Perceived Racism Associated with Declines in Self-rated Oral Health among U.S. Black Women

Abstract

Background:

Racial disparities in oral health are well-documented. Stress has been associated with both perceived racism and oral health, yet little research has directly investigated the association between perceived racism and oral health.

Methods:

We used data from the Black Women’s Health Study, a longitudinal cohort study that includes a geographically diverse sample of Black women across the US. Perceived exposure to racism was assessed via two scales, one assessing lifetime exposure and one everyday exposure. Self-rated oral health was subsequently assessed over multiple time points. We used Cox proportional hazard models to calculate adjusted incidence rate ratios (IRRs) estimating the association between higher levels of perceived racism and incident “fair” or “poor” oral health, and explored potential effect measure modification using stratified models.

Results:

The adjusted IRRs (n=27,008) relating perceived racism to incident fair/poor oral health were 1.50 (95% confidence interval [CI] 1.35,1.66) comparing the highest quartile of everyday racism to the lowest and 1.45 (95% CI 1.31, 1.61) for the highest score of lifetime racism compared to the lowest. We did not see evidence of effect modification.

Conclusion:

Higher levels of perceived racism documented in 2009 were associated with declines in self-rated oral health from 2011 to 2019.

INTRODUCTION

Racial disparities in oral health are widely documented in the United States;(1–4) Blacks experience higher rates of tooth decay, root caries, tooth loss, edentulism,(1, 4) and periodontal disease (2, 4) than Whites. These observed differences are only partially explained by differences in income, education (5–7), and dental insurance coverage(4). Further, the inequitable distribution of poor oral health outcomes, including dental caries, periodontal disease, edentulism, and oral cancers, can be observed throughout the globe and contribute to significant health and economic burdens (8). In addition, conditions such as periodontitis and tooth loss can be painful, adversely impact diet, and have unwanted social consequences (e.g., anxiety) (9). Black women, specifically, experience higher rates of adverse health outcomes associated with periodontal disease, including cardiovascular disease(10), type 2 diabetes(11), and hypertension(12). Thus, exploration of oral health disparities is of particular significance.

Racial disparities in health may be driven in part by inequitable access to needed health care(13), racially biased delivery of the care that is received(14), as well as by the health consequences of enduring racism. Specifically, perceived racial discrimination is an important source of psychosocial stress in the lives of Black women,(15,16) and evidence suggests that stress is a contributor to poor oral health. Finlayson et al. found that chronic stress was associated with poorer self-rated oral health in a probability sample of Black people in the US.(17) In spite of this, little research has directly evaluated a link between racism and poor oral health. Further, research that attempts to elucidate connections between racism and health outcomes often uses self-identified racial or ethnic identity as a proxy for exposure to racism, as opposed to a direct measure, which is a decidedly limited operationalization of racism(18,19). Thus, we sought to assess perceived experiences of racism in relation to declines of self-rated oral health within the Black Women’s Health Study (BWHS) over ten years of study follow-up.

METHODS

The BWHS is a longitudinal observational cohort study that began in 1995 when 59,000 Black women ages 21 to 69 years (median age, 38 years) from across the U.S. completed postal questionnaires mailed to Essence magazine subscribers, members of professional organizations, and friends and family of early enrollees. Over 80% of women resided in California, Georgia, Illinois, Indiana, Louisiana, Maryland, Massa0chusetts, Michigan, New Jersey, New York, South Carolina, Virginia, and the District of Columbia. Participation is maintained through biennial mailed and web health questionnaires, which collected demographic, medical and reproductive history, health behaviors, diet, smoking, and various psychosocial data; deceased subjects are identified through the National Death Index, postal service, and relatives/friends.(20, 21) Long-term follow-up of the baseline cohort has been successful for >85% of potential person-years through 2015. The Institutional Review Board of Boston University Medical Center approved the study. Participants indicate their consent by filling out the questionnaires.

Perceived Racism

Experiences of racism were assessed in 2009 using two sets of questions adapted from an instrument developed by Williams et al. (22) The first consisted of five questions about everyday race-related discrimination, intended to measure chronic, interpersonal experiences of racism (“everyday racism”). Specifically, participants were asked about the frequency in daily life of the following experiences: “You receive poorer service than other people in restaurants or stores,” “People act as if they think you are not intelligent,” “People act as if they are afraid of you,” “People act as if they think you are dishonest,” and “People act as if they are better than you.” Response options were “never,” “a few times a year,” “once a month,” “once a week,” and “almost every day,” coded as 1 through 5. An everyday racism score was created by averaging subjects’ responses to the five questions, and creating quartiles based on the distribution of the averaged scores. The second set consisted of three questions which ascertained exposure to racism over one’s lifetime (“lifetime racism”) by asking whether the participant was ever “treated unfairly due to your race” on the job, in housing, and by the police. Response categories were “yes” and “no.” A lifetime racism score summed the number of positive responses to the three questions (0, 1, 2, or 3).

