It is noteworthy that almost one half of the study participants were current smokers; this is almost three times the 2007 smoking rate of 14% in the province of British Columbia, Canada [28
]. The participants tended to be heavy smokers who were highly dependent on nicotine. Other researchers also have reported very high rates of tobacco dependence among people with serious mental illness [6
], particularly those with schizophrenia [29
]. What is particularly troubling about our findings is that Vancouver is a region that has some of the strongest tobacco control measures in Canada [30
]. Although these measures have been instrumental in reducing the smoking rate to one of the lowest in Canada, a more tailored approach with considerable support, including pharmacological aid, social support and other resources, is needed for community-based people with serious mental illness.
We found that tobacco use rates varied by psychiatric diagnosis (39.2% for those with mood and anxiety disorders and 59.8% for those with schizophrenia), and that diagnosis was only predictive of men's smoking. The overall rate is lower than what has been reported elsewhere. It has been reported that, in Kentucky, the prevalence of current daily smoking for patients with bipolar disorder and schizophrenia were 66% and 74%, respectively [31
]. This may point to the importance of the social context in influencing the tobacco use of people with serious mental illness. Kentucky, a tobacco producing state in the USA, is reported to have the highest current smoking prevalence rate in the USA [32
More men than women reported being current smokers and the predictors of tobacco use varied by gender, in the gender-stratified analysis we found differential predictors of current smoking status. These findings suggest that while strategies need to be found for people with mental health issues, in general, services need to be gender sensitive. Gender has historically been a factor in tobacco use; men have been more likely to smoke than have women. Although the gender gap in the general population's smoking rate is narrowing, there remains a substantial differential in the smoking rates of men and women with serious mental illness. More research is needed of people with serious mental illness to untangle the relationships among gender, psychiatric diagnosis, the social context, and smoking status.
The specific needs of people with a diagnosis of schizophrenia spectrum disorder are unique. For example, they may require more support for cessation and they may need education about how their negative symptoms may interfere with some of the conventional methods of cessation support such as group interaction. The finding that smokers had higher rates of substance use than did the non-smokers echoes the results of other researchers and magnifies the overlap between tobacco use and other substance use. Best practice guidelines recommend that treatment for these co-occurring disorders be integrated [33
]. Although movement towards the integration of mental health and addiction services is gaining momentum, and more settings have begun to successfully incorporate smoking cessation into their practice [34
], there is still much dispute among clinicians about whether tobacco use should be treated as an addiction and considered part of the spectrum of substance use within the context of dual disorder services.
Many of the smokers in this study reported strategically using tobacco to cope with their psychiatric symptoms. Reports published elsewhere have discussed the complicated roles nicotine and tobacco play in the lives of people with mental illness [35
]. The stimulating effect of nicotine is known to modulate social and interpersonal factors to reduce anxiety and to relieve boredom. Nicotine also alters the neurochemistry of the brain and affects the rate at which psychotropic medications are metabolised [35
]. Clearly the use of tobacco has serious implications for psychiatric recovery, which is a compelling reason to advocate strongly for the clinical monitoring of changes in tobacco use in clients.
Tobacco cessation support is a service that should be offered to all clients wanting to stop smoking, and smoking cessation interventions have been shown to be effective in mentally ill clients residing in the community [36
]. The reason for the high smoking rates among persons with mental illness may, in part, be related to mental healthcare providers' reluctance to integrate interventions for tobacco reduction into their practice, and the lack of attention given to tobacco dependence in organizations providing services for the mentally ill. Integrated solutions must include preparing mental health providers to support tobacco reduction and smoking cessation efforts.
It is clear that the economic costs of tobacco use place a significant burden on people with serious mental illness, especially because many rely on government subsistence, which is well below the poverty line [37
]. At the time of this survey, income from a disability pension was capped at $856.42 per month. Social assistance for a single person with a disability, provided by the Government of BC, was 62% of the low-income cut off established by the federal government [38
]. Smokers in this study spent an average of $160 per month on tobacco; almost 20% of their monthly income. In addition, many of the smokers made choices to smoke "butts" and to buy cigarettes instead of food. It is well documented that poverty is associated with poorer health outcomes and the extra burden of tobacco-related effects confounds these people's already compromised health outcomes. Tobacco use treatments have been shown to be highly cost-effective [39
]. Subsidizing nicotine replacement therapy (NRT) is efficacious in significantly increasing cessation rates and the number of cessation attempts by smokers wanting to stop smoking [40
]. In heavy smokers, higher doses of NRT have been shown to increase cessation rates [41
]. A way to reduce both the physical and the economic burden of tobacco is for governments or third-party health insurers to provide nicotine replacement therapeutic products free of charge for people with serious mental illness.
These findings must be considered in light of several methodological limitations. First, the relatively low participation rate limits our ability to generalize to the community-based mental health population as a whole. Other community-based studies of people with mental illness have reported similar response rates [42
]. There are specific factors associated with seriously mentally ill people's willingness to engage in research [44
]. Many of these factors affected our ability to recruit participants, including the lack of a supportive research culture in the study settings and a reliance on mental health team staff for client referral. Client-specific factors included a fear that the information provided would not be kept confidential and would have an impact on their healthcare. The length of the questionnaire may have been a barrier; many people believed that they could not complete a 45-minute interview. The presence of some symptoms (e.g., paranoia) may have had an additional impact on recruitment. Another limitation of the study relates to the accuracy of the medical diagnosis data; 19% of the participants did not permit access to their medical records. Our reliance on self-reported diagnosis, for these case, may have resulted in misclassification bias. Additionally, some confidence intervals for the odds ratios were very wide (i.e., cocaine use, being widowed, and having no housing) indicating a lack of precision in these estimates.