Structural Racism and Use of Labor Neuraxial Analgesia Among Non-Hispanic Black Birthing People

Abstract

Objective:

To assess the association between structural racism and labor neuraxial analgesia use.

Methods:

This cross-sectional study analyzed 2017 US Natality data for non-Hispanic Black and White birthing people. The exposure was a multidimensional structural racism index measured in the county of the delivery hospital. It was calculated as the mean of three Black-to-White inequity ratios (ratios for lower education, unemployment, and incarceration in jails), and categorized into terciles, with the 3^rd tercile corresponding to high structural racism. The outcome was the labor neuraxial analgesia rate. Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) of neuraxial analgesia associated with terciles of the index were estimated using multivariate logistic regression models. Black and White people were compared using an interaction term between race and ethnicity and the racism index.

Results:

Of the 1,740,716 birth certificates analyzed, 396,303 (22.8%) were for Black people. For Black people in the 1^st tercile of the racism index, the labor neuraxial analgesia rate was 77.2%, in the 2^nd tercile 74.7%, and in the 3^rd tercile 72.4%. For White people the rates were 80.4%, 78.2%, and 78.2%, respectively. For Black people, compared to the 1^st tercile of the racism index, the 2^nd tercile was associated with 18.4% (95% CI: 16.9, 19.9) decreased adjusted odds of receiving neuraxial analgesia and the 3^rd tercile with 28.3% (95% CI: 26.9, 29.6) decreased adjusted odds. For White people, the decreases were 13.4% (95% CI: 12.5, 14.4) in the 2^nd tercile and 15.6% (95% CI: 14.7, 16.5) in the 3^rd tercile. A significant difference in the odds of neuraxial analgesia was observed between Black and White people for the 2^nd and 3^rd terciles.

Conclusions:

A multidimensional index of structural racism is associated with significantly reduced odds of receiving labor neuraxial analgesia among Black people and, to a lesser extent, White people.

PRECIS

Structural racism is associated with significantly reduced use of labor neuraxial analgesia among non-Hispanic Black people and, to a lesser extent, among non-Hispanic White people.

INTRODUCTION

Labor is among the most painful experiences a woman may encounter in her lifetime.¹ Neuraxial analgesia (i.e., spinal, epidural or combined spinal/epidural analgesia) is the most effective technique to relieve this pain, and it also contributes to reducing the risk of severe maternal morbidity.² About 75% of people in labor receive neuraxial analgesia. Compared to non-Hispanic White birthing people, non-Hispanic Black people are about 10% less likely to receive neuraxial analgesia.^3–5 Approximately 75% percent of non-Hispanic Black people give birth in hospitals treating mostly Black people, which are characterized by lower resources and performance compared to hospitals treating mostly White people.^6–11

Mechanisms accounting for reduced use of labor neuraxial analgesia among non-Hispanic Black people are multifactorial but can be categorized into patient preference, insufficient prenatal education on pain management options during labor, and reduced access to labor neuraxial analgesia in the delivery hospital (e.g., because of lack of 24/7 dedicated obstetric anesthesia team).¹² However, racism could be a significant contributor to racial and ethnic disparities in labor neuraxial analgesia use (Figure 1).¹³ Racism - including structural, institutional and interpersonal racism - is a root cause of racial and ethnic disparities in healthcare access and utilization for birthing people in the United States.^14–17 Structural racism can be estimated at the county or state level using unidimensional indicators or multidimensional indexes which capture the exclusion of Black people from resources (e.g., unemployment), opportunities (e.g., education), or their unfair judiciary treatment (e.g., incarceration).^18–22 This led us to hypothesize that a multidimensional indicator of structural racism measured in the county of the delivery hospital is associated with reduced use of labor neuraxial analgesia among non-Hispanic Black people.

Figure 1:

Hypothesized pathways linking structural, institutional, and interpersonal racism to reduced use of labor neuraxial analgesia. Structural racism refers to a system where public policies, institutional practices, cultural representations, and other norms work together to maintain and perpetuate racial group inequities in housing, education, employment, earnings, benefits, credit, media, health care, or criminal justice.^{46, 47, 54} LNA, labor neuraxial analgesia. Figure adapted from Matern Child Health J 2022;26:661-9. Used with permission.