Self-Rated Oral Health

In 2011, 2015 and 2019, participants were asked: “Overall, how would you rate the heath of your teeth and gums?” Response options were “excellent”, “very good”, “good”, “fair”, and “poor”. This measure of self-rated oral health is frequently used in large, population-based studies where it is not feasible for participants to be clinically examined; numerous validation studies(23–25) have found that this measure is broadly associated with clinical assessments of dental status, including in this cohort(26). In 2015, we conducted an internal validation of self-reported oral health measures among BWHS participants residing in the state of Massachusetts. A total of 77 women who completed a supplemental oral health questionnaire between November 2014 and May 2015, lived within commuting distance to the Boston University Medical Campus, and reported having at least 8 natural teeth present responded to an invitation to attend the Center for Clinical Research at the Henry M. Goldman School of Dental Medicine for a clinical evaluation of periodontal health status. Among clinically-determined moderate-to-severe cases of periodontal disease, the estimated sensitivity for self-rated oral health was 84% with perfect specificity(26).

Covariates

Data on age, cigarette smoking, current weight, history of type 2 diabetes, geographical region of residence, and neighborhood socioeconomic status were collected at baseline (1995) and on each follow-up questionnaire. Parity was assessed at baseline and on each follow-up questionnaire through 2011. Height was obtained on the baseline questionnaire (1995); years of completed education was measured at baseline and again in 2003. Body mass index (BMI) (weight divided by height² (kg/m²)) was calculated with weight data collected at each follow-up cycle. Participants were asked whether they had a dental cleaning within the previous two-year period in 2007 and on all subsequent questionnaires. Finally, two variables from the 2009 questionnaire assessed participants’ style of coping with racist or discriminatory events (27): 1) “If you feel you have been treated unfairly due to your race, do you”: a) Usually accept it as a fact of life? or b) Usually try to do something about it? (“coping – acceptance”); and 2) “If you have been treated unfairly due to your race, do you”: a) Usually talk to other people about it? or b) Usually keep it to yourself? (“coping – discussion”). Participants were asked to select the single best answer for each question.

Data Analysis

The outcome of interest for this analysis was the incidence of self-rated ‘fair/poor’ oral health. At the analytic baseline in 2011, a total 32,265 women had complete data on oral health and were considered eligible for the outcome. Those who were missing data on either racism question (everyday racism, n=3,714; lifetime racism, n=1,543) were excluded from the analysis, resulting in an analytic sample of 27,008 women.

Person-time was calculated from the start of follow-up in 2011 until the report of fair or poor oral health, loss to follow-up, death, or end of follow-up on December 31st 2020 (the end of the 2019 data collection cycle), whichever happened first. Adjusted Cox models with time-varying covariates, a form of non-proportional hazards(28,29), were used to estimate the incidence of fair/poor self-rated oral health over each data collection cycle interval (2011–2015, 2015–2019). Inclusion of covariates was determined based on the potential for them to be a common cause of the exposure and outcome (e.g., potential confounders), as well as those known to be strong predictors of the outcome (e.g., cigarette smoking). Thus, multivariable models were adjusted by age (years), BMI (<25, 25–29, ≥30, missing/unknown kg/m2), education (≤12, 13–15, 16, >16 years), cigarette smoking (never, past, current), parity (nulliparous, 1, 2, ≥3), neighborhood socioeconomic status (quintiles, missing/unknown), U.S. geographical region (Northeast, South, Midwest, West), and history of diabetes (yes, no). All covariates were treated as time-varying with the exception of education. Covariates were selected for inclusion in the adjusted models if supported by literature or if they changed the effect modeled by more than 10%.

In order to assess potential effect measure modification, we further explored the association within strata of age (<50, ≥50 years), parity (nulliparous, parous), and coping style – accept (accept, do something), and discuss (talk about, keep to oneself). All analyses were performed using SAS, version 9.4, software (SAS Institute, Inc., Cary, North Carolina).

RESULTS

Overall, study subjects had a mean age of 53.8 years (range: 36 – 86). Twenty-one percent and 11% of the sample reported experiencing the highest levels of everyday and lifetime racism, respectively, as shown in Table 1. Higher education, higher BMI, and ever-smoking were associated with more frequent experiences of racism. Independent of the level of racism experienced, a majority of women reported either “trying to do something” or “talking to other people” as coping strategies.

