Items (variables)	Severity*	Discrimination*
	Had no companion
1	In postpartum	4.98 (3.18)	0.72 (0.52)
	Underwent fundal pressure maneuver
2	Yes	4.84 (3.19)	0.45 (0.31)
	Did not have companion
3	At delivery	4.40 (2.37)	1.22 (0.88)
	Did not have companion
4	During labor	3.36 (1.10)	1.26 (0.58)
	Understood information given
5	No or not all	3.12 (1.37)	0.64 (0.31)
	Asked for analgesia and was not assisted
6	Yes	2.55 (0.93)	0.74 (0.31)
	Had privacy during delivery
7	Little or none	2.53 (1.01)	0.65 (0.29)
	Had skin-to-skin contact with newborn
8	No or only after procedures with newborn	1.57 (0.66)	0.57 (0.25)
	Felt comfortable to ask questions and participate in decisions
9	No or more or less	1.50 (0.40)	1.20 (0.43)
	Felt welcomed in the environment
10	More or less, little or nothing	1.10 (0.24)	1.58 (0.51)
	Felt safe in the environment
11	No or more or less	0.95 (0.22)	1.70 (0.62)

* Point Estimation (Standard Error)

With regard to discrimination, both in the general sample of women and for the group of those who went into labor, it was higher for the following items: not having had a companion during labor (2.05 general sample; 1.26 went into labor), not feeling welcomed (1.81 general sample; 1.58 went into labor), and not feeling safe in the maternity environment (1.59 general sample; 1.70 went into labor). In the group that went into labor, the items, did not have a companion during labor (1.22; SI 0.88) and did not feel comfortable to ask questions and participate in decisions (1.20; SI 0.43) also showed greater discrimination.

Discussion

The prevalence of mistreatment of women during childbirth found in this study, measured in two different ways, were 12.5% using a direct question and 23.7% through an instrument containing a set of items (factors). This variability is due to the use of different ways to measure the same phenomenon and reinforces the differences found in the literature regarding the prevalence of mistreatment [13, 14, 20]. In the international scenario, a systematic review with Latin American and Caribbean studies published between 1990 and 2017, identified a prevalence of 43% of disrespect and mistreatment in childbirth care [34] while another, without geographic restriction, found proportions that varied from 15% to 98% [35].

The lack of standardization in the definition of mistreatment of women during childbirth, sometimes also referred as obstetric violence, the different instruments used for its measurement and the methodological weaknesses of the studies introduce potential systematic errors for this measure, aspects that affect the generalization and comparability of estimates [35]. Some authors also suggest that this disparity also occurs due to the non-recognition, by women, of situations considered violent according to the definitions supported by official bodies [19], such as the WHO [5] and the Brazilian Ministry of Health [6], as well as the scientific community [1, 36].

A Brazilian study [19] with over 550 postpartum women showed that 51.7% of them categorized their knowledge about obstetric violence as none, little or reasonable. In this study, 12.6% of women reported having experienced this kind of violence, and this situation was associated with marital status (single/separated women), lower income, delivery in lithotomy position, performance of the pressure maneuver on the uterine fundus, and separation from the newborn soon after birth [19].

The prevalence of non-recommended practices found in the present study is similar to those of another investigation, which identified that lithotomy position was present in 91.7% of deliveries, that oxytocin infusion and amniotomy were used in 40% of women [19]. Despite evidence showing that upright positions are associated with shorter labor time, less intense pain sensation, less use of interventions and greater satisfaction with the birth experience [37], lithotomy position is still almost universal in hospital childbirth in Brazil. Moreover, the fundal pressure maneuver, a painful procedure that has been shown to be associated with increased risk of uterine rupture, perineal injuries, neonatal fractures and brain injuries [16, 19], had a prevalence of 37% in another study [19], higher than that found in the present research (11.0%).

