Abstract
BACKGROUND:
Postpartum hemorrhage is a leading cause of maternal morbidity and mortality in the United States and disproportionately affects pregnant persons of color.
OBJECTIVE:
This study aimed to identify the demographic and obstetrical characteristics of those who received different levels of antihemorrhagic intervention in the setting of severe postpartum hemorrhage requiring blood transfusion.
STUDY DESIGN:
This was a retrospective cohort study of patients with documented postpartum hemorrhage (estimated blood loss of ≥1000 mL) and blood product transfusion. Moreover, 3 levels of antihemorrhagic intervention were defined as follows: level 1, administration of uterotonics only; level 2, performance of a procedure (ie, B-Lynch suture, O’Leary stitch, Bakri balloon, dilation and curettage, laceration repair, or embolization); and level 3, hysterectomy. Maternal demographics, obstetrical characteristics, and comorbidities were extracted from electronic health records. Ordinal logistic regression was used to estimate the odds of higher intervention levels adjusting for maternal demographic and obstetrical characteristics.
RESULTS:
Of note, 365 patients were included in this study, with a racial or ethnic composition of 30% White, 42% Black, 18% Hispanic, and 10% other. Moreover, 233 patients (64%) received level 1 intervention, 98 patients (27%) received level 2 intervention, and 34 patients (9%) received level 3 intervention. Patients receiving higher levels of intervention were more likely to have greater estimated blood loss (P<.001), have more transfusions (P<.001), and be of advanced maternal age (P=.004). Black and Hispanic patients were less likely to have received higher levels of intervention than White patients (P=.034). After adjusting for estimated blood loss, advanced maternal age, placenta accreta spectrum, and fibroids, Black patients remained significantly less likely to receive higher levels of intervention (adjusted odds ratio, 0.55; 95% confidence interval, 0.30–0.98). This difference persisted at an estimated blood loss of ≥3000 mL, with Black and Hispanic patients being significantly less likely to receive higher levels of intervention than White patients (odds ratio, 0.31 [95% confidence interval, 0.10–0.92] and 0.10 [95% confidence interval, 0.01–0.53], respectively).
CONCLUSION:
Among patients experiencing postpartum hemorrhage and receiving transfusion, Black patients are less likely to receive higher levels of antihemorrhagic intervention. This disparity is concerning in this high-risk population and requires further attention and investigation.
Keywords: antihemorrhagic intervention, blood loss, hysterectomy, maternal morbidity, obstetrics, postpartum hemorrhage, pregnancy, racial disparities, retrospective cohort study, transfusion
Introduction
Maternal morbidity affects nearly 1.7 million pregnant patients in the United States annually, and postpartum hemorrhage (PPH) is a leading cause of morbidity and mortality.1,2 The American College of Obstetricians and Gynecologists (ACOG) defines PPH as the cumulative blood loss of ≥1000 mL or signs or symptoms of hypovolemia within 24 hours after birth.3 The rate of PPH has increased in the United States. It is estimated that the prevalence of PPH rose from 2.9% in 2010 to 3.2% in 2014.4,5 Furthermore, a study examining severe PPH, defined as PPH necessitating blood transfusion, hysterectomy, and/or surgical repair of the uterus, identified a more than doubling of the rate from 1999 to 2008.6 The management of PPH requires maintenance of hemodynamic stability and prompt diagnosis of the cause, followed by swift and targeted treatment (Figure 1).3 Care escalation is essential to maternal survival in PPH and intensifies from uterotonics to procedural management to hysterectomy. Hysterectomy is the definitive treatment for severe, intractable PPH.3,7
FIGURE 1. Postpartum hemorrhage management algorithm.
The management of postpartum hemorrhage requires first identifying the source of bleeding, followed by targeted medical management, procedural intervention, and transfusion. For severe intractable bleeding, a hysterectomy is definitive management.
D&C, dilation and curettage; FFP, fresh frozen plasma; pRBC, packed red blood cells.
Maternal morbidity and mortality disproportionately affect Black pregnant patients, and Black patients are twice as likely to experience severe maternal morbidity. This risk remains elevated even after adjusting for sociodemographic factors and comorbidities.8–10 In addition, Black pregnant patients have been found to be at higher risk of PPH-associated severe morbidity and mortality involving transfusion, disseminated intravascular coagulation, and hysterectomy.11
Black pregnant patients are 3 to 4 times more likely than White pregnant patients to die during childbirth, with most of these deaths thought to be preventable.12–14 Appropriate escalation of care is key to the management of PPH, and it may account for an important source of maternal health disparities. Past research has focused on maternal morbidity outcomes, but little is known about levels of care in granular detail and how it affects maternal morbidity and mortality.15 Thus, this study aimed to profile the maternal demographic and obstetrical characteristics that are associated with different levels of antihemorrhagic interventions in the setting of severe PPH requiring transfusion.
Materials and Methods
Study population
This was a retrospective cohort study of pregnant patients with documented PPH and blood product transfusion (Figure 2). The data were collected from 2 urban tertiary care hospitals in an academic health system from December 2015 to June 2020. The exclusion criteria included patients who had not participated in one or more antenatal visits in our hospital system, were outside of the gestational age range (<23 or >43 weeks of gestation), or were missing key covariates. An automated extraction algorithm of the electronic medical record (EMR) was used to identify clinically important factors and supplemented with a medical record review. Appendix A shows the International Classification of Diseases, Tenth Revision [ICD-10] codes.
