Factors influencing women’s perceptions of shared decision making during labor and delivery: Results from a large-scale cohort study of first childbirth

Abstract

Objective

To examine correlates of shared decision making during labor and delivery.

Methods

Data were from a cohort of women who gave birth to their first baby in Pennsylvania, 2009–2011 (N = 3,006). We used logistic regression models to examine the association between labor induction and mode of delivery in relation to women’s perceptions of shared decision making, and to investigate race/ethnicity and SES as potential moderators.

Results

Women who were Black and who did not have a college degree or private insurance were less likely to report high shared decision making, as well as women who underwent labor induction, instrumental vaginal or cesarean delivery. Models with interaction terms showed that the reduction in odds of shared decision making associated with cesarean delivery was greater for Black women than for White women.

Conclusions

Women in marginalized social groups were less likely to report shared decision making during birth and Black women who delivered by cesarean had particularly low odds of shared decision making.

Practice implications

Strategies designed to improve the quality of patient-provider communication, information sharing, and shared decision making must be attentive to the needs of vulnerable groups to ensure that such interventions reduce rather than widen disparities.

1. Introduction

Nearly 4 million women give birth in the U.S. each year; 32% deliver by cesarean, and at least 22% give birth following an induction of labor [1,2]. There is broad agreement that cesarean delivery is currently overused in the U.S., given that maternal and neonatal morbidity and mortality have not decreased as cesarean delivery has become more common [3]. Furthermore, cesarean delivery is a major abdominal surgery and carries increased risk of complications for the mother and baby, as well as increased risk of placenta problems in future pregnancies compared to vaginal delivery [3,4]. There is less consensus on whether labor induction is overused, but the major concern is that induction of labor when it is not medically necessary may lead to a higher risk of delivering by cesarean [3,5–7]. Despite these concerns about overuse of obstetric procedures, little is known about how procedure utilization in childbirth relates to women’s involvement in decision making during labor.

Shared decision making, a process in which clinicians and patients work together to make choices about screening, treatment, and other aspects of care, is a dimension of patient-centered care and a key aspect of care quality [8]. While studies have found mixed results of the effects of shared decision making on outcomes [9], some U.S. research has linked shared decision making with higher quality care in a variety of contexts, including depression treatment [10], diabetes management [11], and cancer treatment [12]. Shared decision making in maternity care could have benefits such as reducing decisional conflict, increasing satisfaction with the delivery experience, and generating more positive feelings toward the newborn and fewer depressive symptoms [13], and some maternity care providers have called for more attention to shared decision making as part of increasing the provision of patient-centered care. Advocates of patient-centered care have noted that not all patients may wish to participate in decision-making to the same degree. Although no studies that we are aware of have assessed women’s desire for shared decision making in maternity care specifically, the proportion of patients reporting a desire for shared decision-making in general has been increasing over time [14]. Additionally, some research has shown that patients who participate in decision making report better quality of care and higher quality physician communication regardless of their stated preferences for decision making role [15,16].

Research on decision making in the context of maternity care has focused on women with a prior cesarean delivery and “elective” or “maternal request” cesarean delivery, and found that decision aids or formal counseling programs can increase knowledge and reduce decisional conflict [17–21]. One study conducted at two northern California hospitals found no relationship between women’s perceptions of shared decision making during prenatal care and odds of delivering by cesarean [22]. However, despite the fact that women are actively involved in decisions during the clinical management of labor, no previous research that we are aware of has examined predictors of shared decision making during the intrapartum period of labor and childbirth.

The quality of clinician-patient communication and shared decision making vary by the race/ethnicity and socioeconomic status (SES) of the patient. High-quality communication may be a necessary prerequisite for patients to be involved in making decisions about their care, and racial/ethnic minority patients and lower-SES patients tend to experience poorer-quality communication with clinicians [23–25]. White patients are more likely to experience shared decision making compared to racial/ethnic minority patients [26–30]. Lower-SES patients may have lower levels of health literacy, which can be a barrier to engaging in shared decision making [31]. There are several potential reasons for these disparities in the quality of clinician-patient interaction. For example, there may be greater social separation between clinicians and racial/ethnic minority patients, given that only about 4% of U.S. physicians identify as Black or African American [32]. Greater social separation may contribute to a greater imbalance of power in the patient-clinician relationship, which has been identified as a barrier to shared decision making [33]. Implicit bias among clinicians may also be a factor in communication quality and decision-making style [34]. Minority and lower-SES patients are more likely to be seen in low-resource settings where time pressures may be more intense, which is another barrier to shared decision making [33].