METHODS

This study analyzing de-identified data was deemed exempt by the Institutional Review Board of the authors’ institution. We report the study according to the STROBE guidelines.²³ Data for this study came from the 2017 restricted access Natality File provided by the National Vital Statistics System (National Center for Health Statistics, Centers for Diseases Control and Prevention). The Natality File is based on the 2003 revised US Standard Certificate of Live Birth, which was gradually implemented across states, from two states in 2003 to fifty states and the District of Columbia (DC) in 2015.²⁴ The Natality File is a census of all live births in the United States and contains comprehensive information on the pregnant people, labor, and delivery.

Hospital data were not collected because they are not available in the Natality file. Furthermore, the Natality file provides no hospital identifier that would allow abstracting hospital data from other data systems (e.g., American Hospital Association Annual Survey). We therefore estimated the characteristics of the delivery hospital at the county level. These characteristics were provided by the Area Health Resource File (year 2017 or 2013-2017 5-year estimate) or by the Vera Institute of Justice (year 2017).^{25, 26} The Area Health Resource File data contain detailed information abstracted from over 50 data sources on health care professions, health facilities, population characteristics, and economics measured at the county level. The Vera Institute of Justice provides the incarceration rate in jails measured at the county level, overall and according to race and ethnicity.

The study sample included birth certificates for vaginal and intra-partum cesarean deliveries from January 2017 to December 2017 (Figure 2). We identified vaginal and cesarean deliveries using a specific checkbox on the birth certificate.²⁷ Since this checkbox did not indicate whether a cesarean delivery was intrapartum or not, we defined a cesarean delivery as intrapartum if associated with at least one of the following elements indicating labor: 1) trial of labor attempted if previous cesarean delivery; 2) induction of labor; 3) augmentation of labor; 4) antibiotics received by the mother during labor; and 5) clinical chorioamnionitis diagnosed during labor or maternal temperature >38°C during labor. In a sensitivity analysis, we limited the study sample to people who had a vaginal birth.

Figure 2:

Flowchart of the study. *A cesarean delivery is defined intrapartum if associated with at least one of the following: 1) trial of labor attempted if previous cesarean delivery; 2) induction of labor; 3) augmentation of labor; 4) antibiotics received by the mother during labor; and 5) clinical chorioamnionitis or maternal temperature >38°C diagnosed during labor. †Reasons for exclusion are not mutually exclusive. ‡Includes 1,381 counties in 45 states and Washington, DC^§¶ and 1,358,143 with labor neuraxial analgesia (78.02%).^§The states included are: AK, AL, AR, AZ, CA, CO, DC, FL, GA, IA, ID, IL, IN, KS, KY, LA, MA, MD, ME, MI, MN, MO, MS, MT, NC, ND, NE, NH, NJ, NM, NV, NY, OH, OK, OR, PA, SC, SD, TN, TX, UT, VA, WA, WI, WV, WY. ^¶1,296,146 women (74.5%) gave birth in their county of residence.

Exclusion criteria were: 1) hospital or residence county with less than 100 non-Hispanic White residents or 100 non-Hispanic Black residents, 2) hospital or residence county with less than 100 annual in-hospital births, 3) missing Black-to-White ratios for low education, unemployment rate, or jail incarceration rate, 3) missing information on anesthesia care, 4) maternal race and ethnicity not corresponding to non-Hispanic White or to non-Hispanic Black, 5) multiple pregnancy, 6) not in-hospital birth, and 7) mother not residing in the United States.

The exposure of primary interest was a multidimensional index of structural racism measured in the county of the delivery hospital, with a higher index value indicating more structural racism.^{18, 19, 28–32} The index was calculated as the mean of three non-Hispanic Black to non-Hispanic White inequity ratios: 1) ratio for lower education level defined as less than high-school diploma; 2) ratio for unemployment rate; and 3) ratio for incarceration rate in jails. We focused on incarceration in jails and not in prisons because there is about one jail per county in the US, which is under local control, but not one prison per county, which is under state or federal control. We did not analyze institutional and interpersonal racism because we have no information on these two forms of racism in the data.