FIGURE 2.
Flowchart for inclusion and exclusion criteria of study population, 2015–2020
AHI, antihemorrhagic intervention; PPH, postpartum hemorrhage; pRBC, packed red blood cells.
We first screened for PPH via EMR based on cumulative estimated blood lost during delivery hospitalization. For patients with potential PPH, a manual medical record review was performed to confirm and quantify the PPH, if present. PPH was defined following the ACOG criteria (ie, estimated blood loss [EBL] of ≥1000 mL). For 2 patients that showed clinical signs of PPH but cumulative blood lost could not be assessed or EBL did not meet the threshold, a maternal-fetal medicine specialist reviewed the records to confirm the inclusion in the cohort. The study population was stratified into 3 EBL levels: level 1, 1000 to 1499 mL; level 2, 1500 to 2999 mL; and level 3, ≥3000 mL. An EBL of 1500 mL was chosen as a cutoff based on clinical guidelines to give tranexamic acid at this level of blood loss and its use to indicate severe PPH and initiation of massive transfusion protocols.7,16,17
From this larger group of patients experiencing PPH, patients requiring transfusion were identified by at least 1 documented unit of packed red blood cells (pRBCs) in their EMR for their delivery admission, which was confirmed by medical record review. In addition, the quantity of each type of blood product (pRBCs, frozen fresh plasma [FFP], platelets, and cryoprecipitate) was recorded.
Measures and covariates
Our primary outcome of interest was an antihemorrhagic intervention, which we defined as 3 levels: level 1, uterotonics; level 2, procedure; and level 3, hysterectomy (Figure 3). This stratification was based on the management algorithm recommended by ACOG, in which uterotonics are first maximized before pursuing procedural and surgical interventions.3 This information was recorded from a manual medical record review.
FIGURE 3.
Definitions of 3 levels of antihemorrhagic intervention
Demographic variables included age (with advanced maternal age defined as ≥35 years old), self-reported maternal race and ethnicity, insurance, and zip code. All patients were categorized into 1 of 3 ethnic groups (Hispanic or Latino, non-Hispanic or Latino, and unknown or missing) and 1 of 6 racial groups (White, Black, Asian, American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander, and other). Given the sparseness of data for many of these groups, data were collapsed into 4 racial and ethnic groups: (1) non-Hispanic White (White), (2) non-Hispanic Black (Black), (3) non-Hispanic Other (Other), and (4) Hispanic. Insurance status was categorized as either private or public, self-pay, or uninsured. Rural-Urban Community Area codes were used to convert zip codes into urban, suburban, and rural classifications.18
Obstetrical conditions were identified using a version of an obstetrical-specific comorbidity index, which was updated to ICD-10 coding.19,20 Relevant diagnoses were extracted from the patient’s pregnancy problem list or delivery hospitalization and/or confirmed by manual medical review. Obstetrical characteristics included parity, gestational age at delivery, preterm delivery (gestational age of <259 days), uterine fibroids, previous uterine surgery (ie, myomectomy, dilation and curettage, and hysteroscopy), history of cesarean delivery, and type of delivery (cesarean delivery vs vaginal delivery). Measured patient comorbidities included obesity, chronic hypertension, gestational hypertension, preeclampsia, pregestational diabetes mellitus, gestational diabetes mellitus, placenta previa, and placenta accreta spectrum. Adverse obstetrical outcomes included chorioamnionitis, endometritis, and intensive care unit (ICU) admission.
Statistical analysis
For unadjusted associations of categorical variables, the chi-square test was used. For continuous variables, we first tested normality via the Shapiro-Wilk test; the analysis of variance test was used for normally distributed variables, and the Kruskal-Wallis test was used for nonnormally distributed variables. The measured demographic, clinical, and obstetrical covariates were used to construct an ordinal multivariable logistic regression model, with the outcome being the level of antihemorrhagic intervention received. Covariates were included in the adjustment set based on hypothesized potential confounding, and goodness of fit was assessed using likelihood ratio tests and Bayesian information criteria tests. The final adjustment set included maternal race and ethnicity, advanced maternal age, placenta accreta spectrum, fibroids, and EBL.
Data were extracted from the EMR by the Johns Hopkins Core for Clinical Research and Data Acquisition. Manual medical record review extracted additional clinical data from the EMR and was entered into a Research Electronic Data Capture database. Data were analyzed using Stata (version 16.1; Stata-Corp, College Station, TX). A 2-sided P value of .05 was prespecified as statistically significant. This study was approved by Johns Hopkins Medicine Institutional Review Board (under protocol number 00163026).
Results
Baseline cohort characteristics
A total of 365 patients (mean age, 30 years; standard deviation, 6 years) were included in this study who experienced PPH requiring transfusion of at least 1 unit of pRBCs (Table 1). The racial or ethnic composition of the cohort was 30% White, 42% Black, 18% Hispanic, and 10% other. Most of the population resided in an urban area (94%). Of note, 53% of patients were covered by private insurance, and 47% of patients were covered by public insurance, uninsured, or self-paid. Furthermore, 83% of patients underwent a cesarean delivery.
