Trends in Education-Related Smoking Disparities Among U.S. Black or African American and White Adults: Intersections of Race, Sex, and Region

Abstract

Introduction

Despite its overall decline in the United States, trends in cigarette smoking could vary by intersection with demographic characteristics. We explored trends in education-related disparities in current smoking among U.S. adults by race (Black or African American and White), sex, and U.S. census region.

Aims and Methods

Data were from U.S. civilian non-institutionalized adults (aged ≥18 years) who self-identified as Black or African American and White and participated in the 1995–2019 Tobacco Use Supplement to the Current Population Survey. We estimated average annual percent changes in current cigarette smoking by the intersections of race, sex, census region, and educational attainment. We calculated educated-related prevalence differences in current cigarette smoking by subtracting the prevalence of bachelor’s degrees from that of <high school in 1995–1996 and 2018–2019, then examined their variations by the intersection of sex, race, and region. Weighted multivariable logistic regression models were used to test education*survey year interactions across sexes, races, and regions.

Results

Education-related disparities in current cigarette smoking increased over time, especially among Black or African American male (PD_1995–1996 = 22.8%; PD_2018–2019 = 27.2%) and female adults (PD_1995–1996 = 12.1%; PD_2018–2019 = 16.5%). By region, Black or African American male adults in the Midwest showed the largest increase in education-related current cigarette smoking disparities, followed by Black or African American male and female adults in the South, and White male and female adults in the Midwest. These findings were because of small to no declines in the prevalence of current cigarette smoking among those with <high school education.

Conclusion and Relevance

The gap in the prevalence of current cigarette smoking by education widened over time, especially among Black or African American adults in certain regions.

Implications

Despite the decline in the prevalence of current cigarette smoking in the U.S. population overall, such public health gain may not benefit all individuals equally. Using the data from a U.S. representative serial cross-sectional survey study during 1995–2019, we found that disparities in current cigarette smoking prevalence between those with <high school versus bachelor’s degree education widened especially among Black or African American adults. Future research to investigate the barriers to progress among Black or African American adults with <high school education could inform interventions to reduce racial and education-related cigarette smoking disparities.

Introduction

Comprehensive tobacco control efforts have contributed to significant declines in cigarette smoking in the United States over the past three decades.¹ Indeed, the prevalence of current cigarette smoking was over 6% points lower in 2019 (14.0%) compared with 2005 (20.9%).² However, data reporting the overall positive change over time in current smoking obscure existing sociodemographic disparities based on sex/gender, racial category, socioeconomic status (SES), and geographic location. Current cigarette smoking is more prevalent among males (vs. females), American Indian or Alaskan Natives (vs. White), individuals with less than a college education (vs. undergraduate degree), those who are with lower annual household income (vs. ≥US$100 000), and certain census regions (Northeast, Midwest, South vs. West).³ The current study sought to move beyond testing independent associations to test interactions of education, race, sex, and geography on the prevalence of cigarette smoking over time. Such an examination has the potential to offer a fuller understanding of the impact of tobacco control efforts toward addressing smoking-related disparities.

The overall racial disparity in cigarette smoking between White and Black or African American people was eliminated roughly 20 years ago.⁴ However, whether this disparity exists depends on education level. Between 1992 and 2018, the prevalence of smoking among White adults who did not complete high school declined by 22.7% points, from 63.9% to 41.2%.⁵ In contrast, during this timeframe, was a 1.2% points increase (37.5% to 38.7%) among Black or African American adults with less than high school education. College graduates, both White and Black or African American, demonstrated declines in the prevalence of smoking—11.7% points among White adults (from 25.1% to 13.4%) and 7% points among Black or African American adults (from 16.5% to 9.5%). These data suggested Black or African American adults with low SES who smoke faced disadvantages in terms of the likelihood of cessation compared to their higher SES counterparts as well as White adults with a smoking history. Over the longer term, the smoking-attributable disease burden is likely compounded when considering race and education together.