To better understand the relationship between procedure use in childbirth, race/ethnicity and SES, and shared decision making, this study aimed to 1) characterize the association between race/ethnicity, SES and shared decision making during childbirth, 2) examine whether shared decision making varies by use of obstetric procedures such as labor induction and cesarean delivery, and 3) assess whether the relationship between obstetric procedure use and shared decision making is different depending on the woman’s race/ethnicity or SES.

2. Methods

2.1 Data and Sample

Data were from the First Baby Study, a cohort of 3,006 women who gave birth to a first, singleton baby in a Pennsylvania hospital between 2009 and 2011. Approval for the First Baby Study was granted by the Penn State College of Medicine Institutional Review Board (IRB) and by the IRBs of participating hospitals. Study participants gave informed consent. Women were interviewed in the third trimester of pregnancy and again one month after the birth, and periodically through 36 months postpartum. The primary purpose of the First Baby Study was to assess the relationship between delivery mode and subsequent childbearing. Interview responses were linked to birth certificate and hospital discharge data. This analysis drew variables from the baseline and 1-month interviews, as well as the hospital discharge and birth certificate data. Details about the data source for each measure are available in Appendix 1. We chose to exclude women who delivered by planned cesarean in order to focus on decision making in the context of labor. The analytic sample consisted of women who gave birth vaginally or by unplanned cesarean, and with no missing values on covariates (N=2,787).

2.2 Measures

Delivery Decision Making Scale

The main dependent variable of interest is shared decision making about delivery care, measured with the Delivery Decision Making Scale (DDMS), a scale developed by the principal investigator of the First Baby Study following qualitative interviews of women shortly after childbirth. The scale contains 6 items (shown in Table 1), with response choices of true or false. Scale items focus on women’s perceptions of involvement and satisfaction with the decision making process, rather than the concrete practices constituting formal definitions of shared decision making. We reversed coding for the one negative item, and assigned one point for each “true” response, yielding a scale ranging from 0–6, where scores indicate higher levels of shared decision making. The scale had a Kuder-Richardson score of .69, which is good for a true-false instrument, particularly one with only 6 items. Each of the 6 items contributed appropriately to the total score. More than half of the respondents (64.0%) obtained the maximum possible total score of 6. Therefore, we dichotomized the total scores, comparing those who scored 6 (totally satisfied with delivery decision making) to those with scores of less than 6.

Table 1.

Delivery Decision Making Scale items.

My opinions and desires about delivery were taken into account I was treated with respect throughout the delivery process. I had an equal say in how my baby was delivered. I had an equal say in when my baby was delivered. I was not given a choice as to when and/or how my baby would be delivered. I was pleased with the way decisions were made as to when and how my baby would be delivered.

Race/ethnicity and SES

This analysis categorized women as White, Black, Latina, or other, based on their self-reported race and ethnicity. We used two measures of socioeconomic status: education level (less than high school, some college, and Bachelor’s degree or higher) and insurance type. Because three-quarters of women in the cohort were privately insured and very few women (<3%) were uninsured, insurance type was measured as private or non-private.

Obstetric procedures: labor induction, cesarean delivery, and assisted vaginal delivery

Women were asked whether a healthcare provider had done anything to cause their labor to begin, and if so, the reason for the induction and what method(s) were used. We initially coded all women as having induced labor if they said that a provider had done something to try to cause their labor to begin, but we used the additional questions to confirm whether it seemed likely that the woman had had a medical induction. Ultimately, women were categorized as induced if they met all of the following criteria: 1) they said they were induced, 2) they reported arriving at the hospital not in labor, 3) if they did not report a reason for induction that suggested that they were augmented rather than induced, and 4) if they reported at least one medical method of induction (artificial rupture of membranes, Pitocin, cervical medication, oral medication, or Foley catheter/balloon methods).

