Temporal Association of United States Tobacco Taxation on Smoking Attributable Bladder Cancer Mortality and Disability Adjusted Life Years

Abstract

PURPOSE:

The causal link between smoking and bladder cancer (BC) development is well-established but the long-term impact of tobacco taxation and health policy on BC mortality and disability-adjusted life years (DALYs) has not been fully elucidated. Given the protracted latency of carcinogenesis, this study examines whether historical changes in tobacco taxation and smoke-free laws are associated with reductions in BC disease burden and mortality in the United States.

METHODS:

Smoking-attributable BC mortality and DALY data, and federal and state tobacco taxation data were differenced as time series to achieve stationarity. Cross-correlation analysis identified optimal lag times. A semi-logarithmic multivariable linear regression was used to estimate the percent change in BC outcomes per 1% increase in tobacco tax. Analyses were adjusted for national health expenditures and stratified by state.

RESULTS:

The median lag time between tobacco tax changes and smoking-attributable BC mortality was 17 years whereas the lag to DALYs was 24 years.. National-level regression showed no significant association between taxation and BC mortality (−0.09%, p = 0.64) or DALYs (1.77%, p = 0.051). However, 22 states exhibited significant reductions in mortality, with the greatest observed in Arkansas (−3.64%, p < 0.001), California, and Indiana. Sixteen states showed significant DALY reductions, led by California (−4.68%). The implementation of smoke-free laws alone did not associate with decreases in smoking-attributable BC mortality and DALY.

CONCLUSION:

Tobacco taxation is significantly associated with long-term reductions in smoking-attributable BC mortality and DALYs at the state level, but not nationally. These findings demonstrate the importance of adjunctive localized public health policy and the delayed impact of tobacco control measures on cancer outcomes. Further investigation is warranted to understand the mechanisms driving state-level variability and to inform targeted prevention strategies.

INTRODUCTION

Cigarette smoking remains the leading modifiable risk factor for bladder cancer (BC), accounting for approximately 50% of all cases.^1,2 Men are disproportionately affected, with a three- to four-fold higher risk of developing BC compared to women. The disease typically evolves over several decades, and cumulative smoking exposure significantly increases the risk. However, smoking history is complex, and a study demonstrated a differential dose–response relationship between smoking intensity and duration. Among individuals with equal pack-years, those who smoked fewer cigarettes per day over a longer period had a greater risk of developing BC than those who smoked more per day over a shorter period.³ Therefore, understanding the temporal dynamics between smoking exposure and BC risk becomes important, particularly as it relates to public health policy.

Our prior study identified a 28-year lag between smoking exposure and BC incidence, suggesting a prolonged latency period between exposure and disease onset.⁴ While the association between smoking and BC incidence is well established, less is known about the long-term impact of tobacco control policies on smoking-attributable BC mortality and overall disease burden. A summary measure of disease burden is disability-adjusted life years (DALY). DALY, as a primary outcome, accounts for both the loss of health caused by disability prior to death and the shortened lifespan resulting from premature mortality. Tobacco taxation is among the most effective public health strategies for reducing smoking prevalence. Evidence consistently demonstrates that increased cigarette taxes lead to decreased tobacco consumption and reductions in smoking-related diseases.⁵ Additionally, behavioral regulations such as the US Tobacco 21 policy, which raised the minimum legal age to purchase tobacco products to 21, are designed to curb youth initiation and long-term nicotine addiction. Early implementation of Tobacco 21 policies is projected to yield substantial reductions in premature mortality through 2100.⁶ Stronger smoke-free ordinances also improve population health and have demonstrated adjunctive effects on lung cancer outcomes.⁷ The goal is to protect people from exposure to secondhand smoke. As of July 2025, 25 states have enacted comprehensive smoke-free laws that cover workplaces, restaurants, and bars. It has been reported 64.8% of increased smoking cessation in the United States can be entirely accounted for by clean indoor air laws and cigarette excise taxes.⁸ The extent to which historical changes in tobacco taxation and smoke-free laws have influenced long-term trends in BC mortality and disability-adjusted life years (DALYs) remains unclear.