Structural racism created highly segregated communities with lower-quality housing, education, and occupational opportunities, coupled with inadequate access to high-quality healthcare in the United States.^6,7 Racial segregation remains prominent across the United States and is conspicuous in cities in the Midwest (e.g. Detroit, Michigan and Cleveland, and Ohio) and South (e.g. Pine Bluff, Arkansas and Albany, and Georgia) regions.⁸ It also set the stage for targeted marketing of menthol brand cigarettes to Black or African American adults, which are harder to quit,⁹ in geographic locations with predominantly Black/ or African American residents.⁸ Access to culturally appropriate smoking cessation treatments¹⁰ as well as access and uptake of pharmacological cessation aids is limited.¹¹ This group also faces unique stressors that are positively related to current smoking, such as discrimination^12,13 and distress¹⁴ and make quitting harder,¹³ despite high interest.¹⁵ Thus, examining the relationships between geography, in combination with education level and the racial group would enhance our understanding of disparities in current smoking over time.

The sociocultural nature of cigarette smoking lends itself to differences in prevalence by sex/gender as well as a racial group. National Health Interview Survey data in 2005 and 2013 indicated a comparable prevalence of current smoking among White and Black or African American males (2005: White 24.0% vs. Black or African American 26.7%; 2013: White 21.2% vs. Black or African American 21.8%, respectively), and that the prevalence among males in both groups was greater than their female counterparts (2005: White 20.0% vs. Black or African American 17.3%; 2013: White 17.8% vs. Black or African American 15.4%, respectively).¹⁶ In both years, the prevalence was greater among White versus Black or African American females. It was also noteworthy that the change in smoking prevalence between 2005 and 2013 exhibited greater declines among males compared with females in both racial groups. That is, most people who smoke identified as male, yet there was less change (decline) in smoking among females. Missing from the available data are longitudinal trends across multiple time points and intersections that go beyond sex/gender and racial group.

The Current Study

Continued progress in tobacco control, and specifically efforts to decrease cigarette smoking, is needed. While previous studies indicated that current cigarette smoking varies by race, sex, region, and education independently, these variables may have intersectional relationships with trends in current cigarette smoking. To the best of our knowledge, no previous studies have reported such analyses, which would be an example of a precision public health approach to surveillance and gaining insight regarding patterns of smoking by key demographics. The purpose of this study is to explore trends in current cigarette smoking among U.S. adults from 1995 to 2019 by the intersections of these four demographic variables. The goal is to signal specific subpopulations that are disproportionately burdened by cigarette smoking, and potentially related health outcomes so that targeted interventions can be developed and delivered to reduce tobacco-related disparities.

Method

Study population

Data were from the Tobacco Use Supplement to the Current Population Survey (TUS-CPS). TUS-CPS is a survey of tobacco use in the U.S. civilian non-institutionalized adult population (aged ≥18 years) who completed the Current Population Survey (CPS). The survey has been administered by the U.S. Census Bureau approximately every 3–4 years. Eligible adults were sampled using a two-stage address-based probability sampling, where in stage 1 substate area primary sampling units (PSUs) were sampled and in stage 2 households were sampled within PSUs. Individuals from sampled households provided informed consent and were interviewed by phone or in-person. Further details about the methodology of TUS-CPS can be found online.¹⁷

In this analysis, we used data from the 1995–1996, 1998–1999, 2000, 2001–2002, 2003, 2006–2007, 2010–2011, 2014–2015, and 2018–2019 data collection cycles, with response rates ranging from 74.5% to 86.9%. We further limited the analysis to self-respondents with known smoking status who identified as non-Hispanic and Black or African American (n = 136 437) and non-Hispanic and White (n = 1 150 107). We excluded other racial/ethnic groups for the following reasons. First, Hispanic and Asian populations have consisted of heterogeneous groups with varying prevalences of current cigarette smoking. Therefore, analyses related to Hispanic and Asian populations need to disaggregate the data to handle such heterogeneity. Second, over 60% of multiracial adults self-identify as American Indians with other races.¹⁸ Thus, analyses for American Indian and multiracial populations need to address this complexity. This analysis did not require review or approval by the National Institutes of Health Institutional Review Board per 45 CFR 46 because it only involved de-identified data and therefore is considered “not human subjects research.”

Measures

Respondents were asked if they smoked at least 100 cigarettes in their lifetime and whether they now smoked cigarettes every day, some days, or not at all. Those who reported having smoked at least 100 cigarettes in their lifetime and currently smoke every day or some days were classified as adults who currently smoke. Otherwise, respondents were classified as adults who do not currently smoke. CPS also collected information on respondents’ sex (male vs. female), census region (Northeast, South, Midwest, and West), and highest level of school completed or degree received (<high school, high school diploma/GED, some college, associate degree, bachelor’s degree, or higher), and annual household income (<$25 000, $25 000–$49 999, $50 000–$74 999, $75 000+, and missing). Because TUS-CPS were not conducted in regular time intervals, to correctly represent the time scale, mid-points of each survey wave were converted into survey years (1996.04, 1999.04, 2000.21, 2001.82, 2003.49, 2006.68, 2010.68, 2014.99, and 2018.99) and analyzed as a continuous variable.