Women were asked whether they delivered vaginally or by cesarean. Women who delivered vaginally were asked whether forceps or a vacuum extractor were used. We created two delivery mode variables. The first had 3 categories: 1) spontaneous vaginal delivery, 2) assisted vaginal delivery, and 3) cesarean delivery. The second compared cesarean delivery to all vaginal births.

Covariates

Covariates included the following socio-demographic characteristics: age (18–24, 25–30, or 31–36), partnership status (married and living together, not married but living with partner, other), nativity (foreign- or US-born), and rural zip code of residence. We also controlled for maternal health conditions, including pre-pregnancy or gestational diabetes, pre-pregnancy or gestational hypertension, and pre-pregnancy obesity. Finally, we included in our models whether the woman was supported in labor by a doula, the baby’s gestational age at birth (less than 37 weeks, 37–38 weeks, 39–40 weeks, 41+ weeks), and whether actual delivery mode aligned with the woman’s prenatal delivery mode preference.

2.3 Analysis

We used two-way tabulation with chi-square tests to assess significant bivariate associations between women’s characteristics and shared decision making. In multivariate analyses, we used logistic regression models to estimate the relationship between race/ethnicity, SES, obstetric procedures, and shared decision making while controlling for covariates. Finally, to explore whether obstetric procedures were differently related to patient shared decision making depending on race/ethnicity or SES (moderation), we added interaction terms to the models. Where the model fit was improved by the inclusion of an interaction term (indicated by a statistically significant coefficient for the interaction term), we calculated predicted probabilities and sample average marginal effects of obstetric procedure use on shared decision making by race/ethnicity, education and insurance type. The marginal effects for each subject in the data were calculated based on their actual covariate values, and then the results were averaged across all subjects [35]. All multivariate models employed clustered standard errors to account for correlation among women giving birth in the same hospital. All analyses were conducted using Stata 14.

3. Results

3.1 Sample Characteristics

Characteristics of participants are shown in Table 2. Eighty-three percent of women were White. About 17% had a high school education or less, 27% had some college, and over 50% had either Bachelor’s degree or higher. Half of the women were between the ages of 25 and 30, and 71% were married. In this cohort of women giving birth for the first time, nearly 25% delivered by unplanned cesarean. About 64% of the participants obtained the maximum score of 6 on the Delivery Decision Making Scale, indicating a high level of shared decision making.

Table 2.

Characteristics of analytic sample, First Baby Study (N=2,787).

	N	%
Race/ethnicity
White	2325	83.4
Black	200	7.2
Latina	150	5.4
Other, multiple race	112	4.0
Education level
High school degree or less	460	16.5
Some college	745	26.7
Bachelor's degree or higher	1582	56.8
Insurance type
Private	2115	75.8
Non-private	672	24.1
Age
18–24	758	27.2
25–30	1365	49.0
31–36	664	23.8
Partnership status
Married and living together	1964	70.5
Not married but living with partner	496	17.8
Other	327	11.7
Born outside United States	171	6.1
Rural zipcode of residence	233	8.4
Pre-pregnancy or gestational hypertension	197	7.1
Pre-pregnancy or gestational diabetes	176	6.3
Pre-pregnancy obesity	549	19.7
Labor support from a doula	185	6.6
Gestational age category
Less than 37 weeks	111	4.0
37 or 38 weeks	527	18.9
39–40 weeks	1708	61.3
41+ weeks	441	15.8
Delivery mode consistent with prenatal preference	2105	75.5
Birth mode
Spontaneous vaginal	1842	66.1
Assisted vaginal	257	9.2
Cesarean	688	24.7
Labor was induced	947	34.0
High level of shared decision making	1775	63.7

3.2 Shared decision making by delivery mode, race/ethnicity and SES

Perceptions of shared decision making varied by race/ethnicity, SES, and obstetric procedure use. A significantly smaller proportion of women in each racial/ethnic minority group reported a high level of shared decision making compared to White women (Table 3). Women with lower levels of education and women without private insurance were also less likely to have a high level of shared decision making, compared to women with more education and those with private insurance. Sixty-nine percent of women with spontaneous vaginal deliveries reported a high level of shared decision making, compared to 61% of women with assisted vaginal deliveries, and 51% of women with cesarean deliveries. Women who had their labor induced were less likely to report a high level of shared decision making compared to women with spontaneous labor (58% vs. 67%).

