Abstract
This cross-sectional study compares self-reported smoking by adults before vs after prohibition of menthol-flavored cigarettes.
Before the US Food and Drug Administration (FDA) proposed a rule to eliminate menthol as a characterizing flavor in cigarettes,1 Massachusetts became the first to implement a statewide menthol flavor ban in June 2020.2 This ban decreased cigarette sales in Massachusetts while disproportionately increasing sales in bordering states.3,4 Whether this decrease translated into a decline in cigarette smoking is unknown. This study estimated the association between the menthol flavor ban and current cigarette smoking in Massachusetts.
Methods
We analyzed data of participants (aged ≥25 years) from nationally representative, cross-sectional Behavioral Risk Factor Surveillance System (BRFSS) surveys of 2017 to 2021, with response rates ranging from 44% to 49%. This cross-sectional study was exempt from institutional review board approval or informed consent requirement per the Common Rule since the analyses used publicly available deidentified data. We followed the STROBE reporting guideline.
A difference-in-differences method was used to compare current cigarette smoking among individuals in Massachusetts who were surveyed before (January 2017-May 2020) and after (June 2020-December 2021) the ban with smoking in comparison states (eMethods in Supplement 1). Comparison states consisted of 42 states and Washington, DC, and excluded Florida and New Jersey (due to gaps in survey years) and California, Colorado, Illinois, Minnesota, and New York (due to local-level menthol flavor bans).5 The analysis controlled for sociodemographic characteristics (eg, age, sex, and educational level), state-level time-varying factors (eg, Medicaid enrollment rate), and survey month and year fixed effects (eMethods in Supplement 1). We collected and controlled for self-identified race and ethnicity to account for differences in menthol cigarette use among groups.
Survey weights were used in all regressions, and SEs were clustered within states. A 2-sided P < .05 was considered statistically significant. Stata, version 17.0 (StataCorp LLC) was used to conduct analyses.
Results
Of the 1 485 589 individuals included, 51.8% were females and 48.2% were males, with a mean (SD) age of 57.9 (16.3) years. Patterns in current cigarette smoking between Massachusetts and comparison states were parallel before the ban and diverged after the ban. There were significant differences in population composition between the Massachusetts’ and comparison states’ samples (Table 1).6 Compared with comparison states, the ban was associated with a decline in current cigarette smoking in Massachusetts by 1.0 (95% CI, −1.4 to −0.7; P < .001) percentage point (Table 2) or 8.1% relative reduction in smoking prevalence. The estimate was robust when Massachusetts’ bordering states were excluded from comparison states but was biased when the 2020 survey was excluded from the sample or when states with local-level menthol flavor bans were included in comparison states (Table 2).
Table 1. Current Cigarette Smoking and Sociodemographic Characteristics for Massachusetts and Comparison Statesa.
