The Impact of Neighborhood Deprivation on Glycemic Control for Patients with Type 2 Diabetes During Pregnancy

Abstract

Background:

The impact of neighborhood level factors on glycemic control and pregnancy outcomes is understudied. The primary objective was to determine whether there is an association between glycemic control during pregnancy and level of neighborhood deprivation, defined by area deprivation index (ADI).

Materials and Methods:

We conducted a retrospective cohort study of women with type 2 diabetes who received care at a tertiary referral center from 2007 to 2017. Patients living in more deprived neighborhoods (ADI >85th national percentile) were compared to those living in less deprived neighborhoods (ADI ≤85th percentile). The primary outcome was change in hemoglobin A1c (HbA1c) over time. Demographic characteristics were compared between groups, and trends in mean A1c through each trimester were tested with repeated measures analysis.

Results:

Of 237 women meeting study criteria, 93 (39.2%) lived in less deprived (low ADI) and 144 (60.8%) lived in more deprived neighborhoods (high ADI). Women living in more deprived neighborhoods were more likely to be Black (86.8% vs. 53.8%, p < 0.01), less likely to be married (11.3% vs. 31.2%, p < 0.01), and had more severe diabetes (p = 0.05). Both groups achieved significant improvement in HbA1c across each trimester using repeated measures analysis. Those living in more deprived neighborhoods had significantly more improvement in HbA1c from their initial visit to the third trimester compared to those in less deprived neighborhoods, (p = 0.01) such that there was no longer a statistically significant disparity in HbA1c by the third trimester (6.69 ± 0.97 Less deprived vs. 6.95 ± 1.22 more deprived, p = 0.19).

Conclusions:

Low-income women living in more deprived neighborhoods enter pregnancy with significantly worse glycemic control than those living in less deprived neighborhoods, but the gap in glycemic control largely closes by the end of pregnancy with similar maternal and neonatal outcomes.

Introduction

The nonobstetric literature shows clear evidence that community resources have an impact on diabetes, obesity, and hypertension.¹ A randomized controlled trial of reproductive aged women and their children showed that those who were given the opportunity to move to an area with a low-poverty census had lower body mass index (BMI) and risk of developing diabetes.² While it is clear that location has implications for health outcomes, there are little data about the impact of neighborhood level factors on pregnancy-specific outcomes among women with diabetes. A recent study showed that there is an association between diet quality in pregnancy and proximity to supermarkets.³ However, the impact of neighborhood level factors on glycemic control and pregnancy outcomes is understudied and, likely, underestimated.

Pregnant patients with type 2 diabetes (T2DM), who lack insurance before pregnancy, gain access to aggressive medical management once they become pregnant. State Medicaid programs provide funding for prenatal care, and a critical goal is to achieve rapid and sustained glycemic control to both optimize maternal health and minimize the neonatal risks associated with maternal hyperglycemia.⁴ However, this strategy may be too little, too late. From a life course perspective, pregnancy provides only a brief window to combat suboptimal chronic disease control and address the impact of place, including social determinants of health, and neighborhood level deprivation.⁵

The objective of this study was to determine whether there is an association between glycemic control and level of neighborhood deprivation, defined by the area deprivation index (ADI), in a sample of newly insured pregnant patients, who became eligible for Medicaid once they were pregnant. The ADI is a validated tool that allows measurement of socioeconomic deprivation within a neighborhood at the census block level based on 17 weighted factors, with poverty, income, and education carrying the greatest relative weights.⁶ The tool was initially validated by Singh et al. in U.S. mortality data and has been subsequently validated in several studies of pregnancy.^7

–10,14 We hypothesized that living in a less deprived neighborhood is associated with better glycemic control during pregnancy among our sample of low-income, previously uninsured pregnant patients.

Materials and Methods

This was a single-center, retrospective cohort study of pregnant patients with T2DM presenting for care at the Center for Diabetes in Pregnancy at Barnes Jewish Hospital from 2007 to 2017. The Washington University School of Medicine Institutional Review Board approved this study (identification no. 201911173).