Table 3.

Unadjusted shared decision making by race/ethnicity, and SES, and obstetric procedure use, First Baby Study (N=2,787).

	High DDMS score
	%	p
Race/ethnicity and SES
Race/ethnicity		<0.001
White	66.4
Black	42.5
Latina	52.0
Other, multiple race	60.7
Education level		<0.001
High school degree or less	50.9
Some college	60.9
Bachelor's degree or higher	68.7
Insurance type		<0.001
Private	67.8
Public, other or uninsured	50.9
Obstetric procedures
Labor induction		<0.001
No	66.7
Yes	57.8
Birth mode		<0.001
Spontaneous vaginal	68.7
Assisted vaginal	61.1
Cesarean	51.3

3.3 Multivariate results estimating odds of high shared decision making

After adjusting for covariates (Table 4), Black women had lower odds of experiencing a high level of shared decision making compared to White women (AOR=0.58, p<0.001), but there was no longer a statistically significant difference in shared decision making between Latina women and White women. The measures of SES were also associated with adjusted odds of high level of shared decision making; women with a Bachelor’s degree had higher odds of experiencing a high level of shared decision making compared to women with a high school degree or less (AOR=1.31, p=0.023), while women with non-private insurance had lower odds (AOR=0.76, p=0.007). Obstetric procedures were associated with lower levels of shared decision making. Women with assisted vaginal delivery (AOR=0.67, p<0.001) and unplanned cesarean delivery (AOR=0.53, p=.008) had lower odds of high level of shared decision making compared to women with spontaneous vaginal delivery. Labor induction was also independently associated with lower odds of high level of shared decision making (AOR=0.77, p=.007).

Table 4.

Adjusted odds of high level of shared decision making by birth mode and other characteristics, First Baby Study (N=2,787).

	OR	p
Race/ethnicity
White	1.00
Black	0.58	<0.001
Latina	0.73	0.196
Other, multiple race	0.82	0.287
Education level
High school degree or less	1.00
Some college	1.16	0.155
Bachelor's degree or higher	1.31	0.023
Insurance type
Private	1.00
Public, other or uninsured	0.76	0.007
Delivery mode
Spontaneous vaginal	1.00
Assisted vaginal	0.67	<0.001
Unplanned cesarean	0.53	0.008
Labor induction	0.77	0.007

odels adjust for age, partnership status, nativity, rurality, diabetes, hypertension, obesity, doula support, gestational age, and consistency of delivery mode with prenatal preference

3.4 Models with interaction terms

Finally, we investigated whether the association between delivery mode and shared decision making varied by race/ethnicity and SES. To simplify these models, we compared cesarean delivery to vaginal delivery without distinguishing between spontaneous and assisted vaginal delivery. There was a statistically significant interaction between Black race/ethnicity and cesarean delivery (not shown), so we calculated predicted probabilities to investigate this interaction (Figure 1). White women who delivered vaginally had a predicted probability of high level of shared decision making of 0.68, while White women who delivered by cesarean had a predicted probability of high level of shared decision making of 0.55 – a difference of 13 percentage points. For Black women, cesarean delivery was associated with a significantly larger reduction in the predicted probability of high level of shared decision making: Black women who delivered vaginally had a predicted probability of high level of shared decision making of 0.61, and Black women who delivered by cesarean had a predicted probability of high level of shared decision making of 0.28 – a difference of 33 percentage points. These findings indicate that the magnitude of the reduction in the chance of reporting a high level of shared decision making associated with cesarean delivery was much larger for Black women than for White women.

Figure 1.

Predicted probability of high level of shared decision making by delivery mode and race/ethnicity, from models including interaction terms.

