Research Critique on Impact of Nursing diversity

Original Research

Nurse workforce diversity and reduced risk of severe adverse maternal outcomes

Jean Guglielminotti, MD, PhD; Goleen Samari, PhD, MPH; Alexander M. Friedman, MD, MPH; Allison Lee, MD, MS; Ruth Landau, MD; Guohua Li, MD, DrPH

BACKGROUND: Racial and ethnic diversification of the physician and nurse workforce is recommended as a leverage point to address the impact of structural racism in maternal care, but empirical evidence sup- porting this recommendation is currently lacking. OBJECTIVE: This study aimed to assess the association between state-level registered nurse workforce racial and ethnic diversity and severe adverse maternal outcomes during childbirth. STUDY DESIGN: This population-based cross-sectional study ana- lyzed 2017 US birth certificate data. Severe adverse maternal outcomes included eclampsia, blood transfusion, hysterectomy, or intensive care unit admission. Proportions of minoritized racial and ethnic registered nurses in each state were abstracted from the American Community Sur- vey (5-year estimate, 2013−2017). This proportion was categorized into 3 terciles, with the first tercile corresponding to the lowest proportion and the third tercile corresponding to the highest proportion. Crude and adjusted odds ratios and 95% confidence intervals of severe adverse maternal outcomes associated with terciles of the state proportion of minoritized racial and ethnic nurses were estimated using logistic regres- sion models. RESULTS: Of the 3,668,813 birth certificates studied, 29,174 recorded severe adverse maternal outcomes (79.5 per 10,000; 95% con- fidence interval, 78.6−80.4). The mean state proportion of minoritized racial and ethnic nurses was 22.1%, ranging from 3.3% in Maine to 68.2% in Hawaii. For White mothers, the incidence of severe adverse out- comes was 85.3 per 10,000 for those who gave births in states in the first tercile of the proportion of minoritized racial and ethnic nurses and 53.9 per 10,000 for those who gave birth in states in the third tercile (risk dif- ference, �31.4 per 10,000; 95% confidence interval, �34.4 to �28.5).

Cite this article as: Guglielminotti J, Samari G, Friedman AM, et al. Nurse workforce diversity and reduced risk of severe adverse maternal outcomes. Am J Obstet Gynecol MFM 2022;4:100689.

2589-9333/$36.00 © 2022 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajogmf.2022.100689

EDITOR'S

It corresponds to a 37% decreased risk of severe adverse maternal out- comes associated with giving birth in a state in the third tercile (crude odds ratio, 0.63; 95% confidence interval, 0.60−0.66). A decreased risk of severe adverse maternal outcomes was observed for Black mothers (crude odds ratio, 0.65; 95% confidence interval, 0.61−0.70), Hispanic mothers (crude odds ratio, 0.51; 95% confidence interval, 0.48−0.54), and Asian and Pacific Islander mothers (crude odds ratio, 0.65; 95% con- fidence interval, 0.58−0.72) but not for Native American mothers (crude odds ratio, 0.89; 95% confidence interval, 0.72−1.09) or mothers with >1 race (crude odds ratio, 1.44; 95% confidence interval, 0.72−1.09). After adjustment for patients and hospital characteristics, giving birth in states in the third tercile was associated with a reduced risk of severe adverse outcomes as follows: 32% for White mothers (adjusted odds ratio, 0.68; 95% confidence interval, 0.59−0.77), 20% for Black mothers (adjusted odds ratio, 0.80; 95% confidence interval, 0.65−0.99), 31% for Hispanic mothers (adjusted odds ratio, 0.69; 95% confidence interval, 0.58−0.82), and 50% for Asian and Pacific Islander mothers (adjusted odds ratio, 0.50; 95% confidence interval, 0.38−0.65). The associations of the proportion of minoritized racial and ethnic nurses with the risk of severe adverse maternal outcomes were not statistically significant for Native American mothers and more than 1 race mothers. Results were similar when blood transfusion was excluded from the outcome measure. CONCLUSION: A diverse state registered nurse workforce was associ- ated with a reduced risk of severe adverse maternal outcomes during childbirth.

Key words: childbirth, epidemiology, healthcare workforce, maternal morbidity, racial and ethnic diversity, structural racism

CHOICE

Introduction

I n 2020, the US federal government recognized addressing racial and eth-

nic disparities in severe maternal mor- bidity as a public health priority.1,2

Compared with non-Hispanic White birthing people, minoritized racial and ethnic groups are up to 3 times as likely to experience life-threatening complica- tions during pregnancy, childbirth, and

the postpartum period.3 Among racial and ethnic minoritized people, non-His- panic Black and Native American people are at particularly high risk of severe adverse maternal outcomes (SAMOs).4,5

Structural racism contributes to these disparities in SAMO, independent of social determinants of health (eg, poverty or education).6−9 Structural racism refers to a system where public policies, institu- tional practices, cultural representations, and other norms work to perpetuate racial group inequities.10−13

