Hospital Characteristics and Other Factors Associated with the Risk of Postpartum Hemorrhage in the United States

Abstract

Background:

Postpartum hemorrhage (PPH) is one of the leading causes of maternal morbidity and mortality in the United States. It occurs when blood loss exceeds 1000mL regardless of the delivery route. Careful assessment of various causes and risk factors of PPH is essential to reduce and prevent further complications, avoid maternal morbidity and mortality, and better manage PPH. This study aimed to examine the associations of hospital characteristics and regions of hospital locations across the United States with PPH risk, as the outcomes of such an assessment may contribute to practice-relevant scientific evidence to improve policies and protocols regarding effective PPH management.

Methods:

This retrospective study used the 2018 National Inpatient Sample database from the Healthcare Cost and Utilization Project (HCUP) to examine the associations of PPH risk with characteristics and regions of hospital lections.

Results:

After controlling for clinical risk factors, the results showed that hospitals owned by private investors had significant associations with decreased risk of PPH. Conversely, large bed size, urban teaching status, and West and Midwest location were associated with an increased risk of PPH.

Conclusion:

Additional research is needed to determine whether these variations across regions and hospital characteristics are due to differences in obstetric practice and management.

Introduction

Postpartum hemorrhage (PPH) is the most common complication of childbirth and represents an obstetric emergency that can be life-threatening if not properly managed.^1,2 PPH can occur within the first 24 hours and up to 12 weeks after childbirth, when blood loss exceeds 1000 mL, regardless of the delivery route.³ The prevalence of PPH is on the rise in the United States as it increased from 2.9% in 2010 to 3.2% in 2017.¹ PPH can occur in patients without risk factors for hemorrhage. However, certain conditions may increase the risk of PPH. These include multiple pregnancies, fibroids, pre-eclampsia, amnionitis, instrumental vaginal delivery, previous cesarean delivery,^3

–6 prolonged labor, obesity,^6,7 infection, and the use of forceps or vacuum-assisted delivery.⁶ Demographic factors such as an age of ≥35 years and ethnicity of African American,⁸ Asian, or Hispanic are also other factors that may increase the PPH risk.^9,10,11

Complications of PPH include anemia, fatigue, dilutional coagulopathy, orthostatic hypotension, myocardial ischemia, postpartum depression, postpartum pituitary necrosis (Sheehan syndrome), the need for blood transfusion, and death.¹⁰ Sheehan syndrome, which occurs rarely, is characterized by lactation failure, amenorrhea, breast atrophy, loss of pubic and auxiliary hair, hypothyroidism, and adrenal cortical insufficiency.¹² PPH is associated with hemodynamic instability symptoms and hypovolemic shock,¹³ and as the patient continues bleeding, it can lead to significant vasodilation and hypotension.^14,15 Severe hemorrhage can cause ischemic injury to the liver, heart, kidney, and brain, and increases the length of hospital stay.^14,15 An increased length of hospital stay is also associated with an increased risk of other complications such as infection and venous thromboembolism, which can lead to increased maternal mortality rates¹⁴ as well as increased care costs.^16,17

Given the impacts of PPH complications and the burden of those conditions on both individuals and health care systems, empirical investigations of factors associated with PPH are imperative, with specific attention to geographic regions and hospital characteristics to determine whether such factors contribute to clinical outcomes of maternal delivery and PPH rates. The identification of regional or hospital-level variation in PPH may help to further inform strategies to lower related maternal morbidity and mortality.

To date, few studies have assessed the relationship of hospital characteristics and geographic region on obstetric practice and pregnancy outcomes; most studies have focused on the causes of PPH. Therefore, this study aims to answer the following research questions: (1) Is there a relationship between hospital characteristics and the risk of PPH? (2) Is there a relationship between the region of hospital location and the risk of PPH? This is the first study that examines the correlation between PPH rates with the region of hospital location and characteristics using Healthcare Cost and Utilization Project's (HCUP) data.