Since people may give birth in a hospital not located in their residence county, we conducted a sensitivity analysis limited to people who gave birth in their residence county.

The outcome measure of primary interest was the labor neuraxial analgesia rate. In the birth certificate, neuraxial analgesia is reported in a specific checkbox (“Epidural or spinal anesthesia during labor”), and defined as the “administration to the mother of a regional anesthetic for control of the pain of labor (i.e., delivery of the agent into a limited space with the distribution of the analgesic effect limited to the lower body)”.²⁷ The reported sensitivity of the checkbox neuraxial analgesia in the birth certificate is greater than 85%.^{33, 34} The Natality Files do not contain detailed information on the type of neuraxial analgesia precluding the analysis of specific techniques (spinal, epidural, or combined spinal-epidural).

Maternal characteristics and comorbidities directly recorded from birth certificate data included: maternal age (≤ 19, 20-29, 30-39, or ≥ 40 years); education level (less than high school, high school with no diploma, high school graduate or general educational diploma, and college or higher); health insurance (Medicaid, private insurance, self-pay, or other); pre-pregnancy body mass index (≤ 18.4, 18.5-24.9, 25.0-29.9, 30.0-34.9, or ≥ 35 kg/m²); and preexisting or gestational diabetes or hypertension (binary, yes/no).

Obstetrical characteristics directly recorded from birth certificate data included: previous cesarean section (yes/no); month prenatal care began (1^st-3^rd, 4^th-6^th, ≥ 7^th, or no prenatal care); number of prenatal visits (continuous); delivery during a weekend (yes/no); mother transferred in (i.e., transfer from another facility for maternal medical or fetal indications for delivery) (yes/no); parity (nulliparous or parous); gestational age at delivery (≤ 33 completed weeks, 34-38 completed weeks, or ≥ 39 completed weeks); non-cephalic presentation (yes/no); induction of labor (yes/no); augmentation of labor (yes/no); antibiotics during labor (yes/no); fever or chorioamnionitis during labor (yes/no); attendant at birth (doctor of Medicine, doctor of Osteopathy, midwife, or other); and birth weight (<2500 grams, 2500-4000 grams, or > 4000 grams).

We estimated the characteristics of the delivery hospital at the county level, including: hospital location (metropolitan, micropolitan, or non-metropolitan and non-micropolitan); number of in-hospital births (continuous); number of physician anesthesiologists (continuous, per 1000 hospital births); number of certified registered nurse anesthetists (continuous, per 1000 hospital births); number of obstetricians and gynecologists (continuous, per 1000 hospital births); proportion of non-Hispanic Black residents (continuous, %); proportion of residents with less than high school diploma (continuous, %); proportion of unemployed residents (continuous, %); and incarceration rate in jails (continuous, per 100,000 residents).

Statistical analysis was performed with R version 4.1.2.³⁵ No clinically meaningful change in labor neuraxial analgesia rate associated with the racism index was defined a priori.

The distribution of the labor neuraxial analgesia rate across counties was examined visually using a caterpillar plot. Then, the rate was compared between non-Hispanic Black and non-Hispanic White people using the absolute standardized difference (ASD). An ASD greater than 0.10 (10.0%) was chosen to indicate a significant imbalance between groups. Comparison of sociodemographic, comorbidities, pregnancy, labor and delivery, and hospital county characteristics between people who received labor neuraxial analgesia and people who did not also used the ASD.

The distribution of the racism index across counties was also examined using a caterpillar plot and the index compared between non-Hispanic Black and non-Hispanic White people using the ASD.

The racism index was categorized into three terciles, with the third tercile corresponding to the highest structural racism. Terciles were calculated independently for non-Hispanic Black people and for non-Hispanic White people.