TABLE 1.
Baseline characteristics of cohort stratified by race and ethnicity
| Characteristics | White (n=109) | Black (n=154) | Hispanic (n=65) | Other (n=37) | Total (N=365) | P value |
|---|---|---|---|---|---|---|
| n (%) or median (IQR) | ||||||
| Demographic characteristics | ||||||
| Zip code | <.001a | |||||
| Suburban | 16 (14.68) | 5 (3.25) | 0 (0.00) | 2 (5.41) | 23 (6.30) | |
| Urban | 93 (85.32) | 149 (96.75) | 65 (100.00) | 35 (94.59) | 342 (93.70) | |
| Insurance | <.001a | |||||
| Private | 78 (71.56) | 77 (50.00) | 12 (18.46) | 26 (70.27) | 193 (52.88) | |
| Public | 31 (28.44) | 77 (50.00) | 53 (81.54) | 11 (29.73) | 172 (47.12) | |
| Advanced age (≥35 y) | 25 (22.94) | 41 (26.62) | 14 (21.54) | 16 (43.24) | 96 (26.30) | .076 |
| Age, mean±SD | 31±5 | 30±7 | 29±6 | 33±5 | 30±6 | .013a |
| Obstetrical characteristics | ||||||
| Preterm (<259 d) | 36 (33.03) | 52 (33.77) | 12 (18.46) | 8 (21.62) | 108 (29.59) | .075 |
| Parity | .002a | |||||
| Nulliparous | 62 (56.88) | 51 (33.12) | 26 (40.00) | 13 (35.14) | 152 (41.64) | |
| 1 | 29 (26.61) | 49 (31.82) | 16 (24.62) | 15 (40.54) | 109 (29.86) | |
| 2+ | 18 (16.51) | 54 (35.06) | 23 (35.38) | 9 (24.32) | 104 (28.49) | |
| Fibroids | 2 (1.83) | 22 (14.29) | 5 (7.69) | 2 (5.41) | 31 (8.49) | .004a |
| Previous uterine surgery | 26 (23.85) | 42 (27.27) | 4 (6.15) | 11 (29.73) | 83 (22.74) | .004a |
| Cesarean delivery | 84 (77.06) | 142 (92.21) | 49 (75.38) | 27 (72.97) | 302 (82.74) | .001a |
| Previous cesarean delivery | 43 (39.45) | 82 (53.25) | 22 (33.85) | 17 (45.95) | 164 (44.93) | .031a |
| Maternal comorbidities | ||||||
| Obesity | 31 (28.44) | 82 (53.25) | 16 (24.62) | 9 (24.32) | 138 (37.81) | <.001a |
| Chronic hypertension | 12 (11.01) | 35 (22.73) | 4 (6.15) | 2 (5.41) | 53 (14.52) | .001a |
| Gestational hypertension | 35 (32.11) | 64 (41.56) | 21 (32.31) | 6 (16.22) | 126 (34.52) | .026a |
| Preeclampsia | 27 (24.77) | 49 (31.82) | 16 (24.62) | 5 (13.51) | 97 (26.58) | .126 |
| Pregestational diabetes mellitus | 9 (8.26) | 11 (7.14) | 2 (3.08) | 1 (2.70) | 23 (6.30) | .416 |
| Gestational diabetes mellitus | 9 (8.26) | 11 (7.14) | 8 (12.31) | 14 (37.84) | 42 (11.51) | <.001a |
| Placenta previa | 15 (13.76) | 14 (9.09) | 9 (13.85) | 6 (16.22) | 44 (12.05) | .498 |
| Placenta accreta spectrum | 10 (9.17) | 9 (5.84) | 2 (3.08) | 3 (8.11) | 24 (6.58) | .429 |
| Adverse obstetrical outcomes | ||||||
| Chorioamnionitis | 12 (11.01) | 12 (7.79) | 11 (16.92) | 6 (16.22) | 41 (11.23) | .182 |
| Endometritis | 6 (5.50) | 12 (7.79) | 3 (4.62) | 7 (18.92) | 28 (7.67) | .042a |
| ICU admission | 10 (9.17) | 13 (8.44) | 2 (3.08) | 5 (13.51) | 30 (8.22) | .284 |
P values were calculated using the analysis of variance test (normally distributed) or Kruskal-Wallis test (nonnormally distributed) for continuous variables and the chi-square test for categorical variables.
ICU, intensive care unit; IQR, interquartile range; SD, standard deviation.
A P value of <.05 is considered significant.
Analysis by race and ethnicity
Black and Hispanic patients were more likely to reside in urban areas and have public insurance (P<.001 for both) (Table 1). Black patients were more likely to have had a previous birth, previous cesarean delivery, and fibroids. In addition, they were more likely to have comorbidities, including gestational hypertension, chronic hypertension, and obesity. Patients in the other racial or ethnic category were more likely to have gestational diabetes mellitus. There was no statistically significant difference in EBL by race and ethnicity, with the median EBL in this cohort being 1800 mL (interquartile range, 1300–2500 mL) (Table 2). However, there was a difference in the units of pRBCs (P=.033) received across different race and ethnicity groups. There was no statistically significant difference among other blood products received across groups, and there was no difference in the timing of transfusion. Finally, the rates of individual types of procedural interventions also did not differ among groups, except for laceration repairs, which were found to be more common among patients in the other category (P<.001).