Racial and ethnic diversification of the physician and nurse workforce has been recommended as a possible remedy for reducing the impact of structural racism

on racial and ethnic disparities in mater- nal health outcomes.14,15 A racially diverse workforce improves access to healthcare for minoritized racial and eth- nic people, reduces provider implicit bias, and increases the likelihood of racial and ethnic concordance between patients and healthcare workers. How- ever, evidence linking physician or nurse workforce diversity to improved mater- nal health outcomes is currently lacking. Registered nurses (RNs) are crucial for comprehensive maternal healthcare and are the frontline healthcare providers responsible for identifying warning signs of maternal complications that require urgent bedside evaluation by clinicians and timely intervention.16 Thus, a racially diverse RN workforce could be

September 2022 AJOG MFM 1

AJOG MFM at a Glance

Why was this study conducted? Racial and ethnic diversification of the physician and nurse workforce is recom- mended as a possible remedy for reducing the impact of structural racism on racial and ethnic disparities in maternal health outcomes. However, evidence linking healthcare workforce diversity to improved maternal health outcomes is currently lacking.

Key findings In this nationwide study in 2017, racial and ethnic diversity in the state nurse workforce was associated with a reduced risk of severe adverse maternal out- comes in White, Black, Hispanic, and Asian and Pacific Islander mothers.

What does this add to what is known? This finding could guide the development of intervention programs to reduce racial and ethnic disparities in maternal health outcomes by diversifying the healthcare workforce.

Original Research

associated with a reduced risk of SAMO. Using 2017 US birth certificate data, we aimed to assess the association between state-level proportion of RNs from minoritized racial and ethnic groups and SAMO.

Materials and Methods The study protocol was deemed exempt by the institutional review board of the authors’ institution. This study was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.

Data system Data for this study were abstracted from the 2017 US birth certificates contained in the restricted access Natality File of the National Vital Statistics System (National Center for Health Statistics, Centers for Diseases Control and Pre- vention). This data system is based on the 2003 revised US Standard Certificate of Live Birth.17 As of January 2015, it was implemented in the 50 US states and the District of Columbia.

Study sample The study sample included all births between January 1, 2017, and December 31, 2017. The exclusion criteria were (1) mother not residing in the United States, (2) maternal state of residence not corresponding to state of delivery, (3) birth not occurring in a hospital, (4) missing information on maternal race

2 AJOG MFM September 2022

and ethnicity, (5) missing information on maternal outcomes, and (6) missing information on county or state identi- fiers for maternal residence or delivery hospital.

Exposure The exposure of interest was the propor- tion of RNs from minoritized racial and ethnic groups in each state. It was calcu- lated as follows: 100£ (total number of RNs�number of non-Hispanic White RNs)/total number of RNs. Total num- ber of RNs and number of non-Hispanic White RNs by state were abstracted from the American Community Survey (5-year estimate, 2013−2017) available in the Area Health Resource File (AHRF).18 The information on other race and ethnicity RNs (eg, non-His- panic Black RNs) was missing for many states, precluding the use of a state race- specific proportion (eg, state proportion of non-Hispanic Black RNs).

Outcome The outcome was SAMO defined as the presence of at least 1 of the 4 following conditions or procedures: eclampsia, blood transfusion, hysterectomy, and intensive care unit (ICU) admission. In a sensitivity analysis, we excluded blood transfusion from the definition of SAMO.

The 4 conditions and procedures are recorded in specific check boxes on the birth certificate. The reported sensitivity

of the individual components in a study conducted in Massachusetts in 2011 to 2013 and using administrative hospital discharge data as the gold standard ranges from 12% for blood transfusion to 51% for hysterectomy.19

Birth certificates do not contain codes of the International Classification of Diseases, precluding the assessment of severe maternal morbidity as defined by the US Centers for Disease Control and Prevention.20

Maternal, hospital, and state characteristics Maternal characteristics and comorbid- ities directly recorded from birth certifi- cate data included age (≤19, 20−29, 30 −39, or ≥40 years), race and ethnicity, education level (less than high school, high school with no diploma, high school graduate or General Educational Diploma, or college and higher), health insurance (Medicaid, private, self-pay, or other); 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 dia- betes mellitus or hypertension. Maternal race and ethnicity were categorized into 6 mutually exclusive groups: (1) non- Hispanic White (hereafter referred to as White); (2) non-Hispanic Black (Black); (3) Hispanic; (4) non-Hispanic Asian, Native Hawaiian, and Other Pacific Islander (Asian and Pacific Islander); (5) non-Hispanic American Indian and Alaskan Native (Native American); and (6) more than 1 race. The following maternal characteris-

tics were estimated at the county of resi- dence level: urban or rural residence, proportion of persons in poverty, and proportion of persons unemployed. Obstetrical characteristics directly

recorded from birth certificate data included previous cesarean delivery, month of gestation prenatal care began (1−3, 4−6, ≥7, or no prenatal care), number of prenatal visits, delivery dur- ing a weekend, mother transferred in, nulliparous, gestational age at delivery (≤33, 34−38, or ≥39 weeks), multiple pregnancy, noncephalic presentation, induction of labor, attendant at birth (doctor of medicine, doctor of osteopa- thy, midwife, or other), delivery mode