The Center for Diabetes in Pregnancy predominantly serves low-income patients typically insured by Medicaid, the majority of whom have no health insurance outside of pregnancy. The income threshold to qualify for Missouri Medicaid outside of pregnancy was extremely low during this study, <$5550 annually for a family of four.^11,*

Patients cared for at the Center for Diabetes in Pregnancy are treated with medical management, including medication titration with oral agents or insulin by a maternal fetal medicine attending/resident physician dyad, and ongoing support through consultation with a diabetes educator and nutritionist at the initial prenatal visit and as needed throughout pregnancy. Our management follows recommendations of the American College of Obstetricians and Gynecologists (ACOG) for glycemic control in pregnancy. This includes titrating medications to a target fasting blood glucose of <95 mg/dL and 1-hour postprandial <140 mg/dL.¹²

Patients were included in the study if they were diagnosed with T2DM before pregnancy and insured by Illinois or Missouri Medicaid. Diagnostic criteria for T2DM included a personal history of T2DM, a random blood glucose >200 mg/dL, a fasting blood glucose >100 mg/dL, or a hemoglobin A1c (HbA1c) >6.2% upon entry into prenatal care. Patients were screened if high risk based on a BMI >30 kg/m² or a significant family history of diabetes.¹³

Exclusion criteria included a diagnosis of a major fetal anomaly, delivery before 24 weeks' gestation, or a multifetal gestation since these factors could independently be associated with poorer glycemic control or neonatal outcomes. We only included pregnant patients with T2DM because the timing of diagnosis for gestational diabetes (GDM) varies greatly, and we wanted to be able to assess improvement in glycemic control over the course of pregnancy, and HbA1c is not checked routinely for patients with GDM. Although patients with T2DM with poor glycemic control upon entering pregnancy have a higher likelihood of anomalies, to avoid significant outliers we chose to exclude those patients. The primary outcome was change in HbA1c throughout pregnancy. Secondary outcomes included antepartum care characteristics, maternal outcomes, and neonatal outcomes.

Trained obstetric research assistants collected demographic, obstetric, and neonatal outcome data from the electronic medical record. ADI scores were determined for each patient based on U.S. census block group utilizing the Singh method.^6,14 The Singh method takes into account 17 socioeconomic indicators included in the 1990 census: educational distribution, median family income, income disparity, occupational composition, unemployment rate, family poverty rate, percentage of the population below 150% of the poverty rate, single-parent household rate, home ownership rate, median home value, median gross rent, median monthly mortgage, and household crowding.¹⁴ Figure 1 is a table from Kind et al., with the relative weights of each component.¹⁵ It is updated at least every 5 years, and most recent data were obtained in 2015.⁶

FIG. 1.

ADI components and weights, from Kind et al.¹⁵ This figure is a table from Kind et al.,¹⁵ demonstrating the 17 components included in the ADI and their relative weights. ADI, area deprivation index.

The ADI national percentile score ranges from 1 to 100, with 1 being the least deprived and 100 being the most deprived neighborhoods.¹⁶ The percentile for each patient was determined by utilizing the University of Wisconsin's Neighborhood Atlas.⁶ Numbers represent national percentiles and take the factors above into account. These percentile groups have remained relatively stable over time.⁶ For this study, high ADI refers to highly or more deprived neighborhoods (ADI >85th national percentile), and low ADI refers to less or least deprived neighborhoods (ADI ≤85th percentile).

Antepartum care characteristics

Antepartum care characteristics included maternal age, prepregnancy BMI, insurance status, self-reported race/ethnicity, marital status, gestational age (GA) at initiation of prenatal care, whether they utilized tobacco products, alcohol, or recreational drugs during the pregnancy, and whether they had chronic hypertension (presence of elevated blood pressure, systolic ≥140 mmHg or diastolic pressure ≥90 mmHg, before 20 weeks' gestation).¹⁷ In addition to HbA1c data, we also collected data regarding their diabetes history, including the age at which they were diagnosed with diabetes, diabetes class by White Classification of disease severity upon entry into pregnancy, whether they initiated prenatal care on insulin, and whether they required hospitalization to optimize glycemic control during pregnancy.¹⁸ GA was determined utilizing the recommendations by the ACOG.¹⁹

Maternal secondary outcomes

Maternal secondary outcomes included number of prenatal visits, GA at delivery, mode of delivery (vaginal delivery or cesarean section), and a maternal morbidity composite, including: hemorrhage (defined as a blood loss of >1000 mL during delivery),²⁰ fever (temperature of >38.0°C), endomyometritis (defined by a fever and fundal tenderness),²¹ receipt of blood transfusion, infection (defined as a fever and a clinical source of infection other than endomyometritis), and wound infection (defined as purulent drainage or evidence of cellulitis at the incision site).²² BMI at the time of delivery and at the postpartum visit was also recorded.