TABLE 2.
Clinical management of PPH stratified by race and ethnicity
| Variables | White (n=109) | Black (n=154) | Hispanic (n=65) | Other (n=37) | Total (N=365) | P value |
|---|---|---|---|---|---|---|
| n (%) or median (IQR) | ||||||
| EBL (L) | 2.0 (1.4–2.5) | 1.8 (1.2–2.5) | 1.7 (1.3–2.2) | 2.0 (1.4–2.5) | 1.8 (1.3–2.5) | .440 |
| Blood products | ||||||
| pRBC (units) | 2 (1–3) | 2 (1–3) | 2 (1–2) | 2 (1–3) | 2 (1–3) | .033a |
| FFP (units) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–1) | 0 (0–0) | .210 |
| Platelets (units) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) | .466 |
| Cryo (units) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–0) | .575 |
| Total (units) | 2 (1–3) | 2 (1–4) | 2 (1–2) | 2 (1–4) | 2 (1–3) | .052 |
| Transfusion timing | ||||||
| Intrapartum | 54 (49.54) | 80 (51.95) | 32 (49.23) | 22 (59.46) | 188 (51.51) | .740 |
| Postpartumb | 71 (65.14) | 102 (66.23) | 43 (66.15) | 24 (64.86) | 240 (65.75) | .997 |
| Early | 51 (46.79) | 61 (39.61) | 25 (38.46) | 15 (40.54) | 152 (41.64) | .628 |
| Late | 35 (32.11) | 57 (37.01) | 22 (33.85) | 11 (29.73) | 125 (34.25) | .781 |
| Procedures | ||||||
| B-Lynch suture | 6 (5.50) | 2 (1.30) | 2 (3.08) | 1 (2.70) | 11 (3.01) | .275 |
| O’Leary stitch | 10 (9.17) | 11 (7.14) | 5 (7.69) | 4 (10.81) | 30 (8.22) | .868 |
| Bakri balloon | 12 (11.01) | 20 (12.99) | 5 (7.69) | 7 (18.92) | 44 (12.05) | .384 |
| D&C | 18 (16.51) | 14 (9.09) | 10 (15.38) | 2 (5.41) | 44 (12.05) | .136 |
| Lac Repair | 5 (4.59) | 1 (0.65) | 1 (1.54) | 6 (16.22) | 13 (3.56) | <.001a |
| Ex Lap | 4 (3.67) | 5 (3.25) | 2 (3.08) | 2 (5.41) | 13 (3.56) | .928 |
| IR embolization | 9 (8.26) | 10 (6.49) | 1 (1.54) | 1 (2.70) | 21 (5.75) | .243 |
| Hysterectomy | 14 (12.84) | 14 (9.09) | 2 (3.08) | 4 (10.81) | 34 (9.32) | .194 |
The P values were calculated using the Kruskal-Wallis test (nonnormally distributed) for continuous variables and chi-square test for categorical variables.
Cryo, cryoprecipitate; D&C, dilation and curettage; EBL, estimated blood loss; Ex lap, exploratory laparotomy; FFP, frozen fresh plasma; IQR, interquartile range; IR embolization, interventional radiology embolization; Lac repair, laceration repair; pRBC, packed red blood cells.
A P value of <.05 is considered significant;;
Postpartum period (any transfusion ≥4 hours after delivery), early postpartum period (4–24 hours after delivery), and late postpartum period (>24 hour after delivery).
Analysis by level of intervention
Overall, 233 patients (64%) required level 1 intervention (uterotonics only), 98 patients (27%) required level 2 intervention (procedure), and 34 patients (9%) required level 3 intervention (hysterectomy). Important baseline differences were present among patients receiving different levels of intervention (Table 3). Patients receiving higher levels of intervention were more likely to be of advanced maternal age (P=.004), have had greater EBL (P<.001), be preterm (P=.006), be multiparous (P=.004), and have had a previous cesarean delivery (P=.003). In addition, they were more likely to have placenta previa (P<.001), have placenta accreta spectrum (P<.001), have received more blood products (P<.001), and have been admitted to the ICU (P<.001). In addition, the rates of cesarean delivery differed among groups (P<.001). There was no maternal mortality in this cohort (Table 3).
TABLE 3.