Methods

This is a retrospective cross-sectional study that utilized the 2018 National Inpatient Sample (NIS) from HCUP databases. The NIS is the largest publicly available all-payer inpatient care database in the United States and consists of data for >7 million hospital stays from U.S. community hospitals each year. The NIS is a hospital inpatient-stays database derived from hospitals' billing data gathered by statewide data organizations across the United States and includes clinical and resource-use information from discharge abstracts. The NIS provides an annual national sample of hospital discharge records that can be used to examine the relationships among hospital outcomes, discharge characteristics, and hospital characteristics at the national, regional, and census division levels. For states participating in the HCUP Central Distributor, each State's Inpatient Database provides an annual census of hospital discharge data for that state.¹⁸

Study population

A low-risk sample of women aged ≤19–54 years in the third stage of labor with an index for PPH was selected using NIS data from January 1, 2018 to December 30, 2018. This study excluded cases with previous c-sections, intrapartum hemorrhage, abruption placenta, and placenta previa and high-risk cases in obstetrics due to their high-risk clinical conditions associated with maternal hemorrhage.

Measures/variables

This study used the International Classification of Diseases, Tenth Revision (ICD-10) from the American Academy of Professional Coders (AAPC) medical coding. The code O72 was identified for PPH. Other codes that were selected for further analysis include O0993, O432, O43213, O43223, O43233, O468X3, O458X3, O4693, O61, O63, O70, O71, O8611, O8612, O8619, Z3800, and Z3801.^14,19,20

The main dependent variable of interest for this study was PPH. Independent variables included hospital characteristics and regions of hospital locations. Hospital characteristics included location, ownership, teaching status, bed size, and regions included Northeast, Midwest, South, and West. HCUP data categorized hospital location/teaching status, coded as (1) rural, (2) urban nonteaching, and (3) urban teaching. The HCUP data element for hospital ownership was coded as (1) government, nonfederal (public); (2) private, not-for-profit (voluntary); and (3) private, investor-owned (proprietary). The variable bed size was coded as (1) small, (2) medium, and (3) large, depending on the number of hospital beds specific to the location, region, and teaching status of the hospital. The variable region had the following attributes (1) Northeast, (2) Midwest, (3) South, and (4) West.²¹

Control variables included maternal age, race/ethnicity, and variables that were predicted to influence the increased risks of PPH, such as failed induction of labor, prolonged labor, puerperal infections, other obstetric trauma, and perineal laceration during delivery.²² Maternal age was categorized as ≤19, 20–34, and 35–54 years of age. Maternal race and ethnicity were defined based on HCUP coding. The variable race and ethnicity in HCUP coding were set as one data element (HCUP, 2018). In HCUP coding race variable was defined as (1) White, (2) Black, (3) Hispanic, (4) Asian/Pacific Islander, (5) Native American, and (6) Other.²¹

Statistical analysis

A multivariable logistic regression model was performed to estimate the adjusted odds ratio and confidence interval (95% CI) of PPH for each type of hospital characteristic—location, ownership, bed size, teaching status, and region—while controlling for age, race/ethnicity, failed induction of labor, other obstetric trauma, perineal lacerations (third and fourth degree), and puerperal infections.^19,23

This study was reviewed by the Georgia Southern University institutional review board (IRB) and was exempted from full board review as it uses de-identified secondary data. All analyses were performed with STATA software version 16.1.

Results

A total of 2,978,244 single live-birth hospitalizations with low-risk pregnancies and ages ≤19–54 years from 3,517,119 were selected. Approximately 4% of the sample was identified as having had PPH. Of the sample population, 53.23% were White, 14.27% Black, 20.73% Hispanic, 6.31% Asian/Pacific Islander, 0.72% Native American, and 4.74% Other. In PPH subpopulation, the majority of women were aged 20–34 years (94.02%), with 5.59% aged ≤19 and 0.39% aged 35–54 (Table 1).

Table 1.

Comparison of Race and Age in the Sample population Versus Postpartum Hemorrhage Subpopulation

Variables	Sample population postpartum hemorrhage
Variables	%	%
Race
White	53.23	47.15
Black	14.27	15.12
Hispanic	20.73	23.74
Asian/Pacific Islander	6.31	8.08
Native American	0.72	1.19
Other	4.74	4.72
Age (years)
≤19	27.29	5.59
20–34	56.91	94.02
35–54	15.79	0.39

Multivariable logistic regressions for PPH

Multiple logistic regression was performed to estimate adjusted odds ratios for PPH risk with characteristics and regions of hospital locations while holding the race, age, failed induction of labor, other obstetric trauma, perineal laceration during delivery, and puerperal infection constant (Table 2).