Neonatal secondary outcomes

Neonatal secondary outcomes included large for GA, defined as a birth weight >90th percentile based on the Alexander growth curve, respiratory distress syndrome as determined by the pediatric resuscitation team at time of delivery and recorded in the electronic medical record, GA at delivery, Apgar score ≤7 at 5 minutes, admission to the neonatal intensive care unit following delivery for >24 hours, shoulder dystocia (defined as need for shoulder dystocia maneuvers recorded in the electronic medical record), birth injury (including brachial plexus injuries and bony fractures), and neonatal hypoglycemia (plasma glucose level of <30 mg/dL in the first 24 hours of life and <45 g/dL thereafter).^23

–27 In addition, we collected birth weight on all infants.

Statistical analyses

The distribution of ADI within the sample of low-income patients with Medicaid insurance was displayed by histogram, which was right skewed. The sample was divided into less deprived (ADI ≤85th percentile) and more deprived (ADI >85th percentile) groups, based on previous literature.⁹ Stratified analyses were conducted by race, as a social construct that may influence study outcomes. Descriptive and bivariate analyses compared demographic characteristics between pregnant patients by ADI group using two sample t-test or Wilcoxon rank-sum test.

We analyzed repeated measurements of HbA1c in each trimester to assess how HbA1c changed over time. Paired t-test was used to assess HbA1c changes over time between more and less deprived groups, and the sign test was used in cases when the distribution was not normal. Differences in secondary maternal and neonatal outcomes between more and less deprived were calculated using crude odds ratios. Backward stepwise multivariable logistic regression models were used to adjust for potential confounders, including race, marital status, and White's classification of diabetes. Two sided tests with p-value <0.05 were considered statistically significant. SAS 9.4 was used for all statistical analyses.

Results

During the study period, 948 pregnant patients received care at the Center for Diabetes in Pregnancy at Barnes Jewish Hospital. Of these, 237 pregnant patients had T2DM and met study criteria and 711 were excluded. ADI in the sample ranged from 30 to 100 by national centiles (Fig. 2), with 93 (39.2%) living in less deprived neighborhoods (ADI ≤85th percentile) and 144 (60.8%) living in more deprived neighborhoods (ADI >85th percentile). Women living in more deprived neighborhoods were more likely to be Black (86.8% vs. 53.8%, p < 0.01), less likely to be married (11.3% vs. 31.2%, p < 0.01), and had more severe diabetes by White's class (p < 0.05).

FIG. 2.

Distribution of ADI of study population. This figure is a histogram of the ADI of the study population. The x-axis represents the ADI deciles, 1 being least deprived, 10 being most deprived. The y-axis represents percentage of patients in each category.

Patients living in more deprived neighborhoods had higher HbA1c levels at baseline (8.71 vs. 7.84, p = 0.03), but there was no difference between groups in the second or third trimesters (Table 2). Both groups achieved significant improvement in HbA1c longitudinally with a significant decrease in their mean HbA1c from the first to second trimester (less deprived: −0.99 ± 1.76, p = 0.01; more deprived: −2.18 ± 2.39, p < 0.01) and from the first to third trimester (less deprived: −0.92 ± 1.3, p < 0.01; more deprived: −2.0 ± 2.46, p < 0.01). However, there was not a significant decrease from the second to third trimester in either group (Table 3). Patients from more deprived neighborhoods had a greater decrease in HbA1c from initial visit to third trimester than those living in less deprived neighborhoods (−1.08 ± 2.10, p = 0.01).

Table 1.