Original Research

(vaginal spontaneous, vaginal assisted [vacuum or forceps], or cesarean), and birthweight (≤2499, 2500−4000, or ≥4000 g). Because birth certificate data do not

provide a hospital identifier, the follow- ing hospital characteristics were esti- mated at the hospital county level: urban or rural location, number of hos- pital births, and number of obstetricians and gynecologists (per 1000 hospital births). The following state characteristics

were abstracted from the AHRF or the State Health Facts of the Kaiser Family Foundation18,21: proportion of racial and ethnic minority residents, proportion of

FIGURE 1 Flowchart of the study

(Asterisk) Reasons for exclusion are not mutually ex ICU, intensive care unit.

Guglielminotti. State workforce diversity and maternal health.

persons below poverty level, proportion of persons unemployed, number of physicians (per 1000 residents), number of RNs (per 1000 residents), and Medic- aid income eligibility threshold.

Statistical analysis Statistical analysis was performed with R (version 4.0.3; R Foundation for Sta- tistical Computing, Vienna, Austria) and the package “lme4” for mixed-effect modeling.22 No study a priori power was performed.

Descriptive statistics The distribution of the state proportion of minoritized racial and ethnic RNs

clusive.

Am J Obstet Gynecol MFM 2022.

across the 50 states and the District of Columbia was examined visually using a caterpillar plot. Moreover, this proportion was compared between mothers with and without SAMO using the standardized difference (SD), with a value >10% indic- ative of a significant imbalance.23

Crude analysis The state proportion of RNs from racial and ethnic minorities was categorized into 3 terciles, with the first tercile cor- responding to the lowest proportion and the third tercile corresponding to the highest proportion. Terciles were calculated for each of the 6 racial and ethnic maternal groups. For each

September 2022 AJOG MFM 3

FIGURE 2 Proportion of minoritized racial and ethnic registered nurses across states

RN, registered nurse.

Guglielminotti. State workforce diversity and maternal health. Am J Obstet Gynecol MFM 2022.

Original Research

maternal racial and ethnic group, the incidence of SAMO was estimated for each tercile of the state proportion of minoritized racial and ethnic RNs. The risk difference was estimated as the dif- ference between the incidence in the third tercile and the incidence in the first tercile (reference). Crude odds ratios (ORs) of SAMO associated with terciles of the state proportion of minoritized RNs were estimated using univariate fixed-effect logistic regression models, with SAMO as the dependent variable and proportion as the indepen- dent variable.

Adjusted analysis For each maternal racial and ethnic group, the adjusted ORs (aORs) of SAMO associated with terciles of the state proportion of minoritized racial and ethnic RNs were estimated using multivariate mixed-effect logistic regression models with the hospital

4 AJOG MFM September 2022

county nested within the hospital state as the random effect (random intercept and constant slope), and adjusted for patients and hospital characteristics. Mixed-effect models consider the corre- lation between women within hospitals and hospitals within states.

To identify variables required to adjust the ORs of SAMO, we developed a multivariable prediction model for SAMO using data from all race and eth- nicity mothers. Candidate variables included in the model were characteris- tics with an SD of >10% in the compari- son of mothers with and without SAMO presented in Appendix 1. Race and eth- nicity were not included as a candidate variable. We used mixed-effect logistic regression modeling with SAMO as the dependent variable, the candidate varia- bles as independent variables, and the hospital county nested within the hospi- tal state as the random effect. Selection was performed using a backward

procedure with a significance threshold of 0.05. We performed a complete case analysis with 200,655 birth certificates (5.5%) excluded for missing values of the candidate variables. The results of the multivariable prediction model are presented in Appendix 2. In a sensitivity analysis, the state pro-

portion of the racial and ethnic group examined was added to the variables used for adjustment (eg, state propor- tion of Black residents when analyzing Black mothers).

Results Of the 3,668,813 birth certificates stud- ied, 29,174 recorded SAMO or 79.5 per 10,000 (95% confidence interval [CI], 78.6−80.4) (Figure 1). The most fre- quent complication recorded was blood transfusion (39.3 per 10,000), followed by eclampsia (28.1 per 10,000), ICU admission (16.3 per 10,000), and hyster- ectomy (4.7 per 10,000) (Appendix 3).

