Identifying Obstetric Mistreatment Experiences in U.S. Birth Narratives

Abstract Background: Traumatic births are those resulting in feelings of distress that persist after the birth experience. Health care providers may play a role in these experiences through various forms of mistreatment. Analyses of global birth experiences have generated several domains of mistreatment. This study applies these evidence-based domains of mistreatment as an a priori coding scheme for analysis of 96 oral narratives of U.S.-based births to describe the nature of perceived mistreatment using participants' own descriptions of experiences. Method: Ninety-six transcripts of oral birth stories from 61 participants were coded using the domains of mistreatment experiences described by the Bohren et al.'s (2015) systematic review of obstetric mistreatment. Results: N = 131 individual experiences of perceived obstetric mistreatment were identified in 41 out of 96 narratives (42.7%). The most frequent types of experiences were Poor Rapport (90 incidences) and Failure to Meet Professional Standards of Care (29). Clinical Implications: Although most women in our study did not perceive any instances of obstetric mistreatment during their childbirth, over 40% of participants noted at least one event that fit one of the typologies we used as a framework for analysis. Visibility and review of the types of perceived mistreatment experiences that occur during birth enables health system leaders to implement prevention and accountability strategies. Most instances of perceived mistreatment during birth may be prevented through intentional implementation of individualized, respectful, supportive care during labor and birth.

In our study, we extend on Bohren et al.'s (2015) and Vedam et al.'s (2019) work in two ways to describe the topography of obstetric mistreatment in a U.S. sample when applying internationally informed obstetric mistreatment domains. First, we apply Bohren et al.'s seven domains of obstetric mistreatment (Poor Rapport, Failure to Meet Professional Standards of Care, Stigma and Discrimination, Verbal Abuse, System Conditions, Physical and Sexual Abuse) as a coding methodology to a U.S. sample of oral birth stories. These original domains were identified using a thematic synthesis approach and therefore represent broad, descriptive themes that describe types of mistreatment phenomena observed across qualitative findings from multiple studies. This domain identification approach used by Bohren et al. is distinct from a review or evaluation of medical records against standards of care to identify objective failures to meet stated care standards identified by licensing or other governing bodies, an approach that would be especially difficult to use given the heterogeneity of the original data sample, which includes studies from multiple countries, each with their own systems for oversight of care standards. Application of these themes as descriptive cate-O bstetric mistreatment, dehumanizing treatment by medical personnel that results in an experience of maternal victimization via the loss of autonomy, is a factor contributing to traumatic birth experiences (Darilek, 2018). The World Health Organization (WHO) has acknowledged the global prevalence of these experiences and the need to mitigate them, asserting that people in labor are entitled to "the highest attainable standard of health, which includes the right to dignified, respectful health care" (WHO, 2014, p. 1) and recommending "respectful maternity care... [that] maintains [women's] dignity, privacy and confidentiality, and ensures freedom from harm and mistreatment [during birth]" (WHO, 2018, p. 3) as the global standard of care. This global call to action is supported by literature on maternal perceptions of mistreatment, though most of these inquiries are based on birth experiences outside the United States (Allen, 1998;Moyzakitis, 2004;Nicholls & Ayers, 2007;Thomson & Downe, 2008), suggesting a need for investigations into the phenomenon of perceived mistreatment in U.S.-based birth experiences.
One of the most robust assessments of domains of obstetric mistreatment was conducted by Bohren et al. (2015). Researchers applied a thematic synthesis approach across 65 studies on experiences of mistreatment during birth from 34 countries, generating an evidence-based classification system of perinatal mistreatment experiences consisting of seven domains: physical abuse, verbal abuse, sexual abuse, stigma and discrimination, professional care failure, poor rapport, and system conditions. A strength of the study is its use of a systematic review yielding an evidence-based typology describing domains of mistreatment that can be applied in future research to identify frequencies and types of mistreatment across populations. Although the study incorporated international experiences, it may not be representative of any specific country. Only one study from the United States met inclusion criteria, compared with, for example, 11 studies based in sub-Saharan Africa, and five each from Asia, the Middle East and North Africa, and Latin America.
Development of an evidence-based typology laid the groundwork for application of the domains in further research. One of the U.S.-based expansions is Vedam et al.'s (2019) application of the mistreatment domains in a 60-item survey as part of the Giving Voice to Mothers project. In their sample of 2,138 U.S. births, 17.3% of participants self-reported experiencing some form of mistreatment in accordance with the original domains, suggesting the value of the domains as a framework for assessing mistreatment in U.S. births in other contexts and through other research approaches.
One limitation of survey-based applications of the domains as an assessment strategy is that this approach relies on participant identification of their experiences matching Vedam et al.'s (2019) item wording, which can overly constrain instances in which participant experiences meet domain criteria. In contrast, this is a strength of a priori coding methodology, where experiential criteria are identified directly from participants' descriptions of their birth expe-Previous studies of traumatic experiences during childbirth have identified perceived mistreatment by caregivers as a serious issue globally, but less attention has been given to the role of mistreatment in birth experiences in the United States.
Shutterstock/DC Studio with multiple birth experiences were invited to share their stories in any order that made the most sense to them.

