Health Gains From US Smoking Decline Fall Short of Expectations

The gap in cigarette smoking prevalence between high- and low-prevalence states is expected to close somewhat in the next 10 years, but not the health consequences from smoking, according to findings from US Tobacco Use Supplement to the Current Population Surveys (TUS-CPS), which were repeated from 1992 to 2022. This is because the gains seen from a faster decline in smoking among young adults in the highest-prevalence states are expected to be counteracted by a slower decline in smoking prevalence among adults older than 50 years.

As a result, a forecasted decline in the adverse health consequences of smoking will also be hindered, according to the findings published in JAMA Network Open.¹ Among the 1.7 million respondents to the surveys, the majority (56.3%) were female and non-Hispanic White (78.0%), and the most common age group represented was 35 to 49 years (29.6%).

“We project that the national smoking prevalence will be under 5% by 2035,” Matthew Stone, PhD, first author and assistant professor at UC San Diego Herbert Wertheim School of Public Health and Human Longevity Science and Moores Cancer Center member, said in a statement.² “However, the much slower decline in smokers over the age of 50, particularly in previously high-smoking states, will mean that the high rates of lung cancer, heart disease, and chronic obstructive pulmonary disease that are caused by lifelong smoking will take longer to decrease.”

Survey Findings

Between 1992 and 2022, 18 TUS-CPS state-representative surveys were repeated to estimate sociodemographic trends in smoking prevalence, with hopes to close the gap between low- and high-prevalence states by 2035. The comparators were tertiles according to 1990s prevalence, and respondents were asked if they had smoked at least 100 cigarettes in their lifetime; those who responded yes were also asked if they currently smoked every day, some days, or not at all.

Overall, the national prevalence fell from 22.8% between 1992 and 2001 to 9.4% by 2022. Also during these times, breaking the totals down by state, smoking prevalence significantly declined in all states and Washington, DC—but there remained a wide range. From 1992 to 2001, prevalence ranged from 14.8% in Utah to 30.6% in Kentucky, and in 2022, from 1.4% in Hawaii to 16.0% in Iowa.

For the states in tertile 1 (lowest prevalence) from 1992 to 2001, prevalence ranged from 14.8% to 22.4%; tertile 2, from 22.4% to 24.2%; and tertile 3, from 26.3% to 30.6%. By 2022, all state tertiles had seen an approximate 13 percentage point decline. The prevalence rate in tertile 1 was 7.4%; in tertile 2, 10.0%; and in tertile 3, 12.7%. Further, looking ahead to 2035, these smoking prevalence rates are expected to continue to decline, but for there to be a lessening of differences among the 3 tertiles: 3.8% in tertile 1, 5.1% in tertile 2, and 6.6% in tertile 3. Also in 2035, the national prevalence is projected to be 4.9%, or the equivalent of differential rates of decline of 0.28 per person per year in tertile 1 to 0.38 in tertile 2 to 0.47 in tertile 3.

When the findings were analyzed by age group and compared between tertiles 1 and 3 for 2001-2002 and 2018-2022, the largest declines in smoking prevalence were seen for adults aged 18 to 24 years in tertile 3 states. Between 2018 and 2022, the rate of smoking fell by 21.3% in tertile 3 states vs 16.4% in tertile 1 states. In contrast, among those aged 25 to 34 years over these same periods, the prevalence dropped by 10.9% and 10.0% in tertile 1 and 3 states, respectively; 11.9% and 12.1% among those ages 35 to 49 years; and 4.7% and 2.3% among those 50 years and older.

No differences were seen according to sex, race, ethnicity, or education level.

Conclusions

“We observed that the most consequential narrowing of the prevalence gap occurred in states with historically high smoking rates,” the study authors wrote. But they issued caution about this, explaining that their 2022 prevalence estimates came in approximately 10% lower than those of the National Health Interview Survey (NHIS) for the same period—and that this is likely related to the TUS-CPS survey being designed to estimate state and national prevalences vs the NHIS gauging just national prevalence.³

Because the TUS-CPS surveillance system does not provide annual estimates, the authors of the present study imputed rates for missing years. Therefore, the full scope of true variability and potential biases may be limited, they noted.

References

1. Stone MD, Pierce JP, Dang B, et al. State and Sociodemographic Trends in US Cigarette Smoking With Future Projections. JAMA Netw Open. 2025;8(4):e256834. doi:10.1001/jamanetworkopen.2025.6834

2. Young adults drive historic decline in smoking. News release. Newswise; April 25, 2025. Accessed April 28, 2025. https://www.newswise.com/articles/young-adults-drive-historic-decline-in-smoking-but-older-smokers-slow-progress/?sc=dwhr&xy=10024803

3. Sánchez-Romero LM, Cadham CJ, Hirschtick JL, et al. A comparison of tobacco product prevalence by different frequency of use thresholds across three US surveys. BMC Public Health. 2021;21(1):1203. doi:10.1186/s12889-021-11283-w