Brief Intervention to Promote Smoking Cessation in DM Smokers

Preventing multiple high-risk behaviors is an important focus in diabetes treatment. The investigators found that cigarette smoking is a leading avoidable cause of death which kills about 6,000 people annually, accounting for one of every five deaths in Hong Kong. The smoking prevalence among males with diabetes (27.0%) is slightly higher than that of the general male population (24.5%) in 2006. Smoking has long been known to worsen the prognosis of diabetic patients in increasing the risks of macro- and micro-vascular complications and there is also evidence that smoking increases insulin resistance, worsens diabetes control [5] and may even induce the disease. Smoking cessation should be a natural choice for smoking diabetic patients as it can reduce the risk of morbidity and mortality.

Smoking is an addictive behavior and quitting is very difficult with high chance of relapse. Only a small percentage of smokers can quit without professional help. The stage-matched smoking cessation intervention is a commonly used approach. Current diabetic treatments are mostly focused on glucose-oriented care strategies, and specialized smoking cessation counseling and support are rarely incorporated in routine practice of diabetes care. This may be due to the already tight schedule of diabetic treatments and health care providers often find it difficult to provide additional smoking cessation intervention to their patients. Furthermore, smokers who have diabetes are more likely to be in the pre-contemplation stage for quitting and they seem to have less interest in their health. Many smokers who are not prepared to quit may reject smoking cessation intervention, hence approaches that highlighting the relevance and salience of health information by making it personally relevant (tailored) and addressing the specific characteristics of diabetic patients (targeted) to promote cessation are needed. The investigators propose a stage-matched smoking cessation intervention by increasing the relevance of health information with respect to readiness to change of the individual and addressing the relationship between smoking and diabetic complications and HbA1c level to increase the relevance to all diabetic patients. It is hypothesized that diabetic patients would be more likely to make change in smoking behavior, if health risks of diabetic complications and a likelihood of an increase in HbA1c level were addressed simultaneously.

The study will consist of 2 phases, Phase I is a qualitative study with focus group interviews and Phase II is a randomized controlled trial. In Phase I, current smoking and ex-smoking patients with diabetes will be recruited in the DM clinics of nine hospitals by purposive sampling for focus group interviews. Smoking diabetic patients will be recruited for modifying and piloting of the survey instruments to be used in Phase II. In Phase II, research assistants will determine the eligibility of patients in the clinics with the following inclusion criteria: (1) smokes at least 2 cigarettes daily in the past 30 days, (2) diagnosed with type 2 diabetes for at least 6 months such that their diabetic conditions and treatments should be stable, (3) age 18 or above, and (4) can communicate in Cantonese; and the exclusion criteria: (1) too sick to receive intervention, (2) poor cognitive state (People with poor cognitive state refers to those with poor cognitive function and unable to communicate logically), (3) mental illness, (4) is undergoing other smoking cessation program, and (5) with unstable diabetic (i.e. those with poor diabetes condition that needed hospitalization) or other medical conditions deemed to be not suitable by the doctor in charge. After obtained written informed consent on-site, the subjects will be randomly allocated either to the tailored and targeted smoking cessation intervention (DM) group, or the control group.

Randomization will be done within each clinic. For each clinic, our trained counselors will contact the consented subjects and administer the baseline questionnaire. The subject will then be randomly allocated either into the intervention group or the control group. Based on a block randomization design, the investigators shall generate a list of computer-based random number using SAS® software for each hospital to allocate consented patients into intervention or control group. Instead of distributing the random number list to the counsellors directly, counsellors have to open a serially-numbered sealed-opaque envelop to ensure allocation concealment. Counsellors will guide patients to complete the baseline questionnaire before they open the sealed envelope and assign patients into the intervention or control group. The envelope will contain an instruction sheet that specifies one of the two groups. All the participants will complete the measurement of exhaled carbon monoxide (CO) level at the site.

Analysis Data The qualitative data will be analyzed by using NVivo software for coding. The quantitative data will be analyzed by using the Statistical Package for Social Science. The baseline characteristics of the two groups will be compared by Chi-square test. The effectiveness in helping smoking patients to quit of the DM intervention to the Control group will be evaluated by comparing the percentages of participants with the outcomes in each group using chi-square tests on an intention-to-treat analysis. Independent t-tests will be used to compare the changes in HbA1c levels of the intervention group to the control. Multiple regression and multiple logistic regressions will be used to calculate the adjusted odds ratios for quitting: stage of readiness and reduction in cigarette consumption; and change in other health-related lifestyle practices; between the intervention and the control groups as well as between quitters and non-quitters. For stopping rule, futility test will be performed. It indicates the cessation of the clinical trial as soon as it becomes clear that a negative outcome (Control better than intervention) or equivalence between treatments is very likely.