Introduction

Preterm birth (PTB) rates among African-American (“Black”) women have been higher than among European-American (“White”) women in the United States for decades [1, 2]; the causes are poorly understood [1, 3]. For at least two decades, researchers have hypothesized a role for maternal stress in PTB [1–10]. A causal role for stress is biologically plausible based on physiologic mechanisms—including neuroendocrine, vascular, inflammatory, and immune processes—thought to be involved both in stress responses [11] and the initiation of labor [12]. O’Campo, Dunkel Schetter and colleagues [13] have observed higher allostatic load scores among postpartum African American women compared with their White counterparts [13]; allostatic load refers to a set of biomarkers believed to reflect cumulative physiologic wear and tear due to chronic stress and the body’s attempts to adapt [11]. Epigenetic phenomena triggered by chronic stress also have been hypothesized as a contributor to PTB [14, 15]. Chronic stress, even without dramatic events, may be particularly damaging to health [11] it predicts adult chronic disease [11] through mechanisms also thought to be involved in preterm birth [12]. Studies of the association between maternal stress and PTB have produced inconsistent results, however [16–21]. This may reflect the diverse measures of stress and its timing (e.g., lifetime vs during pregnancy and when during pregnancy) used by different studies, along with the complexity of links between stress and PTB, which may be highly dependent on the social and environmental context [6]. While some studies have focused on stress across a woman’s lifetime [6, 16, 22–24], most have focused on stress during pregnancy.

Based on a likely role for stress in PTB, several scholars have hypothesized that experiencing racial discrimination (race-based unfair treatment) may contribute to the Black-White disparity in PTB through physiologic responses to stress [2, 8, 12, 25]. Studies of the relationship between experiences of racial discrimination and the racial disparity in PTB have not been consistent, however. Six of nine published studies on perceived racial discrimination and PTB reported positive associations [5, 26–30] while three did not [31–33]. All of these studies examined the effects of incidents of race-based unfair treatment; none measured worry, anxiety, or the related concept, vigilance (heightened alertness anticipating discriminatory treatment). Vigilance has been linked with many adverse health outcomes [34] but, to our knowledge, has not been studied in relation to birth outcomes.

Experiences of racial discrimination can take multiple forms, including not only overt or subtle discriminatory incidents, but also a range of psychological and emotional reactions to incidents or to awareness of the existence of racism. Even without personally experiencing incidents, individuals may worry or be anxious about potential race-based discrimination based on awareness of pervasive and historical discrimination. People may experience vigilance or concern about confirming negative stereotypes about their social group [35]. They may ruminate about past experiences or experiences of loved ones or of their racial/ethnic group in general, or feel they must perform exceptionally to disprove prejudicial assumptions. Diverse psychological responses to specific incidents or to general awareness of racism, e.g., worry, anger or self-doubt, may occur [8, 34, 36, 37]. We conducted an earlier qualitative study to inform the measurement of experiences of racial discrimination in birth outcomes research [38]. Using semi-structured guides, African-American facilitators asked 40 reproductive-age African-American women in six focus groups to describe their experiences of racial discrimination, racism, or being treated unfairly based on their race. While participants described overt, subtle, and ambiguous incidents, they gave at least as much emphasis to pervasive worry that they or loved ones would experience race-based insults or slights and the need to remain perpetually on guard. Building on those findings and the literature, this study investigated the potential role of chronic worry about race-based unfair treatment (“chronic worry about racial discrimination” or simply “chronic worry”) in the Black-White disparity in PTB, with full awareness that chronic worry about racial discrimination represents only one of many kinds of experiences of racial discrimination that could damage health through pathways involving stress. This study focuses on women’s everyday experiences of racial discrimination in any setting, not specifically on experiences with seeking or receiving health care.

