Keywords
stillbirth - health disparities - social determinants of health - ethnic disparities
Intrauterine fetal demise or stillbirth, is defined in the United States as fetal
death at ≥20 weeks' gestation, and complicates approximately one in 160 pregnancies
or approximately 23,000 stillbirths annually.[1]
[2] While stillbirth is significantly more common in low to middle-income countries,
the stillbirth rate in the United States remains higher than many other developed
nations.[3] Improvements in prenatal care led to a sharp decline in stillbirth rates during
the mid-20th century, but have since remained stagnant.[2]
[4] Numerous maternal and fetal factors have been previously identified as contributing
to stillbirth.[1]
[5]
[6] Recent studies have shed light on the association between social determinants of
health (SDoH) and stillbirth. It has been shown that certain SDoH (income, education,
health care access, and air pollution) are associated with higher rates of stillbirth.[7]
[8]
[9]
[10]
[11] SDoH are conditions in the environments in which people are born, live, learn, work,
play, worship, and age that affect a wide range of health, functioning, and quality-of-life
outcomes and risks.[12] Five key areas include: economic stability, education access and quality, social
and community context, health care access and quality and neighborhood, and built
environment.[12] Among nonmodifiable risk factors for stillbirth, non-Hispanic (NH) Black race remains
one of the most prevalent and persistent.[13]
[14] Among NH Black women in the United States, the rate of stillbirth is more than twice
that of other racial groups.[1]
[2]
, Reasons for this disparity have been attributed to clustering of maternal and fetal
risk factors in NH Black women, but when these factors are not present, the disparity
in stillbirth rates remains.[5]
[15]
[16] Understanding and improving these conditions could enhance health outcomes and reduce
racial and ethnic disparities in stillbirth.[17]
There is a paucity of literature examining the association of SDoH with stillbirth
in the United States. Additionally prior studies have been limited by sample size
and types of SDoH factors evaluated.[11]
[15] Available evidence does show a significant association between factors such as gender
equality (as measured by contraceptive access and maternal education) and poverty
with stillbirth, as well as improvement in stillbirth rates, observed in racially
integrated communities.[10]
[18]
[19]
[20] Nonetheless, more research in this area is greatly needed. The aim of this study
was to evaluate the impact of SDoH risk factors on stillbirth among pregnancy-related
hospitalizations in a large U.S. population.
Materials and Methods
We conducted a cross-sectional analysis of delivery-related hospital discharges using
the 2016 to 2017 annual data from the Healthcare Cost and Utilization Project's (HCUP)
Nationwide Inpatient Sample (NIS). The NIS is a large, all-payer, publicly available
inpatient database in the United States (
https://www.hcup-us.ahrq.gov/db/nation/nis/nisdbdocumentation.jsp
). The NIS dataset is created using a sampling technique that stratifies all nonfederal
community hospitals from participating states into groups based on five major hospital
characteristics as follows: (1) rural/urban location, (2) number of beds, (3) geographic
region, (4) teaching status, and (5) ownership. HCUP performs quality control of diagnostic
and procedural codes to minimize coding errors within the database and maintain a
valid and consistent dataset (
https://www.hcup-us.ahrq.gov/db/quality.jsp#editcheck2015
). A 20% sample of hospitals was randomly drawn from each stratum, and all inpatient
discharges from the selected hospitals were included. The NIS database comprises hospital
stratum identifiers and discharge-level sampling weights that can be used to generate
national estimates.
Each hospital discharge record contains International Classification of Diseases,
10th Revision, Clinical Modification (ICD-10-CM) codes for a patient's principal diagnosis
and up to 29 secondary diagnoses. Delivery-related admissions were classified as maternal
and/or neonatal based on the presence of ICD-10-CM diagnosis codes ([Table 1]). Z-codes were utilized to characterize SDoH risk factors and their subtypes which
were the exposure variables in this research. The outcome of interest for this study
was characterized using ICD-10-CM to select women with singleton stillbirth ([Table 1]).