Study Population Antepartum Demographics

Demographic Total cohort = 237	Less deprived (ADI) 93 (39.2%)	More deprived (ADI) 144 (60.8%)	p
Maternal age at first visit, years	29.3 ± 5.5	30.2 ± 6.1	0.30
Maternal age >35 years old, n (%)	14 (15.1)	36 (25.0)	0.07
Prepregnant BMI, kg/m²	37.8 ± 8.6	37.5 ± 13.8	0.91
BMI, kg/m² (at first prenatal visit)	39.2 ± 9.4	39.1 ± 12.2	0.96
First prenatal visit obese (BMI ≥30, kg/m²), n (%)	59 (63.4)	86 (59.7)	0.59
Race/ethnicity, n (%)			<0.01
Black	50 (53.8)	125 (86.8)
White	32 (34.4)	13 (9.03)
Hispanic	6 (6.45)	6 (4.17)
Other	5 (5.38)	0 (0)
Married	29 (31.2)	16 (11.3)	<0.01
Diabetes diagnosis age	24.3 ± 6.6	23.5 ± 7.1	0.38
Diabetes class by White class, n (%)			<0.05
2_B	61 (70.1)	72 (51.4)
3_C	21 (24.1)	51 (36.4)
4_R (retinopathy)	0 (0)	3 (2.14)
5_F (diabetic kidney disease)	3 (3.45)	5 (3.57)
6_Other	2 (2.3)	9 (6.43)
Gestational age at initiation visit	11.3 ± 6.6	11.6 ± 5.7	0.72
Tobacco use in pregnancy, n (%)	18 (19.4)	36 (25.0)	0.34
Alcohol use in pregnancy, n (%)	1 (1.08)	4 (2.78)	0.65
Illicit drug use in pregnancy, n (%)	11 (11.8)	10 (6.9)	0.24
Entered pregnancy on medication, n (%)	47 (50.5)	74 (51.4)	1.0
Chronic hypertension, n (%)	36 (47.4)	76 (58.9)	0.11

Bold values signify statistical significance.

ADI, area deprivation index; BMI, body mass index.

Table 2.

Description of Hemoglobin A1c Throughout Pregnancy in Low and More Deprived Cohorts

Pregnancy time point	Less deprived (Low ADI) Mean ± SD Range (min–max)	More deprived (High ADI) Mean ± SD Range (min–max)	p
HbA1c at first visit, mmol/mol	7.84 ± 2.25(5.3–18.5)	8.71 ± 2.47(5.0–15.0)	0.03
HbA1c at second trimester, mmol/mol	6.75 ± 1.13(5.2–10.1)	6.79 ± 1.20(5.4–11.2)	0.90
HbA1c third trimester, mmol/mol	6.69 ± 0.97(5.2–9.2)	6.95 ± 1.22(4.2–10.3)	0.19

Bold values signify statistical significance.

HbA1c, hemoglobin A1c; SD, standard deviation.

Table 3.

Description of Longitudinal Change of Hemoglobin A1c Across Pregnancy

Change in HbA1c across time points	Longitudinal analysis of HbA1c change from initial visit to delivery				Comparison HbA1c differences between more deprived vs. less deprived
	Less deprived (Low ADI)		More deprived (High ADI)		Mean difference, ± SD (95% CI of difference)	p
	Mean ± SD	p	Mean ± SD	p	Mean difference, ± SD (95% CI of difference)	p
Change in HbA1c from initial visit to second trimester, mmol/mol	−0.99 ± 1.76	0.01	−2.18 ± 2.39	<0.01	−1.19 ± 2.10(−2.42 to 0.04)	0.06
Change in HbA1c from second trimester to third trimester, mmol/mol	−0.10 ± 0.72	0.51	0.14 ± 0.76	0.46	0.25 ± 0.74(−0.25 to 0.75)	0.32
Change in HbA1c from initial visit to third trimester, mmol/mol	−0.92 ± 1.3	<0.01	−2.0 ± 2.46	<0.01	−1.08 ± 2.10(−1.93 to −0.23)	0.01

Bold values signify statistical significance.

CI, confidence interval.

In a subgroup analysis stratifying ADI group by black and white race, black and white patients in less and more deprived groups had a significant improvement in HbA1c across each trimester using repeated measures analysis (p < 0.01) (Table 4). The only significant difference in HbA1c between black and white patients was in the less deprived group in the third trimester (6.96% vs. 6.35%, p = 0.04). There were no differences in maternal or neonatal outcomes between patients in more deprived and less deprived neighborhoods (Tables 5 and 6), with the exception of 5-minute Apgar score <7, which was more frequently noted in the more deprived group (18.10% vs. 9.68%, adjusted odds ratio 2.55, 95% confidence interval 1.03–6.32) (Table 6).

Table 4.

Description of Hemoglobin A1c Between Black and White Cohort by Area Deprivation Index

Pregnancy time point	Less deprived (Low ADI)			More deprived (High ADI)
Pregnancy time point	Black Mean ± SD (n = 33)	White Mean ± SD (n = 22)	p	Black Mean ± SD (n = 91)	White Mean ± SD (n = 9)	p
HbA1c at first visit, mmol/mol	8.20 ± 2.52	7.44 ± 1.81	0.23	8.83 ± 2.49	7.78 ± 2.54	0.23
HbA1c at second trimester, mmol/mol	6.89 ± 1.16	6.33 ± 1.19	0.28	6.85 ± 1.23	—only 1 patient	—
HbA1c at third trimester, mmol/mol	6.96 ± 1.01	6.35 ± 0.87	0.04	6.96 ± 1.26	6.92 ± 0.97	0.93

Bold values signify statistical significance.