For each racial and ethnic group, the labor neuraxial analgesia rate was estimated for each tercile of the racism index. The rate difference was estimated as the difference between the labor neuraxial analgesia rate in the second tercile (or third tercile) and the rate in the first tercile (reference). Crude odds ratios (OR) of labor neuraxial analgesia associated with terciles of the racism index were estimated using univariate logistic regression models, independently for non-Hispanic White people and non-Hispanic Black people. In these models, labor neuraxial analgesia was the dependent variable and the terciles of the racism index the independent variable.

The following variables were used to adjust the odds ratio of labor neuraxial analgesia associated with the terciles of the racism index, independently for non-Hispanic White people and non-Hispanic Black people: body mass index; gestational hypertension; previous cesarean section; parity; induction of labor; augmentation of labor; antibiotics received during labor; chorioamnionitis; and hospital core base statistical area classification. These variables were selected based on the three following criteria (Appendix 1, available online at http://links.lww.com/xxx): 1) variables with an absolute standardized difference > 0.10 (10.0%) in the univariate comparison of people with and without labor neuraxial analgesia (Table 1); 2) variables not on the causal pathway linking structural racism to labor neuraxial analgesia use (e.g., education level) (Figure 1); and 3) variables not occurring after labor neuraxial analgesia placement (e.g., delivery mode). Non-Hispanic Black people and non-Hispanic White people were compared using the regression coefficient (β) of an interaction term between race and ethnicity and the terciles of the racism index in the logistic model. The associated percent change was calculated as $100\times \left(e^{\beta }-1\right)$. A complete case analysis was performed with 2.7% of birth certificates excluded for missing values of the variables used for adjustment.

Table 1:

Univariate comparison of people who received labor neuraxial analgesia and people who did not (45 US states and District of Columbia, 2017).