Given the long latency of carcinogenesis following smoking exposure, it is critical to examine whether shifts in tobacco taxation policy and smoke-free laws correspond with subsequent reductions in BC-related mortality and disease burden. This study aims to evaluate the temporal relationship between tobacco taxation and health policy with BC mortality and DALYs in the US. By examining these long-term trends, we aim to provide evidence on the broader population-level impact of tobacco control policies on BC outcomes.

METHODS

Data for this study were obtained from multiple publicly available sources. Smoking-attributable BC mortality and DALYs, our primary outcomes of interest, were retrieved from the Global Burden of Disease (GBD) database, a comprehensive resource maintained by the Institute for Health Metrics and Evaluation (IHME) at the University of Washington. The GBD provides global estimates of diseases, injuries, and risk factors stratified by age, sex, country, and time (https://vizhub.healthdata.org/gbd-results/?params=gbd-api-2021-public/ca45d304e08fcd8d000df598de7233ad). Data from the GBD ranges from 1990 to 2021. BC deaths and DALY were queried with “smoking” set as the risk variable. US federal and state tobacco taxation data from 1975 to 2019 were obtained from the Centers for Disease Control and Prevention (CDC) (https://data.cdc.gov/Policy/The-Tax-Burden-on-Tobacco-1970-2019/7nwe-3aj9/about_data). Annual US national health expenditure data from 1975 to 2019 were retrieved from the Centers for Medicare & Medicaid Services (CMS) (https://www.cms.gov/data-research/statistics-trends-and-reports/national-health-expenditure-data/historical) and used as a control variable. Information on which states have passed smoke-free laws as of July 2025 was obtained from the Campaign for Tobacco Free Kids. Urologists’ density per state as of June 2025 was obtained from the American Board of Medical Specialties.

To assess the temporal relationships, tobacco taxation data at the national and individual state-level were smoothed using a 15-year moving average, while smoking-attributable BC mortality and DALY data were smoothed using a 5-year moving average, as previously reported.⁴ Time series data were differenced up to the fifth order to achieve stationarity, which was confirmed via the Augmented Dickey-Fuller test (Supplemental Data Table 1). Cross-correlation analysis was then performed to identify the optimal negative lag between taxation and disease outcomes on the national and state level. Taxation data were subsequently lagged based on this analysis, with a minimum lag time of 10 years applied to account for the latency associated with neoplastic disease progression. A standard 28-year lag was applied to national health expenditure data based on prior literature.⁴

In accordance previous ecological studies on temporally associating tobacco consumption as a risk factor of cancer mortality and BC incidence, a semi-logarithmic multivariable linear regression model was selected.^4,9 Modeling was implemented in R Studio to estimate the percent change in smoking-attributable BC mortality and DALYs associated with a 1% increase in tobacco tax. The final multivariate regressions were written:

$$\mathit{Log}{\mathit{BCM}}_t=\alpha +{\beta }_1{\mathit{WTT}}_t+{\beta }_2{\mathit{WHE}}_t$$

$$\mathit{Log}{\mathit{BCDALY}}_t=\alpha +{\beta }_1{\mathit{WTT}}_t+{\beta }_2{\mathit{WHE}}_t$$

where $\mathit{Log}{\mathit{BCM}}_t$ is the natural logarithm of smoking-attributable BC mortality, $\mathit{Log}{\mathit{BCDALY}}_t$ is the natural logarithm of smoking-attributable BC DALY, ${\mathit{WTT}}_t$ is the lag weighted tobacco tax percentage, ${\mathit{WHE}}_t$ is the lag weighted health expenditure, $\alpha$ is the constant. The coefficient value $\beta$ indicates the associated proportional change $\mathrm{e}(e^{\beta }-1)\times 100$.