Table 5.

Pregnancy Outcomes Among Cohort

Pregnancy outcome	Less deprived (Low ADI) 93 (39.2%)	More deprived (High ADI) 144 (60.8%)	p
BMI at delivery, kg/m², mean ± SD	43.1 ± 10.3	41.8 ± 14.0	0.42
BMI postpartum kg/m², mean ± SD	39.5 ± 10.2	40.1 ± 12.3	0.72
Hospitalization for insulin titration, n (%)	33 (35.5)	53 (36.8)	0.89
No. of prenatal care visits, median [IQR]	14 (12–17.5)	13 (9–17)	0.16
GA at delivery, median [IQR]	37 (36–39)	37 (35–38)	0.33
Mode of delivery, n (%)
Vaginal	33 (35.5)	54 (37.5)	0.78
Cesarean	60 (64.5)	90 (62.5)	0.78
Birth weight (grams), mean ± SD	3193 ± 785	3092 ± 982	0.38
Postpartum stay (days), median [IQR]	4 (2–4)	4 (2–4)	0.59

GA, gestational age; IQR, interquartile range.

Table 6.

Maternal and Neonatal Secondary Outcomes

Secondary outcome	Less deprived (Low ADI), n (%)	More deprived (High ADI), n (%)	p	OR	aOR^a
5-minute Apgar score <7	9 (9.68)	26 (18.1)	0.09	2.06 (0.92–4.61)	2.55 (1.03–6.32)
Large for gestational age	19 (20.4)	32 (22.2)	0.87	1.11 (0.59–2.11)	1.18 (0.58–2.44)
Macrosomia	12 (12.9)	24 (16.7)	0.46	1.35 (0.64–2.85)	1.45 (0.62–3.41)
Neonatal hypoglycemia	18 (28.1)	22 (21.0)	0.35	0.68 (0.33–1.39)	0.64 (0.28–1.45)
NICU admission	7 (7.53)	19 (13.2)	0.21	1.87 (0.75–4.64)	2.38 (0.80–7.09)
Shoulder dystocia	6 (6.45)	7 (4.86)	0.77	0.74 (0.24–2.28)	0.76 (0.22–2.60)
Birth injury or complications	35 (54.7)	57 (54.3)	1.0	0.98 (0.53–1.84)	0.98 (0.50–1.93)
Maternal delivery complication^b composite^c	42 (45.2)	72 (50.0)	0.51	1.21 (0.72–2.05)	1.26 (0.68–2.33)

Bold values signify statistical significance.

Backward stepwise regression to select covariables in aOR model, the full variables group was race, marital status, diabetes class, p < 0.05 in Table 1.

Maternal delivery complication included: fever, placental abruption, retained placenta, infection, transfusion, pre-eclampsia w/o severe features, pre-eclampsia with severe features.

Composite included: postpartum hemorrhage, fever, endomyometritis, receipt of blood transfusion, infection, and wound infection.

aOR, adjusted odds ratio; Apgar, scoring system to guide neonatal resuscitation; NICU, neonatal intensive care unit; OR, odds ratio.

Discussion

Low-income, previously uninsured pregnant patients with T2DM and Medicaid insurance during pregnancy who lived in highly deprived neighborhoods entered pregnancy with significantly worse glycemic control than those living in less deprived neighborhoods. However, both groups experienced a significant improvement in glycemic control over the course of pregnancy so that this disparity was no longer present by the third trimester and both groups experienced similar maternal and neonatal outcomes. A notable exception is that black women living in less deprived neighborhoods did not achieve the same level of glycemic control as white women living in more deprived neighborhoods.

Our findings suggest that place, race, and the societal forces of racism that dictate who lives where are important factors that should be considered for patients with chronic diseases as they contemplate and enter pregnancy. Aggressive diabetes management and prenatal care, through newly acquired Medicaid coverage for pregnancy, significantly attenuated the disparity in glycemic control by neighborhood deprivation, except for black women living in less deprived neighborhoods compared to white women living in similar neighborhoods. In addition, it should be noted that postpartum care and prepregnancy care are often synonymous. Current Medicaid policies with short-term postpartum coverage, and limited access to primary care between pregnancies, likely lead to worsening of chronic disease and contribute to the initial pregnancy visit disparities identified in this study.