Procedure
Consent was obtained from participants before completion of the preinterview online survey and at the beginning of their interview. All interviews were audio-recorded for later transcription. Participants were read a debriefing protocol at the end of their interview that summarized the research process, provided contact information for the research team, and offered local and national resources for any participant interested in support after recollecting negative experiences. The narratives were recorded and promptly transcribed with all identifying information (e.g., names and locations) removed. Upon transcription, completion original audio files were deleted. All study materials and protocols were reviewed and approved by our Institutional Review Board.

Analysis
Audio recordings were initially transcribed using Transcribe (Wreally, 2019) and checked for transcription accuracy by researchers. Transcripts were coded in NVivo 12 (QSR, 2018) using the seven obstetric mistreatment domains identified by Bohren et al. (2015). To meet the domain criteria, Bohren et al.'s first-order identification criteria of specific event types were used. For example, the domain of Poor Rapport includes the second-order theme of "ineffective communication," which includes the criteria of poor communication, dismissal of women's concerns, language and interpretation issues, and poor staff attitudes.
We relied on participants' own description of their birth experiences as the primary unit of analysis; therefore, researchers did not correct or edit participant descriptions even when they were likely to contain misapplied or inaccurate medical terminology. Instead, to focus coding on the perceptual experiences of participants and mitigate potential effects of coder bias, incidents were only coded when 1) they met one of Bohren et al.'s (2015) domain criteria and 2) additional evaluative or affective indicators reflected that the participant perceived the incident as mistreatment. Identification of evaluative indicators was guided by the Appraisal Framework (Martin & White, 2005), which describes linguistic features that indicate attitudinal assessments in narratives.
For example, a participant said, I had a panic attack. I was not really aware of certain aspects of the surgery... It was traumatic. In this example, I was not really aware of certain aspects of the surgery met the domain criteria for Poor Rapport (primary domain) via ineffective communication (second-order theme) according to the identifying criteria of poor communication (first-order theme criteria); it was traumatic is an evaluative statement reflecting the participants' assessment of their experience, and therefore meets the coding criteria. Similarly, Nobody responded and nobody did anything [cries] met the domain criteria for Poor Rapport, and the indicator of cries satisfies the paralinguistic criteria. In contrast, a participant stated, [The physician] just wasn't paying attention to us. But for the most part I was fine with that. [The physician] just wasn't gories is therefore appropriate for use in the analysis of narratives, in which descriptions of experiences are the primary data unit. Rather than using a closed-form assessment method (Vedam et al., 2019), we apply the domains to the analysis of oral narratives where evaluative linguistic indicators (e.g., It was traumatic) and affective indicators (e.g., crying during narration) are used to interpret meaning from the participants' own descriptions and perceptions of their experiences during birth to determine whether Bohren et al.'s domain criteria are met.

Participants
Participants over age 18 who had given birth in the United States were recruited via paper and digital flyers. The call for participants read Participate in a research study on birth and infant feeding. If you are at least 18 years old and have experience giving birth, then we want you to interview for a study on your experience. Participants were not required to have experienced obstetric mistreatment to participate. Although birth, postpartum, and infant feeding experiences were collected for this sample, only birth narratives are analyzed in this study. Paper copies of the flyer were distributed to three pediatric offices, and other public spaces (e.g., college campuses, and community message boards) within the New England region of the United States. Digital copies of the flyer were distributed via national maternal health professional ListServs, email contact lists for parenting and infant feeding groups, and via a local provider network contact list of approximately 300 clinical providers for distribution to their networks and patients, and for inclusion in their newsletters or social media pages. Recruitment was open from June 2018 through January 2020. This sample size is three times larger than the recommendation for thematic saturation (Hennink et al., 2017) and was selected to protect against under identification of anticipated low-frequency domain events (e.g., Sexual Abuse).