Methods

Results

Table 1 describes Black and White women’s characteristics and how the prevalence of chronic worry about racial discrimination varied with those characteristics. Over one-third (36.9%, 95% CI 32.9–40.9%) of Black women and 5.5% (CI 4.5–6.5%) of White women reported chronic worry. Among Black (but not White) women, those who were older, married, and had higher income or education levels were more likely to report chronic worry. For White women, lower levels of income and education were associated with reporting chronic worry. Most medical variables were not associated with chronic worry, but among Black women, those who reported being diagnosed with hypertension before pregnancy were more likely to report chronic worry.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Self-reported characteristics, and prevalence of chronic worry about racial discrimination (chronic worry) by those characteristics, among eligible U.S.-born non-Latino Black and White women with singleton live births in California, MIHA 2011–2014.

https://doi.org/10.1371/journal.pone.0186151.t001

Table 2 displays numbers and rates of PTB among Black women and White women overall and according to whether they reported chronic worry. Overall, 9.2% (CI 7.2–11.2%) of Black and 5.8% (CI 4.8–6.8%) of White women had a PTB. Among Black women, PTB occurred among 12.5% (CI 8.4–16.7%) of those reporting chronic worry and 7.2% (CI 5.3–9.2%) who did not. Among White women, 9.9% (CI 2.6–17.1%) of those reporting chronic worry about racial discrimination had a PTB, compared with 5.6% (CI 4.6–6.5%) of those who did not. (S3 Table includes the information in Table 2 and displays below it the numbers and rates of PTBs according to each of the social/demographic, behavioral, and medical/medical care covariates. PTB rates appeared to vary by many characteristics, but confidence intervals generally overlapped. Patterns by age, income, and education tended to mirror observations in the literature, with apparently higher rates of PTB among Black women (but not among White women) with higher levels of income. The most striking differences in PTB rates by characteristics were seen for binge drinking among White women, and for diagnoses of diabetes or hypertension before pregnancy among Black women.)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Prevalence of preterm births overall by whether women reported chronic worry about racial discrimination (chronic worry) among U.S.-born non-Latino Black and White women with singleton live births in California, MIHA 2011–2014.

https://doi.org/10.1371/journal.pone.0186151.t002

Separately for Black women and for White women, Table 3 shows prevalence ratios (PR) and 95% Confidence Intervals (CI) for the association between chronic worry about racial discrimination and PTB, first unadjusted, then adjusted for each cluster of covariates, and then adjusted for all covariates. Without adjustment, the PR comparing PTB rates among Black women reporting chronic worry with those who did not was 1.73 (CI 1.12–2.67); the PR among Black women was similar after adjustment for the social/demographic covariates alone (1.95, CI 1.27–2.97) and all covariates (PR 2.00, CI 1.33–3.01). Corresponding estimates for White women appeared similar but slightly lower than PRs among Black women after adjustment; none was statistically significant. (S4 Table shows the results in Table 3 and in addition for all variables in the model corresponding to Table 3. On examining the PRs for the association between each covariate and PTB adjusting for all other covariates, the only variables that were significantly associated with PTB among Black women were neighborhood poverty (with greater neighborhood poverty rates paradoxically protective), fair or poor prepregnancy health, hypertension before pregnancy, and inadequate weight gain. For White women the only significant variables were parity, marital status, neighborhood poverty, binge drinking, and diabetes before pregnancy.)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Prevalence ratios for preterm birth associated with chronic worry about racial discrimination among U.S.-born non-Latino Black and White women with singleton live births in California, MIHA 2011–2014.

https://doi.org/10.1371/journal.pone.0186151.t003

Table 4 displays results of analyses examining the Black-White disparity in PTB. The unadjusted PR for PTB rates among Black women compared with White women was 1.59 (CI 1.21–2.09). After adjusting for chronic worry about racial discrimination, the Black-White PTB difference appeared attenuated and was no longer statistically significant (PR 1.30, CI 0.93–1.81); it was 1.08 (CI 0.76–1.54) after adding the social/demographic variables, then 1.17 (CI 0.85–1.63) after adding the behavioral and medical variables. (S5 Table includes the information in Table 4 and in addition displays the PRs for each covariate. S6 Table shows that adding only the behavioral and medical variables without the social/demographic variables to the model produced negligible changes in the PR adjusted only for chronic worry.)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Prevalence ratios comparing PTB prevalence among U.S.-born non-Latino Black relative to White women with singleton live births in California, before and after adjustment for chronic worry about racial discrimination and covariates, MIHA 2011–2014.