	Missing	No neuraxial analgesia n = 382,573	Neuraxial analgesia n = 1,358,143	Absolute standardized difference (*)
Sociodemographic characteristics
Age (year)	0			0.055
≤ 19		15,803 (4.1%)	68,089 (5.0%)
20-29		187,109 (48.9%)	680,983 (50.1%)
30-39		169,385 (44.3%)	574,473 (42.3%)
≥ 40		10,276 (2.7%)	34,598 (2.5%)
Non-Hispanic Black race and ethnicity	0	99,915 (26.1%)	296,388 (21.8%)	0.101
Education level	10,672 (0.6%)			0.126
Less than high school		7646 (2.0%)	10,422 (0.8%)
High school with no diploma		30,877 (8.1%)	94,574 (7.0%)
High school graduate or general educational diploma		94,697 (25.0%)	317,267 (23.5%)
College and higher		246,217 (64.9%)	928,344 (68.7%)
Health insurance	10,807 (0.6%)			0.143
Medicaid		156,981 (41.4%)	495,069 (36.7%)
Private		197,251 (52.0%)	785,697 (58.2%)
Self-pay		11,847 (3.1%)	23,708 (1.8%)
Other		13,541 (3.6%)	45,815 (3.4%)
Comorbidites
Body mass index (kg/m2)	41,463 (2.4%)			0.102
≤ 18.4		14,645 (4.0%)	43,851 (3.3%)
18.5-24.9		180,280 (48.7%)	596,345 (44.9%)
25.0-29.9		91,312 (24.6%)	337,706 (25.4%)
30.0-34.9		46,831 (12.6%)	186,225 (14.0%)
≥ 35		37,476 (10.1%)	164,582 (12.4%)
Preexisting diabetes	965 (0.1%)	1988 (0.5%)	10,621 (0.8%)	0.033
Gestational diabetes	965 (0.1%)	16,614 (4.3%)	74,772 (5.5%)	0.054
Preexisting hypertension	965 (0.1%)	5599 (1.5%)	291,46 (2.1%)	0.051
Gestational hypertension	965 (0.1%)	17,024 (4.5%)	102,851 (7.6%)	0.132
Pregnancy, labor, and delivery
Previous cesarean section	965 (0.1%)	18,469 (4.8%)	125,271 (9.2%)	0.173
Month prenatal care began	44,744 (2.6%)			0.140
1st -3rd		279,722 (75.5%)	1,063,855 (80.3%)
4th -6th		62,574 (16.9%)	193,024 (14.6%)
≥ 7th		17,622 (4.8%)	51,188 (3.9%)
No prenatal care		10,532 (2.8%)	17,455 (1.3%)
Number of prenatal visits	43,967 (2.5%)	6.6 (sd, 2.0)	7.0 (sd, 1.9)	0.172
Delivery during a weekend	0	106,141 (27.7%)	367,635 (27.1%)	0.015
Mother transferred in	405 (<0.1%)	2043 (0.5%)	5320 (0.4%)	0.021
Nulliparous	5994 (0.3%)	99,678 (26.2%)	496,843 (36.7%)	0.228
Gestational age at delivery	726 (<0.1%)			0.088
≤ 33 completed weeks		13,641 (3.6%)	29,108 (2.1%)
34-38 completed weeks		120,077 (31.4%)	419,649 (30.9%)
≥ 39 completed weeks		248,363 (65.0%)	909,152 (67.0%)
Non-cephalic presentation	3399 (0.2%)	7319 (1.9%)	36,048 (2.7%)	0.050
Induction of labor	0	79,876 (20.9%)	491,048 (36.2%)	0.343
Augmentation of labor	0	62,865 (16.4%)	379,565 (27.9%)	0.280
Antibiotics received during labor	0	75,318 (19.7%)	443,117 (32.6%)	0.298
Chorioamnionitis	0	1858 (0.5%)	26,170 (1.9%)	0.132
Attendant at birth	690 (<0.1%)			0.306
Doctor of Medicine		277,115 (72.5%)	1,121,466 (82.6%)
Doctor of Osteopathy		30,528 (8.0%)	114,174 (8.4%)
Midwife		70,538 (18.4%)	115,301 (8.5%)
Other		4140 (1.1%)	6764 (0.5%)
Delivery mode	0			0.453
Spontaneous vaginal delivery		343,476 (89.8%)	986,529 (72.6%)
Assisted vaginal delivery		8269 (2.2%)	56,595 (4.2%)
Intrapartum cesarean delivery		30,828 (8.1%)	315,019 (23.2%)
Birth weight (grams)	977 (0.1%)			0.073
<2500 grams		28,664 (7.5%)	78,257 (5.8%)
2500-4000 grams		323,985 (84.8%)	1,165,459 (85.8%)
> 4000 grams		29,433 (7.7%)	113,941 (8.4%)
Characteristics of the county of the delivery hospital
Core base statistical area classification (†)	0			0.137
Metropolitan		334,522 (87.4%)	1,241,147 (91.4%)
Micropolitan		38,702 (10.1%)	100,836 (7.4%)
Non-metropolitan and non-micropolitan		9349 (2.4%)	16,160 (1.2%)
Number of in-hospital births (†)	0	12,515 (sd, 18,891)	13,725 (sd, 18,728)	0.064
Number of physician anesthesiologists (per 1000 hospital births) (†)	0	12.5 (sd, 9.1)	13.4 (sd, 9.1)	0.055
Number of certified registered nurse anesthetists (per 1000 hospital births) (†)	0	15.4 (sd, 14.4)	15.8 (sd, 14.7)	0.031
Number of obstetricians and gynecologists (per 1000 hospital births) (†)	0	10.9 (sd, 6.2)	11.2 (sd, 6.0)	0.053
Proportion of non-Hispanic Black residents (%) (†)	0	14.5 (sd, 14.0)	15.3 (sd, 13.7)	0.061
Proportion of residents with less than high school diploma (%) (†)	0	7.6 (sd, 3.0)	7.5 (sd, 2.8)	0.031
Proportion of unemployed residents (%) (†)	0	3.3 (sd, 0.9)	3.2 (sd, 0.8)	0.044
Jail incarceration rate (per 100,000 residents) (‡)	0	351 (sd, 233)	355 (sd, 281)	0.014

Results are presented as count (%) or mean (1 standard deviation). An absolute standardized difference greater than 0.10 (10.0%) indicates a clinically relevant imbalance between groups.