Baseline characteristics of cohort stratified by level of intervention
| Characteristics | Level 1: uterotonic (n=233) | Level 2: procedure (n=98) | Level 3: hysterectomy (n=34) | Total (N=365) | P value |
|---|---|---|---|---|---|
| n (%) or median (IQR) | |||||
| Demographic characteristics | |||||
| Maternal race | .124 | ||||
| White | 61 (26.18) | 34 (34.69) | 14 (41.18) | 109 (29.86) | |
| Black | 106 (45.49) | 34 (34.69) | 14 (41.18) | 154 (42.19) | |
| Hispanic | 46 (19.74) | 17 (17.35) | 2 (5.88) | 65 (17.81) | |
| Other | 20 (8.58) | 13 (13.27) | 4 (11.76) | 37 (10.14) | |
| Zip code | .673 | ||||
| Suburban | 13 (5.58) | 8 (8.16) | 2 (5.88) | 23 (6.30) | |
| Urban | 220 (94.42) | 90 (91.84) | 32 (94.12) | 342 (93.70) | |
| Insurance | .084 | ||||
| Private | 176 (50.21) | 52 (53.06) | 24 (70.59) | 193 (52.88) | |
| Public | 116 (49.79) | 46 (46.94) | 10 (29.41) | 172 (47.12) | |
| Advanced age (≥35 y) | 56 (24.03) | 23 (23.47) | 17 (50.00) | 96 (26.03) | .004a |
| Age, mean±SD | 30±6 | 30±6 | 34±7 | 30±6 | .004a |
| Obstetrical characteristics | |||||
| Estimated blood loss (L) | 1.5 (1.2–2.0) | 2.0 (1.5–2.8) | 4.1 (2.9–8.8) | 1.8 (1.3–2.5) | <.001a |
| Preterm (<259 d) | 61 (26.18) | 29 (29.59) | 18 (52.94) | 108 (29.59) | .006a |
| Parity | .004a | ||||
| Nulliparous | 99 (42.49) | 48 (48.98) | 5 (14.71) | 152 (41.64) | |
| 1 | 65 (27.90) | 31 (31.63) | 13 (38.24) | 109 (29.86) | |
| 2+ | 69 (29.61) | 19 (19.39) | 16 (47.06) | 104 (28.49) | |
| Fibroids | 23 (9.87) | 4 (4.08) | 4 (11.76) | 31 (8.49) | .175 |
| Previous uterine surgery | 49 (21.03) | 21 (21.43) | 13 (38.24) | 83 (22.74) | .077 |
| Cesarean delivery | 214 (91.85) | 59 (60.20) | 29 (85.29) | 302 (82.74) | <.001a |
| Previous cesarean delivery | 110 (52.79) | 32 (32.65) | 22 (64.71) | 164 (44.93) | .003a |
| Maternal comorbidities | |||||
| Obesity | 93 (39.91) | 31 (31.63) | 14 (41.18) | 138 (37.81) | .334 |
| Chronic hypertension | 40 (17.17) | 9 (9.18) | 4 (11.76) | 53 (14.52) | .152 |
| Gestational hypertension | 81 (34.76) | 33 (33.67) | 12 (35.29) | 126 (34.52) | .977 |
| Preeclampsia | 61 (26.18) | 25 (25.51) | 11 (32.35) | 97 (26.58) | .720 |
| Pregestational diabetes mellitus | 19 (8.15) | 2 (2.04) | 2 (5.88) | 23 (6.30) | .112 |
| Gestational diabetes mellitus | 25 (10.73) | 16 (16.33) | 1 (2.94) | 42 (11.51) | .090 |
| Placenta previa | 17 (7.30) | 18 (18.37) | 9 (26.47) | 44 (12.05) | <.001a |
| Placenta accreta spectrum | 3 (1.29) | 3 (3.06) | 18 (52.94) | 24 (6.58) | <.001a |
| Adverse obstetrical outcomes | |||||
| Chorioamnionitis | 24 (10.30) | 15 (15.31) | 2 (5.88) | 41 (11.23) | .245 |
| Endometritis | 16 (6.87) | 12 (12.24) | 0 (0.00) | 28 (7.67) | .052 |
| Total blood products | 2 (1–2) | 2 (2–5) | 12 (6–27) | 2 (1–3) | <.001a |
| ICU admission | 5 (2.15) | 5 (5.10) | 20 (58.82) | 30 (8.22) | <.001a |
The P values were calculated using the analysis of variance test (normally distributed) or Kruskal-Wallis test (nonnormally distributed) for continuous variables and the chi-square test for categorical variables.
ICU, intensive care unit; IQR, interquartile range; SD, standard deviation.
A P value of <.05 is considered significant.
Of note, insurance status and zip code were not found to be significantly different among patients receiving different levels of intervention. Relevant conditions, such as uterine fibroids, previous uterine surgery, preeclampsia, gestational hypertension, chorioamnionitis, endometritis, pregestational and gestational diabetes mellitus, chronic hypertension, and obesity were not different.
Black and Hispanic patients were less likely to have received higher levels of intervention than White patients (P=.034). After adjusting for advanced maternal age, EBL, placenta accreta spectrum, and fibroids, Black patients remained significantly less likely to receive higher levels of intervention (adjusted odds ratio [aOR], 0.55; 95% confidence interval [CI], 0.30–0.98) (Figure 4). After adding an adjustment for predelivery hemoglobin, Black patients remained less likely to receive higher levels of intervention, but this difference was no longer significant (aOR, 0.64; 95% CI, 0.35–1.17).
FIGURE 4. Adjusted odds ratio for higher level of intervention by race and ethnicity.