Table 1.

Age-Standardized¹ Baseline Characteristics in 2011 (N=27,008)

	Everyday Racism		Lifetime Racism
	Quartile 1 (least racism experienced) n (4955)	Quartile 4 (most racism experienced) n (5637)	No to all 3 racism questions n (9818)	Yes to all 3 racism questions n (3094)
Age (years) (Mean (SD))	56.3 (10.9)	51.4 (8.7)	53.2 (10.3)	54.9 (9.1)
Body Mass Index (kg/m2) (%)
<25	26	19	23	21
25–30	27	24	26	25
≥30	41	51	44	47
Education (years) (%)
≤12	13	10	15	6
13–15	26	26	27	23
16	29	26	27	27
>16	32	38	31	45
Neighborhood Socioeconomic Status (quintiles) (%)
Lowest quintile	19	19	20	16
Middle quintile	18	18	18	18
Highest quintile	18	18	16	21
Missing	7	9	8	9
Geographic Region (%)
Northeast	23	22	22	23
South	41	35	42	30
Midwest	20	22	19	23
West	16	21	16	23
Parity (number) (%)
Nulliparous	25	26	24	25
1	27	27	28	26
2	28	26	28	27
≥3	19	21	19	22
History of Type 2 Diabetes (%)	14	17	16	14
Cigarette Smoking (%)
Current	7	10	8	8
Past	22	26	22	28
Never	71	65	70	64
Recent Dental Cleaning (%)	82	81	81	83
Coping – Accept (%)
Accept racism as a fact of life	23	38	31	33
Try to do something about racism	69	61	63	66
Missing	9	1	6	1
Coping – Discuss (%)
Talk to others about experience(s)	84	86	84	89
Keep experience(s) to oneself	7	13	9	10
Missing	9	1	6	1

¹All variables, except for age, were standardized to the 2011 age distribution of the BWHS sample.

Between 2011 and 2019, 3,820 participants reported incident fair/poor oral health during 203,551 person-years of follow-up (mean=7.5 person-years). As the frequency of both everyday and lifetime racism increased, the incidence of fair/poor self-rated oral health increased (Table 2). Compared to the lowest quartile of everyday racism, the unadjusted incidence rate ratio (IRR) for the highest quartile was 1.57 (95% confidence interval [CI]: 1.41–1.74). Adjustment for covariates slightly attenuated the IRR to 1.50 (1.35 – 1.66). Lifetime racism was similarly positively associated with incident fair/poor self-rated oral health. Compared to women who reported “no” to all three lifetime racism questions, the crude and adjusted estimates for women who reported “yes” to all were 1.36 (95% CI: 1.23–1.51) and 1.45 (95% CI: 1.31–1.61), respectively.

Table 2.

Incidence Rate Ratios (IRRs) and 95% Confidence Intervals (CI) obtained from Cox Regression Models for Measures of Racism in Relation to Self-Rated Oral Health, Black Women’s Health Study (N=27 008)

	IRR (95% CI)
	Cases	Person Years	Crude Model	Adjusted Model1
Everyday Racism
Quartile (Q) 1 (lowest)	601	37 500	1.00 (ref)	1.00 (ref)
Q 2	1143	62 014	1.19 (1.08 to 1.31)	1.19 (1.08 to 1.32)
Q 3	1092	61 959	1.14 (1.04 to 1.27)	1.16 (1.05 to 1.28)
Q 4 (highest)	984	42 078	1.57 (1.41 to 1.74)	1.50 (1.35 to 1.66)
Lifetime Racism
No to All	1266	74 297	1.00 (ref)	1.00 (ref)
Yes to 1	1126	61 899	1.06 (0.98 to 1.15)	1.11 (1.02 to 1.20)
Yes to 2	888	44 339	1.17 (1.07 to 1.27)	1.23 (1.13 to 1.35)
Yes to 3	540	23 014	1.36 (1.23 to 1.51)	1.45 (1.31 to 1.61)

¹Model adjusted for age, history of cigarette smoking, neighborhood SES (Quintiles, missing), body mass index (<25, 25–29, ≥30, missing), geographic region (North East, South, Midwest, West), years of education (≤12 years, 13–15 years, 16 years, >16 years), parity (Nulliparous, 1, 2, ≥ 3), and history of type 2 diabetes (No, Yes).

The association between perceived racism and fair/poor oral health was present within all strata of age, parity, coping/acceptance, and coping/discussion (Table 3).

Table 3.

Incidence Rate Ratios (IRRs) and 95% Confidence Intervals (CI) obtained from Cox Regression Models of Perceived racism (everyday and lifetime) in relation to fair/poor self-rated oral health within strata of age, parity, and approach to coping.