4. Discussion and Conclusion

4.1 Discussion

Most first-time mothers in this study felt very involved in the delivery decision-making process, with 64% reporting the highest possible score on the scale assessing decision-making involvement. However, women reporting lower levels of shared decision making were disproportionately likely to be from racial/ethnic minority groups, less educated, and to lack private insurance. These patterns persisted in multivariate models, revealing a disproportionately less engaged decision experience for pregnant patients in more marginalized social groups. Additionally, after controlling for other characteristics, obstetric procedure use continued to be a strong predictor of women’s perceptions of involvement in decision making, with labor induction, assisted vaginal delivery, and cesarean delivery all independently contributing to lower perceived involvement. An earlier study found no association between women’s perceptions of shared decision making in prenatal care and cesarean delivery [22]; the difference in our findings may have to do with the fact that decision making in prenatal care is more distal from the intrapartum experience and the clinical decision to deliver by cesarean. Furthermore, we used a newly-established measure of shared decision making, which is tailored to the childbirth experience and captured women’s satisfaction with the delivery decision making. Standardized measures and routinely collected data on women’s maternity care experiences, including shared decision making, communication quality, and other aspects of patient-centered care, may enhance the ability of stakeholders in the maternity care system to examine quality of care across patient settings and over time.

Despite attention to shared decision making as an important component of patient-centered care, most research and interventions have focused on ambulatory or outpatient care in the clinic, not on the context of acute care in the hospital, such as the intrapartum period. This study found that women with obstetric interventions including induction of labor, assisted vaginal delivery, and unplanned cesarean delivery reported lower levels of shared decision making. Even when these procedures are medically necessary, many situations are not the type of emergency that would preclude taking the time to involve the woman in the decision-making process. However, increasing women’s shared decision making during childbirth (or patient shared decision making in other types of inpatient care) requires some adaptations to existing procedures to achieve higher levels of shared decision making.

For example, decision aids are an effective way of ensuring that patients are provided balanced information prior to making a decision [36], and have been shown to be effective at increasing women’s knowledge and reducing decisional conflict for several scenarios in pregnancy and birth [20,21,37]. However, decision aids may not be practical for the scenario of an unplanned cesarean delivery during labor, where there may be wide variation in the potential reasons for considering a cesarean delivery and the woman is experiencing labor. Communication skills training for clinicians that is specifically targeted toward intrapartum care and that incorporates principles of shared decision making may be useful to enhance the quality of the decision-making process without the use of formal decision aids [38]. Decision aids might be appropriate for making decisions about common reasons for inducing labor (such as pregnancy continuing beyond 41 weeks), and organizations like the California Maternal Quality Care Collaborative and a partnership between Childbirth Connection and the Informed Medical Decisions Foundation have projects underway to develop and pilot decision aids to promote shared decision making in more maternity care situations [39,40].

These results also indicate that eliminating racial/ethnic and SES disparities in the receipt of patient-centered care and in obstetric procedures should be a priority. Women who delivered by cesarean reported being less involved in decisions about their delivery overall, but Black women who delivered by cesarean reported strikingly lower levels of shared decision making compared to White women. This suggests that the decision-making process leading up to cesarean delivery was more likely to be problematic for Black women. These disparities could have to do with systemic factors that differentially affect the choices and outcomes black women face. For example, it could be related to the quality of the hospitals where women of different racial/ethnic groups give birth (i.e. Black women may be more likely to deliver in lower quality hospitals) [41], or due to differential treatment within the same hospital [34]. This could also relate to the cumulative and intergenerational risks of racism that interact with socioeconomic and educational opportunities across the life course [42]. The reproductive and maternal health decrements that black women, in particular, suffer in the US manifest via disparate access to family planning, preventive services, and prenatal care and result in disproportionate rates of maternal morbidity and mortality [43–45]. These patterns have their roots in a long history of exploitation of black women in reproductive and maternal health research and services access [46,47].