Table 1
ICD-10 codes for social determinants of health
|
Details
|
ICD-10 codes
|
|
Maternal conditions and complications
|
A34, C58, D392, F53, M830, Ox, Z037x, Z3201, Z34x, Z35x, Z36x, Z37x, Z39x, Z3Ax
|
|
Intrauterine fetal demise
|
O364x, Z8759x, P95, Z371
|
|
Any SDoH issue
|
Z55x, Z56x, Z57x, Z59x, Z60x, Z62x, Z63x, Z64x, Z65
|
|
Problems related to education and literacy
|
Z55x
|
|
Problems related to employment and unemployment
|
Z56x
|
|
Occupational exposure to risk factors
|
Z57x
|
|
Problems related to housing and economic circumstances
|
Z59x
|
|
Problems related to social environment
|
Z60x
|
|
Problems related to upbringing
|
Z62x
|
|
Other problems related to primary support group, including family circumstances
|
Z63x
|
|
Problems related to psychosocial circumstances
|
Z64x, Z65x
|
Abbreviations: ICD-10, International Classification of Disease, 10th Revision; SDoH,
social determinants of health.
To explore differences in baseline characteristics, we compared the distribution of
selected sociodemographic and hospital characteristics of delivery-related discharges
among those who experienced stillbirth to those who did not. We grouped maternal age
in years into three categories as follows: (1) 15 to 24 years, (2) 25 to 34 years,
and (3) 35 to 49 years. Race/ethnicity classification was first determined by ethnicity
(Hispanic or NH); with the NH group further subdivided by race (White, Black, or other).
Due to small numbers, the category of other is a combination of the following groups:
Asian or Pacific Islander, Native American, mixed race, and other. Median household
income quartile was estimated by HCUP using the zip code of residence on the mother's
delivery record and zip code–demographic data. Primary payers for each hospital stay
were classified into Medicare, Medicaid, private, self-pay and others. Hospital characteristics
were assessed based on their teaching status (teaching vs. nonteaching), location
(urban vs. rural), and U.S. geographic region (Northeast, Midwest, South, or West).
Hospital bed size was categorized as small, medium, large, or missing using U.S. regions,
the urban–rural designation of the hospital, and teaching status. Rural hospitals
were not split based on teaching status for bed size categorization purpose, because
rural teaching hospitals were rare.
All statistical analyses were performed using R version 3.5.1 (University of Auckland,
Auckland, New Zealand), R Studio Version 1.1.4.2.3 (Boston, MA). We performed descriptive
analyses to evaluate the relationship between patient characteristics and stillbirth.
Bivariate analyses using Pearson's Chi-square test were conducted to determine the
association between each of the SDoH subtypes and the frequency of stillbirth. Next,
we calculated the rates of stillbirth among mothers who experienced SDoH risk factors,
stratified by race/ethnicity. Using adjusted survey logistic regression model, we
generated odds ratios to quantify the association between SDoH risk factors and stillbirth
after adjusting for patient and hospital characteristics. Lastly, the risk of stillbirth
was calculated for each SDoH subtype after adjusting for hospitalization characteristics.
Since the primary outcome of the study, stillbirth, is a rare event, the adjusted
odd ratios approximate the relative risks.
All hypothesis tests were two-tailed with type-I error set at 5%. Since the study
was performed using deidentified publicly available data, the institutional review
board at Baylor College of Medicine classified this study as exempt.
Results
Our study included a total of 8,148,646 hospitalizations of which 91,140 were related
to stillbirth hospitalizations yielding a stillbirth incidence of 1.1%.The total number
of stillbirths and incidence stratified by patient characteristics are presented in
[Table 2].