Descriptive statistics The mean proportion of minoritized racial and ethnic RNs in the 50 US states and the District of Columbia was 22.1%. It ranged from a minimum of 3.3% in the state of Maine to a maxi- mum of 68.2% in the state of Hawaii (Figure 2). Compared with mothers without SAMO, mothers with SAMO gave birth in states with a lower propor- tion of minoritized racial and ethnic RNs (30.6% vs 27.9%, respectively; SD, 17.2%). A lower state proportion of minoritized racial and ethnic RNs for mothers with SAMO was observed for all racial and ethnic groups, except for mothers with more than 1 race (Appen- dix 4).

Crude analysis For White mothers, the incidence of SAMO was 85.3 per 10,000 for those who gave births in states in the first tercile of the proportion of minoritized racial and ethnic RNs and 53.9 per 10,000 for those who gave birth in states in the third tercile (risk differ- ence, �31.4 per 10,000; 95% CI, �34.4 to �28.5) (Table 1). It corresponds to a 37% decreased risk of SAMO associ- ated with giving birth in a state in the

TABLE 1 Incidence of severe adverse maternal outcomes associated with the terciles of the state proportion of minoritized racial and ethnic registered nurses (United States, 2017)

Maternal race and ethnicity Number of women

Number of SAMO cases

Incidence (per 10,000; 95% CI)

Risk difference (95% CI)a Crude OR (95% CI)b Adjusted OR 1 (95% CI)c Adjusted OR 2 (95% CI)d

White

Tercile 1 (3.3%−14.2%) 632,434 5395 85.3 (83.0−7.6) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (14.3%−32.2%) 653,299 5262 80.5 (78.4−82.7) �4.8 (�7.9 to �1.6) 0.94 (0.91−0.98) 0.87 (0.78−0.98) 0.78 (0.65−0.92)

Tercile 3 (32.3%−68.3%) 594,021 3199 53.9 (52.0−55.7) �31.4 (�34.4 to �28.5) 0.63 (0.60−0.66) 0.68 (0.59−0.77) 0.53 (0.39−0.72)

Black

Tercile 1 (3.3%−21.9%) 208,367 2721 130.6 (125.7−135.5) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (22.0%−41.4%) 203,273 2141 105.3 (100.9−109.8) �25.3 (�31.9 to �18.7) 0.80 (0.76−0.85) 0.94 (0.79−1.12) 1.05 (0.87−1.26)

Tercile 3 (41.5%−68.3%) 133,295 1142 85.7 (80.7−90.6) �44.9 (�51.9 to �38.0) 0.65 (0.61−0.70) 0.80 (0.65−0.99) 0.81 (0.65−0.99)

Hispanic

Tercile 1(3.3%−36.6%) 283,422 2844 100.3 (96.7−104.0) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (36.7%−46.2%) 185,946 1185 63.7 (60.1−67.3) �36.6 (�41.8 to �31.5) 0.63 (0.59−0.68) 0.78 (0.64−0.94) 0.71 (0.57−0.88)

Tercile 3 (46.3%−68.3%) 413,657 2112 51.1 (48.9−53.2) �49.2 (�53.6 to �45.0) 0.51 (0.48−0.54) 0.69 (0.58−0.82) 0.50 (0.34−0.73)

Asian and Pacific Islander

Tercile 1 (3.3%−28.6%) 85,673 830 96.9 (90.3−103.4) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (28.7%−46.2%) 66,924 393 58.7 (52.9−64.5) �38.2 (�46.9 to �29.4) 0.60 (0.54−0.68) 0.61 (0.48−0.77) 0.54 (0.43−0.69)

Tercile 3 (46.3%−68.3%) 101,163 636 62.9 (58.0−67.7) �34.0 (�42.2 to �25.8) 0.65 (0.58−0.72) 0.50 (0.38−0.65) 0.33 (0.24−0.46)

Native American

Tercile 1 (3.3%−21.9%) 10,468 202 193.0 (166.6−219.3 0.0 (Reference) 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (22.0%−26.5%) 7599 118 155.3 (127.5−183.1 �37.7 (�76.0 to 0.6) 0.80 (0.64−1.01) 0.97 (0.64−1.45) 0.69 (0.43−1.13)

Tercile 3 (26.6%−68.3%) 10,183 175 171.9 (146.6−197.1 �21.1 (�57.6 to 15.4) 0.89 (0.72−1.09) 0.89 (0.61−1.29) 0.82 (0.56−1.21)

More than 1 race

Tercile 1 (3.3%−17.3%) 26,476 250 94.4 (82.8−106.1) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (17.4%−41.4%) 26,653 218 81.8 (71.0−92.6) �12.6 (�28.5 to 3.3) 0.87 (0.72−1.04) 0.84 (0.64−1.10) 0.82 (0.62−1.07)

Tercile 3 (41.5%−68.3%) 25,960 351 135.2 (121.2−149.3) 40.8 (22.5−59.0) 1.44 (1.22−1.69) 0.89 (0.66−1.21) 0.80 (0.58−1.11)