Our study identified that more than one third of women who went into labor were submitted to an episiotomy. The high prevalence of this procedure was also observed in the survey Birth in Brazil (50%), with a higher occurrence in primiparous women (75%) [36]. Routine episiotomy increases the risk of third- and fourth-degree perineal laceration, infection and bleeding, without reducing the complications of pain and urinary and fecal incontinence in the long term [17]. Unawareness of the risks associated with this procedure and contradictions in its indication cause many women to perceive it as a routine practice. This was seen in the application of the IRT model, in which the inclusion of the variable did not contribute psychometric information, with the caveat that the small sample size may have been responsible for the non-differentiation of this variable in the model. The same might have occurred for other frequent interventions, such as amniotomy, that were eliminated in the exploratory phase (S1 Table).

An interesting finding of the present study is the contrasting rates of c-sections and proportions of women who went into labor according to some key characteristics. Going into labor was more frequent in the public hospital, as well as among younger black or brown women, with lower socioeconomic and education levels. Previous studies have shown that going into labor is a risk factor for mistreatment, as has been evidenced in the research Birth in Brazil: women who went into labor had a 79% higher chance of reporting verbal, psychological or physical violence [15]. It is also relevant to note that c-section rates in our study were higher among women who did not feel mistreated, which points to the apparent convenience of elective procedures, where women do not initiate labor and the process of care usually flows under the control of both patients and assistant professionals. However, this apparent convenience is highly questioned, once the performance of the surgical procedure without or with incomplete clarification of risks and benefits or even without the woman’s deliberate consent is a situation of veiled mistreatment that is usually not recognized as such [38, 39]. These women might feel they were not listened to regarding their preferences, and they commonly feel lonely and vulnerable after the procedure in face of the pain and the limitations to take care of themselves and of the newborn [39]. Women who go through a c-section after previous labor report feeling disrespected more frequently [40], especially due to poor communication with the health team [39].

The items related to the absence of a companion during labor, delivery and postpartum periods were the most important for the MLWC, both for women who went into labor and for those who had elective cesarean sections. This result can be explained, at least in part, by: the wide knowledge of the law in relation to the presence of a companion during delivery—Federal Law No. 11.108/2005, which allows women to recognize such violations of their rights [41]; and the feeling of fear, loneliness and vulnerability, which triggers in women the need for support from someone they trust at the time of delivery. In agreement with the results of this study, the research Birth in Brazil also identified the presence of a companion as a significant inhibitor of mistreatment [15].

Another factor present in the definition of the MLWC was not understanding the information given by health professionals, a situation that reduces the woman’s autonomy to participate in decisions about her care and makes her feel uncomfortable to clarify doubts. Protagonism in childbirth is associated with satisfaction regarding assistance [18, 42]; however, research shows that most women do not feel at ease during care. A study conducted in public maternity hospitals in a northeastern Brazilian state, for example, showed that 29.8% of women were dissatisfied with receiving guidance and the possibility of asking questions, and 49.8% were dissatisfied with the possibility of making complaints about the care provided [20]. Besides not having access to effective communication with health professionals, to be able to express their wishes and beliefs, many women suffer mistreatment and even abandonment, as a form of coercion, for being seen as complaining and demanding beings [1, 7].

The gap between the perception of violence by women and what is considered as the definition of this phenomenon, according to technical-scientific publications, is one of the most relevant aspects of this study. The data depicted on S1 Table allows us to identify several practices that fit the definition contained in national and international documents guiding childbirth care, but which, in the analysis carried out in this study, did not contribute with psychometric information in preparation of the instruments to measure MLWC. This may have occurred for two reasons: non-identification of the intervention or situation as violent by the women; and insufficient sample size to show significance in this model. The high number of women who answered “I don’t know” or “I don’t remember” to questions such as if labor was induced with oxytocin or if they’d rather have another position in labor likely reflects the lack of information about their rights, options and evidence base on intrapartum care. These data also points to communication gaps between birthing women and the health professionals assisting them.