Preinterview Online Survey
The survey gathered participants' demographic information (e.g., birth history, age, race and ethnicity, marital status, and residency). The survey intake collected additional information not used in the current study (e.g., mode and experience of infant feeding).

Interview Protocol
The interview protocol elicited narrative responses about participants' experiences with pregnancy, birth, and the postpartum period; there were no questions specific to mistreatment, victimization, or care quality. During the interview, participants were asked to share their birth story in detail, as [they] experienced it, including any relevant circumstances leading up the birth that [they] consider important. Participants were instructed to include any memory that feels especially vivid or meaningful, including anything that [they] were thinking or feeling at the time. Participants matic. Participants who didn't feel like procedures were explained to them felt unsupported by their providers.
Some participants reported that their concerns were disregarded. A participant recalled, I kept saying like -people aren't listening to me. People aren't hearing me, people areyou know-it's oh she's just another mom. Some mothers also felt like their pain was dismissed, with one recalling, It took me like crying for them to finally realize that [the epidural] wasn't done correctly. Dismissing maternal concerns contributed to maternal distress, with one participant summarizing, I was a just angry and hurt and scared.
Loss of autonomy. Objectification, lack of respect for preferred birth positions and practices, denial of food and mobility, and treating participants as passive participants in their paying any attention to us met the criteria for Poor Rapport, but the participants' own evaluation (I was fine with that) indicates no perceived mistreatment, and therefore this was not coded as Poor Rapport.

Identification of Bohren et al.'s (2015) Obstetric Mistreatment Domains
Two coders reviewed narratives with 60% overlap across codes. Coders' interrater reliability was estimated using Cohen's Kappa and indicated strong agreement (κ = 0.801). Fifty-five narratives (57.3%) did not include any instances of mistreatment. The other 41 (42.7%) narratives had an average of 3.11 (SD = 3.04) unique instances per narrative, totaling 131 unique instances from 37 unique participants in our sample. The most frequently identified domain was Poor Rapport (68.7% of all mistreatment codes), followed by Failure to Meet Professional Standards of Care (22.1%), Stigma and Discrimination (6.9%), Verbal Abuse (1.5%), and System Conditions (.8%); there were no instances of Physical and Sexual Abuse identified.

Poor Rapport Domain
Ineffective communication. Poor communication, dismissal of maternal concerns, and language and interpretation issues met the criteria for ineffective communication, totaling 71 incidences (Table 2). Poor communication between participants and their providers contributed to participant distress. Some experienced anxiety at seeing providers react to emergency situations without understanding the circumstances; one participant described her distress at seeing her physician freaking out while another recalled her provider acting in a panic. Some participants responded to perceived distress among their health care team with their own panic; one recalled, And all of a sudden I had nurses coming in… I had no idea what was going on but people are reaching [all around] and I just started bawling. Another recalled feeling upset by lack of information about circumstances surrounding her emergency cesarean stating, There was a lot of like motion and commotion and like sort of an emergency feel to the whole thing… And I was just like, 'Wait, what?'…It was upsetting and trau-