CI, confidence interval; OR, odds ratio; SAMO, severe adverse maternal outcome. aCalculated as the difference between the incidence in the third (or second) tercile and the incidence in the first tercile (reference). bEstimated using univariate fixed-effect logistic regression. cEstimated using multivariate mixed-effect logistic regression with the hospital county nested within the hospital state as the random effect and adjusted for (1) age, (2) body mass index, (3) residence (rural or urban), (4) preexisting diabetes mellitus, (5) gestational dia- betes mellitus, (6) preexisting hypertension, (7) gestational hypertension, (8) month prenatal care began, (9) number of prenatal visits, (10) mother transferred in, (11) gestational age at delivery, (12) multiple pregnancy, (13) noncephalic presentation, (14) induction of labor, (15) attendant at birth, (16) delivery mode, (17) birthweight, and (18) hospital location (rural or urban). dWith further adjustment for the state proportion of the racial and ethnic group examined (eg, state proportion of Black residents when analyzing Black mothers).

Guglielminotti. State workforce diversity and maternal health. Am J Obstet Gynecol MFM 2022.

Septem ber2022

AJOG M FM

5

O riginalR

esearch

TABLE 2 Incidence of severe adverse maternal outcomes after exclusion of blood transfusion associated with the terciles of the state proportion of minoritized racial and ethnic registered nurses (United States, 2017)

Maternal race and ethnicity Number of women Number of SAMO cases without blood transfusion Incidence (per 10,000; 95% CI) Risk difference (95% CI)a Crude OR (95% CI)b Adjusted OR (95% CI)c

White

Tercile 1 (3.3%−14.2%) 632,434 3145 49.7 (48.0−51.5) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (14.3%−32.2%) 653,299 3257 49.9 (48.1−51.6) 0.1 (�2.3 to 2.6) 1.00 (0.95−1.05) 0.90 (0.78−1.04)

Tercile 3 (32.3%−68.3%) 594,021 1582 26.6 (25.3−27.9) �23.1 (�25.3 to �20.9) 0.53 (0.50−0.57) 0.64 (0.54−0.76)

Black

Tercile 1 (3.3%−21.9%) 208,367 1807 86.7 (82.7−90.7) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (22.0%−41.4%) 203,273 1256 61.8 (58.4−65.2) �24.9 (�30.2 to �19.7) 0.71 (0.66−0.76) 0.86 (0.70−1.05)

Tercile 3 (41.5%−68.3%) 133,295 587 44.0 (40.5−47.6) �42.7 (�48.0 to �37.3) 0.51 (0.46−0.56) 0.60 (0.46−0.79)

Hispanic

Tercile 1(3.3%−36.6%) 283,422 1613 56.9 (54.1−59.7) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (36.7%−46.2%) 185,946 663 35.7 (32.9−38.4) �21.3 (�25.1 to �17.4) 0.63 (0.57−0.68) 0.67 (0.54−0.85)

Tercile 3 (46.3%−68.3%) 413,657 1118 27.0 (25.4−28.6) �29.9 (�33.1 to �26.7) 0.47 (0.44−0.51) 0.59 (0.48−0.73)

Asian and Pacific Islander

Tercile 1 (3.3%−28.6%) 85,673 474 55.3 (50.4−60.3) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (28.7%−46.2%) 66,924 181 27.0 (23.1−31.0) �28.3 (�34.6 to �21.9) 0.49 (0.41−0.58) 0.42 (0.30−0.58)

Tercile 3 (46.3%−68.3%) 101,163 514 50.8 (46.4−55.2) �4.5 (�11.1 to 2.1) 0.92 (0.81−1.04) 0.52 (0.37−0.72)

Native American

Tercile 1 (3.3%−21.9%) 10,468 90 86.0 (68.3−103.7) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (22.0%−26.5%) 7599 70 92.1 (70.6−113.6) 6.1 (�21.7 to 34.0) 1.07 (0.78−1.47) 1.11 (0.64−1.92)

Tercile 3 (26.6%−68.3%) 10,183 84 82.5 (64.9−100.1) �3.5 (�28.4 to 21.4) 0.96 (0.71−1.29) 0.76 (0.45−1.27)

More than 1 race

Tercile 1 (3.3%−17.3%) 26,476 131 49.5 (41.0−57.9) 0.0 (Reference) 1.00 (Reference) 1.00 (Reference)

Tercile 2 (17.4%−41.4%) 26,653 138 51.8 (43.2−60.4) 2.3 (�9.8 to 14.4) 1.05 (0.82−1.33) 0.96 (0.67−1.37)