Statistical Analysis

Data were weighted using self-response survey weights to account for non-response and to be representative of the civilian non-institutionalized population of the United States (aged ≥18 years). For each of the 20 intersections of sex, race, and education, we estimated the prevalence of current cigarette smoking. We then conducted multivariable logistic regression analyses on current cigarette smoking by sex and race to test education*survey year interaction, while adjusting for age, annual household income, education, and survey year. We repeated these analyses within each census region. These analyses were performed using SUDAAN 11.0.1 with Fay-adjusted balanced repeated replication methods to account for the complex sample design of the TUS-CPS.¹⁷

Trends were assessed with joinpoint models using Joinpoint Regression Program 4.8.0.1.,¹⁹ with grid search model fitting and the Bayesian Information Criterion (BIC)3 model selection method to empirically identify potential turning points in these trends. We required at least five observations (survey waves) from a joinpoint to either end of the data, and at least four observations between two joinpoints. We reported average annual percent change (AAPC) from these joinpoint models. Additionally, using the estimated prevalence of current smoking from the joinpoint regression models, we calculated the prevalence difference (PD) between those with <high school education and those who are with bachelor’s degrees (PD = P_{<high school}−P_bachelor’s) by race, sex, and census region in 1995–1996 and 2018–2019 to examine how education-related smoking disparities changed over time. Statistical significance was set at p < .05 for all analyses.

Results

Table 1 shows the demographic characteristics and prevalence of current cigarette smoking among U.S. Black or African American and White adults during 1995–2019. Slightly less than half of the population was male, and most identified as White. Close to a third of the population completed a bachelor’s degree. Over one-third of the population lived in the South, approximately one-quarter in the Midwest, slightly under one-fifth in the Northeast, and one-fifth lived in the West. Current cigarette smoking was more prevalent among males (vs. female), White (vs. Black/African American), less than bachelor’s degree education (vs. bachelor’s degree or higher), Midwest (vs. West), 59 years and younger (vs. ≥60 years), and those who were surveyed before 2018–2019 (vs. 2018–2019) survey waves.

Table 1.

Demographic Characteristics and Weight Prevalence of Current Smoking, Tobacco Use Supplement to the Current Population Survey, 1995–2019

	Unweighted N	Weighted% (95%CI)	Weighted% current smoking (95%CI)
Sex
Male	560 164	47.8(47.7 to 47.8)	21.2(21.0 to 21.3)
Female	726 380	52.2(52.2 to 52.3)	17.9(17.8 to 18.0)
Race/ethnicity
Black/African American	136 437	14.1(14.1 to 14.2)	18.5(18.2 to 18.7)
White	1 150 107	85.9(85.8 to 85.9)	19.6(19.5 to 19.8)
Education completed
<High school	146 035	11.2(11.1 to 11.3)	30.2(29.8 to 30.5)
High school diploma/GED	407 978	31.0(30.9 to 31.2)	25.7(25.5 to 25.9)
Some college	246 578	19.9(19.8 to 20.0)	20.7(20.5 to 20.9)
Associate degree	118 014	8.9(8.9 to 9.0)	17.9(17.7 to 18.2)
Bachelor’s degree or higher	367 939	29.0(28.8 to 29.2)	8.3(8.2 to 8.4)
Region
Northeast	266 896	19.6(19.5 to 19.6)	18.0(17.8 to 18.3)
South	415 821	37.0(36.9 to 37.1)	20.4(20.2 to 20.6)
Midwest	337 659	25.2(25.1 to 25.3)	21.3(21.0 to 21.6)
West	266 168	18.2(18.1 to 18.3)	16.7(16.5 to 17.0)
Age
18–29	193 428	20.1(20.0 to 20.1)	22.9(22.6 to 23.2)
30–39	234 570	18.1(18.0 to 18.1)	23.0(22.8 to 23.3)
40–49	251 994	19.4(19.4 to 19.4)	23.0(22.8 to 23.3)
50–59	226 040	16.8(16.8 to 16.9)	20.6(20.4 to 20.8)
60 or older	380 512	25.6(25.6 to 25.6)	10.8(10.7 to 11.0)
Survey wave
1995–1996	167 213	10.5(10.5 to 10.5)	24.4(24.1 to 24.6)
1998–1999	155 617	10.7(10.7 to 10.7)	23.0(22.7 to 23.3)
2000	108 143	10.8(10.8 to 10.8)	22.8(22.4 to 23.1)
2001–2002	162 780	10.9(10.9 to 10.9)	21.9(21.6 to 22.3)
2003	157 258	11.0(10.9 to 11.0)	20.0(19.6 to 20.3)
2006–2007	145 272	11.2(11.2 to 11.2)	19.7(19.4 to 20.0)
2010–2011	141 975	11.5(11.5 to 11.5)	17.3(17.0 to 17.6)
2014–2015	135 485	11.6(11.6 to 11.7)	15.0(14.8 to 15.2)
2018–2019	112 801	11.8(11.8 to 11.8)	12.6(12.3 to 12.8)

