Addressing social determinants of health has become a critical concern in all areas of medicine. Significant ethnic and racial inequities with respect to both access to care and health outcomes are evident across multiple specialties, including obstetrics.1–3 Among immigrants, barriers to health care access may result from real or perceived social exclusion, which lead to delayed or fewer preventive health services, resulting in relatively worse outcomes when compared with similar individuals with legal resident status.4–7
Routine and early prenatal care is accepted as offering protection against maternal and infant morbidity and mortality and is a well-accepted standard of care8,9 Alongside attending regular prenatal care visits, identifying and treating common conditions of pregnancy (such as iron deficiency anemia) serve as reasonable proxies of access to standard obstetric care.10,11
Current debate over immigration policies, deportation activities, and the accompanying political rhetoric may contribute to a sense of isolation among Hispanic, Latino, and undocumented patients. Immigrant responses to these exclusionary policies have been to adopt a policy of community avoidance.12,13 However, research associating access and utilization of health care, particularly adequacy of prenatal care, to recent immigration policies, and the anti-immigrant rhetoric that accompanies such policies, is limited and has not yet extended to measures of adequacy of prenatal care.14,15
The aim of this study was not to argue for or against any policy nor practice of immigration reform, but rather to identify whether an association exists between anti-immigrant rhetoric and receipt of adequate prenatal care among patients who deliver in a U.S.-based hospital.
This research study was approved by the Baylor College of Medicine institutional review board (H-26364, most recent approval March 2, 2018; H-43085, most recent approval June 5, 2018). This is a retrospective population-based cohort study (August 2011 through July 2017) designed to evaluate rates and timing of prenatal care among women before and after significant increases in anti-immigrant political rhetoric. Hispanic and Latino women (hereafter condensed to “Hispanic”) were the cohort of interest and were stratified by birth country of origin. These patients were compared against all other gravid women in the cohort, consistent with a population-based design. All patients in the cohort delivered at one of our two obstetric hospitals (Ben Taub Hospital and the Texas Children’s Pavilion for Women, Houston, Texas), and all patients delivering at either of these two hospitals were approached and consented and enrolled at the time of delivery for inclusion in our institutional database, PeriBank.16 All patients in PeriBank are directly queried after consent regarding their birth country, parental birth country, duration in the United States, and five-digit zip code of residence and work through each trimester. Additional direct query and abstracted data included patient history (including prior occurrence of preeclampsia and diabetes, smoking status, nicotine and substance use, familial obstetric history, and entry into care, number of prenatal visits, and location of prenatal care clinics and health care providers), socioeconomic status (education, income, immigration status, birth country, time in the United States), and residential and workplace data (each trimester of pregnancy residence and workplace five-digit zip code). Additional abstracted data analyzed in this study included the mean and nadir hemoglobin value as well as multiple comorbidities (eg, gestational diabetes, substance use disorders), which are known to increase or decrease the number of prenatal visits. Given that treatment of iron deficiency anemia is standard obstetric care, both the mean and nadir hemoglobin measures were intended to be reasonable secondary objective proxy measures of attainment of prenatal care.
Patients included in the current study were enrolled in our PeriBank database between August 2011 through July 2017 (N=24,933 deliveries). The index pregnancy was defined as the most recent pregnancy in the database for each patient. All women who delivered were identified, and all patients had detailed prenatal care by gestational age and number of visits verified and abstracted. Women were assigned to the before and after groups based on the date of delivery for the index pregnancy. PeriBank includes all births occurring at both hospitals (Texas Children's Pavilion for Women and Ben Taub Hospital). These two hospitals are distinct from one another with respect to payer mix, race and ethnicity, insurance status, and other sociodemographic variables. When considered together, they are representative of the general population of births in Houston, Texas.
Publicly available Google search trends were mined for the search terms “Make America Great Again,” “Mexico Wall,” and “Deportation” by the greater region, specifically the southern United States (as defined by trends.google.com/explore/subregion, including Texas, New Mexico, Oklahoma, Arkansas, Louisiana, Alabama, Mississippi, the Carolinas, Florida, Georgia, and Tennessee). These search terms terms were chosen for their relation to common themes of the debate and accompanying political rhetoric surrounding immigration in the United States over our study time period and similarity of geographic region. Our choice of terms was further based on their representation of explicit (Deportation, Mexico Wall) and implicit (Make America Great Again) anti-immigrant sentiment. Because our population-based study aimed to determine the significance of association as a large permeation and not just by increased use of the terms, Google trend by a priori defined subregion and limited terms enabled best estimations true to the focus of the study in a contemporaneous region and time period. The time of first deviation from the mode Google search popularity value for each term was ascertained (mode inflection date). A mode inflection date of January 7, 2015, was extrapolated from the Google trend analytics and used to define the period before large-scale change in trends in rhetoric use pre (before rhetoric) and post (after rhetoric) (Fig. 1). No patient-specific data were used in determining these search terms; thus, the terms and their inflection date were naïve to the primary outcome.