Latina women were less likely to report a high level of shared decision making compared to White women, but this difference was no longer significant in multivariate models. Additionally, the association between delivery mode and shared decision making appeared to be weaker among Latina women than among White women, although the difference in association was not statistically significant. There is substantial heterogeneity among individuals who identify as Latina, in ways that can affect childbirth experiences. For example, one study that assessed chances of cesarean among low-risk women by detailed maternal ethnicity found that women of certain Latina ethnic groups had similar odds of cesarean compared to White women, while women in other Latina ethnic groups had elevated odds of cesarean compared to White women [48]. Prior studies explicitly documenting racial/ethnic differences in shared decision making have demonstrated statistically meaningful Black-White differences [26–28], and have also shown differences between racial/ethnic minority groups (combined), in comparison to Whites [29] Future studies should examine delivery decision-making among Latina women, including relevant subgroups of Latinas.”

In this study population, Black and Latina women had elevated odds of cesarean delivery compared to White women, even after accounting for medical risk factors such as diabetes, hypertension, obesity, and gestational age at birth[49] , consistent with prior studies [50,51]. Greater use of obstetric procedures among Black and Latina women, compared with Whites may be another piece of the broader picture of disparities in delivery decision-making.

While this study produced novel and important findings, there are also limitations to consider. First, this study used a new scale developed specifically to assess women’s experiences with decision making during delivery. A 6-item scale with true/false response choices may not be optimal, and we found that the scale was highly skewed, with the majority of women in this study scoring at the top of the scale. Several studies have compared patient reports of shared decision making and observer report, and found that there are substantial differences in the prevalence of shared decision making depending on which measure is used [52–54]. However, disagreement was almost universally in the direction of the patient reporting more involvement in the decision-making process compared to the observer rating. Assuming that women’s reports in the First Baby Study were likewise inflated, this would attenuate the associations with shared decision making, suggesting that the associations identified in this study may be underestimates. Future research should work to develop a valid and reliable instrument to measure shared decision making that is relevant for women’s childbirth experiences specifically. Second, the First Baby Study cohort over-represents women from more socially privileged groups compared to the population of first-time mothers in Pennsylvania [55], and thus it is possible that the results are not generalizable to all childbearing women. Finally, women who delivered by cesarean may have reported less satisfaction with decision making simply because they had hoped for a vaginal delivery. However, our models controlled for the alignment between women’s prenatal delivery mode preferences and actual delivery mode, addressing this concern.

4.2 Conclusions

This analysis found that women in marginalized social groups were less likely to experience shared decision making during birth, as were women who underwent labor induction or cesarean delivery. Black women delivered by cesarean had particularly low odds of experiencing shared decision making.

4.3 Practice Implications

Efforts are needed to improve the decision-making process for women in labor, including in cases where obstetric interventions are medically necessary, and to increase equity in maternity care. Some states have considered or enacted legislation that encourages or requires the use of shared decision making (e.g. through the use of a certified decision aid) as part of informed consent for certain medical decisions [56]; if adapted to the maternity care context, such requirements could help to ensure that shared decision making is more uniformly adopted across healthcare settings. Additionally, strategies designed to improve the quality of patient-provider communication, information sharing, and shared decision making must be attentive to the needs of vulnerable groups to ensure that such interventions reduce rather than widen disparities [57].

Highlights.

• Few studies have examined shared decision making in the childbirth context

• Obstetric procedure use was associated with less shared decision making

• Shared decision making was less common among lower-SES women and women of color

• Black women who delivered by cesarean had strikingly low shared decision making

Appendix 1. Data source for measures used in analysis

Measure	Data source
Delivery decision making scale	1-month postpartum interview
Race/ethnicity	Baseline interview
Education	Baseline interview
Insurance type	Hospital discharge record
Labor induction	1-month postpartum interview
Cesarean delivery	1-month postpartum interview
Age	Baseline interview
Partnership status	Baseline interview
Nativity	Baseline interview
Rural zip code	Baseline interview
Diabetes	Baseline interview
Hypertension	Baseline interview
Obesity	Baseline interview
Labor support from a doula	1-month postpartum interview
Baby's gestational age at birth	Birth certificate
Delivery mode alignment with prenatal preferences	Baseline interview and 1-month postpartum interview