Our finding that pregnant patients living in more deprived neighborhoods had poorer baseline glycemic control is consistent with prior studies that found that neighborhood factors influence glucose intolerance through lack of access to healthy food options.⁸ The over-representation of black patients in more deprived neighborhoods likely reflects systemic racism and both past and present practices, such as redlining, that limit opportunities for upward mobility and relegate certain racial and class groups to areas with limited resources.²⁸

Prior studies found that ADI is a strong predictor of control of diabetes outside of pregnancy.²⁹ These findings may, in part, be explained by the lack of resources available in deprived areas and food insecurity.³⁰ In addition, insurance status likely plays a significant role in access to health care between pregnancies for all low-income pregnant patients, but especially those living in more deprived neighborhoods. A previous study noted a significant deprivation-insurance effect modification in children with asthma, noting that those with public insurance had worse outcomes than those living in the same high deprivation areas with private insurance.³¹

It is reassuring that low-income pregnant patients across the ADI spectrum uniformly achieved improved glycemic control during pregnancy, and the gap in HbA1c from initial prenatal visit to delivery narrowed to a statistically nonsignificant level for most patients. These findings suggest that aggressive medical management during the course of prenatal care improves glycemic control across the entire ADI spectrum. However, our overall findings may be prone to a type 2 error, due to our limited sample size, and suggest that prenatal care alone is insufficient to overcome neighborhood level deprivation for black women living in less deprived neighborhoods. This is underscored by three findings:

Few pregnant patients in either less or more deprived neighborhoods achieved a HbA1c <6.0%, which is recommended by ACOG during pregnancy.²³

A persistent gap between black and white pregnant patients in the most deprived neighborhoods at delivery suggests that more can be done to optimize diabetes outcomes in this population—during, before, and between pregnancies.

Black patients in the less deprived group had higher HbA1c levels than White women in the less deprived group across all three trimesters—a finding that did not achieve statistical significance, but warrants further study.¹³ This final point underscores the systemic racism that is reflected by many reproductive health outcomes. The explanation for these differences occurs at multiple levels, including lack of access to care, differences in the quality of care provided, interpersonal and institutional racism, and neighborhood deprivation, which disproportionately affect black patients.³²

To improve outcomes during pregnancy for low-income pregnant patients with diabetes, our perspective must broaden to incorporate the role of place, race/racism, and the importance of the life course, of which pregnancy is only a small part. These factors likely have a greater impact than prenatal care, which is the focus of most studies and medical resources.

From a health policy perspective, extending postpartum coverage beyond the immediate postpartum period may help to further optimize diabetes and other medical comorbidities, which could reduce the disparities suggested in this study and others.^33

–36 Importantly, it should be noted again that during this study to qualify for Missouri Medicaid, families had to make <$5550 annually.¹¹ This is an exceptionally low threshold; however, many states have similar thresholds, especially those that chose not to expand Medicaid coverage under the Affordable Care Act. This is particularly important because for the women in this study, all of them did not have insurance coverage before becoming pregnant when they qualified for Medicaid. All of the patients in this study lost insurance coverage within 60 days of delivery and had inadequate medical coverage during much of the “4th trimester,” a period when most maternal deaths occur.³⁷ This cycle of losing and gaining coverage only perpetuates and worsens medical conditions for these patients outside of and during pregnancy. Further research is needed to determine additional neighborhood level interventions that could help to overcome this disparity and promote health equity.

There are several strengths of this study. We had access to serial HbA1c data throughout pregnancy, and the patient sample was insured by Medicaid with similar socioeconomic status, differing only by the level of deprivation found at the neighborhood level. This is one of the first studies in the obstetric literature to move the focus from the provision of medical care to the importance of place as we care publicly insured pregnant patients with diabetes.

However, our findings should be considered in the context of the following limitations. First, while HbA1c is utilized as a surrogate for glycemic control, it is not validated in pregnancy and is not as sensitive as continuous glucose monitoring to determine level of control. Second, each patient did not have all three HbA1c's drawn. A post hoc sample size calculation revealed that 354 patients would provide 80% power to see a 0.3% difference in HbA1c (standard deviation 1.3 and 0.7), assuming α = 0.05. Thus, our results are prone to a type 2 error, especially when looking at the differences between white and black patients, as those numbers were particularly small.