^(*)Bolded values indicate absolute standardized difference greater than 0.10.

^(†)Data are from the Area Health Resource File for the year 2017 or the 2013-2017 period.

^(‡)Data are from the Vera Institute of Justice for the year 2017.

To assess the robustness of our findings, we estimated the odds ratios of labor neuraxial analgesia associated with the terciles of the racism index after restricting the study sample to people 1) who gave birth in a hospital located in their residence county, 2) who gave birth through vaginal delivery, and 3) who gave birth in a county with only one hospital providing obstetric services. We also conducted a sensitivity analysis with further adjustment for the hospital county identifier.

RESULTS

During the study period, 1,740,716 birth certificates for vaginal or intrapartum cesarean deliveries in 1381 counties in 45 states and DC were analyzed (Figure 2); of these births, 396,303 (22.8%) were given by non-Hispanic Black people and 1,344,413 (77.2%) by White people.

Labor neuraxial analgesia use was recorded in 1,358,143 birth certificates (78.0%; 95% CI: 77.9, 78.1). The labor neuraxial analgesia rate varied markedly across counties (Appendix 2, available online at http://links.lww.com/xxx) and was lower for non-Hispanic Black people (74.8%; 95% CI: 74.6, 74.9) than for non-Hispanic White people (79.0%; 95% CI: 78.9, 79.1; ASD = 0.10). Comparisons of sociodemographic, comorbidities, pregnancy, labor and delivery, and hospital characteristics between people who did not receive labor neuraxial analgesia and people who did are presented in Table 1.

The median racism index in the county of the delivery hospital was 2.7 (interquartile range (IQR): 2.0, 3.6) but varied markedly across hospital counties (Appendix 3, available online at http://links.lww.com/xxx). The index did not differ between non-Hispanic Black birthing people (median 2.6; IQR: 2.0, 3.4) and non-Hispanic White birthing people (median 2.8; IQR: 2.1, 3.6 ASD = 0.05).

For non-Hispanic White people, the labor neuraxial analgesia rate was 80.4% for people who gave birth in hospitals in the 1^st tercile of the racism index, 78.2% in the 2^nd tercile (rate difference: −2.2%), and 78.2% in the 3^rd tercile (rate difference: −2.2%) (Table 2 and Figure 3). For non-Hispanic White people, compared to the 1^st tercile of the racism index, the 2^nd and 3^rd terciles were each associated with 12.3% (95% CI: 11.4, 13.2) decreased odds of receiving labor neuraxial analgesia.

Table 2:

Crude and adjusted odds ratios of labor neuraxial analgesia associated with the terciles of the structural racism index (45 US states and District of Columbia, 2017).

	Number of people	Number of people with neuraxial analgesia	Rate (%)	Rate difference (%; 95% CI)	Crude OR (95% CI)	Adjusted OR (95% CI) (*,†)
Non-Hispanic Black (n = 396,303)
Structural racism index
Tercile 1 (0.05, 2.21)	132,285	102,079	77.2	0.00 (Ref.)	1.00 (Ref.)	1.00 (Ref.)
Tercile 2 (2.22, 3.18)	138,161	103,200	74.7	−2.5 (−2.8, −2.1)	0.87 (0.86, 0.89)	0.82 (0.80, 0.83)
Tercile 3 (3.19, 61.90)	125,857	91,109	72.4	−4.8 (−5.1, −4.4)	0.78 (0.76, 0.79)	0.72 (0.70, 0.73)
Non-Hispanic White (n = 1,344,413)
Structural racism index
Tercile 1 (0.00, 2.27)	461,658	371,113	80.4	0.0 (Ref.)	1.00 (Ref.)	1.00 (Ref.)
Tercile 2 (2.28, 3.27)	442,730	346,400	78.2	−2.2 (−2.3, −2.0)	0.88 (0.87, 0.89)	0.87 (0.86, 0.88)
Tercile 3 (3.28, 139.00)	440,025	344,242	78.2	−2.2 (−2.3, −2.0)	0.88 (0.87, 0.89)	0.84 (0.83, 0.85)

Abbreviation: CI: confidence interval; OR: odds ratio; Ref.: reference.

