Examining Rural Disparities in Cigarette Smoking Among U.S. Women by Chronological Age: 2002–2022

Abstract

Objective:

This study investigated rural disparities in cigarette smoking among U.S. women by age (18–65+ years) across survey years (2002–2022).

Methods, Data Source:

Data came from the National Survey on Drug Use and Health. Women were categorized by rural-urban residence and age. We examined associations between residence, age, and time on current-smoking prevalence and quit ratios in two-year bins using weighted logistic-regression adjusting for race/ethnicity, education, annual income.

Results:

Effects of residence on current-smoking prevalence interacted with time (t[df=430,180]=4.51, P<.001), with reductions over time among urban (AOR=0.95, 95%CI: 0.94–0.96, P<.001) but not rural residents (AOR=0.99, 95%CI: 0.98–1.01, P=.66). Residence interacted with age (t[df=430,180]=−4.90, P<.001), with greater smoking among rural women in younger (AORs≥1.23, 95%CI: 1.01–1.44, Ps ≤ .008), but not older age brackets (AORs ≤1.04, 95%CI: 0.74–1.35, Ps≥.688). Rural residence predicted lower odds of quitting smoking (AOR=0.80, 95%CI: 0.71–0.91, P<.001).

Conclusions:

There is a growing disparity in smoking prevalence that disproportionately impacts rural women ages 18–49 years raising concerns about multigenerational adverse effects as this demographic is most likely to be pregnant or parenting young children. There is also a rural disparity in quitting smoking across age groups underscoring a need for greater access to smoking-cessation services among rural women.

There has been a striking reduction in U.S. smoking prevalence since the publication of the landmark 1964 Surgeon General’s Report on smoking and cancer (U.S. Department of Health, Education, and Welfare, 1964; U.S. Department of Health, Education, and Welfare, 2020). Unfortunately, this reduction has been unevenly distributed within the U.S. population, with certain sociodemographic groups experiencing smaller declines than others (U.S. Department of Health, Education, and Welfare, 2020). One disparity that is the focus of the present report is among adults residing in rural compared to urban regions (Higgins, 2021; Higgins et al., 2024; Leider et al., 2020; Nechuta & Wallace, 2023; Parker et al., 2022).

Smoking prevalence was greater among urban compared to rural residents when the 1964 Surgeon General’s Report was released. However, subsequent decreases in smoking prevalence over time have been sufficiently steeper among urban compared to rural residents to flip the direction of that disparity such that smoking prevalence and smoking related morbidity and mortality rates are now greater among rural compared to urban residents (Doogan et al., 2017; Cepeda-Benito et al., 2018; Neighbor et al., 2018; Higgins et al., 2024; Roberts et al., 2017). Moreover, this rural versus urban disparity in smoking prevalence is particularly discernible among women (Cepeda-Benito et al., 2018). We have conducted a series of observational studies on this topic using nationally representative samples starting with a report by Cepeda-Benito et al. (2018) that compared trends in smoking prevalence during 2007–2014 among rural and urban men and women using data from the U.S. National Survey on Drug Use and Health (NSDUH). Results from that study noted significant declines over time among rural men and urban men and women, but not among rural women.

Perhaps we should not be surprised that women would be especially impacted by this rural-urban disparity considering that tobacco manufacturers have long targeted women through marketing strategies that pair smoking with empowerment, independence, and sex appeal, developing cigarette brands marketed specifically for women (e.g., Virginia Slims) (U.S. Department of Health, Education, and Welfare, 2001), the use of flavored cigarettes especially menthol which are disproportionately preferred more among women than men (Gil et al., 2021; Centers for Disease Control and Prevention, 2025; Goodwin et al., 2023; Jackler et al., 2022), while also targeting rural communities through marketing involving coupons and price discounts (Brown-Johnson et al., 2014; Centers for Disease Control and Prevention, 2024; Henriksen et al., 2020; Raskind et al., 2021).

A programmatic series of reports after Cepeda-Benito et al. (2018) have shown that this rural-urban disparity extends to women of reproductive age including those who are currently pregnant (Higgins et al., 2024; Nighbor et al., 2018). This disparity among women of reproductive age is especially concerning as smoking increases risk for multi-generational adverse effects. For an excellent example supporting that concern, infant deaths are more prevalent in rural than urban regions, with Sudden Unexpected Infant Death (SUID), an outcome highly associated with maternal smoking, being the largest contributor to that overall rural-urban disparity in infant deaths (Anderson et al., 2019; Mohamoud et al., 2023; Sun et al., 2023). Importantly, these smoking-related health disparities are not limited to reproductive age women or their infants. Rural women are also at greater risk for smoking-related conditions common in older adults including cardiovascular disease and smoking-related cancers (Abrams et al., 2022; Nechuta & Wallace, 2023; Womack et al., 2020). There is also evidence that differences in smoking-cessation rates contribute to these rural-urban disparities. For example, Higgins et al. (2024) reported that quit ratios (i.e., the proportion of women who meet criteria for lifetime smoking but report no smoking in the past year) were greater in urban than rural women of reproductive age.