https://doi.org/10.1371/journal.pone.0186151.t004

Discussion

Results of this large, statewide-representative study suggest that chronic worry about racial discrimination may play an important role in the Black-White disparity in PTB. Although this study cannot establish a causal role for chronic worry in the elevated risk of PTB among Black women, a causal role is plausible based on current understanding of the physiologic mechanisms through which chronic stress could contribute to premature birth [12]. These results add to growing evidence linking adverse health outcomes with a range of experiences of racial discrimination, including diverse psychological and emotional states as well as incidents [34, 46, 47]. The results add to literature linking racial discrimination specifically with PTB [5, 26–30] by suggesting a possible link between PTB risk and a psychological state—namely, chronic worry about the possibility of being treated unfairly based on one’s race.

Among the more than one in three Black women who reported chronic worry about race-based unfair treatment, the prevalence of PTB was approximately twice that among Black women who did not report chronic worry. This difference persisted even after adjustment for many other factors that would tend to diminish the observed associations with chronic worry. If chronic worry about racial discrimination does contribute causally to increased rates of PTB among Black women, the potential impacts—in terms of suffering, unfulfilled human potential because of the developmental sequellae of PTB, and economic costs for society—could be staggering.

A limitation of this study is the fact that our measure of chronic worry about racial discrimination has not been formally tested psychometrically. It was, however, systematically developed based on qualitative research [38] (see Introduction) and pretested cognitively and in focus groups. Another limitation is that we did not have information on the extent to which women worried about other issues, apart from racial discrimination. Our finding, however, that Black women, as expected, reported chronic worry about racial discrimination far more frequently than White women supports the likelihood that this measure captures worry about racial discrimination rather than more generalized anxiety or worry. Furthermore, the observed patterns of association between our measure of chronic worry and other variables in this study (e.g., education and income) are consistent with previous findings linking experiences of discrimination with depressive symptoms [34, 37] smoking [48] and other unhealthy behaviors [37]. In addition, the higher prevalence of chronic worry about racial discrimination observed here among more socioeconomically advantaged Black women is consistent with findings from previous research on perceived incidents of discrimination [34].

Another key limitation is that while our data source included only one question on experiences of discrimination, i.e., worry about racial discrimination, racism-related experiences can encompass diverse experiences [38], as reflected by multi-component scales [36, 49–51]. One cannot be certain whether worry about racial discrimination, as opposed to other unmeasured racism-related psychological states or incidents that are associated with worry, was responsible for the associations observed in this study. It is highly likely, however, that the findings reflect a link between PTB and one or more experiences of racism.

Another potential limitation is that given the retrospective nature of this study, the observed association between chronic worry about racial discrimination and PTB may reflect recall bias. If, however, having experienced PTB made women subsequently more likely to recall or report prior adverse experiences (such as worry about racial discrimination at any time in the past), one might expect to observe associations between PTB and women’s reports of other adverse events or circumstances such as depressive symptoms and major stressors during pregnancy. In this study, we found no association between PTB and depressive symptoms among Black or White women. Of the nine major stressors that were examined, furthermore, none was associated with PTB among Black women, and only two were associated with PTB among White women (shown in S7 Table).

Our findings also suggest that White women who report chronic worry about racial discrimination may be at elevated risk of PTB. One can only speculate about what the U.S.-born non-Latino White women in this study meant when they reported worrying about being treated or judged unfairly based on their race or ethnic group. Did they experience racial tensions or social exclusion in their neighborhoods or workplaces, or resent what they perceived as others’ more favorable treatment? These are not unlikely possibilities given that reporting of chronic worry about racial discrimination was more prevalent among White women of lower educational and income levels. Although experiences of racial discrimination are likely to be different for Black and White women, we believe it is valid not only to examine potential effects of worry about racial discrimination among Black women but also to assess its potential contributions to the Black-White disparity in PTB, which requires including White women in analyses. This analytic approach is supported by previous literature [8, 28, 48, 51].