Table 2.

Logistic Regression of Postpartum Hemorrhage with Hospital Characteristics and Regions

Variables	Postpartum hemorrhage
Variables	AOR and CI	p
Region Reference (Northeast)
Midwest	1.13 (1.01–1.28)	0.025
South	0.80 (0.71–0.89)	<0.001
West	1.20 (1.06–1.35)	0.003
Location Reference (rural)
Urban nonteaching	0.93 (0.83–1.001)	0.054
Urban teaching	1.14 (1.05–1.24)	0.001
Ownership Reference (government)
Private, not-for-profit	0.85 (0.74–0.98)	0.023
Private, investor owned	0.57 (0.49–0.67)	<0.001
Bed size Reference (small)
Medium	0.97 (0.89–1.06)	0. 596
Large	1.11 (1.02–1.21)	0.011
Race Reference (White)
Black	1.15 (1.09–1.22)	<0.001
Hispanic	1.26 (1.18–1.35)	<0.001
Asian/Pacific Islander	1.30 (1.20–1.41)	<0.001
Native American	1.73 (1.50–1.9)	<0.001
Other	1.18 (1.08–1.28)	<0.001
Age	0.88 (0.83–0.93)	<0.001
Perineal laceration	2.04 (1.88–2.22)	<0.001
Other obstetric trauma	2.63 (2.46–2.81)	<0.001
Puerperal infections	5.01 (4.41–5.68)	<0.001
Failed induction of labor	1.97 (1.77–2.19)	<0.001

p Value set at p < 0.01 statistically significant and p < 0.001 highly statistically significant.

AOR, adjusted odds ratios; CI, confidence interval.

Results for regions showed that compared with Northeast hospitals, women who delivered at the South region hospitals had lower adjusted odds of PPH (AOR = 0.80; CI = 0.71–0.89), whereas women who delivered in the West and Midwest regions' hospitals had higher adjusted odds of PPH ([AOR = 1.20; CI = 1.06–1.35] and [AOR = 1.13; CI = 1.01–1.28] respectively), and those differences for West region hospitals were statistically significant. Women who delivered at hospitals owned by private investors and at private not-for-profit hospitals had lower adjusted odds of PPH ([AOR = 0.57; CI = 0.49–0.67] and [AOR = 0.85; CI = 0.74–0.98] respectively) compared with those who delivered in government-owned hospitals.

Women who delivered at urban teaching hospitals had higher adjusted odds of PPH (AOR = 1.14; CI = 1.05–1.24) compared with those who delivered at rural hospitals. Women who delivered at larger hospitals had higher adjusted odds of PPH (AOR = 1.11; CI = 1.02–1.21) compared with those who delivered at small-sized hospitals.

Results displayed significant associations for the variables race, age, and other control variables with PPH (p < 0.001). Compared with White women, all other race groups had increased adjusted odds of PPH, but the adjusted odds were the highest for the Native Americans compared with other races (AOR = 1.73; CI = 1.50–1.9). However, age was associated with decreased adjusted odds of PPH, and showed each one-year increase in age decreased adjusted odds of PPH risk by 12% (AOR = 0.88; CI = 0.83–0.93).

Clinical risk factors, including failed induction of labor, other obstetric trauma, perineal laceration during delivery, and puerperal infections, increased adjusted odds of PPH by 1.97 (CI = 1.77–2.19), 2.63 (CI = 2.46–2.81), 2.04, (CI = 1.88–2.22), 5.01 (CI = 4.41–5.68), respectively. The odds were the highest for puerperal infections (AOR = 5.01). There was no association between PPH risk and prolonged labor.

Discussion

This study investigated the association of PPH with hospital characteristics and the region of location across the United States, using the 2018 NIS from the HCUP database. In general, significant associations of hospital characteristics were observed with PPH risk. Increased risk of PPH was observed in urban teaching hospitals when compared with rural hospitals, which was consistent with the previous study by Gyamfi-Bannerman et al,⁸ but in contrast with studies conducted by Snowden et al²⁴ and Kozhimannil et al,²³ in which researchers found higher odds of PPH in rural low-volume hospitals compared with low-volume urban teaching hospitals.