Tercile 3 (41.5%−68.3%) 25,960 279 107.5 (94.9−120.0) 58.0 (42.9−73.1) 2.18 (1.77−2.69) 0.93 (0.61−1.40) CI, confidence interval; OR, odds ratio; SAMO, severe adverse maternal outcome. aCalculated as the difference between the incidence in the third (or second) tercile and the incidence in the first tercile (reference). bEstimated using univariate fixed-effect logistic regression. cEstimated using multivariate mixed-effect logistic regression with the hospital county nested within the hospital state as the random effect and adjusted for (1) age, (2) body mass index, (3) residence (rural or urban), (4) preexisting diabetes mellitus, (5) gestational dia- betes mellitus, (6) preexisting hypertension, (7) gestational hypertension, (8) month prenatal care began, (9) number of prenatal visits, (10) mother transferred in, (11) gestational age at delivery, (12) multiple gestation, (13) noncephalic presentation, (14) induction of labor, (15) attendant at birth, (16) delivery mode, (17) birthweight, and (18) hospital location (rural or urban).

Guglielminotti. State workforce diversity and maternal health. Am J Obstet Gynecol MFM 2022.

6 AJOG

M FM

Septem ber2022

O riginalR

esearch

Original Research

third tercile (crude OR, 0.63; 95% CI, 0.60−0.66). Moreover, a decreased risk of SAMO associated with giving birth in a state in the third tercile was observed for Black mothers (crude OR, 0.65; 95% CI, 0.61−0.70), Hispanic mothers (crude OR, 0.51; 95% CI, 0.48 −0.54), and Asian and Pacific Islander mothers (crude OR, 0.65; 95% CI, 0.58 −0.72) but not for Native American mothers (crude OR, 0.89; 95% CI, 0.72 −1.09) or mothers with more than 1 race (crude OR, 1.44; 95% CI, 0.72 −1.09). The results were unchanged when excluding blood transfusion from SAMO (Table 2).

Adjusted analysis After adjustment and compared with giving birth in states in the first tercile of the proportion of minoritized racial and ethnic RNs (Table 1), giving birth in states in the third tercile was associ- ated with a reduced risk of SAMO as follows: 32% for White mothers (aOR, 0.68; 95% CI, 0.59−0.77), 20% for Black mothers (aOR, 0.80; 95% CI, 0.65 −0.99), 31% for Hispanic mothers (aOR, 0.69; 95% CI, 0.58−0.82), and 50% for Asian and Pacific Islander mothers (aOR, 0.50; 95% CI, 0.38 −0.65). SAMO risk was not reduced for Native American mothers (aOR, 0.89; 95% CI, 0.61−1.29) or mothers of more than 1 race (aOR, 0.89; 95% CI, 0.66 −1.21). The results were robust with further adjustment for the state propor- tion of the racial and ethnic group examined (Table 1) and when excluding blood transfusion from SAMO (Table 2).

Comment Principal findings In this nationwide study of birth certifi- cate data, a racially diverse RN work- force was associated with a significantly reduced risk of SAMO in both non-His- panic White mothers and minoritized racial and ethnic mothers.

Results in the context of what is known and implications Our study provided robust evidence to support the recommendation to diver- sify the healthcare workforce as a

remedy for addressing racial and ethnic disparities in maternal health outcomes, as shown in the 2018 Consensus State- ment on the Reduction of Peripartum Racial and Ethnic Disparities of the National Partnership for Maternal Safety and in the Surgeon General’s recent call to action to improve mater- nal health.1,2,15 There are at least 3 path- ways linking a racially diverse healthcare workforce to reduced dispar- ities in maternal health outcomes. First, it could improve access to healthcare for underserved patients. For instance, minoritized racial and ethnic physicians are more likely than non-Hispanic White physicians to practice in under- served communities and to treat larger numbers of diverse racial and ethnic patients. Second, it could help reduce provider implicit bias.14 Implicit bias refers to attitudes or stereotypes toward minoritized racial and ethnic groups that affect healthcare workers’ under- standing, actions, and decisions in an unconscious manner, ultimately affect- ing the delivery of care.24−26 Third, it increases the likelihood of racial and ethnic concordance between patients and healthcare workers. Racial and eth- nic concordance has been associated with improvement in patient-physician communication and shared decision- making, greater time spent with physi- cians, and improved patient satisfaction and experience rating.27−30 More recently, racial and ethnic concordance between physicians and newborns has been associated with decreased mortal- ity of non-Hispanic Black infants.31

Although the direction of the effect of the state proportion of RNs from minoritized racial and ethnic groups on SAMO was consistent across the 6 maternal racial and ethnic groups, it was not statistically significant for Native Americans and those of more than 1 race mainly because of the small sample sizes in these 2 groups.

