Smoking increases the risk for developing several chronic diseases, including cancer, heart disease, and diabetes.¹ Currently the leading cause of preventable death,² active cigarette smoking caused almost 450,000 deaths in the U.S. in 2014, resulting in approximately 6.4 million years of life lost.³ To reduce tobacco use, the Guide to Community Preventive Services recommends implementation of quitline programs, reducing out-of-pocket costs for cessation treatment, and smoke-free policies.⁴

Worksites have implemented these practices, with previous studies showing increases in cessation, quit attempts, and perceived support for quitting.^5–7 Small and low-wage worksites are understudied in this context, despite facing several barriers to health promotion.⁸ Small worksites also tend to report paying low wages to a large proportion of their workers.⁹ Workers earning lower wages not only have higher smoking rates,¹ but also smoke with greater intensity and are less likely to quit smoking.¹⁰¹¹ Further, while the current national smoking rate is 16%,¹² employees working for low-wage industries such as retail trade and accommodation and food services have higher smoking prevalence (22% and 29%, respectively).¹³

With noted exceptions,¹⁴ most studies conducted at small and low-wage worksites have focused on cigarette smoking; an examination of smokeless tobacco (SLT) use within these settings warrants further attention. SLT is a form of tobacco that is chewed instead of smoked and includes chewing tobacco, snuff, and Snus. SLT use contributes to the development of chronic diseases such as oral, esophageal, and pancreatic cancers.¹⁵ While the prevalence of SLT use among U.S. adults (3%) is lower than smoking, in contrast to the declines seen in smoking, SLT use has remained stable over time.¹⁶ There is evidence that smoke-free policies reduce SLT use,¹⁷¹⁸ although some studies have found these reductions to be non-significant.¹⁹ Worksite tobacco-free policies may support stronger reductions in SLT use,¹⁶ but additional research is needed to confirm this relationship.

The purpose of this study was to assess how tobacco evidence-based practice implementation is associated with tobacco outcomes among employees at small low-wage worksites. We examined three practice types: policy, program, and communication. Tobacco policies are rules that prohibit tobacco use and protect workers from its harmful effects. Tobacco programs are interventions designed to increase tobacco cessation among current users. Communication focuses on the promotion of tobacco cessation-related information, resources, and programs. Knowing how these practices influence tobacco behavior and perceptions can provide direction for efforts that seek to reduce tobacco-related disparities at the worksite. We proposed the following hypotheses:

This study uses baseline and follow-up data from a three-arm randomized controlled trial (RCT) testing an intervention to increase implementation of evidence-based health promotion practices at small low-wage worksites.²⁰ The RCT was conducted in King County, Washington. Worksites included in the trial employed between 20 and 200 employees and were classified as one of six industries with mean annual wages less than $40,000 in 2010: accommodation and food services; arts, entertainment, and recreation; educational services; health care and social assistance; other services except public administration; and retail trade. (Note: Three exceptions were made for companies close to the 20–200 employee inclusion criteria; these companies employed between 206–235 employees). Worksites were recruited from a purchased list of companies and through an existing partnership network. The University of Washington Institutional Review Board approved all study protocols.

Briefly, the two intervention groups (HealthLinks and HealthLinks+) were assessed on their current implementation of evidence-based practices at baseline. Worksites then received a tailored report with recommendations for practice improvement, a toolkit to help them through the implementation process, and support from a workplace health interventionist. The HealthLinks+ group also received recommendations and support for forming a worksite wellness committee. The delayed control group was eligible to receive the intervention after providing two-year follow-up data. For a more detailed description of the trial and study methods, refer to Hannon et al.²⁰

Table 1 describes the baseline characteristics for worksites and employees. More than half of the worksites were in healthcare & social assistance/educational services industries, and almost all worksites offered health insurance. Most employees were female, non-Hispanic or Latino, white, and had at least some college education. The mean age of employees was 40 years, and about a third of employees reported an annual household income of $75,000 or more. The median household income was $50,000-$74,000 (data not shown), which is lower than King County’s median household income of $78,800 in 2016.

Table 2 describes the change in worksite tobacco evidence-based practice implementation across the three time points (baseline, 15 months, and 24 months). From baseline to 15 months, mean implementation scores increased for all three practices (policy, program, and communication). Program evidence-based practice implementation showed the largest increase, from 0.39 (SD=0.45) to 0.61 (SD=0.46), a 0.22 absolute difference. We also observed increases, from 15 months to 24 months, in implementation of evidence-based policies and programs, although these increases were smaller than we observed from baseline to 15 months. Communication evidence-based practice implementation decreased slightly during this time, from 0.13 to 0.11. Change in implementation over time was significant for all practices (p<0.001). Tobacco policy and program scores increased from baseline in all three arms, while tobacco communication scores only increased in the two intervention arms (data not shown).

Table 3 describes the change in employee tobacco outcomes across the three time points. At baseline, approximately 12% (n=315) of employees were current smokers, and about half of these smokers had attempted to quit smoking within the past six months. The percentage of SLT users was smaller, around 2% (n=47). Only a third of current tobacco users (n=97) strongly agreed or agreed that their worksite supported them in trying to quit using tobacco. For all four outcomes, there were no significant changes over time.