Table 2
Patient characteristics by stillbirth status
|
Stillbirth
|
|
No
n (%)
|
Yes
n (%)
|
Incidence (%)
|
|
N = 8,057,506
|
% = 100
|
N = 91,140
|
% = 100
|
|
Age (y)
|
|
15–24
|
2,080,752
|
25.8
|
21,235
|
23.3
|
1.0
|
|
25–34
|
4,587,105
|
56.9
|
50,075
|
54.9
|
1.1
|
|
35–49
|
1,389,649
|
17.2
|
19,830
|
21.8
|
1.4
|
|
Race/ethnicity
|
|
NH White
|
3,968,510
|
49.3
|
40,215
|
44.1
|
1.0
|
|
NH Black
|
1,205,270
|
15.0
|
20,515
|
22.5
|
1.7
|
|
Hispanic
|
1,580,833
|
19.6
|
15,580
|
17.1
|
1.0
|
|
NH other
|
881,794
|
10.9
|
8,265
|
9.1
|
0.9
|
|
Missing
|
421,099
|
5.2
|
6,565
|
7.2
|
1.5
|
|
Discharge status
|
|
Routine
|
7,876,411
|
97.8
|
89,225
|
97.9
|
1.1
|
|
Transfer
|
42,425
|
0.5
|
650
|
0.7
|
1.5
|
|
Died
|
805
|
0.0
|
85
|
0.1
|
9.6
|
|
Discharged against medical advice
|
25,760
|
0.3
|
490
|
0.5
|
1.9
|
|
Other
|
108,160
|
1.3
|
675
|
0.7
|
0.6
|
|
Missing
|
3,945
|
0.0
|
15
|
0.0
|
0.4
|
|
Zip income quartile
|
|
Lowest quartile
|
2,311,128
|
28.7
|
29,495
|
32.4
|
1.3
|
|
2nd quartile
|
2,002,508
|
24.9
|
23,490
|
25.8
|
1.2
|
|
3rd quartile
|
1,950,823
|
24.2
|
20,815
|
22.8
|
1.1
|
|
Highest quartile
|
1,713,318
|
21.3
|
16,420
|
18.0
|
0.9
|
|
Missing
|
79,730
|
1.0
|
920
|
1.0
|
1.1
|
|
Primary payer
|
|
Medicare
|
67,450
|
0.8
|
1,025
|
1.1
|
1.5
|
|
Medicaid
|
3,528,246
|
43.8
|
42,440
|
46.6
|
1.2
|
|
Private insurance
|
4,018,131
|
49.9
|
41,345
|
45.4
|
1.0
|
|
Self-pay
|
433,289
|
5.4
|
6,130
|
6.7
|
1.4
|
|
Missing
|
10,390
|
0.1
|
200
|
0.2
|
1.9
|
|
Hospital characteristics
|
|
Hospital region
|
|
Northeast
|
1,292,875
|
16.0
|
13,230
|
14.5
|
1.0
|
|
Midwest
|
1,688,008
|
20.9
|
20,160
|
22.1
|
1.2
|
|
South
|
3,139,333
|
39.0
|
38,655
|
42.4
|
1.2
|
|
West
|
1,937,291
|
24.0
|
19,095
|
21.0
|
1.0
|
|
Hospital bed size
|
|
Small
|
1,439,653
|
17.9
|
14,875
|
16.3
|
1.0
|
|
Medium
|
2,470,709
|
30.7
|
25,840
|
28.4
|
1.0
|
|
Large
|
4,147,144
|
51.5
|
50,425
|
55.3
|
1.2
|
|
Hospital location and teaching status
|
|
Rural
|
735,562
|
9.1
|
8,355
|
9.2
|
1.1
|
|
Urban nonteaching
|
1,870,371
|
23.2
|
18,410
|
20.2
|
1.0
|
|
Urban teaching
|
5,451,573
|
67.7
|
64,375
|
70.6
|
1.2
|
Note: Definitions of variable can be found at the following address:
https://www.hcupus.ahrq.gov/db/nation/nis/nisdbdocumentation.jsp
.
Results of our analysis for stillbirth incidence by SDoH risk factors are displayed
on [Table 3]. The incidence of stillbirth among cases with any SDoH risk factor was 2.0% compared
with 1.1% in cases without SDoH risk factors (p < 0.001). Among specific SDoH risk factors, the highest incidence estimates were
related to problems associated with education and literacy (2.9%), occupational risk
(16.7%), housing and economic circumstances (1.7%), upbringing (2.1%), and primary
support group, including family circumstances (6.0%).