Data have been adjusted for maternal race and ethnicity, advanced maternal age, placenta accreta spectrum, fibroids, and estimated blood loss. Asterisk denotes P value of <.05 is considered significant.
aOR, adjusted odds ratio; CI, confidence interval.
Analysis by estimated blood loss
The study population was stratified into three EBL levels: level 1, 1000 to 1499 mL (n=111); level 2, 1500 to 2999 mL (n=189); and level 3, ≥3000 mL of EBL (n=65), to determine if the observed disparity in the escalation of care continued to be present at higher levels of blood loss (Figure 5). At the lowest and intermediate levels of blood loss, there was no significant difference between race and ethnicity groups in the level of intervention received. However, at the highest level of blood loss (ie, ≥3000 mL) ordinal logistic regression analysis demonstrated that the odds of being more likely to receive higher levels of interventions for Black patients was 0.31 times the odds of White patients (odds ratio [OR], 0.31; 95% CI, 0.10–0.92). For Hispanic patients, the odds of being more likely to receive higher levels of intervention was 0.10 times the odds of White patients (OR, 0.10; 95% CI, 0.01–0.53). Further examining those with EBL ≥3000 mL, there was a marginal difference in the level of care received by maternal race and ethnicity (P=.052) (Table 4). Among those continuing to receive only uterotonics at ≥3000 mL, there was no White patient, but Black patients made up 65% (n=11) and Hispanic patients 24% (n=4) of this group.
FIGURE 5. Intervention stratified by race and ethnicity and EBL.
Asterisk denotes P value of <.05 is considered significant.
CI, confidence interval; EBL, estimated blood loss; OR, odds ratio.
TABLE 4.
Level of intervention by race and ethnicity in patients with an estimated blood loss of ≥3000 mL
| Variables | Level 1: uterotonic (n=17) | Level 2: procedure (n=23) | Level 3: hysterectomy (n=25) | Total (N=65) | P value |
|---|---|---|---|---|---|
| Maternal race, n (%) | .052 | ||||
| White | 0 (0.00) | 8 (34.78) | 9 (36.00) | 17 (26.15) | |
| Black | 11 (64.71) | 12 (52.17) | 11 (44.00) | 34 (52.31) | |
| Hispanic | 4 (23.53) | 1 (4.35) | 1 (4.00) | 6 (9.23) | |
| Other | 2 (11.76) | 2 (8.70) | 4 (16.00) | 8 (12.31) |
The P values were calculated using the chi-square test for categorical variables.
Discussion
Principal findings and results
In this retrospective cohort of 365 pregnant patients experiencing PPH requiring transfusion, there was no significant difference in EBL across racial and ethnic groups, but there was a significant difference in total quantity of blood products received. Of note, 64% of patients received uterotonics only, 27% of patients received procedural management, and 9% of patients received a hysterectomy. Although the rates of hysterectomy were much higher than the national average of 80 to 100 per 100,000 births, this is to be expected given the high-risk nature of this population.21
Higher levels of intervention were associated with maternal race and ethnicity, advanced maternal age, higher EBL, preterm birth, placenta accreta spectrum, and greater total blood products. After adjusting for EBL, advanced maternal age, placenta accreta spectrum, and presence of fibroids, Black patients remained significantly less likely to receive higher levels of intervention. This association was still present, although no longer significant, after additionally adjusting for predelivery hemoglobin. Although the predelivery of hemoglobin directly affects the likelihood to transfuse a patient, the management of PPH is more directly related to resuscitation of active bleeding and transfusion based on the most current hemoglobin value. Thus, this result may be secondary to a smaller cohort size of patients with PPH requiring transfusion.
The lower odds of escalation of care among Black patients may seem to contradict past studies, which have shown higher odds of peripartum hysterectomy among Black, Hispanic, and Asian patients than White patients.21,22 However, with the granular data available in this study, the results reflect the spectrum of antihemorrhagic interventions that a patient may receive and offer a more comprehensive picture. Furthermore, in this subpopulation that has experienced severe maternal morbidity, the observed trends may be specific to this high-risk group deserving more in-depth investigation.
This racial disparity remained at an EBL of ≥3000 mL, with Black and Hispanic patients less likely to receive higher levels of intervention for PPH and a notable number of these patients continuing to receive only uterotonics. This qualitatively reflects differences and disparities in the escalation of care. Importantly, there was no maternal death in the cohort.
Clinical and research implications
The findings of this study expose racial disparities in the clinical management and escalation of care of patients experiencing PPH requiring transfusion. A study found that although rates of nonsevere PPH did not change significantly from 2001–2002 to 2011–2012, the rates of PPH requiring blood transfusions more than doubled.23 This paralleled a rise in interventions, such as uterine tamponade, uterine artery embolization, and hysterectomy.23 Thus, it is of utmost importance to investigate and reduce racial disparities in the escalation of care among this growing patient population.