Another point that deserves discussion is the feeling of unsafety reported by women, as well as their perception of not feeling welcomed: approximately one fourth of the participants in this study felt this way, and both aspects were items that reached high discrimination in the IRT analyses. These findings are extremely relevant, because they contrast in some way with the conception that the hospital is a safe environment for obstetric assistance. Our data show that not all women agree with this perspective. Previous studies have suggested that fear and unsafety during childbirth are linked with feelings of uncertainty, loss of control and interventions in their bodies [43, 44]. On the other hand, women who feel welcomed and perceive a positive environment, have a stronger sensation of safety [45]. Understanding more deeply the reasons why many women do not feel safe or welcomed in the hospital when they are giving birth is a matter that deserves further studies, given its importance in the assessment and measurement of mistreatment.

This study analyzes, in an unprecedented way, the mistreatment of women during childbirth using IRT to propose a way to measure this latent variable, through a set of items. The method, besides being statistically robust, evaluates every item of the measurement instrument according to its severity and discrimination capacity, defining for each one a different importance in estimating the MLWC. Moreover, the methodological rigor in conducting the study, with random selection, continuous quality control, and the timely face-to-face interviews also ensure higher quality to the study. Possible disadvantages of face-to-face interviews include embarrassment to address sensitive matters, such as criticizing the care received, more complex logistics, high costs, increased probability of non-participation (exclusions, losses and refusals). As an example, in our study, we had to exclude women living in areas with high occurrence of violent incidents, to preserve the safety of the research team. This represents a potential selection bias, because the women who live in these areas might be the most vulnerable ones, and their absence in the study sample makes the studied population somewhat different from the original one. In the same sense, women who were lost because of contact failure may be another source of selection bias. If we imagine that a more vulnerable profile implies in higher risk of suffering mistreatment during childbirth [46, 47] then the inclusion of these women in the sample would probably enhance the discrimination capacity of the statistical model. However, the findings regarding the association between social vulnerability and mistreatment during childbirth are conflicting in the literature [42, 48].

Among the limitations of our study, one should consider the number of losses and the fact that the sample size is restricted to allow for more powerful conclusions about the set of items defined to measure the MLWC. Literature review points that to adjust an IRT model with high precision of parameter estimations, a sample of at least 500 is needed [49]. A smaller sample size represents a limitation in terms of precision, implying on less accurate estimates (with larger standard errors). However, as the objective of this study was to present a proposal for a measurement instrument and not a definitive instrument, we believe that the information obtained is relevant for the development of a final instrument, produced with larger samples and going through the process of its validation (of the instrument itself and the cutoff point) for the classification of women as to whether or not they have suffered mistreatment.

Conclusion

The conclusions of this study are preliminary, considering that the model developed to measure the MLWC was based on a small sample. As a main finding, it was observed that the mistreatment level identified by IRT was approximately twice the general prevalence of the perception of disrespect, mistreatment or humiliation measured by the direct question to women, demonstrating the divergence between the perception of postpartum women and what is considered mistreatment by health agencies and academic literature. The IRT also made it possible to identify the items with greater severity and discrimination capacity, both for the general sample of women, and for the group that went into labor, pointing out situations that can be avoided in childbirth care, such as the absence of a companion, the lack of understanding of information by the woman, the performance of the fundal pressure maneuver, feelings of not being welcomed and of unsafety.

Harmonizing the speeches, knowledge and perceptions is essential to qualify obstetric care. The development of an instrument to measure mistreatment during childbirth from the perception of women is essential in this context, to standardize the prevalence measures in different national and international scenarios, thus allowing comparability and generalization, as well as to identify practices acknowledged as violent by women who go through the childbirth process and enable a dialogue between the different actors. This study can be considered a first step in this direction, making its own contribution to the proposal of strategies to eliminate obstetric violence.

Flowchart of sample constitution.