| Characteristic | Mean (SE) | Mean difference (SE) | P value | |
|---|---|---|---|---|
| Massachusetts | Comparison statesb | |||
| Current cigarette smoking, % | 12.9 (0.0) | 17.4 (0.6) | 4.5 (0.6) | <.001 |
| Sex, % | ||||
| Female | 52.6 (0.0) | 51.8 (0.1) | −0.8 (0.1) | <.001 |
| Male | 47.4 (0.0) | 48.2 (0.1) | 0.8 (0.1) | <.001 |
| Married and cohabiting, % | 60.0 (0.0) | 60.9 (0.4) | 0.9 (0.4) | .05 |
| Single, % | 40.0 (0.0) | 39.1 (0.4) | −0.9 (0.4) | .05 |
| Race and ethnicity, % | ||||
| Black | 6.4 (0.0) | 12.9 (1.5) | 6.5 (1.5) | <.001 |
| Hispanic | 10.8 (0.0) | 11.3 (3.4) | 0.5 (3.4) | .70 |
| White | 74.4 (0.0) | 69.6 (0.4) | −4.8 (3.6) | .20 |
| Otherc | 8.4 (0.0) | 6.2 (0.6) | −2.3 (0.6) | .001 |
| Educational level, % | ||||
| No high school diploma | 10.0 (0.0) | 12.3 (0.8) | 2.3 (0.8) | .009 |
| High school diploma | 23.3 (0.0) | 27.5 (0.9) | 4.2 (0.9) | <.001 |
| Some college | 24.4 (0.0) | 30.7 (0.6) | 6.3 (0.6) | <.001 |
| ≥College degree | 41.9 (0.0) | 29.3 (0.7) | −12.5 (0.7) | <.001 |
| Missing data | 0.3 (0.0) | 0.2 (0.0) | −0.1 (0.0) | <.001 |
| Age, % y | ||||
| 25-29 y | 9.0 (0.0) | 9.1 (0.2) | 0.1 (0.2) | .30 |
| 30-34 y | 10.2 (0.0) | 10.8 (0.2) | 0.6 (0.2) | .001 |
| 35-39 y | 8.6 (0.0) | 9.2 (0.2) | 0.6 (0.2) | .004 |
| 40-44 y | 9.2 (0.0) | 9.4 (0.1) | 0.2 (0.1) | .10 |
| 45-49 y | 7.4 (0.0) | 8.0 (0.2) | 0.6 (0.2) | .001 |
| 50-54 y | 10.3 (0.0) | 9.6 (0.1) | −0.7 (0.1) | <.001 |
| 55-59 y | 10.0 (0.0) | 9.2 (0.1) | −0.7 (0.1) | <.001 |
| 60-64 y | 10.1 (0.0) | 10.0 (0.2) | −0.1 (0.2) | .60 |
| ≥65 y | 25.2 (0.0) | 24.7 (0.6) | −0.5 (0.6) | .40 |
| Mean COVID-19 infection cases per 1000 peopled | 22.8 (0.0) | 25.8 (0.9) | 3.0 (0.9) | .001 |
| Medicaid enrollments per 1000 persons | 254.5 (0.0) | 218.6 (11.3) | −35.9 (11.3) | <.001 |
| Electronic cigarette, % | ||||
| Tax | 27.5 (0.0) | 17.2 (0.4) | −10.3 (4.4) | .03 |
| State-level sale restrictions | 6.3 (0.0) | 0.3 (0.2) | −6.0 (0.2) | <.001 |
| Household income, % | ||||
| <$10 000 | 2.1 (0.0) | 3.6 (0.2) | 1.4 (0.2) | <.001 |
| $10 000-14 999 | 2.7 (0.0) | 3.6 (0.1) | 0.9 (0.1) | <.001 |
| $15 000-19 999 | 4.1 (0.0) | 5.4 (0.2) | 1.3 (0.2) | <.001 |
| $20 000-24 999 | 3.7 (0.0) | 7.1 (0.2) | 3.4 (0.2) | <.001 |
| $25 000-34 999 | 6.1 (0.0) | 8.8 (0.1) | 2.8 (0.2) | <.001 |
| $35 000-49 999 | 8.6 (0.0) | 11.5 (0.2) | 2.9 (0.2) | <.001 |
| $50 000-74 999 | 11.0 (0.0) | 13.9 (0.3) | 2.8 (0.3) | <.001 |
| ≥$75 000 | 40.5 (0.0) | 31.3 (0.8) | −9.2 (0.8) | <.001 |
| Missing data | 21.3 (0.0) | 14.9 (0.3) | −6.4 (0.3) | <.001 |
| Observations, No. | 30 340 | 1 455 249 | NA | NA |
Abbreviations: BRFSS, Behavioral Risk Factor Surveillance System; NA, not applicable.