Table 3
Prevalence of social determinants of health[a]
[b]
|
Stillbirth
|
p-Value
|
|
No
n (%)
|
Yes
n (%)
|
Incidence
|
|
N = 8,057,506
|
C% = 100
|
N = 91,140
|
C% = 100
|
|
|
Any SDoH issue
|
|
No
|
8,028,651
|
99.6
|
90,550
|
99.4
|
1.1
|
<0.001
|
|
Yes
|
28,855
|
0.4
|
590
|
0.6
|
2.0
|
|
Problems related to education and literacy
|
|
No
|
8,057,341
|
100.0
|
91,135
|
100.0
|
1.1
|
0.02
|
|
Yes
|
165
|
0.0
|
“–”
|
0.0
|
2.9
|
|
Problems related to employment and unemployment
|
|
No
|
8,055,501
|
100.0
|
91,110
|
100.0
|
1.1
|
0.127
|
|
Yes
|
2,005
|
0.0
|
30
|
0.0
|
1.5
|
|
Occupational exposure to risk factors
|
|
No
|
8,057,481
|
100.0
|
91,135
|
100.0
|
1.1
|
<0.001
|
|
Yes
|
25
|
0.0
|
“–”
|
0.0
|
16.7
|
|
Problems related to housing and economic circumstances
|
|
No
|
8,046,391
|
99.9
|
90,945
|
99.8
|
1.1
|
<0.001
|
|
Yes
|
11,115
|
0.1
|
195
|
0.2
|
1.7
|
|
Problems related to social environment
|
|
No
|
8,055,821
|
100.0
|
91,125
|
100.0
|
1.1
|
0.335
|
|
Yes
|
1,685
|
0.0
|
15
|
0.0
|
0.9
|
|
Problems related to upbringing
|
|
No
|
8,051,306
|
99.9
|
91,010
|
99.9
|
1.1
|
<0.001
|
|
Yes
|
6,200
|
0.1
|
130
|
0.1
|
2.1
|
|
Other problems related to primary support group, including family circumstances
|
|
No
|
8,054,601
|
100.0
|
90,955
|
99.8
|
1.1
|
<0.001
|
|
Yes
|
2,905
|
0.0
|
185
|
0.2
|
6.0
|
|
Problems related to psychosocial circumstances
|
|
No
|
8,050,666
|
99.9
|
91,070
|
99.9
|
1.1
|
0.404
|
|
Yes
|
6,840
|
0.1
|
70
|
0.1
|
1.0
|
Abbreviations: SDoH, social determinants of health.
a Based on ICD-10-CM/PCS: International Classification of Diseases, 10th Revision,
Clinical Modification/Procedure Coding System.
b Counts containing a value of 10 or less were suppressed to ensure confidentiality
of patients, these values are noted as “–.”
Of the total study population, 29,445 (1%) of patients had at least one SDoH risk
factor. Racial stratification of patients with SDoH risk factors was as follows: NH
White = 12,535 (42.6%), NH Black = 7,330 (24.9%), Hispanic = 5,635 (19.1%), NH other = 2,330
(7.9%), and 1,615 (5.5%) had missing race/ethnicity data. Among patients with SDoH
risk factor, the highest rates of stillbirth were observed in those identifying as
NH other (3.0%) or Hispanic (2.1%) ([Fig. 1]).
Fig. 1 Rates of stillbirth by race/ethnicity among pregnant women with SDOH risk factors.
NH, non-Hispanic; SDoH, social determinants of health.
The results of the evaluation for the association between SDoH risk factor and patient
characteristics and risk of stillbirth are summarized in [Table 4]. Mothers that presented with SDoH risk factors had an approximately 60% greater
risk of stillbirth compared with those without (odds ratio [OR] = 1.61 [95% confidence
interval (CI) = 1.33–1.95], p < 0.001). Our results showed notable racial/ethnic disparities after adjustment for
other characteristics as outlined in [Table 4]. Compared with NH Whites, NH Blacks experienced more than 50% greater risk for stillbirth
(OR = 1.54 [95% CI = 1.46–1.62], p < 0.001), while Hispanic mothers exhibited a slightly lower likelihood for stillbirth
(OR = 0.94 [95% CI = 0.88–0.99], p = 0.03).