Furthermore, results showed that hospitals located in the West region had the highest magnitude of increased odds of PPH, whereas hospitals owned by private investors had the highest magnitude of decreased odds of PPH. This study also found the increased risk of PPH for all other races when compared with White, but the risk was particularly higher for Native Americans and Asian/Pacific Islanders. These findings were consistent with previous studies conducted by Harvey et al (2017)²⁵ and Kozhimannil et al,²³ but in contrast with findings by Gyamfi-Bannerman et al⁸ and Cabacungan et al.⁹ Gyamfi-Bannerman et al⁸ found an increased risk of PPH for African Americans, and Cabacungan et al⁹ found an increased risk of PPH for African Americans and Native Americans, compared with White women.

Although this study found significant associations for the risk of PPH across hospital characteristics and regions, the exact reasons for such associations need more investigation. For instance, it was noteworthy that the South region hospitals, compared with other regional hospitals, had the lowest adjusted odds of PPH. Such variations might be explained by the characteristics of patients seen, and to some extent, might reflect differences in the quality of obstetric practice and management by region.²⁶

Furthermore, the reason for significant associations in decreased odds of risk of PPH in hospitals owned by private investors might be explained by the availability of resources and the fact that the number of care providers in such hospitals is higher than in hospitals owned by the government and private-not-for profit.²⁷ In addition, government and not-for-profit hospitals rely on government allocations and funding and may sometimes encounter financial constraints and obstacles in funding, especially when tax revenue declines or other competing interests demand public support, and that creates barriers in funding to improve the provision of human and financial resources.²⁸

Also, this study found that delivery in urban teaching hospitals increased adjusted odds of PPH compared with other types of hospital characteristics. Further study is needed to investigate whether scheduled cohort turnover contributed to our findings of higher PPH incidence in urban teaching hospitals. Cohort turnover happens when experienced employees depart. The arrival of new workers can disrupt team function and quality of care delivery. In the case of teaching hospitals, it can be expected that the departure of experienced residents and fellows, and the arrival of new residents who are not productive at the beginning of their tenure, may impact their productivity and functions regarding the quality-of-care services and patient health outcomes.²⁹

Finally, this study observed race as a significant factor in increasing the risk of PPH. Racial disparities in the clinical quality and experience may potentially play a significant role; however, the vast literature suggested that racial disparities in maternal outcomes, including hemorrhage, may be due to unconscious bias and/or social determinants of health.^10,30

Limitations

This study has several limitations. First, only the NIS 2018 database was analyzed; therefore, caution should be considered in generalizing the study results. Second, the sample included low-risk delivery cases and deliveries performed in the hospitals because the study focus was on obstetric practice and management across the available U.S. hospitals within the data set. Other potential limitations of the study regarding NIS data include differences in diagnosis, coding, missing data, and mismatched data, which lead to error and the misclassification of information.

Conclusion

Although early studies evidenced significant regional variations in health care delivery and utilization, there is limited evidence on whether maternity outcomes are impacted by the geographic region of hospital locations or hospital characteristics where delivery takes place.^31
–33 This study found an association between hospital geographic location and characteristics and PPH. However, it should be emphasized that regional and hospital characteristic differences could be influenced partly by variations in resources, obstetric practices, and the processes of operative-care management, as well as factors that may not be identified by the current database such as regional variations in data collection and reporting, geographic differences, and patient characteristics.

Therefore, to improve future management and evaluation of operative processes regarding PPH, additional revision on data collection and the addition of variables related to care management and obstetric protocols are suggested to elucidate such discrepancies in variations of regional and hospital characteristics.

To conclude, the highest and lowest risks of PPH with hospitals located in the West and South regions, teaching hospitals, and hospitals owned by private investors need further investigation regarding resources, obstetric practice, hospital management, and population or sociodemographic characteristics. It is essential to identify race as a potential risk factor for the increased risk of PPH.