Past research has demonstrated that disparities in maternal morbidity and mortality are driven by individual-, community-, provider-, and system-level factors.24 This study found statistically significant differences in patient-specific factors, including sociodemographic variables, such as zip code, insurance, and age, and comorbidities, such as obesity, hypertension, and gestational diabetes mellitus. In addition, this study investigated differences in the escalation of care, which involves factors on the provider (eg, knowledge, implicit bias, cultural competence, and communication) and system level (eg, access to high-quality care, social and political policies, and healthcare institutions). Implicit bias has been shown to directly correlate with lower quality of patient care by impacting patient-provider interactions and treatment decisions and is routinely underrecognized by healthcare providers.25 These factors likely all contributed to the disparities observed in this study and offer opportunities for improvement.
Improving the clinical awareness of severe maternal morbidity (ie, severe blood loss) and the corresponding escalation of care can lead to better maternal outcomes and reduce disparities. This can be achieved by standardization of care through safety bundle implementation, maternal early warning triggers, and clinical decision-making support tools, such as protocols and checklists.26–28 There is growing evidence that these quality improvement tools improve health outcomes and aid in timely diagnosis and treatment to prevent or limit morbidity.27 In addition, implicit bias training, cultural competency education, and diversity and inclusion strategies can address provider-and system-level inadequacies.24
Another key area of improvement is in expanding the measurement of differences in both the care received and outcomes at an institutional level. Thus, this study serves as an example in further highlighting the need for granular data in the fields of disparities research and maternal-fetal medicine. The description, analysis, and interpretation of granular data in larger cohorts can reveal further pitfalls in maternal care and reflect opportunities for broader structural improvements.
Strengths and Limitations
This analysis of patients with PPH requiring transfusion adds important data to the literature on racial disparities in PPH management. The limitations of the study include an overall small sample size of patients (n=365), especially at higher levels of blood loss, which limited statistical power for further adjusted analyses. Furthermore, the 2 hospitals were urban, tertiary, and within proximity of each other. Thus, data on zip codes should be interpreted with caution. Another limitation of the study is the possibility of reverse causality between EBL and level of intervention, in that patients who do not receive the appropriate level of care will consequently lose more blood. To address this concern, additional analyses stratified by EBL were performed.
The strengths of this study include the level of granularity achieved, allowing for the detailed characterization of patients and the clinical management of their PPH. Furthermore, this subset of patients is at high risk of morbidity, and thus, an expanded understanding of their management is particularly important to help support optimal health outcomes. All research was conducted at a tertiary care hospital system that manages large volumes of high-acuity patients. Data were collected both through automated data extraction and manual medical record review to minimize missing or inaccurate information from the EMR.
Conclusions
Significant racial disparities in the escalation of care were identified in this cohort of patients experiencing PPH requiring transfusion. Black patients were less likely to receive higher levels of intervention, even at higher levels of blood loss. Our result warrants further research into this high-risk population, with the ultimate goal of implementing structural improvements to better address the continued disparities in maternal morbidity and mortality. ■
Supplementary Material
AJOG MFM at a Glance.
Why was this study conducted?
This study aimed to characterize disparities by race and ethnicity in antihemorrhagic intervention in the setting of severe postpartum hemorrhage (PPH) requiring blood transfusion.
Key findings
After adjusting for demographic and obstetrical covariates, Black patients were less likely to receive higher levels of intervention in severe PPH. Patients receiving higher levels of intervention are more likely to have greater blood loss, be of advanced maternal age, and have placenta accreta spectrum.
What does this add to what is known?
Black patients are less likely to experience an escalation of care during PPH, which may contribute to maternal morbidity and mortality. This study characterizes at a granular level various antihemorrhagic interventions in a high-risk population (PPH with transfusion).
Acknowledgments
T.M.B. was supported by the National Heart, Lung, and Blood Institute (grant number T32HL007024). The funding source had no involvement in the conduct or preparation of this manuscript.
Footnotes
Supplementary materials
Supplementary material associated with this article can be found in the online version at doi:10.1016/j.ajogmf.2023.100938.
The authors report no conflict of interest.
