Examining respect, autonomy, and mistreatment in childbirth in the US: do provider type and place of birth matter?

We conducted a secondary analysis of data from the Giving Voice to Mothers US (GVtM) study. The cross-sectional survey was developed using a community-based participatory research process to document care experiences of people who were pregnant in the U.S. from 2010 to 2017. This study was designed to document the experiences of pregnant/parenting people of color and those who planned to give birth at home or in a freestanding birth center. The survey explored preferences for care, interactions with providers, experiences of respect and safety, autonomy in decision-making, and access to care options. The full GVtM survey instrument (218 items) was available in English and Spanish. Scale items had pre-defined Likert response options. Details on survey development and recruitment are available in a previous publication [31, 32]. A full report of the findings is also publicly available. The Behavioural Research Ethics Board at the University of British Columbia approved the study (H15-01524).

Surveys of childbearing people’s experiences with pregnancy and birth care are often restricted to those giving birth in hospitals and overrepresent white participants [33]. To address this under-representation, people of color and people who planned a birth either at home or at a freestanding birthing center were oversampled. Invitations to participate were shared through NGO memberships and listservs, social media, doulas, midwives, physicians, and at professional conferences. Recruitment was also embedded in an ongoing statewide maternity care evaluation project led by an NGO partner in New York state. The survey was administered online from March 2016 to June 2017. To ensure ample inclusion of historically marginalized groups the survey was extended for a longer time frame for this group.

In the present analysis, we included respondents who answered questions about place of birth and ultimately gave birth in their planned setting (n = 1771). To reduce confounding, individuals who gave birth in a different setting than planned or were transferred (i.e. from a home or a birth center to a hospital) were excluded, as they report higher rates of disrespect and mistreatment than those with planned hospital births [31].

Three categories of birth setting, and intrapartum provider type were created for analysis: midwife in community setting (MW + comm), midwife in hospital setting (MW + hospital), and physician in hospital setting (MD + hospital). Community births attended by midwives included both births at home and births at a freestanding birth center, while hospital births included births at a hospital intrapartum unit or an in-hospital birth center.

Experiential outcomes were measured in four domains: (1) communication and decision-making autonomy, (2) respect, (3) mistreatment, and (4) time spent during visits.

Experiences with patient-provider communication and decision-making were assessed using the Mothers Autonomy in Decision-Making (MADM) scale, a 7-item instrument measuring the agency and autonomy a person experiences when participating in decision-making with a maternity care provider. This scale documents responses on six-point agree/disagree Likert scales, (scores ranging from 7 to 42) and was assessed for validity and reliability [34]. Scores from the MADM were used to create three categories. Individuals scoring in the bottom 33rd percentile were categorized as having low autonomy, while scores from the 34th through 65th percentiles indicated moderate autonomy, and scores in the 66th percentile and above indicated high autonomy. These cut offs enable clear delineation of low and high scores, with scores in the upper third constituting the optimum experiential outcomes. Respectful care was measured using the validated 14-item Mothers on Respect index (MORi), which assesses the nature of patient-provider interactions, and their impact on one’s sense of respect during maternity care [35]. Using a six-point Likert scale, participants rate their level of comfort when discussing care with their maternity care provider, the impact of provider interactions on their willingness to ask questions, and perceptions of racism or discrimination when receiving care. Scores from the MORi were collapsed into categories of low, medium, and high respect using the same approach as for the MADM scale.

Mistreatment was measured using the Mistreatment Index (MIST), a set of patient designed indicators of mistreatment, abuse, neglect, and health human rights violations [31]. Participants reported whether (yes/no) they experienced: “personal information shared without consent, violation of physical privacy, being shouted at or scolded by provider, provider threatened to withhold treatment or force acceptance of a treatment, provider threatened you in any other way, provider ignored or refused request for help or failed to respond in a timely manner, and experience of physical abuse”. Items were collapsed into a dichotomous mistreatment variable to reflect any negative experience versus no negative experiences. Time spent during prenatal visits (‘I felt I had enough time during my prenatal visits’; 6-point agree/disagree Likert scale) was also included as literature on what matters to service users often includes time as a critical factor in building trusting care relationships [36, 37]. The “had enough time” variable was collapsed into yes (strongly agree/agree) or somewhat/no (somewhat agree/somewhat disagree/disagree/strongly disagree) to compute the dichotomous variable.