Strengths and limitations We analyzed a national census of birth data that facilitated meaningful statisti- cal comparisons for relatively rare out- comes across racial and ethnic groups. However, our findings should be

interpreted in the context of several lim- itations. First, our study was observa- tional, and the association between RN workforce diversity and SAMO has not been necessarily causal. Second, we did not analyze physician workforce diver- sity. Non-Hispanic Black physicians account for only 5% of US physicians, whereas 13% of the US population is non-Hispanic Black, and Hispanic physicians account for 6% of US physi- cians, whereas Hispanics make up 18% of the US population.32 We chose to focus instead on RNs because they are the frontline healthcare providers involved in the early recognition of warning signs for maternal complica- tions; delayed recognition of these com- plications has been repeatedly identified as a major contributor to preventable maternal deaths.16,33 In addition, diver- sifying the RN workforce may be a faster process than diversifying the phy- sician workforce because of the shorter duration of their curriculum (up to 4 years vs up to 8 years, respectively). Third, we analyzed the workforce diver- sity at the state-level but not at the site of delivery level because hospital-level data on racial and ethnic diversity of the healthcare workforce remain unavailable. Fourth, the sensitivity of using birth certificate data to detect SAMO is low.19 However, the underre- porting should be nondifferential across racial and ethnic groups and thus be unlikely to bias the estimated ORs.34

Conclusions A racially and ethnically diverse RN workforce was associated with a reduced risk of SAMO. This finding could guide the development of inter- vention programs to reduce racial and ethnic disparities in maternal health outcomes by diversifying the healthcare workforce. &

Supplementary materials Supplementary material associated with this article can be found in the online version at doi:10.1016/j.ajogmf.2022. 100689.

September 2022 AJOG MFM 7

Original Research

References 1. U.S. Department of Health and Human Serv- ices. The Surgeon General’s call to action to improve maternal health. Available at: https:// www.hhs.gov/sites/default/files/call-to-action- maternal-health.pdf. Accessed March 18, 2022. 2. U.S. Department of Health and Human Serv- ices. Healthy women, healthy pregnancies, healthy futures: action plan to improve maternal health in America. Available at: https://aspe. hhs.gov/system/fi les/aspe-fi les/264076/ healthy-women-healthy-pregnancies-healthy- future-action-plan_0.pdf. Accessed March 18, 2022. 3. Guglielminotti J, Wong CA, Friedman AM, Li G. Racial and ethnic disparities in death associ- ated with severe maternal morbidity in the United States: failure to rescue. Obstet Gynecol 2021;137:791–800. 4. Admon LK, Winkelman TNA, Zivin K, Terplan M, Mhyre JM, Dalton VK. Racial and ethnic dis- parities in the incidence of severe maternal mor- bidity in the United States, 2012-2015. Obstet Gynecol 2018;132:1158–66. 5. Kozhimannil KB, Interrante JD, Tofte AN, Admon LK. Severe maternal morbidity and mortality among indigenous women in the United States. Obstet Gynecol 2020;135:294– 300. 6. Taylor JK. Structural racism and maternal health among Black women. J Law Med Ethics 2020;48:506–17. 7. Crear-Perry J, Correa-De-Araujo R, Lewis Johnson T, Mclemore MR, Neilson E, Wallace M. Social and structural determinants of health inequities in maternal health. J Womens Health (Larchmt) 2021;30:230–5. 8. Janevic T, Zeitlin J, Egorova N, Hebert PL, Balbierz A, Howell EA. Neighborhood racial and economic polarization, hospital of delivery, and severe maternal morbidity. Health Aff (Millwood) 2020;39:768–76. 9. Liu SY, Fiorentini C, Bailey Z, Huynh M, Mcveigh K, Kaplan D. Structural racism and severe maternal morbidity in New York State. Clin Med Insights Womens Health 2019;12. 1179562X19854778. 10. Yearby R. Structural racism and health dis- parities: reconfiguring the social determinants of health framework to include the root cause. J Law Med Ethics 2020;48:518–26. 11. Bailey ZD, Feldman JM, Bassett MT. How structural racism works - racist policies as a root cause of U.S. racial health inequities. N Engl J Med 2021;384:768–73. 12. Bailey ZD, Krieger N, Ag�enor M, Graves J, Linos N, Bassett MT. Structural racism and health inequities in the USA: evidence and inter- ventions. Lancet 2017;389:1453–63. 13. Yearby R, Clark B, Figueroa JF. Structural racism in historical and modern US health care policy. Health Aff (Millwood) 2022;41:187–94.