Therefore, to develop better policies and decision-making, further research is needed to provide evidence regarding associations between such factors and increased risk of PPH, and whether they are influenced by population or sociodemographic characteristics or related to variations in obstetric practice and management of regional and hospital types.³⁴

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

Reale

, Easter

, Xu

, et al. Trends in postpartum hemorrhage in the United States from 2010–2014. Obstet Anesth Dig, 2020; 40(1):12; doi: 10.1097/01.aoa.0000652796.48760.fd

Sheikh

, Najmi

, Khalid

, et al. Evaluation of compliance and outcomes of a management protocol for massive postpartum hemorrhage at a tertiary care hospital in Pakistan. BMC Pregnancy Childbirth, 2011; 11:28; doi: 10.1186/1471-2393-11-28

Carvalho

, Hoque

, Oliver

, et al. Cost-effectiveness of inhaled oxytocin for prevention of postpartum haemorrhage: A modelling study applied to two high burden settings. BMC Med, 2020; 18(1):201. doi: 10.1186/s12916-020-01658-y

Ford

, Patterson

, Seeho

SKM

, et al. Trends and outcomes of postpartum haemorrhage, 2003–2011. BMC Pregnancy Childbirth, 2015; 15(336):334. doi: 10.1186/s12884-015-0788-5

Kramer

, Berg

, Abenhaim

, et al. Incidence, risk factors, and temporal trends in severe postpartum hemorrhage. Am J Obstet Gynecol, 2013; 209(5):449.e1–e7; doi: 10.1016/j.ajog.2013.07.007

Edhi

, Aslam

, Naqvi

, et al. “Post partum hemorrhage: Causes and management.” BMC Res Notes, 2013; 6(1):1–6; doi: 10.1186/1756-0500-6-236

Blomberg

. Maternal obesity and risk of postpartum hemorrhage. Obstet Gynecol, 2011; 118(3):561–568; doi: 10.1097/AOG.0b013e31822a6c59

Gyamfi-Bannerman

, Srinivas

, Wright

, et al. Postpartum hemorrhage outcomes and race. Am J Obstet Gynecol, 2018; 219(2):185.e1–e185.e10; doi: 10.1016/j.ajog.2018.04.052

Cabacungan

, Ngui

, & Mcginley E. Racial/ethnic disparities in maternal morbidities: A statewide study of labor and delivery hospitalizations in Wisconsin. Matern Child Health J, 2012; 16(7):1455.

10.

Green

, Zapata

, Brown

, et al. Rethinking bias to achieve maternal health equity: Changing organizations, not just individuals. Obstet Gynecol, 2021; 137(5):935–940; doi: 10.1097/AOG.0000000000004363

11.

Sebghati

, Chandraharan

. An update on the risk factors for and management of obstetric haemorrhage. Women's Health, 2017; 13(2):34–40; doi: 10.1177/1745505717716860

12.

Howell

. Reducing disparities in severe maternal morbidity and mortality. Clin Obstet Gynecol, 2018; 61(2):387–399; doi: 10.1097/GRF.0000000000000349

13.

Wormer

, Jamil

, Bryant

. Acute postpartum hemorrhage. StatPearls; 2020. Available from: https://www.ncbi.nlm.nih.gov/books/NBK499988 [Last accessed: May 8, 2022].

14.

Marshall

, Durani

, Bartley

, et al. The impact of postpartum hemorrhage on hospital length of stay and inpatient mortality: A national inpatient sample-based analysis. Am J Obstet Gynecol, 2017; 217(3):344.e1; doi: 10.1016/j.ajog.2017.05.004

15.

Nyfløt

, Stray-Pedersen

, Forsén

, et al. Duration of labor and the risk of severe postpartum hemorrhage: A case-control study. PLoS One, 2017; 12(4):1–10; doi: 10.1371/journal.pone.0175306

16.

Law

, McCoy

, Lynen

, et al. Costs of newborn care following complications during pregnancy and delivery. Matern Child Health J, 2015; 19(9):2081–2088; doi: 10.1007/s10995-015-1721-2

17.

Vesco

, Ferrante

, Chen

, et al. Costs of severe maternal morbidity during pregnancy in US commercially insured and medicaid populations: An observational study. Matern Child Health J, 2020; 24(1):30–38; doi: 10.1007/s10995-019-02819-z

18.