(TIF)

Click here for additional data file.

(XLSX)

Click here for additional data file.

Development of the IRT model from a set of situations and interventions usually considered in the definition of mistreatment of women in childbirth care.

(DOCX)

Click here for additional data file.

Description of the items of the measurement instruments for the Mistreatment Level of Women during Childbirth—MLWC1 (all women) and Mistreatment Level of Women during Childbirth—MLWC2 (women who went into labor).

(DOCX)

Click here for additional data file.

14 Dec 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Carla Betina Andreucci, M.D., P.h.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please amend your current ethics statement to address the following concerns:

a) Did participants provide their written or verbal informed consent to participate in this study?

b) If consent was verbal, please explain i) why written consent was not obtained, ii) how you documented participant consent, and iii) whether the ethics committees/IRB approved this consent procedure.

3. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

Reviewer #3: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Dear authors,

The manuscript "Development of an instrument to measure obstetric violence through Item Response Theory" brings significant and innovative knowledge to the study of obstetric violence as a public health problem. The discipline context and the research questions are well described, and the development of a tool to measure obstetric violence occurrence in maternity services is a relevant initiative.

I would like to suggest that you check the language (maybe you could hire a professional proofreader). I would also like to suggest some additional work in the manuscript, as follows.

1) Please describe the settings in which studied women gave birth. I suppose the two facilities are different in many aspects, for example, number of live births per year, health professionals usually responsible for (or involved in) childbirth care, training or teaching activities, c-section rates, etc. Additionally, those who do not know the Brazilian health system and the Brazilian typical obstetric care may find it difficult to understand the public/private contrast.

2) I understand the motivation to exclude women "who had formal contraindication for breastfeeding" and those who lived in places considered risky by the research team. However, I wonder if those women are not the most vulnerable ones, and therefore, the most likely to be subjected to obstetric violence. To my view, this is a potential source of bias that should be addressed in the study.

3) "Each day, two women from the public maternity hospital and one from the private hospital were included in the study until the intended sample was reached" (lines 124-126): please, what was the intended sample? How was it calculated? I am afraid I missed this information, and it is essential to discuss the study's limitations.

4) All information was collected during face to face interviews? How was socioeconomic level assessed? Please explain the categories' grouping.

5) Maybe you could attach a supplementary figure, or table, in order to clarify the number of women drawn in each hospital, the ones who were eligible, included, interviewed, etc. This would be an opportunity to clarify non-participation at each stage, for instance, the refusals and women who were excluded because they lived in a territory assessed as risky by the research team (it is not clear when or how this evaluation and exclusion was made).

6) Please, clarify why there are missing data (table 1).

7) I would strongly suggest that you add a table with more detailed information about women's outcomes, such as c-section rates, according to key characteristics, such as by type of hospital (public x private), women's skin color, etc. From previous studies we can suppose these results will have an impact in women's perception of mistreatment. Additionally, it is not clear if women who went into labor but had a c-section were included in the "women who went into labor" group or not.

8) In the same direction, I would suggest that you add some categories in table 1, for instance, women who did not use pain relief methods. Did they feel they were subjected to obstetric violence? I am eager to know that! Moreover, it was a bit difficult to me to understand why some categories showed were recommended practices ("encouraged to walk", "yes"), while others were the practices that should be abolished ("enema", "yes").

9) In the discussion section, I would suggest to include some thoughts on the data collection strategy. The face to face interview made outside the health facility may have advantages and disadvantages.

10) I would suggest you explore more deeply the feelings of unsafety reported by women. The hospitalization of childbirth care is often justified as the historical "evolution" of the obstetric assistance, understanding it as a pathway to safer care for women and babies. However, it seems that women do not agree with this. Additionally, we suppose that the health facility exists to provide care, but what does it mean when those who should be cared for say they don't feel welcomed? These reflections could reinforce the need for further studies on obstetric violence and the possibilities to measure it.