Table 4
Independent association between SDoH risk factors and patient characteristics (exposure)
and the likelihood of stillbirth (outcome)
|
OR (95% CI)
|
p-Value
|
|
Any SDoH issues
|
|
No
|
Reference
|
<0.001
|
|
Yes
|
1.61 (1.33–1.95)
|
|
Race/ethnicity
|
|
NH White
|
Reference
|
|
|
NH Black
|
1.54 (1.46–1.62)
|
<0.001
|
|
Hispanic
|
0.94 (0.88–0.99)
|
0.03
|
|
NH other
|
0.92 (0.86–0.97)
|
<0.001
|
|
Age (y)
|
|
< 24
|
Reference
|
|
|
25–34
|
1.17 (1.13–1.22)
|
<0.001
|
|
35–49
|
1.60 (1.52–1.68)
|
<0.001
|
|
Discharge status
|
|
Routine
|
Reference
|
|
|
Transfer
|
1.30 (1.09–1.55)
|
<0.001
|
|
Died
|
7.61 (4.56–12.68)
|
<0.001
|
|
Discharged against medical advice
|
1.42 (1.15–1.75)
|
<0.001
|
|
Other
|
0.52 (0.43–0.64)
|
<0.001
|
|
Zip income quartile
|
|
Lowest quartile
|
Reference
|
|
|
2nd quartile
|
0.99 (0.94–1.04)
|
0.62
|
|
3rd quartile
|
0.91 (0.86–0.97)
|
<0.001
|
|
Highest quartile
|
0.83 (0.76–0.91)
|
<0.001
|
|
Primary payer
|
|
Private Insurance
|
Reference
|
|
|
Medicare
|
1.24 (1.07–1.43)
|
<0.001
|
|
Medicaid
|
1.13 (1.09–1.18)
|
<0.001
|
|
Self-pay
|
1.37 (1.27–1.47)
|
<0.001
|
|
Hospital characteristics
|
|
Hospital region
|
|
|
|
Northeast
|
Reference
|
|
|
Midwest
|
1.11 (0.98–1.26)
|
0.09
|
|
South
|
1.14 (1.03–1.27)
|
0.01
|
|
West
|
0.98 (0.88–1.09)
|
0.69
|
|
Hospital bed size
|
|
Small
|
Reference
|
|
|
Medium
|
1.02 (0.94–1.11)
|
0.67
|
|
Large
|
1.18 (1.08–1.30)
|
<0.001
|
|
Hospital location and teaching status
|
|
Rural
|
|
Urban nonteaching
|
0.93 (0.85–1.02)
|
0.12
|
|
Urban teaching
|
1.09 (0.99–1.19)
|
0.09
|
Abbreviations: CI, confidence interval; NH, non-Hispanic; OR, odds ratio; SDoH, social
determinants of health.
The age groups of 25–34 (OR = 1.17 [95% CI =1.13–1.22], p <0.001) and 35–49 (OR = 1.60 [95%CI =1.52–1.68], p <0.001) years also showed elevated risk of stillbirth when compared with the age
group <24 years.
The adjusted odds of stillbirth declined progressively with increasing income quartile
in a dose–response pattern reaching a nadir for mothers in the highest income quartile.
Health insurance status and hospital region of hospitalization were also predictive
of risk for stillbirth. Mothers who self-paid for hospital services had the greatest
risk of stillbirth, while those on private insurance experienced the lowest risk.
Compared with health facilities in the Northeast, those located in the South exhibited
a 14% (OR = 1.14 [95% CI = 1.03–1.27], p = 0.01) greater likelihood of stillbirth occurrence, while health facilities in the
Midwest and the West did not show stillbirth risk elevation. The risk of stillbirth
was also significantly increased with hospitalizations in large compared with small
hospitals. However, rural or urban location of the health facility did not influence
the risk of stillbirth.
Certain SDoH risk factors showed a stronger association with stillbirth than others
([Table 5]). The SDoH risk factors that displayed the strongest linkage to stillbirth were
problems related to occupational risks, (OR = 7.05 [95%CI = 3.54–9.58], p < 0.001), issues of upbringing, (OR = 1.87 [95% CI = 1.23–2.82], p < 0.001), and primary support group, including family circumstances (OR = 5.45 [95%
CI = 3.84–7.76], p < 0.001).
