Discussion

Among this population of healthcare workers at a military community hospital in California, the overall rate of tobacco use (19.8%) was consistent with rates among civilian working adults (Agaku et al., Reference Agaku, King, Husten, Bunnell, Ambrose and Hu2014b; US Centers for Disease Control, 2011a). The overall prevalence, however, does not reflect marked differences in tobacco use by different types of workers. Military enlisted members, also known as Navy hospital corpsmen, were almost three times more likely to use tobacco than other workers. In addition, men were more than twice as likely to use tobacco compared to women. Overall, 37% of male hospital corpsmen used tobacco, compared to less than 10% of all other workers. Heterogeneity of tobacco use by healthcare worker type is consistent with findings from other studies (Sarna et al., Reference Sarna, Bialous, Nandy, Antonio and Yang2014; Shahbazi et al., Reference Shahbazi, Arif, Portwood and Thompson2014; US Centers for Disease Control, 2011a).

The high rate of tobacco use among hospital corpsmen may not be surprising when one considers their very close alignment with the Marine Corps at this command. Almost all Navy hospital corpsmen had deployed to war, or trained to deploy, with the Marine Corps (US Department of Defense & Office of the Chief of Naval Operations, 2010). Their rate of tobacco use was nearly identical to rates reported by Marines in large military surveys (US Department of Defense, 2009; 2013). While the influence of peers on tobacco use has been well established in young adults (Kirst, Mecredy, Borland, & Chaiton, Reference Kirst, Mecredy, Borland and Chaiton2014), it may be very important for public health professionals to recognise that hospital corpsmen identify more closely with their Marine Corps brethren (O'Donnell, Reference O'Donnell2006) than with other healthcare professionals.

Most surveyed workers agreed with a local policy that prohibited tobacco anywhere on the hospital compound. Consistent with other studies of workers (Hehir, Indig, Prosser, & Archer, Reference Hehir, Indig, Prosser and Archer2013; Unrod, Oliver, Heckman, Simmons, & Brandon, Reference Unrod, Oliver, Heckman, Simmons and Brandon2012), tobacco users were significantly less likely than non-users to support restrictive policies. Nonetheless, nearly 50% of tobacco users still agreed with the policy of this military worksite. Increasing levels of support for tobacco bans, even among populations with historically high rates of smoking, have been reported by other investigators (Maguire, Brinkley, & Mansfield, Reference Maguire, Brinkley and Mansfield2010). Some have hypothesised that smokers support smoke-free laws in order to help themselves quit smoking (Nagelhout, Zhuang, Gamst, & Zhu, Reference Nagelhout, Zhuang, Gamst and Zhu2014).

Despite support for tobacco-restrictive worksite policies, less than one-third of survey respondents thought these policies were effective in reducing tobacco use in this healthcare worker population. This perception was consistent between tobacco users and non-users. Similarly, more than 40% of survey respondents reported being aware of policy violations. The local policy was introduced with very little pre-ban publicity and multiple post-ban announcements explaining the closure of the hospital's smoking area. Enforcement of the tobacco-free facility was performed by simple reminders from command staff on the premises; enforcement of tobacco-free vehicles in parking lots was less consistently enforced. Some investigators have suggested that workplace smoking bans, even when strongly supported, may have the unintended consequence of increasing stress among workers (Azagba & Sharaf, Reference Azagba and Sharaf2012). Others have explored whether workplace smoking bans push tobacco users to increase smoking in other venues; but such increases have not been found (Martinez-Sanchez et al., Reference Martinez-Sanchez, Blanch, Fu, Gallus, La Vecchia and Fernandez2014). When a cancer care centre in Florida recently implemented an outdoor smoking ban, employees’ pre-ban expectations of negative impacts on job performance and increases in smoking outside of work were not consistently found in post-ban surveys (Unrod et al., Reference Unrod, Oliver, Heckman, Simmons and Brandon2012).

The most interesting, and perhaps most important, finding of this project was that the tobacco-restrictive workplace policy did appear to impact tobacco use in this healthcare worker population. Among the 33% of tobacco users who successfully decreased their use between 2011 and 2012, only changing habits in response to worksite policy was significantly associated with success. Prior access to tobacco cessation resources, and being motivated by other personal and external factors, were not associated with successful reduction in tobacco use in multivariable models. Notably, use of cessation resources in the past year might have been considered an outcome of policy change, but the construct of this question (i.e., ‘ever’ use) made it more appropriate as a descriptor of tobacco users’ histories. While all factors that may have influenced behaviour change could not be addressed, only changing behaviour in response to the workplace tobacco ban was strongly associated with reduced tobacco use.