11) Maybe you could use the STROBE checklist before submitting the revised manuscript.

Reviewer #2: This cross-sectional study has a great potential to contribute to maternal health measures, as authors propose an instrument with 11 important components for quality of care.

According to Strobe Checklist, Introduction and Abstract are adequate. Methods sections would be improved if authors include a brief statement of any efforts carried on to address potential sources of bias; and also if they describe how data was collected, registered and organized. Results ans discussion are well described.

I suggest some background information update as the manuscript refers to "Rede Cegonha" which is now not available anymore, as it is not a public policy but a governamental program/project. Also, the findings that led to Itens exclusion as described on Supplementary Table 1 are very interesting and maybe authors could consider including one paragraph about those itens on the manuscript results and discussion, as they bring along interventions and situations that are highly prevalent in the country.

I'm affraid I couldn't check for the quality of the instrument development through the IRT Model, but the authors use literature fairly. I also suggest some minor revision on English overall, in order to improve understanding for readers.

Reviewer #3: This is a very important study for research on maternity care and reproductive health. However, some adjustments are necessary for this article to be published:

1. Please consider having the article proofread by an English expert so that we can ensure a better understanding and standardization of the text and the terms.

2. "Humanized" and "obstetric violence" are not the terms used by international researchers. I understand that these are terms used mainly in Brazil and Latin America, but considering that this article will be read by international readers, and that there are already well-known terms for both, it is important to use them or explain to the reader why you chose to use this definition instead of the well-established one.

3.Although I am not an expert in quantitative studies, it seems to me that the statistical aspects are well described. On the other hand, I missed more details about the questionnaire in the Data Collection section.

4. Lines 322 to 336 are presenting Results instead of Discussion. Consider moving up.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Ana Carolina Arruda Franzon

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

6 Feb 2022

Dear Dr. Andreucci and reviewer,

We were happy to receive your feedback about our manuscript entitled Development of an instrument to measure obstetric violence through Item Response Theory [PONE-D-21-31080] and we are thankful for the careful reviews that have been carried out. We have thoroughly and carefully reviewed the manuscript and handled all of the points, one by one. We are sure to be now resubmitting to your appreciation an improved version of our article.

Please find below the point-by-point descriptions and responses to the reviews provided. Each point is addressed with a response in this letter and a description of the amendments that have been made to the manuscript and where in the text they can be viewed. In the manuscript, all amendments are marked in yellow for prompt identification.

Ethics statement has been amended, as well as data availability.

As suggested by all three reviewers, language has been checked by a professional (see attached Declaration) and many changes were made to the text.

�  Comments by Reviewer #1:

1. Please describe the settings in which studied women gave birth. I suppose the two facilities are different in many aspects, for example, number of live births per year, health professionals usually responsible for (or involved in) childbirth care, training or teaching activities, c-section rates, etc. Additionally, those who do not know the Brazilian health system and the Brazilian typical obstetric care may find it difficult to understand the public/private contrast.

Response: We have included the descriptions of the hospitals in the 1st paragraph under Study Design and Population (lines 124-132).

2. I understand the motivation to exclude women "who had formal contraindication for breastfeeding" and those who lived in places considered risky by the research team. However, I wonder if those women are not the most vulnerable ones, and therefore, the most likely to be subjected to obstetric violence. To my view, this is a potential source of bias that should be addressed in the study.

Response: Concerning the women who had formal contraindication for breastfeeding, we have made it clearer in the text that this was an elegibility criterion (not an exclusion one), because in the original research brestfeeding prevalence was a main outcome. This clarification was made in the 2nd paragraph under Study Design and Population (Lines 134-138). Concerning the women who lived in risky areas, that were actually excluded, we have included a new passage and added references to discuss this as a potential source of bias (before last paragraph of the Discussion – lines: 462-472).