Table 5
Association between subtypes of SDoH and stillbirth
|
Group
|
OR
|
Lower CI
|
Upper CI
|
p-Value
|
|
Any SDoH issue
|
1.60
|
1.37
|
2.00
|
<0.001
|
|
Problems related to education and literacy
|
2.43
|
0.31
|
4.62
|
0.4
|
|
Problems related to employment and unemployment
|
1.21
|
0.53
|
2.76
|
0.64
|
|
Occupational exposure to risk factors
|
7.05
|
3.54
|
9.58
|
<0.001
|
|
Problems related to housing and economic circumstances
|
1.34
|
0.97
|
1.85
|
0.08
|
|
Problems related to social environment
|
0.74
|
0.25
|
2.20
|
0.59
|
|
Problems related to upbringing
|
1.87
|
1.23
|
2.82
|
<0.001
|
|
Other problems related to primary support group, including family circumstances
|
5.46
|
3.84
|
7.76
|
<0.001
|
|
Problems related to psychosocial circumstances
|
0.84
|
0.48
|
1.45
|
0.52
|
Abbreviations: CI, confidence interval; OR, odds ratio; SDoH, social determinants
of health.
Discussion
Stillbirth is one of the most common adverse pregnancy outcomes and has a profound
impact on women and their families.[21]
[22] Recent studies have shed light on the association between SDoH and stillbirth.[7]
[8]
[9]
[10]
[11]
[12] Although the rate of stillbirth in the United States has been stable, significant
racial/ethnic disparities persist.[2]
[23] Understanding the causes of these disparities can provide insight into opportunities
to reduce stillbirth rates. Data presented here supports these prior findings and
demonstrates the incidence of stillbirth among cases with any SDoH risk factor to
be greater than in those without SDoH risk factors (2.0 vs. 1.1%; p <0.001). SDoH issues are associated with a 60% increased risk of stillbirth and 1%
of our study population had at least one risk factor present. The rate of stillbirth
when SDoH issues were present was highest among those identified as NH other (3.0%)
or Hispanic (2.1%). The reasons for this are not clear but may be impacted by factors
such as access and quality of health care or associated medical comorbidities in these
populations.[2]
[9]
[24]
[25] Additionally factors such as the impact of immigration status cannot be assessed
using these generalized categories. For example, it has been shown that more favorable
birth outcomes are typically observed among Hispanic immigrants, but this same advantage
is not consistently seen in U.S. born Hispanic individuals.[9]
When examining the impact of individual SDoH risk factors, most striking was a seven-fold
increased risk of stillbirth when occupational exposures were present, a nearly two-fold
increased risk when problems with upbringing were present, and a more than five-fold
increased stillbirth risk associated with problems related to primary support groups.
Additionally, there is likely an additive effect with the presence of more than one
SDoH factor, which is supportive of prior evidence suggesting that multiple issues
relating to interpersonal or environmental stressors increase risk of stillbirth in
a dose-dependent manner.[26]
Additionally, results presented here support prior studies showing the association
between stillbirth and environmental factors in both geography and facility type where
care was received. We demonstrate a 14% greater likelihood of stillbirth in health
care facilities in the South, consistent with prior data showing a disproportionately
higher incidence of stillbirth compared with other regions.[10]
[27] Proposed explanations of this increased incidence of stillbirth in the southern
United States include historical exposure to racial segregation, higher rates of poverty,
and higher ambient temperatures by which heat exposure and dehydration result in placental
compromise, all of which have been previously shown to be associated with increased
rates of stillbirth.[10]
[12]
[19]
[28]
We also show an association between stillbirth and maternal transfers. Transfers of
care are high risk by nature and are more likely to end up at tertiary care center.