Table 2 shows the prevalence of current cigarette smoking by the intersection of sex, race, education, and survey year. Across race and sex, declines in current cigarette smoking were faster as education increased (education*year interaction p < .01). Among Black or African American males with <high school education, AAPC was −2.03% during 1995–2007 and 0.89% during 2007–2019, while those with a bachelor’s degree showed AAPC of −1.19% during 1995–2000 and −5.40% during 2000–2019. For Black or African American females, the decline in the prevalence of current cigarette smoking was also slower among those with <high school education (AAPC_1995–2019 = −0.69%) than those with a bachelor’s degree (AAPC_1995–2019 = −4.19%). Similarly, among White males, those with <high school education showed an AAPC of −0.83%, while those with a bachelor’s degree showed an AAPC of −2.80% during 1995–2007 and hastened to −5.35% during 2007–2019. Finally, among White females, the decline in the prevalence of current cigarette smoking was slower among those with <high school education (AAPC_1995–2019 = −0.52%) than those with a bachelor’s degree (AAPC_1995–2019 = −4.91%). Increases in educational-related current smoking disparities were highest among Black or African American males (PD_1995–1996 = 22.8%; PD_2018–2019 = 27.2%; a 4.4% increase) and females (PD_1995–1996 = 12.1%; PD_2018–2019 =16.5%; a 4.4% increase), followed by White females (PD_1995-1996=18.1%; PD_2018-2019=21.8%; a 3.7% increase), and White males (PD_1995–1996 = 24.1%; PD_2018–2019 = 25.6%; a 3.7% increase).

Table 2.

Prevalence of Current Cigarette Smoking by Sex, Race, and Education, Tobacco Use Supplement of the Current Population Survey 1995–2019

