One year after Massachusetts’ FTR was implemented, there were no statistically significant changes in current cigarette use among Massachusetts adults overall (relative to Connecticut adults) or among Massachusetts African Americans (relative to Connecticut African Americans). There was evidence that the policy had an inequitable effect on sexual minority cigarette use. Specifically, the policy led to a 14% greater reduction in cigarette use in Massachusetts compared to Connecticut among non-sexual minorities but no greater decrease in cigarette use in Massachusetts compared to Connecticut among sexual minorities.

This study is among the first state-level evaluations of a FTR with a control group. The lack of impact overall aligns with the findings from an evaluation conducted in San Francisco that found no change in cigarette use after a menthol ban was implemented [32]; however, the sample comprised clients in a substance use disorder treatment program, not the general adult population. A separate study of San Francisco’s comprehensive FTR that included young adults found changes in tobacco use after policy implementation [33]; among adults ages 18–24, there was a significant decrease in cigarette use after policy implementation. However, among adults ages 25–34, there was no decrease in cigarette use after policy implementation. Importantly, neither study had a control group of participants who were not under an FTR. As cigarette use continues to decrease among adults [34, 35] it is important to have a control group so that the effects of that specific policy can be isolated from the underlying secular trends.

In a national sample that did compare prevalence between those under an FTR and a control group, there was a significantly lower prevalence of current tobacco use among those under a policy (20.55%) compared to those not under a policy (23.98%) [36]. However, this study included participants aged 15–36 years and did not disaggregate the type of tobacco products, making it hard to compare the study findings to our own. The present study adds to the mixed findings on whether FTRs decrease adult tobacco use.

When we disaggregated the data, we found evidence that Massachusetts’ policy had a significant impact on sexual minority cigarette use. Contrary to the original hypothesis, cigarette use among sexual minority adults in Massachusetts increased slightly after policy implementation, while cigarette use decreased among sexual minority adults in Connecticut. Although this study did not investigate potential reasons for this finding, it is possible that there were unique stressors to sexual minority adults in Massachusetts that resulted in increased cigarette smoking. Additionally, because of the higher prevalence of cigarette smoking (menthol and non-menthol) among sexual minorities compared to the general population [37] this policy may have nudged this group toward smoking non-menthol cigarettes instead of cessation. However, more work is needed to explore these possibilities and understand this result.

Although there were non-significant results for African American cigarette use, these null results are important to highlight. Menthol cigarettes are a driver of tobacco-related health disparities among African Americans [16, 17]. There are many who advocate for a menthol ban to reduce health inequities among African Americans [38, 39], and modeling and simulation studies predict menthol bans would save hundreds of thousands of lives, many of them African American, over the span of a few decades [40, 41]. Yet, in the present study, there was no indication that Massachusetts’ FTR reduced cigarette smoking among African American adults.

One potential reason for the unexpected effects is the continued access to menthol cigarettes in bordering states. A study among Massachusetts adults who smoked menthol cigarettes found two thirds of respondents accessed menthol cigarettes in another state after the FTR was implemented [42]. Additionally, young adults in Massachusetts reported easily accessing flavored products through social sources despite the FTR [43]. Another potential explanation is policy enforcement challenges. Because policy implementation coincided with the COVID-19 pandemic, there were disruptions and competing interests for public health officials. The quantity of seized illicit tobacco products increased each year after implementation, suggesting enforcement activities have ramped up since the first year [44]. If there were broader geographic restrictions on menthol cigarettes (e.g., a national menthol ban) that more widely reduced access to the product and had effective enforcement, perhaps the hypothesized effect would have been observed.

This study has several strengths that should be noted. First, this study used a DID approach, which allows us to estimate causal effects of the policy using a pre-post analysis and a control state. Second, we looked at the policy’s impact both overall and among groups with higher rates of flavored tobacco use (African Americans and sexual minorities) to determine if the policy has differential impacts on specific populations. Finally, we used BRFSS data, giving us large, representative samples of the adult population in Massachusetts and Connecticut.

There are also several study limitations to note. First, there was only one post-policy timepoint and it was one year after the policy was implemented. Therefore, in addition to this being a short period to observe a behavior change such as cigarette cessation or decreased initiation, we were unable to look at trends after the policy was implemented. Second, the trends in tobacco use prior to policy implementation were not perfectly parallel between Massachusetts and Connecticut, suggesting the parallel trends assumption may not have been satisfied for all models; if there were other factors influencing tobacco use trends between 2019 and 2021, this may lead to biased results. Finally, categorization of sexual minorities was done in such a way that if a respondent provided an answer (not missing data or a refusal to answer) and did not explicitly say they were straight/heterosexual (they responded “don’t know/not sure”), they were classified for this analysis as a sexual minority. This was done to be as inclusive as possible; however, by categorizing sexual minorities this way, it is likely that some people who do not identify as a sexual minority were erroneously classified as such. This potential misclassification bias may result in an underestimation of the true effect size.