Third, the patient population served by the Center for Diabetes in Pregnancy is, at baseline, a highly deprived population, as evidenced by the distribution of ADI for our patient population (Fig. 3). In addition, while ADI accounts for neighborhood level disadvantage, it does not necessarily explain disadvantages at the individual patient level. Therefore, we were actually comparing moderately deprived patients to highly deprived patients. Finally, we were not powered to determine difference in maternal or neonatal secondary outcomes.

FIG. 3.

Hemoglobin A1c values throughout pregnancy by ADI group. This is a graphical representation of the hemoglobin A1c values throughout pregnancy by ADI group. The x-axis is the time period in pregnancy when the hemoglobin A1c was collected. The y-axis is hemoglobin A1c values, starting at 5.0%. The standard deviation of each time period and group is represented by the hatch marks on each column.

Conclusions

ADI is a proxy for what is happening in communities at the neighborhood level. While excellent prenatal care is necessary and closed the gap for many patients living in more deprived neighborhoods in this study, our findings suggest that it may be insufficient to overcome neighborhood level deprivation for black patients with T2DM living in less deprived neighborhoods during pregnancy. Place is important, and comprehensive diabetes management during pregnancy should be optimized by providing additional support to pregnant patients at the neighborhood level.

Footnotes

Disclaimer

The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official view of the American Diabetes Association or the Robert Wood Johnson Foundation.

Author Disclosure Statement

The authors report no conflicts of interest or financial disclosures.

Funding Information

Dr. Carter is funded by the American Diabetes Association Pathway to Stop Diabetes Award (1-19-ACE-02), the Robert Wood Johnson Foundation Grant No. 74250, and a NIH/NICHD K23 grant (HD095075-03) and the Washington University Diabetes Research Center (NIDDK/P30DK020579). Dr. Herrick is funded by NIH/NICHD K23 (HD096204).

References

Young

, Laurent

, Chung

, Wu

. Geographic distribution of healthy resources and adverse pregnancy outcomes. Matern Child Health J, 2016; 20:1673–1679.

Ludwig

, Sanbonmatsu

, Gennetian

, et al. Neighborhoods, obesity, and diabetes—A randomized social experiment. N Engl J Med, 2011; 365:1509–1519.

Laraia

, Siega-Riz

, Kaufman

, Jones

. Proximity of supermarkets is positively associated with diet quality index for pregnancy. Prev Med, 2004; 39:869–875.

Alexopoulos

, Blair

, Peters

. Management of preexisting diabetes in pregnancy: A review. JAMA, 2019; 321:1811–1819.

Goldenberg

, McClure

, Harrison

, Miodovnik

. Diabetes during pregnancy in low- and middle-income countries. Am J Perinatol, 2016; 33:1227–1235.

University of Wisconsin School of Medicine and Public Health. Area Deprivation Index v2.0. Secondary Area Deprivation Index v2.0 2020. Available at: https://www.neighborhoodatlas.medicine.wisc.edu/ Accessed September 15, 2021 .

Andrews

, Tamura

, Claudel

, et al. Geospatial analysis of neighborhood deprivation index (NDI) for the United States by County. J Maps, 2020; 16:101–112.

Headen

, Laraia

, Coleman-Phox

, Vieten

, Adler

, Epel

. Neighborhood typology and cardiometabolic pregnancy outcomes in the maternal adiposity metabolism and stress study. Obesity (Silver Spring), 2019; 27:166–173.

Kind

AJH

, Buckingham

. Making neighborhood-disadvantage metrics accessible—The neighborhood atlas. N Engl J Med, 2018; 378:2456–2458.

10.

Lian

, Struthers

, Liu

. Statistical assessment of neighborhood socioeconomic deprivation environment in spatial epidemiologic studies. Open J Stat, 2016; 6:436–442.

11.

Missouri Foundation for

Health

. Missouri Medicaid Basics: Spring, 2019. https://mffh.org/wp-content/uploads/2020/11/Missouri-Medicaid-Basics-2020-web.pdf Accessed September 15, 2021 .

12.

ACOG Practice Bulletin No. 190 Summary: Gestational Diabetes Mellitus. Obstet Gynecol, 2018; 131:406–408.

13.

ACOG Practice Bulletin No. 201: Pregestational Diabetes Mellitus. Obstet Gynecol, 2018; 132:e228–e248.

14.

Singh

GK.

Area deprivation and widening inequalities in US mortality, 1969–1998. Am J Public Health, 2003; 93:1137–1143.