Sociodemographic covariates included race, number of previous births, maternal age at last birth, nativity, highest level of education, total household income before taxes, source of payment for care, and whether respondents experienced elevated pregnancy risk. Since social risk factors may impact access to different models of maternity care and quality of care, while also being proxies for systemic inequities, we created five composite measures of social risk reported during pregnancy or in the year leading up to it by grouping variables under five categories. These included (1) primary care concerns (My children’s health, access to women’s health services, lack of health insurance) (2) mental health or substance use concerns (peace of mind/stress/mental health, depression, counseling for mental health, treatment for depression, problems with drug dependency, daily alcohol use, drug/alcohol treatment program, smoking (tobacco) and help to quit smoking; (3) financial and food concerns (access to healthy food, Inability to find work, inability to meet financial obligations, heat or electricity turned off, inability to buy enough food, during your recent pregnancy, need for food subsidies, state sponsored health plan, Temporary Assistance for Needy Families (TANF), assistance from Indian Health Services, Public child-care subsidies; (4) housing concerns (housing subsidies or assistance, access to safe housing, housing instability), and (5) safety concerns (Violence against family or community and violence in neighborhood, intimate partner violence, police violence, yourself or someone in your family, safe house or shelter from abuse (See Appendix for specific factors captured in each social risk factor category). Creating these 5 categories also allowed us to offer further dimension to our analysis which would have been otherwise unavailable due to small cell sizes for individual questions.

We computed descriptive statistics for the sample. We conducted chi-square analyses and independent t-tests in STATA Statistics version 17 to assess differences between the three provider types and setting exposure groups across socio-demographic and clinical characteristics and self-reported experiences of care. Outcomes were dichotomized in order to calculate the odds of different care experiences. Using bivariate logistic regression, the odds of experiencing each outcome were estimated to compare community-based midwifery and hospital-based midwifery care with hospital-based physician care (reference group). Outcome variables with statistically significant relationships were included in multivariate logistic regression models adjusted for socio-demographic and clinical characteristics as well as social risk factors that differed significantly between the three exposure groups. The significance level for all analyses was set a priori at 0.05. To further elucidate the impact of setting on experiences of care, we ran a secondary logistic regression analysis comparing midwives providing care in community settings (home and freestanding birth centers) to midwives providing care in hospital-based settings (reference group) for four primary outcomes: autonomy (MADM), respect (MORi), mistreatment (MIST), and time spent.

Research conducted by Vedam et al. [22] and Yang et al. [23] demonstrated the potential for midwifery care to improve population level health outcomes if buttressed by structural policies that support autonomous midwifery practice, such as licensing, full scope of practice, access to midwifery care, fair insurance reimbursement and enhanced regulations. In states where midwives practiced with greater autonomy, Yang et al. [23] demonstrated lower odds of cesarean delivery, preterm birth, and low birth weight, compared with states with more limited midwifery autonomy. Vedam et al. [22] also provided strong evidence that in states where the regulatory environment for midwifery practice facilitated autonomous, full scope of practice, there were far better outcomes for women and childbearing people—including higher rates of spontaneous vaginal delivery, vaginal birth after cesarean (VBAC), and breastfeeding, and significantly lower rates of cesarean, preterm birth, low birth weight infants, and neonatal death. Both studies suggest that midwives practicing to the fullest expanse of their scope is critical to achieving optimum clinical outcomes. Subsequent investigations demonstrate the effect of midwifery care integration on experiential outcomes. Our findings support that there is a potent influence of the context for midwifery practice and enabling practice environments in institutional settings on experiential outcomes [27‐29, 48].

Given the hierarchical structure and organization of maternity care in the US—as an obstetrician-led, hospital-centric, technocratic space, more robust methods are needed to also differentiate the impact of the models and settings of care, a deliberate move away from analysis by provider type alone. Our findings suggest that midwives providing their care outside of institutions—in homes and freestanding birth centers—are better able to support core principles of person-centered care rooted in a human rights approach which centers respect, relationship, and autonomy in decision-making. However, when the midwifery care model is within an institution such as a hospital, the constraints and culture of that environment are challenged to support the core values of midwifery [12, 49]. Global inclusion of person-centered care, which is organized to support a person’s autonomy, provide a respectful approach, and support for informed and culturally relevant decision-making is scarce in institutional settings [26, 50]. Our analysis of intragroup differences among midwives (community and hospital) demonstrates that the implementation of the midwifery care model is deeply influenced by the setting itself.

Our findings add to evidence showing the model itself seems to be strongly influenced by the setting in which care is given—with community settings (home and freestanding birth centers) offering greater likelihood of support and the hospital settings being limited by the constraints of a medical approach to care which deprioritizes experiential outcomes [51]. As presently enacted, hospital maternity settings do not provide an enabling environment for full realization of the midwifery model of care. An enabling environment is defined as the social system within which people function [52]. All the rules, laws, policies, power relations and social norms that govern engagement in these spaces are considered impactful to the overall functionality and culture of a system [53]. This confirms global understanding that the culture and organization of the care environment itself serves as a limiting or enabling factor in ensuring high quality care [48, 49]. Singular focus on provider behaviors fails to address the complex and intersecting factors that determine the quality of experiential care.