Sex	Race	Education	Estimates	Survey year									PD		Education*year interaction
				1995–1996	1998–1999	2000	2001–2002	2003	2006–2007	2010–2011	2014–2015	2018–2019	1995–1996	2018–2019
Male	Black/African American	<High school	Prevalence	35.3%	33.8%	33.5%	30.6%	30.4%	29.3%	29.2%	30.0%	33.2%	22.8%	27.2%	p<0.01
			AAPC	−2.03%*						0.89%
		High school	Prevalence	32.8%	30.8%	30.0%	29.5%	25.4%	24.3%	23.9%	23.5%	19.4%
			AAPC	−2.09%*
		Some college	Prevalence	25.1%	23.4%	23.3%	21.5%	20.3%	20.1%	16.2%	15.1%	13.5%
			AAPC	−2.73%*
		Associate degree	Prevalence	28.0%	24.8%	20.2%	16.9%	15.1%	19.9%	13.5%	13.7%	10.4%
			AAPC	−3.82%*
		Bachelor’s degree	Prevalence	12.9%	13.0%	13.8%	12.2%	11.4%	8.8%	8.2%	5.5%	4.8%
			AAPC	1.19%			−5.40%*
Male	White	<High school	Prevalence	37.0%	35.7%	35.4%	35.1%	32.8%	34.3%	32.8%	32.8%	28.5%	24.1%	25.6%	p<0.01
			AAPC	−0.83%*
		High school	Prevalence	33.2%	32.8%	32.4%	31.6%	30.0%	29.5%	26.7%	24.2%	21.6%
			AAPC	*−1.23%*						−2.45%*
		Some college	Prevalence	26.8%	26.0%	25.7%	25.2%	23.3%	23.4%	21.1%	18.7%	16.3%
			AAPC	−1.36%*						−2.74%*
		Associate degree	Prevalence	23.7%	23.7%	23.5%	23.2%	20.6%	21.3%	18.0%	14.7%	13.3%
			AAPC	−1.35%*						−3.83%*
		Bachelor’s degree	Prevalence	12.3%	11.7%	11.2%	10.7%	9.9%	9.2%	7.6%	5.7%	4.8%
			AAPC	−2.80%*						−5.35%*
Female	Black/African American	<High school	Prevalence	24.5%	23.2%	22.2%	22.6%	21.2%	21.8%	19.7%	22.4%	20.8%	12.1%	16.5%	p<0.01
			AAPC	−0.69%*
		High school	Prevalence	21.2%	22.0%	22.5%	21.4%	17.9%	17.9%	15.4%	14.9%	12.8%
			AAPC	−2.42%*
		Some college	Prevalence	18.3%	16.9%	14.8%	15.3%	14.9%	13.0%	13.2%	9.7%	10.1%
			AAPC	−2.61%*
		Associate degree	Prevalence	21.1%	16.9%	15.9%	16.8%	15.4%	13.6%	10.3%	11.8%	8.9%
			AAPC	−3.39%*
		Bachelor’s degree	Prevalence	11.3%	10.6%	9.5%	9.4%	7.9%	5.0%	5.0%	4.0%	4.6%
			AAPC	−4.91%*
Female	White	<High school	Prevalence	30.7%	28.8%	29.8%	28.3%	25.8%	29.0%	28.1%	27.9%	26.1%	18.1%	21.8%	p<0.01
			AAPC	−0.52%
		High school	Prevalence	27.7%	26.3%	26.1%	25.7%	23.3%	23.8%	21.7%	20.1%	17.9%
			AAPC	−1.75%*
		Some college	Prevalence	24.0%	22.5%	21.9%	22.5%	21.1%	20.6%	18.9%	17.3%	14.1%
			AAPC	−1.87%*
		Associate degree	Prevalence	21.0%	19.5%	20.0%	19.3%	18.2%	18.0%	16.2%	14.4%	13.1%
			AAPC	−1.95%*
		Bachelor’s degree	Prevalence	11.0%	10.4%	10.6%	9.6%	8.5%	8.3%	6.4%	5.3%	4.4%
			AAPC	−3.98%*

AAPC = Average annual percent change, estimated with Joinpoint regression models. * = p < .05. PD = prevalence differences were calculated by subtracting prevalence of current smoking among those with bachelor’s degrees from that of those with <high school education, using estimated prevalence from Joinpoint regression models.

Table 3 summarizes the findings from region-stratified analyses. Among Black or African American males in the Northeast, declines in current smoking were observed similarly across education (education*year interaction p = .10; Supplementary Figure 1A). This resulted in a reduction in PD between those with <high school education and those with bachelor’s degrees over time (PD_1995–1996 = 24.3%; PD_2018–2019 = 10.9%). However, PDs increased among Black or African American males in the Midwest (PD_1995–1996 = 19.6%; PD_2018–2019 = 34.6%), South (PD_1995–1996 = 20.1%; PD_2018–2019 = 28.4%), and West (PD_1995–1996 = 13.4%; PD_2018–2019 = 27.1%; Supplementary Figures 1B, 1C, and 1D). This was because current smoking among those with <high school education did not change over time in the Midwest and West during the observation period or in the South during 2003–2019 (education*year interaction: Midwest p < .01, South p < .01, West p = .07).

Table 3.

Average Annual Percent Change in Prevalence of Current Cigarette Smoking by Sex, Race, Region, and Education, Tobacco Use Supplement to the Current Population Survey 1995–2019