^(*)Adjusted for: 1) body mass index, 2) gestational hypertension, 3) previous cesarean section, 4) parity, 5) induction of labor, 6) augmentation of labor, 7) antibiotics during labor, 8) chorioamnionitis, and 9) hospital core base statistical area classification.

^(†)The adjusted regression coefficient β of the interaction term between race and the racism index indicated a significant difference in labor neuraxial analgesia rate between non-Hispanic White and non-Hispanic Black people for the 2^nd tercile (β: −0.028; 95% CI: −0.049, −0.006; P-value = 0.011; Percent change: −2.8%; 95% CI: −4.8, −0.6) and for the 3^rd tercile of the racism index (β: −0.181; 95% CI: −0.203, −0.160; P-value < 0.001; Percent change: −16.6%; 95% CI: −18.4, −14.8).

Figure 3:

Labor neuraxial analgesia rate according to the terciles of the structural racism index for non-Hispanic White (grey) and for non-Hispanic Black (orange). The difference is, for each tercile of the structural racism index, the difference in the labor neuraxial analgesia rate between non-Hispanic Black and non-Hispanic White. After adjustment and compared to White, Black in the second tercile of the structural racism index had 2.8% (95% CI: 0.6, 4.8) decreased odds of receiving labor neuraxial analgesia and in the third tercile 16.6% (95% CI: 14.8, 18.4) decreased odds.

For non-Hispanic Black people, the labor neuraxial analgesia rate was 77.2% for people who gave birth in hospitals in the 1^st tercile of the racism index, 74.7% in the 2^nd tercile (rate difference: −2.5%), and 72.4% in the 3^rd tercile (rate difference: −4.8%). For non-Hispanic Black people, compared to the 1^st tercile of the racism index, the 2^nd tercile was associated with 12.7% (95% CI: 11.1, 14.2) decreased odds of receiving labor neuraxial analgesia, and the 3^rd tercile with 22.4% (95% CI: 21.0, 23.8) decreased odds.

For non-Hispanic White people after adjustment, compared to the 1^st tercile of the racism index, the 2^nd tercile was associated with 13.4% (95% CI: 12.5, 14.4) decreased odds of receiving labor neuraxial analgesia and the 3^rd tercile with 15.6% (95% CI: 14.7, 16.5) decreased odds (Table 2 and Appendix 4, available online at http://links.lww.com/xxx).

For non-Hispanic Black people after adjustment, compared to the 1^st tercile of the racism index, the 2^nd tercile was associated with 18.4% (95% CI: 16.9, 19.9) decreased odds of receiving labor neuraxial analgesia and the 3^rd tercile with 28.3% (95% CI: 26.9, 29.6) decreased odds.

The adjusted regression coefficient β of the interaction term between race and the racism index indicated a significant difference in labor neuraxial analgesia rate between non-Hispanic White and non-Hispanic Black people for the 2^nd tercile (β: −0.028; 95% CI: −0.049, −0.006; P-value = 0.011; Percent change: −2.8%; 95% CI: −4.8, −0.6) and for the 3^rd tercile (β: −0.181; 95% CI: −0.203, −0.160; P-value < 0.001; Percent change: −16.6%; 95% CI: −18.4, −14.8).

Results of the sensitivity analyses restricting the study sample to people who gave birth in a hospital located in their county of residence, who gave birth through vaginal delivery, and who gave birth in a county with only one hospital providing obstetric services were similar to the results of the main analysis (Appendixes 5–7, available online at http://links.lww.com/xxx). Results were also robust with further adjustment for the hospital county identifier (Appendix 8, available online at http://links.lww.com/xxx).