A CDC-configured tercile hextile map was created to visualize geographic patterns in public health policy, and a scatterplot was generated to display the bivariate association between smoking-attributable BC mortality and DALYs. Additionally, three-dimensional visualizations of mortality and DALY rates were developed in relation to previously reported BC incidence rates.⁴ Post-hoc analysis of smoking-attributable BC cancer burden rates was associated with state presence of National Cancer Institute (NCI)-designated clinical cancer center or comprehensive cancer center and analyzed using Fishers exact test. Similarly, this was used to evaluate the role of smoke free laws at the state level. State-level smoking consumption between 1990 and 2019 were averaged to evaluate whether this correlated with calculated state-level mortality rates and DALY rates using Pearson correlation and linear regression. This was similarly done for urologist density per state. A p-value < 0.05 was considered statistically significant. We hypothesized that tobacco taxation is associated with long-term reductions in smoking-attributable BC mortality and DALYs. All statistical analyses and visualizations were conducted using GraphPad Prism 9 and R Studio.

RESULTS

The smoking-attributable BC mortality rate in the US has steadily declined from 1.085 per 100,000 in 1990 to 0.802 per 100,000 in 2021 (Figure 1A). A similar trend was observed in DALYs, which decreased from 25.98 years in 1990 to 18.13 years in 2021 (Figure 1B). The optimal lag time associated with changes in the mortality rate on a national level was 17 years (Table 1). The lag time for DALYs was 24 years.

Figure 1:

A) Trend in US bladder cancer death rate from 1990 to 2021. B) Trend in US bladder cancer DALY from 1990 to 2021.

Table 1: Cross-correlation–calculated lag times and state- and national-level multivariable regression for the association between tobacco taxation and bladder cancer mortality and disability-adjusted life years (DALYs).