Sex	Race	Region		Year													PD		Education*Year Interaction
			Education	1995	1997	1999	2001	2003	2005	2007	2009	2011	2013	2015	2017	2019	1995–1996	2018–2019
Male	Black/African American	Northeast	<High school	−3.30%*													24.3%	10.9%	p=0.10
			High school	−2.34%*
			Some college	−1.39%*
			Associate degree	−6.69%*
			Bachelor’s degree	−3.09%*
		Midwest	<High school	0.46%													19.6%	34.6%	p<0.01
			High school	−2.42%*
			Some college	−2.22%*
			Associate degree	−3.88%*
			Bachelor’s degree	−4.74%*
		South	<High school	−3.21%*					1.03%								20.1%	28.4%	p<0.01
			High school	−1.83%*
			Some college	−2.76%*
			Associate degree	−2.84%*
			Bachelor’s degree	−5.36%*
		West	<High school	0.35%													13.4%	27.1%	p=0.07
			High school	−2.31%*
			Some college	−0.31%							−11.62%*
			Associate degree	−4.89%*
			Bachelor’s degree	−4.35%*
	White	Northeast	<High school	−1.13%*													19.0%	19.4%	p<0.01
			High school	−1.99%*
			Some college	−2.46%*
			Associate degree	−0.56%							−5.65%*
			Bachelor’s degree	−1.14%				−5.10%*
		Midwest	<High school	−0.31%													23.5%	29.0%	p<0.01
			High school	−1.07%*							−2.48%*
			Some college	−1.50%*
			Associate degree	−2.57%*
			Bachelor’s degree	−2.78%*
		South	<High school	−0.80%*													25.4%	27.8%	p<0.01
			High school	−1.68%*
			Some college	−1.12%*									−4.43%*
			Associate degree	−2.38%*
			Bachelor’s degree	−4.25%*
		West	<High school	−1.87%*													26.5%	19.7%	p=0.22
			High school	−0.40%*									−2.86%*
			Some college	−2.46%*
			Associate degree	−1.11%							−4.45%*
			Bachelor’s degree	−3.64%*
Female	Black/African American	Northeast	<High school	−1.42%													10.1%	12.6%	p=0.12
			High school	−2.97%*
			Some college	−2.95%*
			Associate degree	−3.98%*
			Bachelor’s degree	−4.79%*
		Midwest	<High school	−0.43%													18.9%	22.4%	p=0.06
			High school	−2.97%*
			Some college	−2.95%*
			Associate degree	−3.98%*
			Bachelor’s degree	−4.79%*
		South	<High school	−0.45%													9.1%	15.5%	p<0.01
			High school	−2.12%*
			Some college	−2.00%*
			Associate degree	−3.21%*
			Bachelor’s degree	−7.25%									−1.54%
		West	<High school	−1.73%													19.6%	15.7%	p=0.92
			High school	−3.87%*
			Some college	−3.13%*
			Associate degree	−13.10%*					−1.37%
			Bachelor’s degree	−3.71%
	White	Northeast	<High school	−0.73%*													12.4%	16.2%	p<0.01
			High school	−2.38%*									−0.30%
			Some college	−2.59%*
			Associate degree	−/2.46%*
			Bachelor’s degree	−4.18%*
		Midwest	<High school	−0.06%													17.8%	24.0%	p<0.01
			High school	−1.47%*
			Some college	−1.68%*
			Associate degree	−1.45%*
			Bachelor’s degree	−3.61%*
		South	<High school	−0.40%													19.0%	23.7%	p<0.01
			High school	−1.76%*
			Some college	−0.36%									−4.86%*
			Associate degree	−1.91%*
			Bachelor’s degree	−3.87%*
		West	<High school	−2.24%*													23.5%	16.7%	p<0.01
			High school	−2.24%*
			Some college	−2.37%*
			Associate degree	−2.54%*
			Bachelor’s degree	−4.61%*

PD = prevalence differences

In contrast, among White males, a reduction in the PD between those with <high school education and those with bachelor’s degrees over time was observed in the West (PD_1995–1996 = 26.5%; PD_2018–2019 = 19.7%; Supplementary Figure 2D). Similar declines in current smoking were observed across education in this region (education*year interaction p = .22). The PD was stable over time among White males in the Northeast (PD_1995–1996 = 19.0%; PD_2018–2019 = 19.4%; Supplementary Figure 2A), despite the faster relative annual declines in those who were more educated than those who were less educated (education*year p < .01). Increases in PDs were observed among White males in the Midwest (PD_1995–1996 = 23.5%; PD_2018–2019 = 29.0%; Supplementary Figure 2B) and South (PD_1995–1996 = 25.4%; PD_2018–2019 = 27.8%; Supplementary Figure 2C). In these regions, only small declines (relative annual percent change <1%) in current smoking were observed among those with <high school education, while the declines were faster in those who were more educated (education*year interaction: Midwest p < .01, South p < .01).