Failure to Meet Professional Standards of Care Domain
Exams and procedures. Physical exams and procedures that are regarded as typical courses of care sometimes resulted in perceived mistreatment from participants, totaling 12 incidences (Table 3). Inductions were a common source of maternal distress in this category, due to both physical and psychological impacts. One participant described their induction as torture; another noted that their induction made them feel totally out of control. Some participants attributed the pain of their procedures directly to their care being rough or aggressive; one participant recalled her water breaking during a routine exam, prior to the start of labor, stating, It was pretty forced from the doctor. In one particularly significant incident, a participant recalled her experience of failed pain management during her cesarean birth, summarizing, After the spinal [block] -it wasn't tested and I had a Csection with no pain medication and so I was screaming that I could feel everything like very like traumatic to your whole body right? As well as your brain [cries]. Lack of informed consent. Some participants reported that they had not provided informed consent for procedures, totaling 9 incidences. Cervical checks were a common procedure associated with perceived mistreatment due to lack of consent. One participant recalled, The doctor comes in and literally just inserts his fingers inside of my vagina to check me without any preamble at all births met the criteria for loss of autonomy, totaling 11 incidences. Some participants felt mistreated after having their mobility restricted in ways they felt were unnecessary. One participant described having to fight for [her]    Discrimination based on medical conditions. Two participants experienced perceived mistreatment due to their health conditions. One participant felt discriminated against due to her herpes diagnosis, stating, The thing that disturbed me the most was I had signs on my door for all these precautions. Like everybody was afraid they were going to get it. … so I was just upset. Another participant felt her care quality was diminished due to her weight, remarking, [They] see a fat woman [and] assume… and stereotype, like-this isn't evidence-based medicine. In these instances, participants noted that this discrimination negatively impacted their trust in their providers.

Loss of Autonomy
[pause]…I started panicking. Another participant disagreed with her provider about whether she was informed about a cervical check, stating, [She claimed] she said, 'I'm going to check you.' I didn't hear that at all. Next thing you know I'm being checked…and I freaked. I screamed and yelled and jerked.
Although some experiences of lack of informed consent processes were associated with perceived pressure or coercion to comply (e.g., I kept saying numerous times that I did not want [a Foley] …more doctors would come in and pressure me), other experiences were more explicit. One participant was told she had lost her right to refuse procedures: I said that I didn't want Pitocin and he told -he patted my hand and said, 'Well you lost your right to refuse when you came [to the hospital]. Some participants noted that, upon reflection, they had had the right to limit procedures, but felt pressured to comply at the time.

Physical Exams and Procedures
After the spinal [block] it wasn't tested and I had a c-section with no pain medication and so I was screaming that I could feel everything like very like traumatic to your whole body right? As well as your brain [cries].

12
My water broke during the exam on the table. Um so I mean I think it was pretty forced from the doctor. I was like why not just give me the meds that do work? Why-why do this to me? There was a ton of bleeding from it. It was awful.

Neglect and Abandonment
I was definitely suffering. And my midwife wasn't really there, she kind of left me alone, that became a source of a lot of sort of feelings of abandonment for me, you know. 8

So they left me [in the hospital room] for two days. [I knew] something's not right … needless to say [after two days] it turned out to be an emergency C-section … it was horrible. It was traumatizing.
TOTAL 29 mistreatment in global samples, but are less prevalent in U.S. birth experiences. Systems Conditions were similarly infrequent in this sample, which is likely due to the current study's focus on birth experiences, whereas most of the subcategory items (i.e., billing issues) in System Conditions would be present in postpartum settings. Researchers who use the original typology to assess birth experience should consider this in their methodology and analysis.
Our study also reported few instances of Stigma and Discrimination. Given the known frequency of these experiences in U.S. birth experiences (Attanasio & Hardeman, 2019;Davis, 2019), this is likely a reflection of the coding standard applied in our study, which would require a participant to explicitly link, through evaluative statements, their mistreatment experience to a specific experience of stigma or discrimination to meet the coding criteria. Studies seeking to describe the nature of discrimination as a form of mistreatment during birth should consider a different coding or methodological approach to accurately capture the scope of these experiences.
Using Bohren et al.'s (2015) internationally informed typology of domains as an a priori coding scheme contributes to both our understanding of the nature of mistreatment in the United States, as well as provides a novel strategy for analyzing birth narratives. Our study required presence of either evaluative or affective indicators participants' distress to meet coding criteria, which may alleviate methodological concerns regarding coder bias in narrative data projects. Ability to apply a range of strategies to identify mistreatment is an additional benefit of using oral narrative, in particular, in which paralinguistic indicators can be used.

Limitations
One potential limitation of our study is the unknown impact of time since birth experience on participants' recall.