8 AJOG MFM September 2022

14. Gillispie-Bell V. The contrast of color: why the Black community continues to suffer health disparities. Obstet Gynecol 2021;137:220–4. 15. Howell EA, Brown H, Brumley J, et al. Reduction of peripartum racial and ethnic dis- parities: a conceptual framework and maternal safety consensus bundle. Obstet Gynecol 2018;131:770–82. 16. Mhyre JM, D’oria R, Hameed AB, et al. The maternal early warning criteria: a proposal from the National Partnership for Maternal Safety. Obstet Gynecol 2014;124:782–6. 17. Centers for Disease Control and Preven- tion. U.S. standard certificate of live birth. Avail- able at: https://www.cdc.gov/nchs/data/dvs/ birth11-03final-ACC.pdf. Accessed March 18, 2022. 18. Health Resources and Services Adminis- tration. Area Health Resources Files (AHRF). data.HRSA.gov. Available at: https://data.hrsa. gov/topics/health-workforce/ahrf. Accessed March 18, 2022. 19. Luke B, Brown MB, Liu CL, Diop H, Stern JE. Validation of severe maternal morbidity on the US certificate of live birth. Epidemiology 2018;29:e31–2. 20. Centers for Disease Control and Preven- tion. How does CDC identify severe maternal morbidity? 2019. Available at: https://www. cdc.gov/reproductivehealth/maternalinfan- thealth/smm/severe-morbidity- ICD.htm. Accessed March 18, 2022. 21. Kaiser Family Foundation. Medicaid and CHIP Income eligibility limits for pregnant women. Available at: https://www.kff.org/med- icaid/state-indicator/medicaid-and-chip- income-eligibility-limits-for-pregnant-women/? c u r r e n t T i m e f r a m e = 0 & s o r t M o d e l =%7B%22colId%22:%22Location%22,%22 sort%22:%22asc%22%7D. Accessed March 18, 2022. 22. R Core Team. The R Project for Statistical Computing. Available at: https://www.R-proj- ect.org/. Accessed March 18, 2022. 23. Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simul Comput 2009;38:1228–34. 24. Chapman EN, Kaatz A, Carnes M. Physi- cians and implicit bias: how doctors may unwit- tingly perpetuate health care disparities. J Gen Intern Med 2013;28:1504–10. 25. Hall WJ, Chapman MV, Lee KM, et al. Implicit racial/ethnic bias among health care professionals and its influence on health care outcomes: a systematic review. Am J Public Health 2015;105:e60–76. 26. Hoffman KM, Trawalter S, Axt JR, Oliver MN. Racial bias in pain assessment and treat- ment recommendations, and false beliefs about biological differences between blacks and whites. Proc Natl Acad Sci U S A 2016;113: 4296–301.

27. Cooper-Patrick L, Gallo JJ, Gonzales JJ, et al. Race, gender, and partnership in the patient- physician relationship. JAMA 1999;282:583–9. 28. Saha S, Arbelaez JJ, Cooper LA. Patient- physician relationships and racial disparities in the quality of health care. Am J Public Health 2003;93:1713–9. 29. Takeshita J, Wang S, Loren AW, et al. Association of racial/ethnic and gender concor- dance between patients and physicians with patient experience ratings. JAMA Netw Open 2020;3:e2024583. 30. Jetty A, Jabbarpour Y, Pollack J, Huerto R, Woo S, Petterson S. Patient-physician racial concordance associated with improved health- care use and lower healthcare expenditures in minority populations. J Racial Ethn Health Dis- parities 2022;9:68–81. 31. Greenwood BN, Hardeman RR, Huang L, Sojourner A. Physician-patient racial concor- dance and disparities in birthing mortality for newborns. Proc Natl Acad Sci U S A 2020;117:21194–200. 32. Ly DP, Jena AB. Trends in diversity and representativeness of health care workers in the United States, 2000 to 2019. JAMA Netw Open 2021;4:e2117086. 33. Building US. Capacity to Review and Prevent Maternal Deaths. Report from nine maternal mortality review committee.Centers for Disease Control and Prevention; 2018. Available at: https://stacks.cdc.gov/view/cdc/ 51660. Accessed March 18, 2022. 34. Lash TL. Bias analysis. In: Lash TL, VanderWeele TJ, Haneuse S, Rothman KJ, eds. Modern epidemiology, Philadelphia, PA: Wolters Kluwer; 2021.

Author and article information From the Department of Anesthesiology, Columbia Uni- versity Vagelos College of Physicians and Surgeons, New York, NY (Drs Guglielminotti, Lee, Landau, and Li); Heil- brunn Department of Population and Family Health, Columbia University Mailman School of Public Health, New York, NY (Dr Samari); Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Columbia University Vagelos College of Physicians and Surgeons, New York, NY (Dr Friedman); Department of Epidemiology, Columbia University Mailman School of Public Health, New York, NY (Dr Li). Received Mar. 18, 2022; revised June 30, 2022;

accepted June 30, 2022. The authors report no conflict of interest. J.G. is supported by an R21 from the National Institute

on Minority Health and Health Disparities (grant number R21 MD016414) and by an R21 from the National Insti- tute of Mental Health (grant number R21 MH126096). This study was presented at the 54th annual meeting

of the Society for Obstetric Anesthesia and Perinatology, Chicago, IL, May 11−15, 2022. Corresponding author: Jean Guglielminotti, MD, PhD.

[email protected]