References
- 1.Danel I, Berg C, Johnson CH, Atrash H. Magnitude of maternal morbidity during labor and delivery: United States, 1993–1997. Am J Public Health 2003;93:631–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Main EK, Goffman D, Scavone BM, et al. National partnership for maternal safety: consensus bundle on obstetric hemorrhage. Anesth Analg 2015;121:142–8. [DOI] [PubMed] [Google Scholar]
- 3.Committee on Practice Bulletins-Obstetrics. Practice Bulletin No. 183: postpartum hemorrhage. Obstet Gynecol 2017;130:e168–86. [DOI] [PubMed] [Google Scholar]
- 4.Berg CJ, MacKay AP, Qin C, Callaghan WM. Overview of maternal morbidity during hospitalization for labor and delivery in the united states: 1993–1997 and 2001–2005. Obstet Gynecol 2009;113:1075–81. [DOI] [PubMed] [Google Scholar]
- 5.Reale SC, Easter SR, Xu X, Bateman BT, Farber MK. Trends in postpartum hemorrhage in the United States from 2010 to 2014. Anesth Analg 2020;130:e119–22. [DOI] [PubMed] [Google Scholar]
- 6.Kramer MS, Berg C, Abenhaim H, et al. Incidence, risk factors, and temporal trends in severe postpartum hemorrhage. Am J Obstet Gynecol 2013;209. 449.e1–7. [DOI] [PubMed] [Google Scholar]
- 7.Evensen A, Anderson JM, Fontaine P. Postpartum hemorrhage: prevention and treatment. Am Fam Physician 2017;95:442–9. [PubMed] [Google Scholar]
- 8.Creanga AA, Bateman BT, Kuklina EV, Callaghan WM. Racial and ethnic disparities in severe maternal morbidity: a multistate analysis, 2008–2010. Am J Obstet Gynecol 2014;210. 435.e1–8. [DOI] [PubMed] [Google Scholar]
- 9.Howell EA, Egorova NN, Janevic T, et al. Race and ethnicity, medical insurance, and within-hospital severe maternal morbidity disparities. Obstet Gynecol 2020;135:285–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Howell EA, Egorova NN, Balbierz A, Zeitlin J, Hebert PL. Site of delivery contribution to black-white severe maternal morbidity disparity. Am J Obstet Gynecol 2016;215:143–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gyamfi-Bannerman C, Srinivas SK, Wright JD, et al. Postpartum hemorrhage outcomes and race. Am J Obstet Gynecol 2018;219. 185.e1–10. [DOI] [PubMed] [Google Scholar]
- 12.Creanga AA, Berg CJ, Syverson C, Seed K, Bruce FC, Callaghan WM. Race, ethnicity, and nativity differentials in pregnancy-related mortality in the united states: 1993–2006. Obstet Gynecol 2012;120:261–8. [DOI] [PubMed] [Google Scholar]
- 13.Berg CJ, Harper MA, Atkinson SM, et al. Preventability of pregnancy-related deaths: results of a state-wide review. Obstet Gynecol 2005;106:1228–34. [DOI] [PubMed] [Google Scholar]
- 14.Geller SE, Cox SM, Callaghan WM, Berg CJ. Morbidity and mortality in pregnancy: laying the Groundwork for Safe Motherhood. Womens Health Issues 2006;16:176–88. [DOI] [PubMed] [Google Scholar]
- 15.Cantwell R, Clutton-Brock T, Cooper G, et al. Saving Mothers’ Lives: reviewing maternal deaths to make motherhood safer: 2006–2008. The eighth report of the confidential enquiries into maternal deaths in the United Kingdom. BJOG 2011;118(Suppl1):1–203. [DOI] [PubMed] [Google Scholar]
- 16.Schorn MN, Phillippi JC. Volume replacement following severe postpartum hemorrhage. J Midwifery Womens Health 2014;59:336–43. [DOI] [PubMed] [Google Scholar]
- 17.Nyfløt LT, Sandven I, Stray-Pedersen B, et al. Risk factors for severe postpartum hemorrhage: a case-control study. BMC Pregnancy Childbirth 2017;17:17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Parikh-Patel A, Bates JH, Campleman S. Colorectal cancer stage at diagnosis by socioeconomic and urban/rural status in California, 1988–2000. Cancer 2006;107:1189–95. [DOI] [PubMed] [Google Scholar]
- 19.Gavin NR, Federspiel JJ, Boyer T, et al. Mode of delivery among women with maternal cardiac disease. J Perinatol 2023. [Epub ahead of print]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Bateman BT, Mhyre JM, Hernandez-Diaz S, et al. Development of a comorbidity index for use in obstetric patients. Obstet Gynecol 2013;122:957–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Govindappagari S, Wright JD, Ananth CV, Huang Y, DʼAlton ME, Friedman AM. Risk of peripartum hysterectomy and center hysterectomy and delivery volume. Obstet Gynecol 2016;128:1215–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lawson SM, Chou B, Martin KL, Ryan I, Eke AC, Martin KD. The association between race/ethnicity and peripartum hysterectomy and the risk of perioperative complications. Int J Gynaecol Obstet 2020;151:57–66. [DOI] [PubMed] [Google Scholar]
- 23.Ahmadzia HK, Grotegut CA, James AH. A national update on rates of postpartum haemorrhage and related interventions. Blood Transfus 2020;18:247–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Howell EA, Brown H, Brumley J, et al. Reduction of peripartum racial and ethnic disparities: a conceptual framework and maternal safety consensus bundle. Obstet Gynecol 2018;131:770–82. [DOI] [PubMed] [Google Scholar]
- 25.Saluja B, Bryant Z. How implicit bias contributes to racial disparities in maternal morbidity and mortality in the United States. J Womens Health (Larchmt) 2021;30:270–3. [DOI] [PubMed] [Google Scholar]
- 26.Howell EA, Zeitlin J. Improving hospital quality to reduce disparities in severe maternal morbidity and mortality. Semin Perinatol 2017;41:266–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Arora KS, Shields LE, Grobman WA, D’Alton ME, Lappen JR, Mercer BM. Triggers, bundles, protocols, and checklists–what every maternal care provider needs to know. Am J Obstet Gynecol 2016;214:444–51. [DOI] [PubMed] [Google Scholar]
- 28.Shields LE, Wiesner S, Klein C, Pelletreau B, Hedriana HL. Use of maternal early warning trigger tool reduces maternal morbidity. Am J Obstet Gynecol 2016;214. 527.e1–6. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.