3. "Each day, two women from the public maternity hospital and one from the private hospital were included in the study until the intended sample was reached" (lines 124-126): please, what was the intended sample? How was it calculated? I am afraid I missed this information, and it is essential to discuss the study's limitations.

Response: We have included a new paragraph (1st paragraph under Sample and Data Collection: lines 143-146) with information about sample size and calculation, with the

citation of the former research that originated the present study. We have also included in the last paragraph of the Discussion more information about the sample size needed to adjust an IRT model, with a new reference. This way, the study’s limitations concerning the sample size are clarified.

4. All information was collected during face to face interviews? How was socioeconomic level assessed? Please explain the categories' grouping.

Response: All information was collected during face to face interviews. This has been made clearer in the text, in the third paragraph under Sample and Data Collection (lines 160-165). In this same paragraph, we have explained how socioeconomic level was assessed and it’s categories grouped.

5. Maybe you could attach a supplementary figure, or table, in order to clarify the number of women drawn in each hospital, the ones who were eligible, included, interviewed, etc. This would be an opportunity to clarify non-participation at each stage, for instance, the refusals and women who were excluded because they lived in a territory assessed as risky by the research team (it is not clear when or how this evaluation and exclusion was made).

Response: A flowchart was included as supplementary Figure 1, cited in the manuscript in the 1st paragraph under Results (line 243).

6. Please, clarify why there are missing data (table 1).

Response: The reasons for missing data were clarified with footnotes in Table 1.

7. I would strongly suggest that you add a table with more detailed information about women's outcomes, such as c-section rates, according to key characteristics, such as by type of hospital (public x private), women's skin color, etc. From previous studies we can suppose these results will have an impact in women's perception of mistreatment. Additionally, it is not clear if women who went into labor but had a c-section were included in the "women who went into labor" group or not.

Response: The women who went into labor but had a c-section were included in the "women who went into labor" group. This information was made clearer in the second paragraph under Results. Also, Table 2 was added, presenting c-section rates and proportion of women who went into labor according to key characteristics. The findings of this new table are described in the Results session and some of them are discussed in a whole new paragraph in the Discussion (6th paragraph under that session), with new reference citations.

8. In the same direction, I would suggest that you add some categories in table 1, for instance, women who did not use pain relief methods. Did they feel they were subjected to obstetric violence? I am eager to know that! Moreover, it was a bit difficult to me to understand why some categories showed were recommended practices ("encouraged to walk", "yes"), while others were the practices that should be abolished ("enema", "yes").

Response: We have included in Table 1 the complementary category of each variable related to the care of women who went into labor. We believe this clarifies the issue, making it easier to understand those interventions and situations in light of being recommended or not.

9. In the discussion section, I would suggest to include some thoughts on the data collection strategy. The face to face interview made outside the health facility may have advantages and disadvantages.

Response: We have included some thoughts on potential disadvantages of face to face interviews in the before last paragraph of the Discussion (lines: 459-462).

10. I would suggest you explore more deeply the feelings of unsafety reported by women. The hospitalization of childbirth care is often justified as the historical "evolution" of the obstetric assistance, understanding it as a pathway to safer care for women and babies. However, it seems that women do not agree with this. Additionally, we suppose that the health facility exists to provide care, but what does it mean when those who should be cared for say they don't feel welcomed? These reflections could reinforce the need for further studies on obstetric violence and the possibilities to measure it.

Response: A new paragraph was added in the Discussion, with new reference citations, exploring this issue. This is the 10th paragraph under Discussion (lines: 441-452).

11. Maybe you could use the STROBE checklist before submitting the revised manuscript.

Response: The STROBE checklist was used and added as a supplementary file.

�  Comments by Reviewer #2:

1. According to Strobe Checklist, Introduction and Abstract are adequate. Methods sections would be improved if authors include a brief statement of any efforts carried on to address potential sources of bias; and also if they describe how data was collected, registered and organized. Results and discussion are well described.