Complex maternal medical comorbidities, high-risk obstetric conditions, and complex
fetal disorders require multidisciplinary subspecialty care typically only available
in large referral centers with the resources to care for both mother and fetus. Examples
of such conditions include poorly controlled maternal hypertension or diabetes, pre-existing
maternal cardiac or renal disease, hemolysis, elevated liver enzymes, low platelets
(HELLP syndrome), placenta accreta spectrum disorders, complex fetal anomalies, maternal
trauma, and pregnancies affected by acute respiratory failure secondary to viral infection,
such as seen during the current COVID-19 pandemic. These conditions create higher
acuity due to their associated complications, such as hemorrhage, disseminated intravascular
coagulopathy, stroke, organ failure, and cardiac arrest, any of which carry an associated
higher risk of antepartum or intrapartum fetal death in addition to higher risk of
maternal death.[1]
[8]
NH Black race has consistently been identified as a nonmodifiable risk factor for
stillbirth and it has also been associated with increased risk for stillbirth recurrence
in subsequent pregnancies.[9]
[29]
[30] In our study we have shown that NH Black race is associated with the highest incidence
of stillbirth at 1.7%. Interestingly, unlike the other racial/ethnic groups in our
study which showed a clear increase in stillbirth rates in the presence of SDoH issues,
NH Black patients had similar stillbirth rates with and without the presence of SDoH
issues. After adjusting for SDoH issues NH Black patients had a 54% higher risk of
stillbirth when compared with NH White patients. Reasons for the increased rate of
stillbirth in the NH Black population are likely numerous and compounding. These may
range from disparities occurring at the individual level, such as poverty and access
to care, to population-level inequities including effects of racial segregation relating
to housing and employment.[22] Delays in seeking, initiating, and provision of adequate health care are also associated
with adverse maternal outcomes, and NH Black women are likely to be impacted by all
three types of delay.[25]
[31]
[32] Another study showed that communities with decreasing levels of segregation decreased
the rate of stillbirth among NH Black women, suggesting that reducing structural racism
can improve health outcomes.[10]
[19]
[33] Recognizing the complex interaction of social determinants in the lives of pregnant
women of minority status, and, specifically, NH Black women, may be crucial to an
improved understanding of disparities in pregnancy outcomes.
Strengths and Limitations
Strengths and Limitations
This study has several strengths. This is an adequately powered study, using a large
database representing a diverse patient population from multiple settings across the
United States, which makes our study results generalizable across many populations.
Additionally, the methodology employed here using Z-codes to examine SDoH risk factors
as it relates to stillbirth is novel. Use of Z-codes was introduced with the adoption
of ICD-10 in 2015 and widespread uptake has been slow.[34] Given this, it is likely that the relationship between stillbirth and presence of
SDoH risk factors presented here is an underestimation of their association.
There are some limitations to this study given that it is an observational study and
cannot imply causality. Use of a large national database, while beneficial for generating
a large and diverse sample size, analyses could be unduly influenced by missing or
incomplete data. This is particularly true when examining race data which are often
limited to broad predetermined categories and binary selection of race or ethnicity,
and excludes biracial or multiracial participants and/or forces inaccurate labeling.
Furthermore, since our data only includes information on hospitalized patients and
not those presenting to outpatient settings and emergency rooms, the incidence of
stillbirth in our study is higher than the general population since hospitalized patients
may have other comorbidities or may have been admitted due to other causes such as
accidents, which would have increased their risk of stillbirth. SDoH factors are only
identified using ICD-10 coding and not from patient demographic characteristics such
as patients' educational background, environmental, or occupational factors impacting
their health, etc., and there is a possibility that not all SDoH factors were enlisted
in patients' discharge records. There are also inherent limitations in using ICD-10
coded definitions for SDoH, given that we may not be capturing more nuanced issues,
such as health care access and quality. Use of ICD-10 codes, particularly Z-codes,
is subjected to the biases of the individual performing the coding. In addition, it
has been shown that SDoH Z-codes are utilized more often in larger, private not-for-profit,
and urban teaching hospitals, whereas the utilization in pregnant population is unknown.
This could have resulted in biases due to underreporting and may undermine the actual
burden of social needs experienced by hospitalized pregnant patients.[34] Additionally, these codes are not pregnancy specific. We are also unable to analyze
associated factors that may influence the risk of stillbirth such as parity, fetal,
and obstetric comorbidities; gestational age; and prior history of stillbirth.
Conclusion
To the best of the authors' knowledge, this study is the first population-level analysis
examining specifically SDoH and stillbirth in a developed nation. Prior examinations
in the areas of stillbirth and social determinants either differ in scope of analysis,
are limited to smaller populations, or fail to examine other determinants outside
race or ethnicity. These data highlight the need for further detailed study on the
impact of specific SDoH risk factors on stillbirth risk, as well as the interplay
between race and SDoH. Future studies should target a variety of risk reduction strategies
aimed at modifiable SDoH risk markers that can be widely implemented at both the population
health level, as well as in the direct clinical setting. Emphasis on the use of a
multidisciplinary health care model incorporating clinical case management and social
work specialists into routine care may help to identify and intervene on at-risk patients.