The BRFSS data used to compute the statistics were available at the individual level. Individuals in the 18- to 24-year age category were excluded because of Tobacco 21 laws that prohibited individuals under 21 years from purchasing tobacco products.6 The BRFSS provided the age categories, which made it impossible to exclude only individuals under 21 years. The means and differences with their SEs were computed using the svy command in Stata, version 17.0 (StataCorp LLC).
Comparison states were the 42 states and Washington, DC, that did not implement any local-level menthol flavor bans. Aside from Massachusetts, some jurisdictions in 5 other states (except California, Colorado, Illinois, Minnesota, and New York) implemented local menthol flavor bans. Florida and New Jersey were excluded from the sample due to gaps in the BRFSS survey in these states.
Other included the following races and ethnicities as identified in the BRFSS survey: American Indian or Alaska Native, Asian, Pacific Islander, and any other race not listed.
The monthly mean infection cases per 1000 persons represent the data for the entire sample period and not only the COVID-19 pandemic period.
Table 2. Estimates of the Association of the Massachusetts Menthol Flavor Ban With Current Cigarette Smoking Among Adults 25 Years or Older From 2017 to 2021.
| Difference-in-differences estimate (95% CI), percentage pointsa | Current smoking prevalence in Massachusetts in 2019, % | Changes in relative current smoking prevalence in Massachusetts, %b | P value | BRFSS unweighted sample size, No. | |
|---|---|---|---|---|---|
| Full sample analysis | |||||
| Unadjusted estimate | −1.0 (−1.3 to −0.8) | 13.0 | −7.9 | <.001 | 1 485 589 |
| Adjusted estimate | −1.0 (−1.4 to −0.7) | 13.0 | −8.1 | <.001 | 1 485 589 |
| Sensitivity analysis | |||||
| Sample excluding 2020 survey data | −1.7 (−2.2 to −1.2) | 13.0 | −13.2 | <.001 | 1 203 465 |
| Comparison states, excluding Massachusetts’ bordering states | −1.1 (−1.5 to −0.7) | 13.0 | −8.3 | <.001 | 1 364 459 |
| Comparison states, including states with local menthol flavor bans | −0.7 (−1.2 to −0.3) | 13.0 | −5.7 | .001 | 1 752 716 |
| Analysis by demographic characteristics | |||||
| Sex | |||||
| Male | −2.2 (−2.7 to −1.7) | 15.4 | −14.4 | <.001 | 659 030 |
| Female | −1.0 (−1.4 to −0.6) | 10.8 | −9.3 | <.001 | 826 559 |
| Educational level | |||||
| ≤High school diploma | −1.1 (−1.7 to −0.4) | 20.6 | −5.3 | .001 | 488 446 |
| Some college degree only | −0.9 (−1.7 to −0.2) | 16.0 | −5.9 | .01 | 409 072 |
| ≥College degree | −0.7 (−1.0 to −0.5) | 5.3 | −14.2 | <.001 | 585 347 |
| Age groups, y | |||||
| 25-64 | −1.2 (−1.7 to −0.7) | 14.5 | −8.3 | <.001 | 907 250 |
| ≥65 | −0.7 (−1.1 to −0.3) | 8.1 | −8.6 | <.001 | 578 339 |
| Race and ethnicity | |||||
| Male | |||||
| Black | −8.4 (−9.6 to −7.3) | 14.9 | −56.8 | <.001 | 45 518 |
| Hispanic | −5.3 (−7.1 to −3.5) | 19.5 | −27.2 | <.001 | 39 986 |
| White | −0.9 (−1.5 to −0.4) | 15.3 | −6.1 | .001 | 525 383 |
| Otherc | −1.1 (−3.9 to 1.6) | 12.1 | −9.5 | .41 | 48 143 |
| Female | |||||
| Black | 8.7 (5.7 to 11.7) | 14.9 | 58.6 | <.001 | 73 166 |
| Hispanic | −3.3 (−4.4 to −2.2) | 8.4 | −39.5 | <.001 | 48 023 |
| White | −0.2 (−0.6 to 0.2) | 11.1 | −2.0 | .32 | 652 968 |
| Otherc | 2.1 (−0.9 to 5.2) | 8.3 | 25.8 | .17 | 52 402 |
Abbreviation: BRFSS, Behavioral Risk Factor Surveillance System.