The primary outcome in this analysis was adequacy of prenatal care. This was defined by multiple measures and means, including the gestational age at the first prenatal visit, mean number of prenatal visits, and incidence of inadequate prenatal care in the index pregnancy. To categorically assign patients as having inadequate compared with adequate or intermediate prenatal care, the previously validated and “gold standard” Adequacy of Prenatal Care Utilization Index was used.17–19 Briefly, the Adequacy of Prenatal Care Utilization Index has been shown to be superior over the Kessner Institute of Medicine Adequacy of Prenatal Care Index when compared with its capacity for incorporating the month that prenatal care began, the proportion of the number of visits as recommended by the American College of Obstetrician Gynecologists, and the gestational age at delivery. Inadequate prenatal care is thus defined as prenatal care begun after the fourth month or less than 50% of recommended prenatal visits received as stratified by gestational age at delivery. Adequate and intermediate were collapsed into a single category and defined as prenatal care begun by the fourth month and 50% to greater than 100% of recommended visits received. All patients were categorized into either “inadequate” or “adequate/intermediate” for subsequent analysis.
All clinical and demographic data for pregnancies enrolled into our study were obtained from PeriBank as previously described.16 Preterm delivery was defined as delivery before 37 weeks of gestation. Groups of interest were compared by χ2 or Fisher exact test for categorical variables, t test for normally distributed continuous variables, and Wilcoxon rank-sum test for nonnormal continuous variables. Given the large sample number, the normality of each data category was visually determined using histograms and normal Q-Q plots. Linear regression was performed for the trend in the before and after rhetoric groups.
In an effort to control for potential confounders, multivariate modeling was undertaken. Generalized linear models (GLMM, Proc GENMOD in SAS 9.4) were used for estimation of odds ratio (OR), adjusted OR, and corresponding 95% CIs for binomial outcome variables (inadequate prenatal care in this pregnancy [primary outcome]). Generalized linear mixed models (Proc GLIMMIX in SAS 9.4) was used for estimation of least square means for nonnormally distributed continuous outcome variables and comparison between study groups. Adjustments were made for maternal age, Hispanic ethnicity, method of payment, gestational age at first visit, type 2 diabetes, chronic hypertension, and gestational age at diagnosis of preeclampsia in the previous pregnancy. P<.05 was considered statistically significant. Because the exposure (change in anti-immigrant rhetoric) was measured over time and in relation to the primary outcome (sufficiency of prenatal care), there was no intent to ascribe causality per se and an E-value (as an estimate of the level of unmeasured confounding) was not calculated.20
Baseline characteristics for the included cohort are shown in Table 1 and further divided by time relative to the identified date of rhetoric increase for Hispanic patients only in Tables 2 and 3. Of the 23,580 patients included, 12,646 (46.3%) were not born in the United States with a majority originating from countries in Central America or Mexico. Compared with U.S. non-native patients, U.S. native patients had greater average body mass indexes (BMIs, calculated as weight (kg)/[height (m)]2), higher rates of preterm birth, higher rates of cesarean delivery, and more frequently used tobacco, alcohol, and other illegal or illicit substances. Patients born in Mexico, Central America, and South America were generally more common in their exposures and outcomes, although individuals from South America were older, of a lower parity, and had a greater rate of cesarean deliveries than their other non–U.S.-born Hispanic counterparts. For Hispanic patients, baseline characteristics of the enrolled cohort before and after rhetoric change remained the same, except for maternal age, maternal education, substance use disorder, and gestational diabetes (all of which were less common; P<.05). Overall, U.S. non-native patients had lower BMIs and lower rates of preterm birth, cesarean delivery, and polysubstance use disorder.