Healthcare Cost and Utilization Project. National inpatient sample (NIS). Agency for Healthcare Research and Quality; 2012. Available from: www.hcup-us.ahrq.gov/nisoverview.jsp [Last accessed: March 7, 2022].

19.

, Fridman

, Korst

, et al. Variations in the incidence of postpartum hemorrhage across hospitals in California. Matern Child Health J, 2005; 9(3):297–306; doi: 10.1007/s10995-005-0009-3

20.

Pregnancy, childbirth and the puerperium ICD-10-CM code range O00-O9A. https://www.aapc.com/codes/icd-10-codes-range/O00-O9A

21.

Garba

, Abdullahi

, Yusuf

, et al. Appropriate documentation of the timing of events in the management of women with postpartum hemorrhage in Aminu Kano Teaching Hospital: A 2-year audit. Niger Med J, 2019; 60(1):9–12; doi: 10.4103/nmj.NMJ_5_19

22.

Dionne

, Deneux-Tharaux

, Dupont

, et al. Duration of expulsive efforts and risk of postpartum hemorrhage in nulliparous women: A population-based study. PLoS One, 2015; 10(11):e0142171; doi: 10.1371/journal.pone.0142171

23.

Kozhimannil

, Thao

, Hung

, et al. Association between hospital birth volume and maternal morbidity among low-risk pregnancies in rural, urban, and teaching hospitals in the United States. Am J Perinatol, 2016; 33(6):590–599; doi: 10.1055/s-0035-1570380

24.

Snowden

, Cheng

, Emeis

, et al. (2015). The impact of hospital obstetric volume on maternal outcomes in term, non-low-birthweight pregnancies. Am J Obstet Gynecol, 2018; 212(3):380.e1–e9; doi: 10.1016/j.ajog.2014.09.026

25.

Harvey

, Lim

, Gandhi

, et al. Racial-ethnic disparities in postpartum hemorrhage in Native Hawaiians, Pacific Islanders, and Asians. Hawai'i Journal of Medicine & Public Health, 2017; 76(5):128–132.

26.

Otolorin

, Gomez

. Currie S, et al. Essential basic and emergency obstetric and newborn care: From education and training to service delivery and quality of care. Int J Gynaecol Obstet, 2015;(Suppl 2):S46–S53; doi: 10.1016/j.ijgo.2015.03.007

27.

Cram

, Bayman

, Popescu

, et al. Uncompensated care provided by for-profit, not-for-profit, and government owned hospitals. BMC Health Serv Res, 2010; 10:90; doi: 10.1186/1472-6963-10-90

28.

Thomas

, Orav

, Brennan

. Hospital ownership and preventable adverse events. J Gen Internal Med, 2000; 15(4):211–219; doi: 10.1111/j.1525-1497.2000.07003.x

29.

Huckman

, Barro

. Cohort turnover and productivity: The July phenomenon in teaching hospitals. National Bureau of Economic Research Working Paper. 2005; No. 11182, DOIN 10.3386/w11182

30.

Driessen

, Bouvier-Colle

, Dupont

, et al. Postpartum hemorrhage resulting from uterine atony after vaginal delivery: Factors associated with severity. Obstet Gynecol, 2011; 117(1):21–31; doi: 10.1097/AOG.0b013e318202c845

31.

Chassin

, Brook

, Park

, et al. Variations in the use of medical and surgical procedures by the Medicare population. N Engl J Med, 1986; 314:285–289.

32.

Quin

, Sheng

, O'Brien

, et al. Regional variation in patient risk factors and mortality after coronary artery bypass grafting. Ann Thorac Surg, 2011; 92(4):1277–1282l; doi: 10.1016/j.athoracsur.2011.05.062

33.

Leap

, Pare

, Solomon

, et al. Does inappropriate use explain small-area variations in the use of health care services?. JAMA, 1990; 263(5):669–672.

34.

Bryant

, Mhyre

, Leffert

, et al. The association of maternal race and ethnicity and the risk of postpartum hemorrhage. Anesth Analg, 2012; 115(5):1127–1136; doi: 10.1213/ANE.0b013e3182691e62