Response: We have included a new paragraph describing how potential sources of bias were addressed, as well as included more details about data collection, registering and organization. This is the 5th paragraph under Sample and Data Collection (lines 171-178).

2. I suggest some background information update as the manuscript refers to "Rede Cegonha" which is now not available anymore, as it is not a public policy but a governamental program/project.

Response: We have reviewed updated information about “Rede Cegonha” and other programs related to prenatal and childbirth care in Brazil. We have modified the 4th paragraph of the Introduction (Lines 75-94) and inserted new sentences with updated references. In this review process we were able to certify that “Rede Cegonha” is still available and adherent institutions send their indicators monthly for monitoring. The ordinance supporting the program today is: Consolidation Ordinance n.3, 28 September 2017 (available at:

https://bvsms.saude.gov.br/bvs/saudelegis/gm/2017/prc0003_03_10_2017.html#ANEXOIITITIICAPII). Even though the program still exists, it has clearly weakened in terms of institutional strength, so we have included this information to clarify the context.

3. Also, the findings that led to Itens exclusion as described on Supplementary Table 1 are very interesting and maybe authors could consider including one paragraph about those itens on the manuscript results and discussion, as they bring along interventions and situations that are highly prevalent in the country.

Response: We have drawn more attention to episiotomy and amniotomy, which are very frequent practices in our country. We have included new passages in the 3rd paragraph under Results and in the 5th paragraph under Discussion. Also, the paragraph beginning with “The gap between the perception of violence” mentions some of the findings of Supplementary Table 1.

�  Comments by Reviewer #3:

1. "Humanized" and "obstetric violence" are not the terms used by international researchers. I understand that these are terms used mainly in Brazil and Latin America, but considering that this article will be read by international readers, and that there are already well-known terms for both, it is important to use them or explain to the reader why you chose to use this definition instead of the well-established one.

Response: we have reviewed these terms throughout the whole text. The term “obstetric violence” was replaced with “mistreatment of women during childbirth”. The term “level of obstetric violence” was replaced with “mistreatment level of women during childbirth”. Therefore the acronym NVO changed into MLWC.

2. Although I am not an expert in quantitative studies, it seems to me that the statistical aspects are well described. On the other hand, I missed more details about the questionnaire in the Data Collection section.

Response: We have provided more details about the questionnaire in the 3rd paragraph under Sample and Data Collection (lines 158-160).

3. Lines 322 to 336 are presenting Results instead of Discussion. Consider moving up.

Response: These lines were moved up to the Results session (lines: 256-259) and some remaining passages in the Discussion were adjusted accordingly.

We believe we have fully addressed all the points raised by the reviewers and we thank you again for the opportunity of qualifying our manuscript.

Yours sincerely, the author.

Submitted filename: Response to Reviewers.pdf

Click here for additional data file.

26 Apr 2022

19 Jun 2022

Dear Dr. Andreucci,

We were happy to receive your second feedback about our manuscript entitled Development of an instrument to measure obstetric violence through Item Response Theory [PONE-D-21-31080] and we are thankful for the careful reviews that have been carried out. We have once more handled all of the points, and we are thus resubmitting to your appreciation an improved version of our article.

Please find below the point-by-point descriptions and responses to the reviews provided. Each point is addressed with a response in this letter and a description of the amendments that have been made to the manuscript and where in the text they can be viewed. In the manuscript, all amendments are marked in yellow for prompt identification.

�  Comments by Reviewer #1:

1. I would like to bring up a few more questions just to deepen the discussion around obstetric violence scholarship. Indeed, I believe "obstetric violence" has a concept of its own, and it has gained ground in academic studies from various disciplines in the last decade, as can be verified from the growing number of articles published using this term. In my view, the use of "obstetric violence" poses a political stance, as it implies the recognition of structural problems that negatively affect women's experiences, jeopardizing their human rights.