The purpose of the current study is to extend this series of studies of rural-urban disparities in smoking prevalence and quit ratios among adult women by assessing patterns from early through late adulthood using data from NSDUH survey years 2002–22. We used NSDUH to facilitate comparisons with the earlier studies in this series and different age brackets to provide a more fine-grained characterization of these disparities in adult women by chronological age. More specifically, the present study aims to assess whether this rural-urban disparity is especially impacting certain age groups of women or is evident throughout adulthood. We included multiple years to also examine potential time trends of interest including possible decreasing disparities over time in the younger age brackets corresponding to the recent overall decreases in smoking among youth (Centers for Disease Control and Prevention, 2024; Erath et al., 2024) or possibly more stable patterns among older women as has been reported in samples that included both sexes (Parker et al., 2022). To assess whether the size of this disparity was continuing to increase over time as reported for reproductive age women from 2002–2019 (Higgins et al., 2024), we extended the time-period assessed out to 2022.

Methods

Data Source

Data were obtained from the public use data files of NSDUH, 2002–2022, excluding the 2020 survey because of considerable challenges in conducting the survey during the COVID pandemic which could influence data validity. NSDUH is a nationally representative annual cross-sectional survey of the civilian, non-institutionalized U.S. population aged 12 years and older (see https://www.samhsa.gov/data/data-we-collect/nsduh-national-survey-drug-use-and-health). Detailed descriptions of survey procedures have been provided previously (e.g., Doogan et al., 2017; Cepeda-Benito et al., 2018; Nighbor et al., 2018; Parker et al., 2022). We used survey-provided participant weights in all analyses. Data were analyzed based on consecutive two-year periods to increase sample size and the precision of the estimates. Population weights were included with the survey data to obtain results representative of the U.S. population by correcting for selection probabilities, non-response, and post-stratification. References to “adjusted” or “unadjusted” models refer to covariate adjustment. This study was based on publicly available anonymized databases and is thus exempt from ethical compliance review.

Measures

The study included two dependent variables: (1) Current smoking and (2) quit ratios. Smoking prevalence was defined as self-reported smoking of at least one cigarette in the past 30 days and at least 100 cigarettes lifetime. Quit ratios were defined as the proportion of survey participants who reported smoking at least 100 cigarettes lifetime but no smoking in the past year.

The study included three independent variables: (1) Rurality, defined using the U.S. Office of Management and Budget Rural-Urban Continuum Codes. For the 2002–2014 NSDUH survey years, definitions were based on 2003 metropolitan or nonmetropolitan statistical area county-level groupings; for the 2015–2022 NSDUH survey years, updated 2013 groupings were used. (2) Chronological age, using the following five age brackets (years) used in NSDUH: 18–25; 26–34; 35–49; 50–64, and 65+. (3) Time, organized into ten two-year periods between 2002–2003 to 2021–2022 omitting data from 2020.

Statistical Analyses

Twenty years of NSDUH data with women 18 years of age and above were analyzed (2002–2019 and 2021–22). Data were organized in two-year bins with population weights applied. Descriptions of the two outcome variables (smoking prevalence and quit ratios) by sociodemographic characteristics were obtained using the Proc SurveyMeans procedure in SAS. Weighted logistic regression models were used to test linear time trends in smoking prevalence and quit ratios by rural or urban residence and then further by chronological age in five brackets, using time in two-year bins as a continuous variable with the Proc SurveyLogistic procedure. The key predictor variables included rural-urban residence, chronological age (five brackets), and time (two-year periods), and interactions between those variables. For each analysis, P < .05 (2-tailed) was considered statistically significant. Models were conducted unadjusted and then adjusted for race and ethnicity, educational attainment, and annual income.