The difference-in-differences estimates show the association between current smoking and the Massachusetts menthol flavor ban that was implemented on June 1, 2020 (eMethods in Supplement 1).
Changes in current smoking prevalence were calculated by dividing the point estimate in the second column by the current smoking prevalence in Massachusetts in the third column and multiplying by 100.
Other included the following races and ethnicities as identified in the BRFSS survey: American Indian or Alaska Native, Asian, Pacific Islander, and any other race not listed.
Changes in current cigarette smoking in Massachusetts after the ban differed between males and females (−14.4% vs −9.3%; P < .001). However, no significant difference was observed by educational level (≤high school diploma: −5.3%; some college: −5.9%; ≥college degree: −14.2%; P = .13), age group (25-64 years: −8.3%; ≥65 years: −8.6%; P = .84), or among males by race and ethnicity (Black: −56.8%; Hispanic: −27.2%; White: −6.1%; other: −9.5%; P = .88) (Table 2). Changes did differ significantly by female race and ethnicity. Smoking increased among Black females (58.6%; P < .001), remained unchanged among White females (−2.0%; P = .32) and other racial groups (25.8%; P = .17), and declined among Hispanic females (−39.5%; P < .001).
Discussion
According to statewide BRFSS surveys, the Massachusetts comprehensive menthol flavor ban was followed by a greater reduction in current cigarette smoking in the state than comparison states overall and among males and Hispanic females while increasing among Black females. As the FDA plans to eliminate menthol as a characterizing flavor in cigarettes, interventions should address possible increases in cigarette smoking among Black females.
Study limitations included recollection error in self-reported BRFSS data and unmeasured differences between Massachusetts and comparison states that could bias the estimates. Furthermore, estimates should be interpreted only as indicative of associations and not as causation.
eMethods. Difference-in-Difference Specification
Data Sharing Statement
References
- 1.US Food and Drug Administration . FDA proposes rules prohibiting menthol cigarettes and flavored cigars to prevent youth initiation, significantly reduce tobacco-related disease and death. June 27, 2022. Accessed July 21, 2022. https://www.fda.gov/news-events/press-announcements/fda-proposes-rules-prohibiting-menthol-cigarettes-and-flavored-cigars-prevent-youth-initiation
- 2.Massachusetts Department of Public Health . Public Health Council approves regulations restricting access to vaping and tobacco products. December 11, 2019. Accessed July 21, 2021. https://www.mass.gov/news/public-health-council-approves-regulations-restricting-access-to-vaping-and-tobacco-products
- 3.Asare S, Majmundar A, Westmaas JL, et al. Association of cigarette sales with comprehensive menthol flavor ban in Massachusetts. JAMA Intern Med. 2022;182(2):231-234. doi: 10.1001/jamainternmed.2021.7333 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Asare S, Majmundar A, Westmaas JL, et al. Spatial analysis of changes in cigarette sales in Massachusetts and bordering states following the Massachusetts menthol flavor ban. JAMA Netw Open. 2022;5(9):e2232103. doi: 10.1001/jamanetworkopen.2022.32103 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Truth Initiative . Local restrictions on flavored tobacco and e-cigarette products. December 30, 2020. Accessed June 1, 2021. https://truthinitiative.org/research-resources/emerging-tobacco-products/local-restrictions-flavored-tobacco-and-e-cigarette
- 6.Preventing Tobacco Addiction Foundation, Tobacco 21. State by state tobacco laws, taxes, and statistics. Accessed October 6, 2022. https://tobacco21.org/state-by-state/
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods. Difference-in-Difference Specification
Data Sharing Statement