Challenges in operationalizing the midwifery model of care in hospitals are often due to the dominant physician-based practice model, which is pathology-focused and technocratically driven [54, 55]. The tendency for practitioners in hospitals to rely heavily on interventions creates a challenge to practice autonomy for both the service user and the midwife [12, 49, 56]. The confines of the practice policies, approaches to risk stratification and interprofessional dynamics are all factors that limit the midwifery model from flourishing in hospital and institutional settings [48, 57]. As Newnham and Kirkham argue “large institutions that prioritize a midwife–institution relationship over a midwife–woman relationship are in themselves unethical and inimical to the midwifery philosophy of care.” [58, p. 2147]. Our findings suggest that while ‘how’ care happens matters, ‘where’ that care happens is equally important.

Deep systemic and organizational changes are necessary to support the integration of midwifery across the range maternity care services. Two strategic directions and impact investments could enhance perinatal care services: (1) integration of a human rights-based framework across all care delivery settings and (2) restructuring hospital-based care to allow for more optimal midwifery care integration to build facility policies to enable the midwifery care model to thrive.

Maternal health care often happens without explicit commitment to a human rights-based approach to care. A limited framing of professional ethics as solely dependent on interpersonal interactions and biases, fails to consider the structural and institutional factors that also threaten human rights and dictate the social nature of human interactions [59, 60]. Our findings indicate a clear need to improve experiences of care in hospital settings and shift institutional approaches to better align with person-centered models of care. A pragmatic example of how healthcare praxis can incorporate rights-based principles is offered in the ‘Black Birthing Bill of Rights’ [61]. The resource visually outlines core principles of autonomy and respect in birth. It is also intended to provide guidance to hospitals, health care providers, government health agencies and others to “change/improve their ethic, policies, and delivery approach to serving Black women and persons throughout the birthing process” [61]. This is a clear and direct example of orienting healthcare practice and relationships to uphold the core principles of human dignity, autonomy and self-determination.

Given the prevalence of midwives working within the hospital setting—their philosophical approach to care can be difficult to actualize as institutional values are focused on profitability and risk profiling designed to locate pathology. These values ultimately drivs divestment from supporting physiologic labor and birth and the psychosocial components of care [49, 57]. Maternity care in the US is a medical model of care led by physicians which predominantly operates through a hierarchical approach over a collaborative approach [62]. For a person-centered care model to flourish in all birth settings, a horizontal approach that promotes active collaboration built on principles of professional regard, mutual respect, and trust in the expertise offered by each provider type is essential [63]. For example, care bundles with simulations solely focused on clinical management without consideration of person-centeredness and interprofessional collaboration perpetuate the gap in our ability to measure and address the psychosocial dynamics that impact outcomes. To address the comprehensive needs of women and childbearing people, healthcare quality improvement initiatives must incorporate affective and social dimensions of care as system level drivers of care—not only relegated to the responsibility of individual providers but embraced to orient the organization and implementation of care. Research continues to show that respect, autonomy and informed decision making are highly valued by perinatal care service users—both locally and globally [25, 64] and are key to ensuring anti-oppression, culturally respectful care becomes the new standard of care [25, 65]. Increased integration of midwifery care into all care pathways—at home, in freestanding birth centers and in hospitals, may ensure greater ability to offer high quality and culturally safe experiences of care to all childbearing people.

The sampling strategy for this study was not designed to be representative, limiting the generalizability of the findings. Participants of color were oversampled to explore factors associated with wide disparities in U.S. birth outcomes by race/ethnicity and the systematic exclusion of racialized populations from research. This survey also intended address the lack of data on experiences of childbearing care outside hospital settings. In previous studies, analyses stratified by both birth setting and provider were limited due to small sample sizes. Our sampling strategy allowed for rigorous comparisons by birth setting and provider type.

A large proportion of our sample reported living in New York state. However, the strength of findings even after controlling for variation in sociodemographic characteristics, and pregnancy risk factors suggest that differences in experiences of care based on setting are unlikely to be limited to a particular state or demographic but may represent a large-scale, systemic problem.

It is also possible that multiple participants gave birth in the same hospitals (especially for births in New York) and we were unable to control for clustering of births within hospitals. Hence, our analysis does not account for the well-documented variations in the quality of care between hospitals [66]. Additionally, we note that our analysis did not assess the potential mitigating effects of continuity of provider across the arc of prenatal care to labor and birth care on the care experience.