Results

Participants

Briefly, the study sample included 430,181 self-identified female participants 18–65+ years of age (86,345 and 343,836 rural and urban residents, respectively), who provided responses to all pertinent survey items. Table 1 details characteristics of this study sample with population weighted percentages shown in parentheses. Women ranged in age from 18 to 65+ years with non-Hispanic White the majority race/ethnicity in the overall, rural, and urban samples. Educational attainment varied across less than high school (10.92%), high school diploma (26.65%), some college (30.61%), and college degree (31.83%); family annual income varied across <$20,000 (18.22%), $20,000-$49,999 (30.49%), $50,000-$74,999 (15.81%), and $75,000 or more (35.49%).

Table 1.

Sample descriptive characteristics and population-weighted percentages of the combined 2002–2022 sample, National Survey on Drug Use and Health.

	Overall (N = 430,181)		Rural (N = 86,345)		Urban (N = 343,836)
	Sample N	%a	Sample N	%a	Sample N	%a
Age
18–25	174,488	13.3	34,269	12.0	140,219	13.5
26–34	75,372	15.4	13,908	12.7	61,464	15.9
35–49	102,618	24.8	20,264	22.5	82,534	25.3
50–64	45,166	24.3	10,037	26.5	35,219	23.9
65+	32,537	22.1	7,867	26.3	24,670	21.4
Race
White	268,586	63.2	66,799	80.6	201,787	60.1
Black/African Am	55,509	12.6	6,297	8.6	49,212	13.3
Native Am/AK Native	6,022	0.6	3,066	1.5	2,956	0.4
Native HI/Other Pac Isl	1,988	0.3	369	0.2	1,619	0.3
Asian	17,661	5.8	1,158	0.9	16,503	6.6
More than one race	12,514	1.8	2,407	1.6	10,107	1.9
Hispanic	67,901	15.8	6,249	6.7	61,652	17.3
Education
Less than high school	58,353	10.9	13,156	13.9	45,197	10.4
High school diploma/GED	123,117	26.6	29,584	35.8	93,533	25.0
Some college	137,942	30.6	28,228	31.3	109,714	30.5
College degree	110,769	31.8	15,377	19.0	95,392	34.1
Income
Less than $20,000	108,446	18.2	25,235	24.1	83,211	17.2
$20,000 – $49,999	144,647	30.5	32,100	36.6	112,547	29.4
$50,000 – $74,999	66,327	15.8	13,196	15.7	53,131	15.8
$75,000 or More	110,761	35.5	15,814	23.6	94,947	37.6
Current Smoking 1
No	336791	84.4	62349	77.4	274442	85.6
Yes	93390	15.6	23996	22.6	69394	14.4
Quit Ratio 2
	46,983/150,623	48.3	9,926/36,186	40.4	37,057/114,437	50.2

^aWeighted to the US population.

¹Current smoking is defined as smoked part or all of at least 1 cigarette in the past 30 days and 100 cigarettes lifetime.

²Quit ratio is defined as the proportion of participants who meet criteria for lifetime smoking (i.e., smoked at least 100 cigarettes in lifetime) but report no smoking in the past year.

Regarding smoking characteristics in the overall survey sample (population weighted %), 150,623 women met criteria for lifetime smoking (33.72%), 93,990 for current smoking (15.60%), and 46,983 for former smoking (48.34%). That same breakdown for rural residents was 36,186 (41.63%) for lifetime smoking, 23,996 (22.62%) for current smoking, and 9,926 (40.35%) for former smoking; for urban residents the breakdown was 114,437 (32.33%) for lifetime smoking, 69,394 (14.37%) for current smoking, and 37,057 (50.15%) for former smoking.

Smoking Prevalence

There were significant main effects of each of the three independent variables on smoking prevalence with rural residence associated with greater adjusted odds of current smoking (Adjusted Odds Ratio [AOR] = 1.18, 95% CI: 1.08–1.29, P < .001), time associated with lower adjusted odds of current smoking (AOR = 0.96, 95% CI: 0.95–0.97, P < .001), and older age with lower adjusted odds of current smoking (AOR = 0.88, 95% CI: 0.86–0.90, P < .001).

The effect of residence interacted significantly with time (t[df=430,180]=4.51, P<.001), with no significant reduction in the odds of current smoking among rural residents over time (AOR = 1.00, 95% CI: 0.98–1.01, P = .656) while the odds decreased significantly among urban residents (AOR = 0.95, 95% CI: 0.94–0.96, P < .001) (Figure 1, Panel A). Residence also interacted significantly with chronological age (t[df=430,180]=−4.90, P<.001), with significantly greater odds of smoking among rural compared to urban residents in the age brackets 18–25 years (AOR = 1.43, 95% CI: 1.26–1.64, P < .001), 26–34 years (AOR = 1.23, 95% CI: 1.06–1.42, P = .008), and 35–49 years (AOR = 1.26, 95% CI: 1.10–1.44, P = .001), but not 50–64 years (AOR = 1.04, 95% CI: 0.85–1.00, P = .688) or 65+ years (AOR = 1.00, 95% CI: 0.74–1.35, P = .999) (Figure 1, Panel B).