There was a significant reduction of total prenatal visits for U.S. non-native patients after the increase in anti-immigration rhetoric; in the same interval time period, there was a significant increase in the total number of prenatal visits for U.S. native patients (Fig. 2A; P<.05). However, there was no significant difference in the delay to the first prenatal visit for either group (Fig. 2B). The average and nadir hemoglobin was significantly increased for U.S. native patients over the study interval, but this was not true for U.S. non-native patients (Fig. 2C, P<.05; Table 4). Comparisons by region of birthplace origin indicate the reduction of total of prenatal visits after rhetoric change was significant for patients from Mexico and Central America, but not South America (Fig. 3A). Conversely, for U.S. native Hispanic gravid women, there was instead an increase in the total of prenatal visits (Fig. 3A). However, there was no difference by region of birthplace in the delay before the first prenatal visit (Fig. 3B) nor the average hemoglobin (Fig. 3C), although U.S. native Hispanic women saw a slight but likely clinically insignificant increase in the after rhetoric change period (Fig. 3C).
To evaluate for trending changes to prenatal care over time, patient data were averaged over 15 days sliding windows relative to the rhetoric date inflection point and a linear regression was fitted. There was a significant trending reduction of prenatal visits and an increase in days until first prenatal visits for Hispanic U.S. non-native, but not their U.S. native counterparts over time (Figs. 4 and 5). There was similarly a significant trend for a greater delay until the first prenatal visit for non-Hispanic U.S. native patients, although there was no significant trend in the total number of visits, which remained the same before and after rhetoric change (Fig. 6). For all patients, the hemoglobin remained constant during the time periods considered (Figs. 4–6).
Adequate prenatal care was assessed using the Adequacy of Prenatal Care Utilization Index, which accounts for gestational age at delivery, the gestational age at the first prenatal visit, and the total number of visits.19,21 After rhetoric change, significantly fewer non-Hispanic U.S. native and Hispanic U.S. native patients had inadequate prenatal care (Fig. 7; 33.8–21.6%; 41.9–35.5%; P<.001). However, there was no significant change in the rate of inadequate prenatal care for Hispanic U.S. non-native patients (Fig. 7; 41.3–42.3%, P=.28). Thus, although all other patients observed a drop in the rate of inadequate and insufficient prenatal care, this same advantage was not seen among Hispanic U.S. non-native patients. A logistic regression model was fitted to formally evaluate the association between immigrant status and inadequate prenatal care after controlling for maternal age, education, gestational diabetes, and substance use (Table 5; Fig. 8). In the time period before the rhetoric increase, less than a high school education (OR 2.214, CI 1.931–2.541) and substance use (OR 2.046, CI 1.533–2.742), but neither ethnicity nor immigrant status was significantly associated with inadequate prenatal care. However, in the period after rhetoric change, both Hispanic U.S. native patients (OR 1.328, CI 1.174–1.502) and Hispanic U.S. non-native patients (OR 1.581, CI 1.407–1.777) were significantly more likely to have inadequate prenatal care (Table 5; Fig. 8). Lastly, the adjusted ORs for Hispanic U.S. native and Hispanic U.S. non-native patients were statistically significant from each other (P=.006), indicating that Hispanic immigrant status withstood as a significant predictor of inadequate prenatal care in the interval after rhetoric change.
Some recent studies support public funding of prenatal care for undocumented immigrants.22–24 However, there are arguments to withhold care to this population, and some of these arguments stem from reform of immigration policies and anti-immigrant biases, not from informed public health interests per se.25 Currently, many immigrants access prenatal care through services such as the Children’s Health Insurance Program and locally funded state programs, and service is provided regardless of documentation status. In particular, the safety net of care provided by the Children’s Health Insurance Program is responsible for coverage of 36,000 pregnant women in Texas alone.26 The Further Continuing Appropriations Act, 2018, passed into law December 21, 2017, provides $2.85 billion in funding allotments for Children’s Health Insurance Program for the first half of 2018.27
Our data demonstrate, coincident with increasing anti-immigration statements as determined by increasing popularity of web searches of “Make America Great Again,” “Mexico Wall,” and “Deportation,” that there have been significant decreases in the numbers of routine prenatal care visits and a delay in prenatal care establishment for pregnant women of Central American and Mexican origin. This trend is not observed for women of South American origin, who represent a much smaller percentage of the currently recognized 11 million undocumented population in the United States.28 However, given the much smaller number of South American women in our cohort, we likely lack the adequate power necessary to make comparisons to or generalizations about this demographic. It is important to note that over this study interval, and temporally coincident with the Affordable Care Act, we observed a decrease in the prevalence and incidence of insufficient and inadequate prenatal care among all cohorts except Hispanic U.S. non-native patients. Whether it is fear or social isolation driving these women away from seeking timely and sufficient prenatal care, numerous studies have demonstrated a relationship between lack of prenatal care and an increased risk for poor prenatal outcomes such as low birth weight and preterm delivery.19,21,29
A trend to a decrease in the hemoglobin nadir after the increased shift in rhetoric was also noted within the Hispanic U.S. non-native population. This suggests that this population is likely not obtaining appropriate routine iron supplementation recommended for patients with hemoglobin less than 11 g/dL during pregnancy. Regardless of the availability and accessibility of services, our research adds to a growing body of evidence showing that recent political anti-immigration sentiments are being heard by our patients and that immigrant populations are avoiding, not seeking, and ultimately not receiving recommended care during pregnancy.