		Everyday Racism Adjusted IRR1 (95%CI)				Lifetime Racism Adjusted IRR1 (95%CI)
	Q 1 (lowest)	Q 2	Q 3	Q 4 (highest)	No to all	Yes to one	Yes to two	Yes to all
Age
<50 years	1.00 (ref)	1.12 (0.92,1.38)	1.08 (0.88,1.33)	1.49 (1.22,1.81)	1.00 (ref)	1.05 (0.91,1.22)	1.26 (1.07,1.48)	1.47 (1.21,1.80)
≥50 years	1.00 (ref)	1.22 (1.09,1.37)	1.18 (1.05,1.33)	1.47 (1.30,1.67)	1.00 (ref)	1.13 (1.02,1.24)	1.22 (1.10,1.36)	1.45 (1.28,1.63)
Parity 2
Nulliparous	1.00 (ref)	1.23 (1.10,1.37)	1.18 (1.05,1.32)	1.52 (1.35,1.71)	1.00 (ref)	1.10 (1.01,1.21)	1.22 (1.01,1.35)	1.41 (1.26,1.59)
Parous	1.00 (ref)	1.08 (0.86,1.35)	1.09 (0.87,1.36)	1.44 (1.15,1.81)	1.00 (ref)	1.12 (0.93,1.33)	1.30 (1.07,1.57)	1.71 (1.37,2.14)
Coping – Acceptance
Accept	1.00 (ref)	1.04 (0.87,1.25)	0.99 (0.82,1.19)	1.42 (1.18,1.71)	1.00 (ref)	1.23 (1.07,1.41)	1.36 (1.17,1.57)	1.36 (1.13,1.63)
Do something	1.00 (ref)	1.26 (1.11,1.43)	1.23 (1.09,1.40)	1.45 (1.27,1.66)	1.00 (ref)	1.03 (0.94,1.15)	1.16 (1.04,1.30)	1.49 (1.31,1.69)
Coping – Discuss
Talk to others	1.00 (ref)	1.20 (1.07,1.33)	1.16 (1.04,1.29)	1.45 (1.29,1.62)	1.00 (ref)	1.10 (1.01,1.21)	1.24 (1.13,1.36)	1.45 (1.30,1.62)
Keep to self	1.00 (ref)	1.02 (0.72,1.43)	1.05 (0.75,1.47)	1.50 (1.08,2.08)	1.00 (ref)	1.01 (0.87,1.42)	1.15 (0.88,1.50)	1.49 (1.10,2.02)

¹Model adjusted for age, history of cigarette smoking, neighborhood SES (Quintiles, missing), body mass index (<25, 25–29, ≥30, missing), geographic region (North East, South, Midwest, West), years of education (≤12 years, 13–15 years, 16 years, >16 years), parity (Nulliparous, 1, 2, ≥ 3), and history of type 2 diabetes (No, Yes).

²Model included all variables in Model 1 except for parity.

DISCUSSION

Using prospective, longitudinal data from a large cohort study, we observed a positive association between experiences of racism and worsening self-rated oral health. Our results are suggestive of a potential “dose response” type relationship, with greater exposure to everyday and lifetime racism having a stronger association with worsening oral health. The observed relationship between experiences of racism and oral health persisted after adjustment for many confounders and stratification by several potential effect measure modifiers.

Our results are generally consistent with existing literature evaluating the relationship between perceived racism and oral health. A study of pregnant Aboriginal Australian women reported a positive association between reporting more experiences of racial discrimination and experiencing a toothache in the past year (30), while a study of pregnant Indigenous Canadian women reported a positive association between experiences of racial discrimination and tooth loss (31). A longitudinal study of immigrants in Canada found a positive association between experiences of discrimination due to race, ethnicity, culture, language or accent, or religion and self-reported presence of an oral health problem (32). On the other hand, one study of Hispanic/Latino adults in the United States reported no association between experiencing unfair treatment due to ethnic identity and clinically confirmed periodontitis (33).

There are multiple potential pathways that could underlie the association between experiences of racism and poorer oral health outcomes. One potential pathway involves oral health being adversely affected by chronic stress caused by social adversity.(34) Experiences of racism are associated with chronic stress.(15, 16) Higher perceived stress has been associated with worse self-rated oral health in a cross-sectional study of community-dwelling adults in Toronto (35) and a national probability study of Black adults in the US.(15) Markers of higher allostatic load, a measure of accumulated chronic stress over the life course, have been positively associated with periodontal disease in data from the National Health and Nutrition Examination Survey.(36, 37) On the other hand, stress due to perceived discrimination has been associated with anxiety and depression(38), which are associated with poorer oral health(39). However, it is also possible that the observed association represents an association between perceived discrimination and perception of oral health status, as opposed to true oral health status. Self-perception may be adversely influenced by the psychological consequences of experiencing racism(38). We believe that this is likely not the primary explanation of our findings due to the high validity of our self-report measure of oral health(23–26); however, we cannot know for sure.