We considered the possible role of multiple racial identifications among women whose first-listed race was White. Overall, 3.2% of White women reported a second or third non-White race; for less than 1% (n = 71), Black/African-American was listed as a second or third race (data on maternal and paternal race among White women. Preterm birth rates did not differ between White women who also reported Black or other races and those who reported only White race. (S8 Table) We also considered whether White (as first-listed race) women whose babies had fathers who were Black or of other racial/ethnic groups, may have worried more frequently about discrimination because of their inter-racial relationships. Based on birth certificate data on paternal as well as maternal race/ethnicity, the 281 out of 8,122 White (as first-listed race) women with partners whose first, second, or third race was listed as Black, had higher rates of chronic worry about racial discrimination than White women with White partners; however, these women did not have higher rates of preterm birth than did women with non-Latino White partners.(shown in S8 Table)

The socioeconomic variation observed in the prevalence of chronic worry about race-based unfair treatment may shed light on a perplexing finding from some earlier studies. While PTB rates among White women generally decrease with more income or education, this has not been observed consistently among Black women [52–56], suggesting that Black women have not always reaped the expected health benefits (e.g., improved birth outcomes) of greater socioeconomic resources. This is consistent with the findings of other researchers [57–59] who have observed that African Americans appeared to derive less health benefit from higher levels of educational attainment than Whites. Our findings of both higher rates of chronic worry about racial discrimination among more socioeconomically-advantaged Black women and an association between chronic worry about racial discrimination and PTB may help explain this apparent paradox. It is conceivable that Black women of higher income or education levels could worry more about racial discrimination than their less socioeconomically-advantaged Black counterparts because they may interact more frequently with Whites; they may be more likely to be in the minority in their workplaces and potentially where they shop, study, reside, and obtain services [60]. They may feel–and may have felt for a long time—that they constantly need to prove themselves against negative stereotypes. Black women of higher socioeconomic status may therefore paradoxically experience more chronic racism-related stress. Cole and Omari [61] discussed the “hidden costs of upward mobility” for middle-class African Americans, including encountering glass ceilings to advancement at work, potential alienation from other African Americans, and facing frequent demands for financial support from less well-off family and friends, all of which could be stressful.

While this study was not designed to reveal the underlying mechanisms for the relationship between chronic worry about racial discrimination and higher risks of PTB among Black women, our findings indicate that this relationship was not explained by the health-related behaviors or medical factors we measured. These findings are limited, however, by concerns about the reliability of self-reports of medical diagnoses. It is possible that unmeasured differences in factors such as economic or social hardships during childhood, stressful experiences before pregnancy, neighborhood characteristics apart from poverty rates, accumulated wealth, or unmeasured aspects of any covariate may be important.

The persistence in some studies of a Black-White disparity in PTB after adjustment for socioeconomic variables has sometimes been interpreted by default to reflect underlying biological differences. Our findings, however—that the magnitude and statistical significance of Black-White differences in PTB prevalence appeared markedly attenuated after adjusting for chronic worry about racial discrimination, with and without other covariates—suggest an important role for social factors, many of which are rarely measured. There is a growing consensus among researchers that the causes of PTB are likely to be complex and multifactorial, and this study’s findings do not challenge that assumption; they do, however, indicate a potentially important role in the Black-White disparity in PTB for Black women’s experiences of racial discrimination. While neither causally definitive nor indicative of an easy medical or policy “fix,” our results add to mounting evidence of the likely toll of racism-related stress on important physical health outcomes in general [46, 47, 62] and specifically suggest a potentially major role for racism-related stress in the racial disparity in PTB, which should be further studied. The findings underscore the need, when studying health disparities by racial or ethnic group, to consider unmeasured social differences, including differences in potentially powerful chronic stressors such as experiences of racial discrimination. Given the difficulty of measuring all relevant social variables in any research, this study highlights the need for awareness that underlying biological differences should not be a default explanation of racial differences in health.

Supporting information

Acknowledgments

The authors wish to thank Melanie Dove, ScD, MPH, for her assistance with the analyses, Amy Edmonds and Julia Acker for their assistance with the research, and Connie Mitchell, MD, MPH, and Shabbir Ahmad, DVM, MS, PhD, for their helpful comments on a draft and ongoing support for the Maternal and Infant Health Assessment and its use to inform policy and practice.