Location	Cross-correlation Lag Time (Years)		Bladder Cancer Mortality			Bladder Cancer DALYs			NCI-Cancer Center Presence	Comprehensive Smoke-free Laws	Urologists per 10,000 People
	Death Rate	DALY	Percent Change (%)	Adjusted R2 Value	p-value	Percent Change (%)	Adjusted R2 Value	p-value
United States	17	24	−0.09	0.9897	ns	1.77	0.9879	ns
Alabama	17	17	−0.89	0.9857	***	−0.87	0.9903	***	†		0.3432
Alaska	13	13	1.00	0.9283	ns	0.54	0.9495	ns			0.3378
Arizona	14	12	−0.21	0.9974	*	−0.63	0.9935	**	†	†	0.3258
Arkansas	21	26	−3.64	0.925	***	−2.26	0.9723	ns			0.2785
California	18	18	−3.16	0.966	**	−4.68	0.9736	***	†	†	0.3464
Colorado	22	22	0.65	0.9937	*	0.83	0.9948	**	†	†	0.3609
Connecticut	11	22	0.09	0.9844	ns	1.26	0.9894	*	†	†	0.4354
Delaware	21	11	0.61	0.8112	ns	1.88	0.9638	***		†	0.3612
Florida	19	19	3.77	0.9924	***	3.58	0.9929	***	†		0.3979
Georgia	14	14	2.26	0.9569	***	3.01	0.9738	***	†		0.3014
Hawaii	23	23	−1.44	0.9773	ns	−1.03	0.9374	ns	†	†	0.3665
Idaho	18	18	1.02	0.9828	*	−0.05	0.9833	ns			0.3147
Illinois	18	18	0.84	0.9938	**	0.48	0.9958	***	†	†	0.3257
Indiana	11	11	−1.05	0.9074	***	−1.69	0.9478	***	†		0.3004
Iowa	21	15	−2.16	0.9816	**	−1.95	0.8625	ns	†	†	0.2715
Kansas	21	18	1.29	0.8014	ns	−0.56	0.9408	***	†	†	0.3737
Kentucky	17	17	−0.23	0.8723	*	−1.09	0.9611	ns	†	†	0.2572
Louisiana	11	11	−0.35	0.9016	***	−0.08	0.9527	***			0.4459
Maine	13	13	−1.40	0.9637	**	−1.55	0.9312	ns		†	0.3630
Maryland	13	13	−1.14	0.959	ns	−0.84	0.9782	ns	†	†	0.4327
Massachusetts	18	25	0.12	0.9959	**	0.57	0.9975	ns	†	†	0.4316
Michigan	15	15	0.62	0.9446	ns	0.10	0.9782	ns	†	†	0.3067
Minnesota	21	21	−0.34	0.9764	ns	−0.21	0.9766	ns	†	†	0.3297
Mississippi	17	17	0.29	0.8779	***	0.29	0.8917	**			0.2616
Missouri	20	20	−1.49	0.9863	***	−1.35	0.9919	***	†	†	0.3026
Montana	18	16	0.85	0.9373	***	1.84	0.9831	***			0.4221
Nebraska	24	24	−1.05	0.9139	**	−1.18	0.9546	*	†	†	0.3291
Nevada	21	14	−1.54	0.9549	*	−1.53	0.9789	***			0.2479
New Hampshire	16	12	2.44	0.987	***	2.91	0.9977	***	†		0.5110
New Jersey	20	20	0.68	0.9939	***	1.08	0.9947	***	†	†	0.4231
New Mexico	18	18	0.94	0.8752	ns	1.40	0.8883	ns	†	†	0.2441
New York	10	10	0.56	0.9951	**	0.93	0.9911	ns	†	†	0.4409
North Carolina	20	20	−1.15	0.9809	ns	−1.12	0.9888	*	†		0.3775
North Dakota	15	13	0.43	0.8756	ns	0.64	0.9141	**		†	0.1883
Ohio	20	20	−1.51	0.9767	*	−1.79	0.9808	ns	†	†	0.3518
Oklahoma	20	20	−0.45	0.8648	*	−0.28	0.8964	ns	†		0.3052
Oregon	25	25	−0.86	0.9551	ns	−0.66	0.931	ns	†		0.4190
Pennsylvania	17	26	−0.63	0.9912	***	−0.81	0.9804	ns	†		0.3548
Rhode Island	13	13	−0.70	0.9873	ns	2.31	0.9849	ns		†	0.4675
South Carolina	11	11	0.45	0.9727	*	0.20	0.9774	*	†		0.3285
South Dakota	22	22	−0.60	0.9618	ns	−0.30	0.9778	ns		†	0.4001
Tennessee	12	24	−0.67	0.82	***	−0.45	0.9663	ns	†		0.3583
Texas	21	21	−1.20	0.9942	**	−0.03	0.9898	***	†		0.2669
Utah	10	10	1.23	0.9913	***	2.42	0.9957	***	†	†	0.2797
Vermont	17	23	−1.29	0.9951	***	−1.14	0.9949	***		†	0.4009
Virginia	23	23	−1.04	0.9854	ns	−0.67	0.9968	*	†		0.3677
Washington	12	15	−0.34	0.9892	ns	0.54	0.9911	ns	†	†	0.3569
West Virginia	19	19	0.12	0.8723	***	0.08	0.926	***			0.2429
Wisconsin	14	14	−1.00	0.99	ns	−1.28	0.9886	ns	†	†	0.3305
Wyoming	18	18	−0.30	0.9031	ns	−0.33	0.9164	ns			0.2383

^*p < 0.05

^**p < 0.01

^***p < 0.001.

^†denotes presence

Abbreviation: ns, not significant

On a national level, a 1% increase in US federal and state tobacco tax was not associated with a reduction in smoking-attributable BC mortality (−0.09% [(e^−0.0008744 − 1) × 100], R² = 0.9897, p = 0.6381). Tobacco taxation was also not associated with a reduction in US smoking-attributable BC DALY (1.77% [(e^0.01751 − 1) × 100], R² = 0.9879, p = 0.051). However, at the state level, 22 states demonstrated a statistically significant reduction in smoking-attributable BC mortality when associated with increased tobacco taxation (Table 1). This indicates that the magnitude of tax increase does not consistently predict the degree of mortality reduction. Arkansas showed the greatest reduction in mortality (−3.64%, p = 5.69 × 10⁻⁷), followed by California (−3.16%, p = 0.00263) and Indiana (−2.16%, p = 3.07 × 10⁻⁵) (Figure 2). In contrast, Nevada had the lowest reduction in deaths (2.44%, p = 0.0108), followed by Florida (2.26%, p = 3.41 × 10⁻⁷) and Iowa (1.29%, p = 0.001144). A total of 16 states showed a significant reduction in DALYs (Table 1). California had the largest decrease (−4.68%, p = 0.000218), followed by Indiana (−1.95%, p = 5.97 × 10⁻⁵) and North Dakota (−1.79%, p = 0.0035) (Figure 2). Florida had the lowest reduction in DALYs (3.01%, p = 1.66 × 10⁻⁷), followed by Nevada (2.91%, p = 4.12 × 10⁻⁵) and Texas (2.42%, p = 0.000621).