DISCUSSION

In this nationwide study, we confirm our hypothesis that a multidimensional indicator of structural racism measured in the county of the delivery hospital is associated with significantly reduced odds of receiving labor neuraxial analgesia among non-Hispanic Black people and, to a lesser extent, non-Hispanic White people.

Previous research on the effects of structural racism on maternal and child health among Black people has consistently reported that unidimensional indicators of residential segregation or unidimensional Black-to-White inequity ratios are associated with increased odds of preterm birth, low birth weight, infant mortality, and child mortality.^{29, 31, 36–39} More recently, these unidimensional indicators have also been associated with increased odds of severe maternal morbidity and mortality.^{21, 28, 32, 40–42} Racial and ethnic disparities in the use of neuraxial analgesia during childbirth, and disparities in pain management during the postpartum period have been reported but their association with racism has not been investigated so far.^{4, 43–45} In the present study, we found empirical evidence that structural racism is significantly associated with lower utilization of labor neuraxial analgesia among both non-Hispanic Black and non-Hispanic White people, independent of sociodemographic and clinical characteristics.

It is noteworthy that our analysis was limited to structural racism and that we did not assess institutional or interpersonal racism, which may also affect the use of labor neuraxial analgesia. This limitation may explain the relatively small effect sizes of the structural racism index reported in our study on use of labor neuraxial analgesia among non-Hispanic Black and non-Hispanic White people. Institutional racism occurs where institutional policies and practices result in discrimination based on race.^{46, 47} It may manifest in the delivery hospital as differential treatment of White and Black people and can be related to lack of anti-racist training or lack of healthcare workforce diversity. Interpersonal racism, conscious or unconscious, such as implicit bias, occurs between individuals and may manifest in the delivery hospital as communication failures or stereotyping by health care professionals .^{48, 49} We also have to acknowledge that we did not assess discrimination that is not uniquely based on race but can be based on insurance type, marital status, sexuality, disability, language barriers, or gender identity.⁵⁰

Our findings suggest that interventions addressing structural racism in anesthesia care could benefit both Black and White people, especially in counties with the highest racism index. In 2021, the American Society of Anesthesiologist issued recommendations for reducing racial and ethnic disparities and mitigating the effects of racism on obstetric anesthesia care, targeting anesthesiologists, nurse anesthetists, and department chairs.⁵¹ These recommendations included: ensuring accurate documentation of race and ethnicity and primary spoken language; creation of disparities dashboards to track changes over time; education of attending anesthesiologists and nurse anesthetists on racial and ethnic disparities in anesthesia care and the roles of bias, institutional, and structural racism; development of best practices for shared decision-making when discussing labor neuraxial analgesia; and diversifying the anesthesia workforce in their department.^{52, 53}

Our findings should be interpreted in the context of several limitations. First, we limited our analysis to non-Hispanic Black people and excluded other minoritized racial and ethnic people such as Hispanic people. This choice relied on the fact that the racism index we used captured Black-to-White inequities but not Hispanic-to-White inequities. Second, we have no information in the Natality file on whether obstetric anesthesia service was available in the delivery hospital. Third, we estimated characteristics of the delivery hospital characteristics at the hospital-county level as a proxy for the hospital because Natality data do not provide a hospital identifier. This approach may not be accurate for counties with more than one hospital. However, our sensitivity analysis limited to women who gave birth in a county with only one hospital providing obstetric services implies that estimating hospital characteristics at the county level did not introduce any significant bias into our main analysis. Fourth, because birth certificate data do not provide a patient identifier, the analysis could not account for people who had more than one childbirth during the study period. Nevertheless, the short duration of the study period (1 year) makes this possibility less likely. Last, we analyzed data before the COVID 19 epidemic. Since COVID 19 has exacerbated the effects of racism on racial and ethnic disparities, we could have observed different results using more recent data.

A multidimensional index of structural racism is associated with significantly reduced odds of receiving labor neuraxial analgesia among birthing people, especially non-Hispanic Black people. It highlights the need to implement interventions to address racism such as education of healthcare professionals or diversification of the workforce that could benefit both Black and White people.