I understand you changed to "mistreatment" to meet one of the reviewer's request, however, I was wondering which term or concept you used in the Brazilian Portuguese questionnaire, as I feel there may be differences in understanding between them. Unfortunately, I couldn't access the file "database_02_22.xlsx” to verify this. In any case, I would suggest that you reflect about the use of terms in your empirical material and in the manuscript before making a final decision between "obstetric violence" and "mistreatment".

Response: We are totally aligned with the view of this Reviewer. Although we have decided to keep the term “mistreatment” in the majority of the text, we have included the term “obstetric violence” in a few points, such as in the abstract, in order to point to the wider context reflected by this term, its poltical stance and the recognition of the structural issues involved. We have also included the term “obstetric violence” where the original study cited used this same term on its data collection. Our decision to keep the term “mistreatment” was for the following reasons: In the Brazilian Portuguese questionnaire we used the following question: “Have you ever (during labor and childbirth care) felt disrespected, humiliated or mistreated by health professionals?”, we have not actually used the term “obstetric violence”; the term mistreatment is more frequent in the international literature, so the use of this term can increase access and visibility of the article. Considering the relevance of the term “obstetric violence”, we decided to add a paragraph in the Introduction (lines 71-77).

2. You mention that "All information considered in this study was collected during face-to-face interviews and the responses were referred by the participants. Women who were not found for the interview, after at least three attempts of contact by telephone and one in person, were considered a loss". However, in the results, you state that "The women not interviewed due to losses and refusals differed in terms of education and skin color,

showing less education (p<0.01) and a higher prevalence of white skin color compared to those interviewed (p=0.032)". How did you get information about education and skin color from women who refused to participate?

Response: Name, age, education and skin color were drawn from the hospital chart, so these were the only data available for losses and refusals. We modified the text for clarification (Lines 169-171).

3. Table 1, you added the information that "*Missing data correspond to responses 'I don’t know' or 'I don’t remember'" – for some variables, would it be relevant to consider these answers? For example, for "Induction with oxytocin", when women don't know whether or not they underwent this procedure, is it possible to state that they didn't receive appropriate information? Can we say the same about "would like another position in labor"? If 54 women answered "I don't know" or "I don't remember", it means that more than 25% of the birthing women didn't receive adequate information about mobility during labor – however, this variable was excluded from the model because of the high number of missing data. I don't mean that you have to re-process your data, but then again, I wonder what does it mean when we face such a high level of women who declare that they do not know if they would like another position in labor. It seems to me that they lack information about their rights, options and about evidence, as well.

Response: We agree with the reviewer's suggestion and have therefore included a paragraph in the manuscript to briefly discuss these findings (lines 444-448).

�  Comments by Reviewer #2:

1. I believe an editorial and authorial decision will be need in order to comprehend or not the very recently published "PORTARIA GM/MS Nº 715, DE 4 DE ABRIL DE 2022" by Brazilian Ministry of Health, which alters the Rede Cegonha program into Rede de Atenção Materna e Infantil (Rami).

Response: We have included the recent “portaria 715/2022”, which alters the Rede Cegonha program, briefly presenting this still very recent context change (lines: 92-101).

We believe we have fully addressed all the points raised by the reviewers in this new round and we thank you again for the opportunity of qualifying our manuscript.

Yours sincerely, the authors.

Submitted filename: Response to Reviewers.docx.pdf

Click here for additional data file.

28 Jun 2022

Development of an instrument to measure mistreatment of women during childbirth through Item Response Theory

PONE-D-21-31080R2

Dear Dr. Paiz,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Carla Betina Andreucci, M.D., P.h.D.

Academic Editor

PLOS ONE

4 Jul 2022

PONE-D-21-31080R2

Development of an instrument to measure mistreatment of women during childbirth through Item Response Theory

Dear Dr. Paiz:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Mrs. Carla Betina Andreucci

Academic Editor

PLOS ONE