This community-based participatory study evaluating the effect of a change in smoking status on asthma symptom control and health services use among adults living with asthma over a 12
month period shows that asthma patients who quit smoking experience significantly lower (80%) night-time symptoms and chest tightness than individuals who continue to smoke. In contrast, non-smokers who took up smoking during follow-up had significantly higher risks of asthma symptoms, 1.4-fold higher risk of having asthma attacks and nearly 5-fold risk of having visits to walk-in clinics for asthma.
It is important for health care providers to encourage asthma patients who smoke to quit. The emphasis for health care providers to take an active role and use strategies to educate asthma patients about the benefits of smoking cessation is well documented [15
]. Unfortunately, the success rates for smoking cessation in those with asthma appears to be disappointingly low and evidence on the effectiveness of smoking cessation in asthma is rather limited [4
]. Although health care providers may have tools to educate their patients about the impact of quitting smoking, they may still be lacking points of reference and achievable benefits of behavioural change that can be discussed with patients. In order to positively impact efforts of smoking cessation, a system that supports both patients and their health care team is required to successfully change patient behaviour. According to the Centers for Disease Control in the US, [18
] the majority of cigarette smokers quit without using evidence-based cessation treatments. However, there are other proven effective treatments: 1) brief clinical interventions (i.e. when a doctor takes 10 minnutes or less to deliver advice and assistance about quitting), 2) counselling, 3) behavioural cessation therapies, 4) treatments with more person-to-person contact and intensity, and 5) cessation medications found to be effective for treating tobacco dependence. To ensure the effectiveness of these approaches, such a system should ideally facilitate health care providers to make referrals and to be reimbursed accordingly for counselling. Furthermore, such a system would ensure patients while quitting smoking receive ongoing support in their routine asthma care follow-up from health care providers.
Some providers may inform smoking patients that they will experience improvement in lung function as early as one week following smoking cessation with a further improvement up to six weeks after quitting, as shown by Chaudhuri et al. [5
]. However, patients may not associate lung function improvement with “feeling better” and having “well-controlled” asthma; instead, improvements in asthma control may be more sought after. While we did not have lung function data to indicate pulmonary improvement, our asthma symptom control data provides similar positive evidence that smoking cessation is significantly associated with lower asthma symptoms within 12
months following cessation.
Our study findings offer health care providers measurable clinical benefits that can be imparted upon asthma patients about the benefits of quitting smoking. An example of a patient education message may include, “After you have stopped smoking for 12months, the odds of feeling chest tightness is lowered by 80% and the risk of night-time symptoms by 75% compared to an individual who is still smoking.
” It is equally important to emphasize to current non-smokers that smoking is one of the significant risk factors for asthma morbidity. Asthma patients who are currently non-smokers should be encouraged to remain so by their health care providers. For example, communicate to non-smokers that, “If you start smoking, your odds of wheezing will be increased by 30%, night-time symptoms by 50% and having an asthma attack (within a 6-month period) by 40%”.
Health care providers can convey the clinical benefits or risks associated with asthma and smoking to patients using such powerful and practical messages.
There is a large body of literature suggesting different methods to present treatment effects to patients that may affect health care decisions. Previously, it has been demonstrated that presenting benefits of specific treatments or behaviour to patients as a relative risk reduction with a comparator is a successful tactic [19
]. Fagerlin et al [20
] found that providing average risk information was valuable for physicians promoting patient behaviour with smoking cessation. Our suggested messages to health care providers agree with what was suggested in the literature about using relative risk reduction information and referencing it with a control comparator. Whether over time this tactic yields affective change will have to be demonstrated in a long-term follow-up study and is beyond the scope of this current study. Interestingly, we found that baseline non-smokers who were smoking at 12-month follow-up had an odds ratio of 4.57 of a walk-in clinic visit compared to patients that remained non-smokers. Previously, studies have found that smoking was associated with a greater longitudinal risk of hospitalization for asthma [21
] and found that persons who visited an emergency room for asthma in the past 12
months were 60% more likely to be smokers [3
While baseline smokers who quit by the 12-month follow-up visit experience improvements in asthma symptom control, individuals who continue to smoke might not show improvement due to the use of inhaled corticosteroids (ICS). Smokers with asthma compared with non-smokers with asthma are less sensitive to the short-term and medium-term effects of ICS on symptoms and lung function [22
], suggesting that ICS doses may need to be adjusted to attain asthma control in smokers.