Figure 2:

Tercile Bivariate Hex-Tile classification of lagged tobacco tax-associated changes in state level smoking-attributable bladder cancer mortality and DALY

No states exhibited a combination of high DALYs with low death rates, or low DALYs with high death rates, based on the tercile bivariate model, suggesting that DALY and mortality move in parallel, rather than indicating a disproportionate burden in one domain. Reductions in smoking-attributable BC mortality and DALY rates were not associated with the presence of NCI-designated cancer centers or smoke-free laws alone across states (Table 1). Of the states that had negative changes in smoking-attributable mortality, 50% had smoke-free laws, while 50% did not (p = 0.3869). Of the states that had negative changes in smoking-attributable DALY, 58.6% had smoke-free laws, while 41.4% did not (p = 0.7752). States that had the largest negative change in smoking-attributable mortality and DALY were not associated with ones with higher smoking prevalence (Deaths: R² = 2.773 x 10⁻⁶, DALY: R² = 0.0389). Furthermore, a greater density of urologists was not associated with a negative change in smoking attributable mortality and DALY (Deaths: R² = 0.0154, DALY: R² = 0.0151).

DISCUSSION

This study evaluated the temporal relationship between tobacco taxation and smoking-attributable BC mortality and DALYs, at both the national and state levels. By incorporating lag analysis into our approach, we aimed to capture the delayed effects of tobacco control policies on long-term smoking-attributable BC mortality and disease burden. No significant associations at the national level were observed, however, several states demonstrated notable reductions in mortality and disease burden following increases in tobacco taxation. These findings indicate that local public health strategies may have a measurable impact on smoking-attributable BC outcomes over time. The application of lag-adjusted models allows for a more accurate understanding of how policy changes influence cancer trends, particularly for diseases with extended latency periods such as BC.

In a prior study by Lortet-Tieulent et al., which examined the proportion of cancer deaths attributable to cigarette smoking across US states, many of the highest rates were found in the South.¹⁰ Our findings did not reflect these same regional patterns. Previously, we observed a smaller reduction in BC incidence associated with tobacco taxation in Southern states;⁴ Arkansas ranked second in smoking-attributable cancer deaths. Discordantly, our analysis herein showed that Arkansas had the largest reduction in smoking-attributable BC mortality. While the state’s tobacco tax increased from $0.59 to $1.15 per pack in 2009 under Act 180 and has not increased since, this contrasting result might reflect improved smoking intensity in Arkansas. Investigation into the effects of Master Settlement Agreement (MSA) spending on smoking disparities in Arkansas found that MSA-funded programs effectively reached populations with the highest smoking rates and predominantly men.¹¹ This suggest a behavioral impact for the large reductions observed in smoking-attributable BC mortality in this state.