Another potential pathway is that experiences of racism may negatively influence dental health utilization (40). A large body of research has evaluated the relationship between experiences of racism and patient experiences in the healthcare setting and a recent meta-analysis reported that experiences of racism are associated with negative patient experiences in healthcare settings and lower healthcare-related trust. (41) However, there is little research specific to racism in the dental healthcare context. One exploratory study of Indigenous Australians reported a strong positive association between race-related discrimination and never having visited a dentist.(42) A more recent study using a nationally representative sample of United States adults reported no association between healthcare discrimination and having had a dental visit, but a positive association between reporting an emotional impact of experienced discrimination and not having visited a dentist in the past year.(43) The conflicting findings on associations of healthcare discrimination with having a dental visit may reflect the fact that a relationship between discrimination and healthcare utilization may represent a long-term feedback loop, in which dental fear and anxiety results in delayed care, more interventions, pain, and cost, which in turn cause more fear. Discriminatory practices by dentists, including recommending different procedures (44) or pain management strategies (45) based on patients’ race, may contribute to this cycle by exacerbating fear and anxiety and yielding more painful or extensive interventions. In fact, in a recent analysis within a subset of BWHS participants, we observed associations between perceived racism and high levels of dental fear and anxiety, and between dental fear and anxiety and poor oral health outcomes. (46) Among those with high dental anxiety, having an ‘unkind dentist’ or feeling a ‘lack of control’ were more commonly cited as fear-inducing stimuli compared to those with low dental anxiety, supporting the potential presence of the aforementioned cycle. Though the current analysis involves longitudinal data, we may not be able to fully capture this ongoing cycle due to only having information about a dental visit in the past year.

Our study has some important strengths. Due to the longitudinal design, we were able to capture incident fair/poor oral health, which removes some of the temporal issues associated with cross-sectional evaluations of experiences of racism and health outcomes. Additionally, we assessed a large sample of Black women across different regions of the United States, which is novel. Nonetheless, our findings should be interpreted in the context of some important limitations. Our oral health outcome was based on a self-rating of oral health status, as opposed to the gold standard of clinical evaluation. Although self-rated oral health status has been clinically validated in many different populations, including participants in the BWHS, the possibility of outcome misclassification remains. (26) Additionally, because reported oral health status could fluctuate over time, it is possible that BWHS participants were included in/excluded from the analytic cohort based on random fluctuations in self-report. However, such fluctuations could also be representative of true fluctuations in oral health conditions, which are largely chronic and progressive. Thus, the potential for bidirectional changes in self-reported oral health status are not necessarily a limitation of self-report but an expected limitation of studying chronic conditions. For these reasons, we collapsed the 5-point Likert response scale to two commonly used categories of ‘Excellent/Very Good/Good’ vs ‘Fair/Poor’ and our chosen analytic approach considered only the first report of ‘Fair/Poor’ oral health. Racism and discrimination are complex, multi-level phenomena that interact with myriad personal, societal, and institutional structures and impact individuals in diverse ways.(47,48) Our measures of perceived experiences of racism were captured at a single point in time, and also may not fully capture every way in which people can experience racism relevant to oral health outcomes, which may result in misclassification of the exposure. Additionally, there exists the potential for bias due to dependent misclassification owing to the exposure and outcome measures both being self-reported(49). However, dependent error requires both the exposure and outcome to be misclassified. Though theoretically possible, it is difficult to argue that self-reported experiences of racial discrimination could be misclassified, as there is no standard by which to measure selfperception other than asking the individual. Finally, the BWHS is not a representative sample of US Black women, with participants generally having higher education levels than Black women nationally, so our findings may not be generalizable to all Black women in the US.

Despite its limitations, our study contributes important evidence to the growing body of literature assessing a connection between experiences of racism and oral health outcomes. In a large, geographically diverse cohort of US Black women, we found an association between higher levels of everyday and lifetime experiences of racism and incident fair/poor self-rated oral health over approximately ten years of follow up. Our findings suggest an important connection between experiences of racism and oral health outcomes that should continue to be investigated as it has the potential to illuminate generative mechanisms of racial disparities in oral health. Such mechanisms may include inequitable access to oral health care and racially biased oral health care practices, both of which could inform potential interventions aimed at improving racial equity in oral health.