Our study's primary limitation is the relative fewer number of cases since the rhetoric inflection point. Thus, detection for morbidity and mortality is likely relatively underestimated and underpowered, therefore limiting our conclusions. Our study is additionally limited by its inability to determine causation and inability to capture patients who go on to deliver in other hospitals or have no prenatal care. There may be unmeasured and unaccounted for confounding, which we have not considered, and we thus make no statements regarding causality and rather present our findings as temporal associations. Influence of prior prenatal care experience and outcomes that may have in subsequent pregnancy fall outside the scope of this article and would be of interest for investigation. Here, prior prenatal experience and access to care must be acknowledged as a potential occult confounder.
Our study is strengthened by application of a rigorous tool for determination of adequate compared with inadequate prenatal care, the Adequacy of Prenatal Care Utilization Index. This accounts for initiation of care and gestational age at delivery and thus appropriately measures prenatal visits relative to the total number of weeks in which it could be maximally received. Our study has additional strengths, including use of 15 days sliding windows relative to the rhetoric date inflection point and fit of both linear regression and application of logistic regression. This enabled us to optimally detect temporal associations with rigorous multivariate statistical measures. Finally, we benefit from direct questioning of patients under consent. Our consent rate approximates 96%, and we use research coordinators native and fluent in Spanish.
In sum, in this study we demonstrate the temporal co-occurrence of measurable health inequities in association with political rhetoric. To improve and broaden access during pregnancy, thereby limiting potential social disparities, we would concur with recent suggestions aimed at universally adopting and promoting a climate of engagement and promotion of health care regardless of nationality and documentation status. Given the close association with political rhetoric, platforms for positive promotional activities could include social media, written materials, and public health addresses in multiple community, regional, and national forums aimed at the populations most affected.
This study makes no statements nor does it address the immigration debate per se, and it is not the goal nor intent of the investigators to proffer opinion nor statement of fact regarding immigration reform. Rather, it is our intent to bring to light the associated effects of such discussions and their accompanying political rhetoric on timing and receipt of prenatal care among a population that will stay in the United States to deliver their newborn. Additionally, generalizability of our conclusions to Hispanic or Latino populations outside of Texas would require further investigation. With that said, it is evident that there is an imminent need for immigration reform yielding transparent and acceptable policies by the public at large. Until such policies are constructed and implemented, there is an evident risk that political rhetoric will continue to bear a significant influence on health disparities in the United States. In this report, we have documented the effect of such rhetoric as a significant decrease in the numbers of routine prenatal care visits and a delay in prenatal care establishment for pregnant women of Central American and Mexican origin. As had been documented for decades, insufficient access and receipt of prenatal care only increase the occurrence and severity of maternal and infant morbidity and mortality and propagates health disparities. In so much as obstetrician–gynecologists are acknowledged advocates of equitable care for women and their infants, it is incumbent on us to provide objectively acquired scientific data to better inform public discourse. It is our hope that ours and others’ data may be used by public health policy experts to design, enable, implement, and enact well-informed policies with a unified goal of not allowing rhetoric to dictate health outcomes and to mitigate health disparities whenever possible.
1. Bryant AS, Worjoloh A, Caughey AB, Washington AE. Racial/ethnic disparities in obstetrical outcomes and care: prevalence and determinants. Am J Obstet Gynecol 2010;202:335–43.
2. MacDorman MF, Callaghan WM, Matthews TJ, Hoyert DL, Kochanek KD. Trends in the preterm-related infant mortality by race and ethnicity, United States 1994–2004. Int J Health Serv 2007;37:635–41.
3. World Health Organization. Social determinants of health. Available at: http://www.who.int/social_determinants/sdh_definition/en/
. Retrieved October 2, 2018.
4. Vargas Bustamente A, Fang H, Garza J, Carter-Pokras O, Wallace SP, Rizzo JA, et al. Variations in healthcare access and utilization among Mexican immigrants: the role of documentation status. J Immigr Minor Health 2012;14:146–55.