Table 4 describes the results from the multivariable GEE analyses. In support of H3(d), greater tobacco communication evidence-based practice implementation was associated with an increased odds of strongly agreeing or agreeing that their worksite provides support for cessation (OR=3.15; 95% CI: 1.14, 8.71; p=0.027). Communication evidence-based practice implementation was not significantly associated with smoking status, SLT use, or quit attempts, therefore H3(a), H3(b), and H3(c) were not supported.

Tobacco policy evidence-based practice implementation was not significantly associated with smoking status, quit attempt, or perceived worksite support for cessation, therefore we did not find support for H1(a), H1(c), and H1(d). Counter to H1(b), an increase in policy implementation was associated with an increased odds of being a current SLT user (OR=3.19; 95% CI: 1.12, 9.13, p=0.031). Similarly, and counter to H2(b), an increase in program implementation was associated with a nearly two-fold increase in the odds of being a current SLT user (OR=1.84; 95% CI: 1.10, 3.07; p=0.019). Program evidence-based practice implementation was not significantly associated with any other tobacco outcome, therefore we did not find support for H2.

Greater tobacco communication implementation was associated with better perceived worksite support for cessation. However, in contrast to previous studies,^23–25 communication practice implementation was not associated with tobacco-related behavior. While implementation significantly increased from baseline to 15 months, communication scores remained fairly low across all three time points. Stronger implementation of these practices may therefore be needed to change behavior. More frequent promotion of tobacco cessation resources and use of multiple promotional methods (e.g., posters, newsletters, websites) are two examples of increased implementation.

Policy and program evidence-based practice implementation were associated with increased odds of being an SLT user, contrary to our hypotheses and previous studies.^17–1926 It is possible that employees switched from cigarettes to SLT following greater practice implementation, although most studies have not found evidence to suggest that this switch in tobacco use behavior widely occurs.^17–192728 A more probable explanation is that our findings are due to random variability. The sample size for current SLT users across the three time points was small (range: 47 to 59), and only three worksites had five or more SLT users at any given time. Since the data are cross-sectional and not linked across time, changes in the workforce composition could have also contributed to our findings. Future research efforts should seek to examine these relationships longitudinally and among a larger sample of SLT users.

Despite the fact that tobacco program evidence-based practices were not promoted in the RCT intervention, implementation of this practice increased significantly over time across both intervention and control groups. One possible reason for this increase is the enactment of new Affordable Care Act (ACA) requirements, which coincided with the timing of data collection for the trial. Specifically, employers with 50 or more full-time employees were required to begin offering health insurance coverage that includes evidence-based preventive services such as tobacco cessation treatment.²⁹³⁰ The largest increase in implementation occurred between baseline and 15 months, when employer requirements for several worksites went into effect. At 15 months, all employers who offered a tobacco program indicated that their program was provided by an insurer. These points provide evidence in favor of the ACA driving increases in tobacco program evidence-based practice implementation.

However, program implementation was not significantly associated with positive changes in any of the tobacco outcomes we observed. Given the low communication implementation scores observed across all three time points, it is possible that some employees lacked knowledge and awareness of program changes. As noted by McAfee et al., “The Affordable Care Act has the potential to dramatically increase coverage of evidence-based cessation treatments and make these treatments available to millions of Americans. However, the potential benefits of this Act will only be realized if both smokers and physicians are aware of the opportunities it affords.”³¹ Employers must therefore promote awareness of cessation programs to ensure employees know what opportunities are available to them to help them quit.

There are several limitations to this study. Of the six industries eligible to participate in the trial, some were over-represented and others were under-represented, which could reduce generalizability of our findings. As mentioned earlier, we did not link employee data across the three time points, which limited our ability to assess change in behavior over time within individuals. We did not collect data on electronic cigarette (e-cigarette) use, a behavior that has attracted greater attention in recent years due to its potential for harm.³² This study also has several strengths, among which include a strong research design, large sample size, and use of data from multiple time points that enabled us to assess trends in worksite tobacco control practices over time.

This study highlights the potential benefit of increasing tobacco communication evidence-based practice implementation to increase perceptions of cessation support. Communication practices include promoting evidence-based interventions (e.g., quitline) and should be considered a complement to other tobacco control initiatives to improve employee health. This study also highlights the potential impact that federal laws such as the ACA may have made on tobacco program availability at small low-wage worksites. However, additional efforts may be needed to change employee tobacco use behavior, including stronger implementation of evidence-based practices and increased awareness of practice changes among employees.

Future research should explore the relationship between tobacco-free practice implementation and SLT use among a larger sample of users. Qualitative research could provide a better understanding of how SLT use may change with increased practice implementation. Studies that assess the impact of the ACA on tobacco control at small low-wage worksites would provide clearer information on employee awareness and use of these programs, and could help to inform intervention efforts in tobacco control. The current study expands our knowledge of how tobacco evidence-based practice implementation can influence employee tobacco outcomes, and contributes to our understanding of how to reduce tobacco use within small low-wage worksites to improve employee health.