Black or African American females had a lower prevalence of current smoking and lower PDs than Black or African American males in general (Supplementary Figures 3A–D). Among those in the Northeast, Midwest, and West, changes in current smoking were similar across education (education*year interaction: Northeast p = .12, Midwest p = .06, West p = .92). However, trends in current smoking differed by education in the South, with no changes observed among those with <high school education but the significant decline among those with ≥high school education (education*year interaction p < .01), resulting in an increase in the PD between those with <high school education and those with bachelor’s degrees over time (PD_1995–1996 = 9.1%; PD_2018–2019 = 15.5%).

White females (Supplementary Figure 4A–D) had a lower prevalence of current smoking than White males but a higher prevalence than Black or African American females in general. Among those in the West, a decline in the PD was observed (PD_1995–1996 = 23.5%; PD_2018–2019 = 16.7%), despite the faster decline in current smoking in those with higher education (education*year interaction p < .01). However, PDs increased among those in the Northeast (PD_1995–1996 = 12.4%; PD_2018–2019 = 16.2%), Midwest (PD_1995–1996 = 17.8%; PD_2018–2019 = 24.0%), and South (PD_1995–1996 = 19.0%; PD_2018–2019 = 23.7%). In these regions, those with <high school education showed <1% relative annual percent change in current smoking while those with more education showed significantly larger relative annual percent changes (education*year interaction p < .01 for all three regions).

Discussion

It has been well documented that the prevalence of cigarette smoking persistently varies by race, sex, geographical region, and education level.¹⁰ Nguyen and colleagues examined changes in education-related cigarette smoking disparities among Black or African American, White, and Hispanic adults in the United States, and found that such disparities worsened over time, especially among Black or African American and Hispanic adults.⁵ The current study extended the understanding on how intersecting demographic characteristics are associated with changes in the prevalence of smoking by adding sexes and geographic regions to the analyses. While Nguyen and colleagues found that White adults have the highest prevalence of current cigarette smoking across all levels of education level except college educated,⁵ our analyses revealed that among those with <high school education, prevalence of current cigarette smoking was higher among Black or African American than White males, but higher among White than Black or African American females. Additionally, we found that changes in education-related cigarette smoking disparities varied greatly by census region. Notably, among Black or African American males, education-related cigarette smoking disparities decreased in the Northeast but increased in the Midwest, South, and West regions over time, with the education*year interaction reaching statistical significance in the Midwest and South regions. Among White males, education-related cigarette smoking disparities increased in the Midwest and South regions, but these increases were smaller than those of Black or African American males. Much of the increase in education-related cigarette smoking disparities was because of a slower or no decline in cigarette smoking among those with less than high school education.

The reasons for these regional differences are unclear but signal underlying inequities. It is quite likely that racism, not race, is a primary explanatory factor underlying the stability of current smoking among Black or African American adults with low SES.²⁰ A 2015 study found that many areas in the Midwest and South regions have higher area racism than the United States average.²¹ Previous studies have reported an association between racial discrimination and cigarette smoking among Black or African American adults,^12,13 which was shown to be mediated through heightened stress and depressive symptoms among Black or African American adults.¹⁴ Therefore, it is plausible that regions with higher levels of racial discrimination (i.e. Midwest and South regions) also have a higher prevalence of cigarette smoking among Black or African American male adults. Another potential explanation could be because of differential access to smoking cessation programs. While the 2010 Patient Protection and Affordable Care Act (ACA) increased access to health insurance,²² Black or African American adults and those who lived in the South remained less likely to be insured compared to their White counterparts and the U.S. average, respectively.^23,24 Thus, many adults who smoke, especially those who are uninsured, may have limited access to smoking cessation programs that require health insurance in these regions.²⁵ Moreover, the availability of community-based smoking cessation resources, as well as cultural competence smoking cessation treatments may be limited in the Midwest and the South regions. Future research testing these hypotheses could inform potential interventions that might reduce education-related disparities in subpopulations for whom smoking prevalence has not decreased over time.