The approach used in our current study is novel. To the best of our knowledge, the only published Canadian study comparing adult asthma patients’ symptom control among smoking status groups was based on a telephone survey conducted by Boulet et al. [1
]. The study assessed self-reported asthma control and health services use among individuals with asthma who were current smokers, had quit smoking or were never smokers. Patients with asthma who smoked were found to have worse control of their asthma than those who had stopped smoking or who had never smoked. However, this study was limited by the cross-sectional design and was unable to measure changes in health outcomes associated with the change in smoking status. In contrast, our study was a community-based face-to-face prospective study where asthma outcome measures were collected and assessed by health care practitioners (family doctors, nurses and asthma educators). The current study measured multiple outcomes including asthma medication use, work/school absenteeism, health status and asthma-related health services use. Most previous published studies have mainly focused on only one of these outcome measures [4
Some limitations to the present study must be noted. First, there is potential for smoking group misclassification due to categorization of smoking status at baseline and 12-month follow-up. The “new smokers” group may contain intermittent smokers should individuals not be smoking for the first time. After adjusting for age, sex, baseline symptoms and health services use, the new smokers had significantly higher odds of wheeze, chest tightness, night-time symptoms, asthma attacks and walk-in clinic visits compared to the non-smokers. Although not included in Table , we also conducted comparisons between the new smokers and all baseline smokers using multivariable mixed effect analysis. If these new smokers were indeed intermittent smokers being misclassified as new smokers, one would anticipate the odds ratios to be closer to one. The odds of asthma outcomes while lower were in fact similar to those between new smokers and non-smokers reported in Table . While there may be some level of potential misclassification, the observed differences between the groups cannot be completely explained by misclassification.
Since there may be important clinical differences between someone who stopped smoking a long time ago versus someone who stopped recently, it would be helpful to know time of smoking cessation. Collecting smoking history and more specifically, the number of packs of cigarettes smoked by smokers, ex-smokers, new- and intermittent smokers, would be useful. However, collecting smoking history retrospectively may not be ideal as it is prone to recall bias. The number of baseline non-smokers who were former ex-smokers or never-smokers may also influence significant findings; nevertheless, the benefit of adding a ”former” or “never” smokers group is limited by the number of individuals who fit the criteria.
Second, due to the design of the study, there is potential for interviewer bias. To avoid interviewer bias, one may consider blinding the interviewer to the status of the patients (case or control) which then would avoid differential probing. In our study, however, it was impractical to blind the interviewers (nurses and asthma educators) since these professionals recruited asthma patients into the study. The smoking status of the patients was, however, unknown to the interviewers prior to the interview and the information about smoking was collected after all other information was completed. In our study, we used standardized questionnaires to collect information on asthma symptoms, health services use and smoking status. The interviewers were not allowed to deviate from the scripts and thus, we did what was possible in a community-based study.
Finally, additional information about lung function and second-hand smoke was not included in the current study. We did not collect lung function data such as patient spirometry tests, and therefore, could not adjust for asthma severity when evaluating patients’ symptoms and asthma control. However, our findings are consistent with results reported by other studies that demonstrate correlation between the improvement in lung function measured objectively with “feeling better” after smoking cessation [6
]. Additionally, although we collected second-hand smoke information in the current study, it was not a variable accounted for in our findings. The absolute number of subjects not exposed to second-hand smoke in the ex-smokers group was 4, whereas it was 2 in the new smokers group. We attempted to include the second-hand smoke variable in our multivariable analysis, however, the small numbers did not allow the model to converge and we were not able to obtain the risk estimate on smoking status. Incorporating these variables as covariates in future larger studies would be an asset. In the current study, there were few individuals who changed smoking status in a 12-month period and thus, there may be a lack of power to detect an effect. A larger study would need to be conducted in order to have adequate power to adjust for the spectrum of asthma-related symptom and exposure variables.