Paradoxically, those who decreased tobacco use were not more likely to cite tobacco policy as a motivation to quit, yet they were more than four times more likely to report changing their tobacco habits in direct response to policy. This apparent disconnect between tobacco users’ perception of what drives their behaviour and what actually drives their behaviour is consistent with research that automatic, cognitive-perceptual processes are responsible for many behaviours which occur outside of conscious awareness (Huang & Bargh, Reference Huang and Bargh2014). By extension, it is common for individuals to fail to recognise the determinants of their behaviour (Schall, Reference Schall2005), and this can be especially true with regard to determinants of tobacco use (Song, Brown, & Glantz, Reference Song, Brown and Glantz2014). Findings are also consistent with other studies of tobacco policy, demonstrating that many policies perceived as ineffective are actually impactful (Cummings et al., Reference Cummings, Fong and Borland2009; Hyland et al., Reference Hyland, Barnoya and Corral2012; Unrod et al., Reference Unrod, Oliver, Heckman, Simmons and Brandon2012; US Institute of Medicine, 2009).

Tobacco policies in the military have been a topic of much public health debate (US Institute of Medicine, 2009). The tobacco industry has historically, and successfully, promoted tobacco use among service members (Jahnke, Haddock, Poston, Hyder, & Lando, Reference Jahnke, Haddock, Poston, Hyder and Lando2011a; Offen, Arvey, Smith, & Malone, Reference Offen, Arvey, Smith and Malone2011). High rates of tobacco use among experienced service members, as well as new military recruits, show that the culture of tobacco use remains strong in this population (US Department of Defense, 2013). Despite overwhelming evidence that tobacco use degrades military health and readiness, well-meaning policymakers have historically protected the ‘rights’ of military members and veterans to easily and inexpensively use tobacco (Offen, Smith, & Malone, Reference Offen, Smith and Malone2013; Smith & Malone, Reference Smith and Malone2012; Reference Offen, Smith and Malone2013). Public health professionals have recently challenged these perceptions and suggested that it is time for the military to make tobacco use incompatible with service, in the same way that obesity and poor physical fitness are incompatible with service (Smith et al., Reference Smith, Jahnke, Poston, Williams, Haddock and Schroeder2014; US Institute of Medicine, 2009). Findings from this project would support the concept that strong tobacco restrictions are both acceptable and effective in a military healthcare workforce.

Tobacco policies in the healthcare industry have also been subject to some debate. While tobacco use among healthcare workers in the US is much lower than in many European countries (La Torre, Reference La Torre2013), some argue than any tobacco use by healthcare providers is unacceptable. The basis for such arguments includes the concepts of providers modelling healthy behaviour and maintaining personal health for good work performance. Some private healthcare employers have recently excluded tobacco users as new-hires, although such policies are unlikely to readily achieve public and legal support throughout the US (Asch, Muller, & Volpp, Reference Asch, Muller and Volpp2013; Huddle, Kertesz, & Nash, Reference Huddle, Kertesz and Nash2014). Findings from the current project may be important in illustrating that healthcare workers, including civilians, at a military command appear to support and accept fairly strong tobacco restrictions at work.

Several limitations should be considered when interpreting the results of this project. Because findings are based on one military healthcare institution in southern California, results may not be generalisable to other populations. Assessments of tobacco use in the workforce were based on occupational health records. Although these records covered the full population, they could not define worker types more distinctly than military officer, military enlisted, and civilian personnel. Similarly, while records comprehensively recorded up-to-date tobacco use as a ‘vital sign’, the type, quantity, and duration of tobacco use was not consistently found in health records. Health records may under-represent true tobacco use since, when stigma is perceived, up to 10% of smokers will non-disclose their smoking history to their healthcare providers (Curry, Richardson, Xiao, & Niaura, Reference Curry, Richardson, Xiao and Niaura2013). In this project, surveys evaluated tobacco use in more detail but were subject to their own limitations. Response bias may have been important, especially with the very low response rate among tobacco non-users. Tobacco users, perhaps understandably, showed much higher interest in the survey, with a response rate closer to 40%. Recall and reporting biases may have been minimised by the anonymous nature of the survey. Finally, survey construct features should be considered when evaluating the relationship between responses to different questions. Notably in this project, questions on tobacco use, including increases or decreases in use, were found on the first page of the survey instrument, and questions on changing habits in response to policy were found on the last page of the survey. This may have mitigated the potential for an association based on survey construct alone.

In summary, this project demonstrated a moderate level of tobacco use among healthcare workers at a military community hospital. Tobacco use varied markedly by demographic characteristics; male hospital corpsmen used tobacco at very high rates that paralleled tobacco use among Marine Corps members with whom they served. Implementation of strong tobacco-restrictive worksite policies was well accepted in this workforce. And, although most workers did not believe restrictive policies would reduce tobacco use, survey findings suggested a strong and statistically significant association between worksite policy and decreased tobacco use. In an era of increasing focus on tobacco use in the military, tobacco use in the healthcare profession, and disparities in tobacco use in the US overall (Warner, Reference Warner2011), these findings may be valuable to public health professionals and policymakers committed to eliminating tobacco-related health challenges.