Tobacco control efforts beyond taxation alone may influence cancer trends. We report that associations of increased taxation and decreased smoking attributable BC mortality and DALY occurred in 22 states. Given that this pattern was not uniform across all jurisdictions, it suggests the role of complementary efforts such as cessation programs and smoke-free laws. In 2012, Indiana’s Smoke-Free Air Law enforced at least moderate smoke-free ordinances. Areas with comprehensive or moderate smoke free ordinances have been associated with a 1.2% reduction in smoking prevalence and an 8.4% decline in lung cancer incidence compared to areas with weaker policies.⁷ In our analysis, Indiana ranked third in reduced smoking-attributable BC mortality and second in reduced DALY. These improvements may be partially attributable to statewide smoke-free legislation, despite Indiana falling below the national average in cigarette taxation and allocating only 10.2% of the CDC-recommended budget for tobacco control in 2018.⁷ 64.8% of reduced smoking in the United States could be attributed to higher cigarette taxes and clean indoor air laws.⁸ However, given that smoke-free laws are not comprehensively implemented in Indiana currently and our data do not show a clear association between these laws and reduced smoking-attributable BC mortality or DALYs, the observed improvements are more likely driven by additional multifactorial components of the public health infrastructure. Thus, uneven adoption of smoke-free policies negatively affects health outcomes and increases health disparities in the United States.¹²

Nevada was found to have one of the highest burdens of smoking-attributable BC deaths. Recent investigation into casino air quality found that particulate matter levels in gaming areas were 5.4 times higher in casinos allowing smoking compared to a smoke-free casino, resulting in nearly 100,000 casino workers and millions of visitors in Las Vegas being regularly exposed to high concentrations of secondhand smoke.¹³ While policies such as the Nevada Indoor Clean Air Act have been effective in reducing the incidence and mortality of head and neck as well as lung cancers in Nevada,¹⁴ BC remains a significant burden, demonstrating the heightened need for stricter tobacco control measures. California continues to demonstrate a strong association with reductions in smoking-attributable BC burden. We also report that 58.6% of states with negative changes in smoking-attributable DALY had strong smoke-free laws. Robust policies at the state-level have led to significant decreases in smoking prevalence at the population level.¹⁵ Models of smoking-attributable deaths estimated that the Tobacco 21 policy could avert approximately 27,000 premature deaths by the year 2100 in California.⁶

Treatment for BC remain effective and continues to advance across all disease stages.¹⁶ Advances in BC management may partially explain some of the observed regional variation, particularly in states where tobacco control policies alone do not fully account for mortality trends. However, overall survival from BC has remained largely unchanged despite treatment advances. Our analysis did not show a significant association between reductions in smoking-attributable BC burden and the presence of NCI-designated cancer centers. While our study primarily examined the preventive impact of tobacco taxation on smoking-attributable BC outcomes, it is important to recognize that reductions in disease burden may also reflect differences in access to timely and effective treatment.

An inherent limitation of our study is due to its ecological design. Analysis at the individual level was not possible, which raises the potential for ecological fallacy. Another limitation is the inability to account for variations in BC stage at presentation, which can influence mortality and DALY outcomes. More intensive tobacco exposure has been demonstrated to be associated with more advanced disease at presentation and could potentially attenuate treatment efficacy and worsen adverse effects from treatment.^17,18 However, because we examined smoking-attributable BC at the population level and adjusted for national health expenditure, some of these differences may be accounted for indirectly. Additionally, although we used age-adjusted models and cross-correlation analysis to explore delayed associations between tobacco taxation and smoking-attributable BC outcomes, these methods cannot fully resolve temporal ambiguity. A strength of our study is the use of an extensive global dataset from the GBD database, which provides comprehensive population-level data from a national and state perspective.

Taken together, these regional differences suggest that health policy effectiveness is not solely dependent on tax increases but also on broader public health infrastructure, enforcement, and complementary efforts such as cessation programs, and smoke-free laws. While some states have leveraged tobacco taxation as part of a comprehensive control strategy, others may have seen limited benefit due to weaker implementation or insufficient supportive measures. This heterogeneity suggests tailoring tobacco control approaches to local contexts and ensuring sustained political and financial commitment to reduce the long-term burden of tobacco-related disease.

CONCLUSION

The long-term effects of taxation on smoking-attributable BC outcomes may be limited or difficult to detect, given the extended latency period of the disease. Tobacco tax policy is necessary but not sufficient on its own to reduce BC mortality and DALY. Efforts on early detection, prevention, and treatment remain essential for reducing the impact of smoking-attributable BC. Variability at the state level indicates that complementary public health measures play a critical role in influencing outcomes over time.