Fig. 1:

Upper panel: mean current-smoking prevalence among rural versus urban adult women ages 18–65+ years in 2-year bins from 2002–2022 excluding 2020 due to concerns about potential influence of the COVID pandemic. Center panel: mean current-smoking prevalence collapsed across survey years plotted as a function of chronological years in five age brackets. Bottom panel: mean current-smoking prevalence among women in five age brackets collapsed across rural versus urban residents and plotted as a function of survey years in two-year bins. Estimates are weighted to reflect the U.S. population during the years 2002–2022. Brackets around data points represent 95% confidence intervals (CIs); where no bracket is discernible, CIs fell within the data point.

The only other significant interaction between these three independent variables was a two-way interaction of age and time (t[df=430,180]=17.19, P<.001) with the odds of current smoking decreasing over time in the age brackets 18–25 years (AOR = 0.81, 95% CI: 0.80–0.82, P < .001), 26–34 years (AOR = 0.93, 95% CI: 0.92–0.94, P < .001), and 35–49 years (AOR = 0.98, 95% CI: 0.97–0.99, P = .002), while remaining unchanged in the 50–64 years (AOR = 1.02, 95% CI: 1.00–1.04, P = .092) and increasing in the 65+ age bracket (AOR = 1.04, 95% CI: 1.01–1.07, P = .004) (Figure 1, Panel C).

Quit Ratios

There were significant main effects of two of the three independent variables on quit ratios with rural residence being associated with lower adjusted odds of meeting criteria for former smoking (AOR = 0.80, 95% CI: 0.71–0.91, P < .001) (Figure 2, Panel A) and older age being associated with greater adjusted odds of meeting criteria for former smoking (AOR = 1.94, 95% CI: 1.86–2.02, P < .001).

Fig. 2:

Upper panel: mean smoking quit ratios among rural versus urban women collapsed across survey years 2002–2022 excluding 2020 due to concerns about potential influence of the COVID pandemic. Bottom panel: mean smoking quit ratios collapsed across rural and urban residents plotted as a function of survey years in two-year bins. Quit ratios represent the proportion of women who reported no smoking in the past year among those who reported having smoked 100 or more cigarettes lifetime. Estimates are weighted to reflect the U.S. population during the years 2002–2022.

The only significant interaction between these three independent variables was a two-way interaction between age and time (t[df=150,622]=8.03, P<.001), with odds of meeting criteria for former smoking increasing over time in the age brackets 18–25 years (AOR = 1.17, CI: 1.13–1.20, P < .001), 26–34 years (AOR = 1.07, CI: 1.04–1.09, P < .001), and age 35–49 years (AOR = 1.02, CI: 1.01–1.04, P = .008), while decreasing in the age brackets 50–64 years (AORs ≤ 0.97, 95% CI: 0.95–0.99, Ps ≤ .011) and 65+ years (AOR = 0.96, CI:0.93–0.99, P = .008) (Figure 2, Panel B).

Discussion

These results extend current knowledge regarding how this well-established rural disparity in smoking prevalence is impacting adult women in at least two important ways. First, the results demonstrate that the size of the disparity in smoking prevalence is continuing to increase over time. Second, that this disparity in smoking prevalence is largely expressed among women between the ages of 18 and 49 years. That is, among the women who are most likely to be pregnant or parenting young children and thus increasing risk for fetal and childhood smoke exposure in utero and through second- and third-hand exposure. We previously reported that the disparity is impacting women of reproductive age including those currently pregnant (Higgins et al., 2024; Nighbor et al., 2018) and now confirm that not only are reproductive-age women impacted, but they are also the demographic that is shouldering the brunt of this disparity. This pattern bodes poorly for the rural disparities in smoking-related impacts on reproductive and infant health including SUID, and later-in-life disparities in cardiovascular and pulmonary disease, and lung and other smoking-related cancers that are disproportionately impacting rural women and can be expected to continue for several decades beyond any resolution of the disparities in smoking prevalence (Abrams et al., 2022; Nechuta & Wallace, 2023; Womack et al., 2020). One encouraging observation regarding smoking prevalence in the present study is the significant age-by-time interaction showing a steep reduction in smoking prevalence over the past several decades in the 18–25 years age bracket that extends across rural and urban populations suggesting a future less ravaged by smoking-related morbidity and mortality if this pattern sustains. Less encouraging are the stable patterns of smoking prevalence across years among rural and urban women 50 years of age and above, which has been reported previously regarding samples that included both sexes (Kleykamp et al., 2025; Kleykamp & Kulak, 2023; Meza et al. 2023). What accounts for these age-related differences remains unclear. However, if at least some older adults believe that they are beyond the age of gaining health benefits, then broad dissemination through healthcare providers and public-service announcements that even those 65 years and above can gain meaningful increases in life expectancy by quitting smoking may be helpful in countering that misconception (Le et al., 2024).