Additionally, the larger increases in education-related cigarette smoking disparities among Black or African American male adults compared to White male adults could be related to menthol cigarette use. It has been documented that menthol cigarette use is associated with nicotine dependence^26,27 and lower likelihood of smoking cessation.^28–33 A recent study showed that the prevalence of menthol cigarette use was persistently higher among Black or African American (vs. White), lower income, and less educated individuals over time.³⁴ Since advertising for menthol cigarettes is more prevalent in neighborhoods with a higher concentration of Black or African American and low-income individuals,³⁵ less-educated Black or African American individuals may be more likely than their White counterparts to initiate cigarette smoking with menthol cigarettes and continue to use these products once started.^36,37 In 2017, Ontario Canada banned menthol-flavored tobacco products, and this ban was associated with higher prevalence of cigarette smoking cessation among individuals who smoke menthol cigarettes than non-menthol cigarettes 2 years after policy implementation.³⁸ Since several localities and/or states have prohibited the sales of menthol cigarettes, and the U.S. Food and Drug Administration has proposed to prohibit menthol as a characterizing flavor in cigarettes and all flavors in cigars,³⁹ future research evaluating how these policies affect education-related disparities among Black or African American adults could shed light on efficacious interventions to reduce education-related cigarette smoking disparities in this population.

Structural racism and housing discrimination⁴⁰ have resulted in Black or African American adults being more likely than White adults to live in lower-income neighborhoods, and to experience lower-quality education through the public school system. These factors are also tied directly to disparities in income level and employment status. Lower-income neighborhoods have higher tobacco outlet density than high-income neighborhoods,⁴¹ which is also associated with cigarette smoking.⁴² Thus, Black or African American adults with lower education levels are disproportionately affected by these social and environmental factors that influence cigarette smoking. Future research evaluating the impacts of structural interventions to reduce the racial educational gap, housing discrimination, and tobacco retailer density could shed light on how they may reduce education-related cigarette smoking disparities by race.

The patterns in cigarette smoking among Black or African American and White females were different from those of their male counterparts. In general, Black or African American females had lower cigarette smoking prevalence than White females, perhaps in part because of the later initiation of cigarette use among Black or African American females compared to White females.¹⁵ Notably, education-related cigarette smoking disparities increased over time for both Black or African American and White females in the Northeast, Midwest, and South regions, largely because of the lack of changes in current smoking among females with less than high school education. Tobacco industry internal documents showed that the industry has used targeted marketing strategies to reach low socioeconomic women for at least four decades, which could at least partially explain the lack of progress in the reduction in cigarette smoking in this population.⁴³ Future research to investigate the factors that promote the decline in cigarette smoking among females with less than high school education in the West region may shed light on how to reduce the burden of cigarette smoking in the population in other U.S. regions. This is particularly important because although they have lower cigarette smoking prevalence than males, females are more likely to experience smoking-related morbidity.⁴⁴

Strengths of this study include large sample size and national representativeness. However, there are limitations. The data collected relied upon participant self-report, which is subject to social desirability bias and potential under-reporting of cigarette smoking behavior. However, previous studies have found self-report smoking behaviors to be reliable.⁴⁵ Additionally, the current analyses did not account for the current use of other tobacco products. Nonetheless, despite the increasing numbers and types of tobacco products available in the U.S. marketplace over the observation period, cigarettes remain the most commonly used tobacco products among U.S. adults.² Adults in the military services and those who are institutionalized (e.g. incarcerated and hospitalized populations) were excluded in TUS-CPS. Given that imprisonment rates are higher among males (vs. female) and Black or African American individuals (vs. White),⁴⁶ and high prevalence of cigarette smoking in both populations,^47,48 intersectional effects of demographics on cigarette smoking may be different from the non-institutionalized, civilian population.

In conclusion, despite the overall decline in the prevalence of cigarette smoking from 1995 to 2019 among non-Hispanic Black or African American and White adults in the United States, education-related disparities persist, and are worsening in some geographic regions depending on race and sex. These increases in education-related disparities are likely because of the lack of progress in reducing the prevalence of cigarette smoking among adults with less than high school education in those regions. Future research to investigate the barriers to progress among Black or African American adults with <high school education could inform targeted interventions to reduce racial and education-related cigarette smoking disparities.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

This research is supported by the National Institute on Minority Health and Health Disparities and the U.S. Food and Drug Administration.

Acknowledgment

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the U.S. Food and Drug Administration, National Institutes of Health, or Department of Health and Human Services.

Declaration of Interests

None to report.

Data Availability

The data used in this analysis are available at this website: https://cancercontrol.cancer.gov/brp/tcrb/tus-cps.