15.

Kind

, Jencks

, Brock

, et al. Neighborhood socioeconomic disadvantage and 30-day rehospitalization: A retrospective cohort study. Ann Intern Med, 2014; 161:765–774.

16.

Health Innovation Project. 2000 Area Deprivation Index. Secondary 2000 Area Deprivation Index. 2014. Available at: http://www.hipxchange.org/ADI Accessed September 15, 2021 .

17.

ACOG Practice Bulletin No. 203: Chronic Hypertension in Pregnancy. Obstet Gynecol, 2019; 133:e26–e50.

18.

Gilmartin

, Ural

, Repke

. Gestational diabetes mellitus. Rev Obstet Gynecol, 2008; 1:129–134.

19.

ACOG Committee Opinion No. 700: Methods for Estimating the Due Date. Obstet Gynecol, 2017; 129:e150–e154.

20.

ACOG Practice Bulletin No. 183: Postpartum Hemorrhage. Obstet Gynecol, 2017; 130:e168–e186.

21.

Faro

Postpartum endometritis. Clin Perinatol, 2005; 32:803–814.

22.

Zuarez-Easton

, Zafran

, Garmi

, Salim

. Postcesarean wound infection: Prevalence, impact, prevention, and management challenges. Int J Womens Health, 2017; 9:81–88.

23.

Schmidt

, Norman

, Misselwitz

, et al. Mode of delivery and mortality and morbidity for very preterm singleton infants in a breech position: A European cohort study. Eur J Obstet Gynecol Reprod Biol, 2019; 234:96–102.

24.

Carter

, Tuuli

, Odibo

, Macones

, Cahill

. Prenatal visit utilization and outcomes in pregnant women with type II and gestational diabetes. J Perinatol, 2017; 37:122–126.

25.

Boulet

, Salihu

, Alexander

. Mode of delivery and birth outcomes of macrosomic infants. J Obstet Gynaecol, 2004; 24:622–629.

26.

Committee on Practice

B-O.

Practice Bulletin No. 178: Shoulder Dystocia. Obstet Gynecol, 2017; 129:e123–e133.

27.

Duryea

, Hawkins

, McIntire

, Casey

, Leveno

. A revised birth weight reference for the United States. Obstet Gynecol, 2014; 124:16–22.

28.

Peek

, Odoms-Young

, Quinn

, Gorawara-Bhat

, Wilson

, Chin

. Racism in healthcare: Its relationship to shared decision-making and health disparities: A response to Bradby. Soc Sci Med, 2010; 71:13–17.

29.

Durfey

SNM

, Kind

AJH

, Buckingham

, DuGoff

, Trivedi

. Neighborhood disadvantage and chronic disease management. Health Serv Res, 2019; 54(Suppl. 1):206–216.

30.

Laraia

, Vinikoor-Imler

, Siega-Riz

. Food insecurity during pregnancy leads to stress, disordered eating, and greater postpartum weight among overweight women. Obesity (Silver Spring), 2015; 23:1303–1311.

31.

Nkoy

, Stone

, Knighton

, et al. Neighborhood deprivation and childhood asthma outcomes, accounting for insurance coverage. Hosp Pediatr, 2018 [Epub ahead of print]; DOI: 10.1542/hpeds.2017-0032.

32.

Gavin

, Adams

, Hartmann

, Benedict

, Chireau

. Racial and ethnic disparities in the use of pregnancy-related health care among Medicaid pregnant women. Matern Child Health J, 2004; 8:113–126.

33.

Callaghan

WM.

Overview of maternal mortality in the United States. Semin Perinatol, 2012; 36:2–6.

34.

Daoud

, O'Campo

, Minh

, et al. Patterns of social inequalities across pregnancy and birth outcomes: A comparison of individual and neighborhood socioeconomic measures. BMC Pregn Childbirth, 2015; 14:393.

35.

Howell

EA.

Reducing disparities in severe maternal morbidity and mortality. Clin Obstet Gynecol, 2018; 61:387–399.

36.

Gifford

, Walls

, Reanji

, Salganicoff

, Gomez

Medicaid coverage of pregnancy and perinatal benefits: Results from a state survey. Kaiser Family Foundation Report. Womens Health Policy 2017.

37.

Creanga

, Syverson

, Seed

, Callaghan

. Pregnancy-related mortality in the United States, 2011–2013. Obstet Gynecol, 2017; 130:366–373.