Verbal Abuse Domain
Harsh language. Only one instance of rude language met the domain criteria (Table 5). A participant recalled a provider who made several remarks about her daughter's anticipated size, He made a lot of really jerky comments. He called her an elephant at some point while … he was rude.
Threats and blaming. One participant recalled abusive language and blame from her provider regarding choosing a cesarean, stating, [The nurse] was like-and I'm never going to forget this-she said, 'So what? Are you going to wuss out on us and have a C-section or are you just going to tough it out to see if you can give birth correctly?' Never forget that.

System Conditions Domain
Lack of resources. One participant experienced distress over the perceived lack of resources at her hospital (Table  6). She recalled not being able to use a shower due to a monitor not being available, summarizing, The pain is progressing… then they finally get me a unit. Nine hours later. It was ridiculous.

Discussion
Although most women in our study did not perceive any instances of obstetric mistreatment during their childbirth, over 40% of participants noted at least one event that fit one of the typologies we used as a framework for analysis. The methodological approach used in this study is an internationally informed series of obstetric mistreatment domains first presented by Bohren et al. (2015) applied as a priori codes in a qualitative analysis of U.S-based oral birth stories. We identified five of Bohren et al.'s seven domains of obstetric mistreatment in our sample of U.S. birth narratives. There were no instances of Physical or Sexual Abuse identified in this sample of oral narratives, suggesting that these categories may describe the phenomena of perceived using phenomenological methods to identify instances of obstetric racism are especially useful in this area (Davis, 2019). Vedam et al. (2019) also used the Bohren et al. (2015) domains as the framework for their survey study, which included a more representative proportion of Black participants, and found that it captured obstetric mistreatment experiences of racial minority mothers, suggesting the Bohren at al. domains could be an appropriate methodology in more representative samples of birth narratives.

Clinical Nursing Implications
Understanding perceptions of mistreatment in birth experiences supports development of evidence-based practices for nursing practice. Interactions with providers are reported to be common factors driving trauma after birth (Reed et al., 2017). This is an opportunity to elevate the high-stakes role of health care providers, particularly nurses, as potential mediators of perceived mistreatment during birth. Perceptions of women that they were not treated with respect or listened to during the childbirth process are evidence that maternity care needs improvement. In the United States, nurses provide the majority of hands-on care during childbirth (National Academies of Sciences, Engineering, and Medicine, 2020), thus nurses are positioned to make sure each woman receives respectful maternity care. Our findings are further evidence that U.S. births are at risk of including an experience of perceived obstetric mistreatment. Many of the instances of perceived mistreatment may be prevented with relatively simple mitigation strategies. Most of the codes for mistreatment were within the domain of Poor Rapport. Some participants indicated that feeling their concerns were heard or being given the opportunity to connect with their providers would have improved their experiences, with one summarizing, I wish that people would have listened to me. Participants felt that simple changes in communication, like more communication in advance of procedures, may have prevented their distress. Nurses can facilitate clear, compassionate communication on shared decision-making as a useful tool for preventing perceived mistreatment within this domain. More robust interventions to address systemic change may consider evidence-based quality improvement protocols used in the context of identification, prevention, and response to perceived mistreatment experiences.
Our study contributes to the growing body of research on maternal perceptions of mistreatment during their birth experiences, as well as offers a framework for analyzing oral narratives as a valuable source of data. Research that centers the lived experiences of those who have labored and given birth is critical for ensuring that the future development of evidence-based best practices prioritizes the mitigation of potential harm and psychological distress from perceived mistreatment. <

Acknowledgment
This research was supported in part by grant R50180000038539 from the College of Fine Arts, Humanities and Social Sciences and the Office of the Vice Chancellor for Research and Innovation at the University We were interested in participants' perceptions of their birth experiences. Thus, our research methodology was not designed to address questions concerning the veracity of recollections. Some evidence suggests time-from-event memory decay results in less accurate recollections (Lacy & Stark, 2013), whereas some does not (Diamond et al., 2020). Although it is possible that the recall quality of these significant life events may decay over time, there is evidence that affective recall quality is reliable despite the passage of time (Porter & Birt, 2001).
Our participants were predominantly white (88%) and our results are limited in understanding of U.S. birth experiences of underrepresented minority groups. In the United States, the disproportionately high rate of maternal and infant mortality for Black mothers (Krisberg, 2019) and the disproportionate risk of Black women experiencing medical discrimination (Attanasio & Hardeman, 2019;Vedam et al., 2019) are urgent health priorities. Studies