The quit-ratio results demonstrate a sizeable disparity of approximately 20% lower odds of quitting among rural compared to urban women over the past several decades that extends across those from 18 to 65+ years. As we discussed regarding women of reproductive age (Higgins et al., 2024), this disparity in quitting smoking across the adult lifespan underscores the need to increase access to evidence-based cessation interventions to rural women. As we also noted in our 2024 report, we believe that recent advances in digital health offer promise for meeting this need. We previously reported positive results with a smartphone-based contingency-management (CM) cessation intervention in a sample of 90 pregnant women that included 23% rural residents who appeared to benefit from the intervention at levels comparable to urban residents. CM is an approach where patients receive small financial incentives for objectively verified success in abstaining from smoking. Results from a recent Cochrane systematic review and meta-analysis (Notley et al., 2025) provide strong evidence for the efficacy of CM for smoking cessation while the incentives are available and after they are discontinued among pregnant women as well as the general population of people who smoke. Although an empirical question, we see no reason why this same CM model or other evidence-based remote cessation interventions would not be helpful for reducing the sizeable and unsettling rural disparity in quitting smoking and associated disparities in smoking-related adverse health outcomes.

Another aspect of the results that merits comment is regarding the absence of a significant interaction of age and residence for quit ratios paralleling the interaction that was observed for smoking prevalence where the disparity was most evident among women between the ages of 18–49 years. That would seem to implicate greater smoking initiation rates among rural compared to urban women as a contributor to these prevalence disparities underscoring a need for greater access to evidence-based smoking prevention efforts in rural communities in addition to the need for greater access to smoking-cessation interventions.

Limitations

This study shares limitations with those noted in prior reports in this series on rural smoking disparities (Cepeda-Benito et al., 2018; Doogan et al., 2017; Higgins et al., 2024; Nighbor et al., 2018; Parker et al., 2022). NSDUH is a cross-sectional observational study that cannot support causal inferences. When comparing different age groups in a cross-sectional study, there is risk of confounding effects of chronological age with cohort differences. NSDUH uses a valid method for classifying geographic residence but there are also other methods for doing so and results on rural-urban health disparities can vary depending on the methods used (Hirko et al., 2022).

Conclusions

These limitations notwithstanding, the present study advances knowledge by further documenting a U.S. rural disparity in smoking that is disproportionately impacting women. They extend prior knowledge by demonstrating that the size of the disparity in prevalence of current smoking is increasing across time and that it is largely shouldered by women between the ages of 18–49 years, that is, among those most likely to be pregnant or mothering young children. The results also further document a disparity in quitting smoking with lower quit ratios in rural compared to adult women of all ages along with results suggesting disparities in smoking initiation rates as well. These findings further underscore a need for more national, state, and local tobacco control and regulatory efforts to reduce smoking among rural women generally, but especially those in those18–49 years age who are shouldering the brunt of this disparity and are at greatest risk for multi-generational adverse effects from smoking during pregnancy or while parenting young children. Towards that goal, we encourage further increases in the price of cigarettes coupled with laws prohibiting price-discounting strategies to neutralize price increases (Brown-Johnson et al., 2014; Centers for Disease Control and Prevention, 2024; Henriksen et al., 2020; Raskind et al., 2021; Ribisl et al., 2022), bans on menthol and other cigarette flavorings (Gil et al., 2021; Centers for Disease Control and Prevention, 2025; Goodwin et al., 2023; Jackler et al., 2022), and greater access to evidence-based smoking prevention and cessation interventions in rural communities.

Highlights.

Rural disparities in smoking prevalence increased over the 20 years examined.

Rural disparities are evident in smoking prevalence and quit ratios.

These disparities in smoking prevalence are most discernible in reproductive age women.

Tobacco control and regulatory interventions targeting rural women are sorely needed.