5. Ku L, Matani S. Left out: immigrants access to health care and insurance. Health Aff (Milwood) 2001;20:247–56.
6. Su D, Pratt W, Stimpson JP, Wong R, Pagán JA. Uninsurance, underinsurance and health care utilization in Mexico by US border residents. J Immigr Minor Health 2014;16:607–12.
7. Wolff H, Epiney M, Lourenco AP, Costanza MC, Delieutraz-Marchand J, Andreoli N, et al. Undocumented migrants lack access to pregnancy care and prevention. BMC Public Health 2008;8:93.
8. WHO sustainability health related targets. 2016. SDG health and health-related targets. Available at: http://www.who.int/gho/publications/world_health_statistics/2016/EN_WHS2016_Chapter6.pdf
. Retrieved October 2, 2018.
9. U.S. Department of Health and Human Services. Child health USA 2013. Available at: https://mchb.hrsa.gov/chusa13/introduction.html
. Retrieved October 2, 2018.
10. Cantor AG, Bougatsos C, Dana T, Blazina I, McDonagh M. Routine iron supplementation and screening for iron deficiency anemia in pregnancy: A systematic review for the U.S. Preventive Services Task Force. Ann Intern Med 2015;162:566–567.
11. USPSTF review. Screening for iron deficiency anemia in childhood and pregnancy: update of the 1996 U.S. Preventative Services Task Force review. Available at: file:///C:/Users/megwh/Downloads/ironscrev.pdf
12. Garcia AS, Keyes DG. Life as an undocumented immigrant: how restrictive local immigration policies affect daily life. Washington, DC: Center for American Progress; 2012.
13. Rhodes SD, Mann L, Simán FM, Song E, Alonzo J, Downs M, et al. The impact of local immigration enforcement policies on the health of immigrant hispanic/latinos in the United States. Am J Public Health 2015;105:329–37.
14. Amnesty International. Deadly delivery: the maternal health care crisis in the USA. Available at: http://www.amnestyusa.org/sites/default/files/pdfs/deadlydelivery.pdf
. Retrieved October 2, 2018.
15. MacDorman MF, Declercq E, Cabral H, Morton C. Recent increases in the U.S. maternal mortality rate: disentangling trends from measurement issues. Obstet Gynecol 2016;128:447–55.
16. Antony KM, Hemarajata P, Chen J, Morris J, Cook C, Masalas D, et al. Generation and validation of a universal perinatal database and biospecimen repository: PeriBank. J Perinatol 2016;36:921–9.
17. Alexander GR, Kotelchuck M. Quantifying the adequacy of prenatal care: a comparison of indices. Public Health Rep 1996;111:408–18.
18. Kotelchuck M. The Adequacy of Prenatal Care Utilization Index: its US distribution and association with low birthweight. Am J Public Health 1994;84:1486–9.
19. Debiec KE, Paul KJ, Mitchell CM, Hitti JE. Inadequate prenatal care and risk of preterm delivery among adolescents: a retrospective study over 10 years. Am J Obstet Gynecol 2010;203:122.e1–6.
20. VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med 2017;167:268–74.
21. Loftus CT, Stewart OT, Hensley MD, Enquobahrie DA, Hawes SE. A longitudinal study of changes in prenatal care utilization between first and second births and low birth weight. Matern Child Health J 2015;19:2627–35.
22. Lu MC, Lin YG, Prietto NM, Garite TJ. Elimination of public funding of prenatal care for undocumented immigrants in California: a cost/benefit analysis. Am J Obstet Gynecol 2000;182:233–9.
23. Brown B. Cutting public funding for undocumented immigrants’ prenatal care would raise the cost of neonatal care. Fam Plann Perspective 2000;32:145–6.
24. Dubard CA, Massing MW. Trends in emergency Medicaid expenditures for recent and undocumented immigrants. JAMA 2007;297:1085–92.
25. Bixby R. Choosing our words and futures wisely: political rhetoric and prenatal care policy for inclusion of women with undocumented status. J Midwifery Womens Health 2011;56:91–3. Retrieved October 2, 2018.
26. Choi M, Livingston A. More than 400,000 Texans' insurance at risk after Congress fails to renew CHIP. The Texas Tribune; 2017. Retrieved October 2, 2018.
27. Medicaid.gov. Medicaid CHIP funding. Available at: https://www.medicaid.gov/chip/financing/index.html
28. Migration Policy Institute. Profile of the unauthorized population: United States. Available at: https://www.migrationpolicy.org/data/unauthorized-immigrant-population/state/US
29. Guillory VJ, Lai SM, Suminski R, Crawford G. Low birth weight in Kansas. J Health Care Poor Underserved 2015;26:577–602.