Supporting smoking cessation in chronic obstructive pulmonary disease with behavioral intervention: a randomized controlled trial

The study was a randomized controlled trial conducted form from January 2008 to May 2012, which involves three months patients’ recruitment, two years’ intervention, two years’ monitoring. Recruitment of practices took place in 28 communities based on our previous epidemiological study [21]. Fourteen healthcare centers enrolled in the study; General healthcare centers in the intervention group received support to implement the behavioral intervention program, whereas the control healthcare centers delivered usual care. Randomization took place on healthcare center level. The healthcare centers were classified in two classes: with high or low task delegation from general practitioners to nurses. The healthcare centers in the classes were then randomly allocated to the groups (See: consort Figure 1).

A two-sided P value <0.05 was used. Based on a two-sided Type 1 error (α) = 0.05, with an 80% power to detect a 25% relative reduction in quitting rates, allowing 20% loss during follow-up, we need that each group should contain a minimum of 7 healthcare centers s and at least 50 patients with COPD per group.

This study was approved by the Ethics Committee of the Xuzhou Center for Disease Control and Prevention and the Regional Ethical Vetting Board, Xuzhou, China. In addition, agreement was received from all of the relevant health centers. Informed consent was obtained from all participants.

Patients were recruited by their family physicians from 14 healthcare units in rural area of Xuzhou city, China, from January to March 2008. Patients had to meet the following criteria at baseline: COPD diagnosed by the standards set forth by the Global Initiative for Chronic Obstructive Lung Disease (GOLD) [22]. Which means all participants underwent standard measurements of lung function (post-bronchodilator forced expiratory volume in 1 s (FEV₁)/forced vital capacity (FVC) <70%) by the County People’s Hospital. Reversibility of airways obstruction was measured according to standard America Thoracic Society guidelines (over 15% and at least 400 ml improvement in FEV₁ after 400 μg salbutamol via a spacer) [23]. The study population consisted of current smokers with all stage COPD. All patients were aged 35 years or older, had smoked 1 cigarette or more per day for the previous year, and had not stopped smoking for more than 3 months during that year. We excluded participants who had any serious or unstable medical disorders such as psychiatric that might affect lung function or they had a current diagnosis of major depression.

Participants were considered as loss to follow-up if they could not contact, died, move to another place, withdraw consent, refuse to proceed, invalid data, or unable to complete the study due to other reasons.

During the screening phase, individuals selected a date to stop smoking (target day) and were told not to attempt to stop before this day. Patients were also told the aim of the study.136 general practitioners, working at 7 different healthcare centers of the intervention group, followed 6 hours training in behavioral interventions for quitting smoking. The training included: ingredients of tobacco, the potential harmful of smoking, smoking and COPD, harmful of exposure to tobacco smoke, smoking's impact on the economy, weigh the pros and cons of smoking and smoking cessation, why the COPD patients need quit smoke, benefits of quitting smoking, how to deal with smoking cravings, how to preventing relapse smoking, how to develop smoking cessation programs, how to help them quit smoking and the motivational interview.

The teams of general practitioners (including assistants and nurses) were established in the healthcare centers of intervention group. They were responsible for supervising and advising patients with COPD registering in their healthcare centers to quit smoke. The tasks of them included home visiting patients with COPD at least once a week, to obtain the current condition of quitting smoking, tell the patients how to do next, and record the time of quitting smoking. If a patient has quitted smoking, the general practitioners were to follow-up the patient once a week at the first month and afterwards once a month until the end the study. Meanwhile, they also were responsible for examining the exhaled carbon monoxide (CO) level for COPD patients. Every month, the participants were asked to meet together once to discuss relevant questions for quitting smoking and share the experience of quitting smoking. The professional group (including respiratory, rehabilitation, nutrition, sports and psychology specialists) routinely visited the COPD patients in healthcare centers every two months. The general practitioners in the healthcare centers reported the follow-up status of patient with COPD and smoking cessation conditions once a month. The professional group assessed smoking cessation conditions, developed the focus of the next follow-up period for each patient, and sent the focus to the teams of the general practitioners. The focus was how to improve patients’ quitting smoking during next follow-up period. The professional group also provided other physiological supporting, including the benefits of smoking cessation and how to deal with obesity. This study was lasted four years. During the follow-up period, personnel support was offered with easily access for all participants. All participants were tested for exhaled carbon monoxide (CO) level at the baseline and every six months in the follow-up period.

At the baseline, all patients were interviewed by the general practitioners and the professional group at healthcare centers. All participants were asked if they were willing to quit smoking. The participants who would like to quit smoking were given a booklet that included:

For the patients in intervention group, we strongly encouraged them to stop smoking, and received brief smoking cessation advice after the baseline interview, which consist of 5–8 minutes’ discussion with member of professional group about their smoking habits. Smoking cessation advice was focused on the risk of developing COPD, lung cancer, coronary artery disease, and the harmful effects to other family members. If the participants wanted to stop smoking, we provided them a plan to quit smoking. For example, we asked the participants to set date to stop smoking (target day). If the patients currently smoked twenty cigarettes per day, they could decrease two cigarettes per day or two days. If the patients did not want to stop smoking, we encouraged them to stop smoking within six months. They could decrease one cigarette per day or one week. It was suggested for all participants to postpone the time as long as possible when they want to smoke. Drinking water, talking with someone, or turning their attention on other things were recommended as effective ways to postpone smoking time. No pharmacological treatment for smoking addiction was provided in the current study.

An exhaled carbon monoxide (CO) measurement was taken at baseline and every six month in the follow-up period to confirm smoking status at healthcare centers. The primary efficacy measure was continuous abstinence from smoking for 2 years from the start of month 24 to the end of month30. We defined continuous abstinence as a participant report of zero cigarettes per day for at least 6 months and confirmed by exhaled carbon monoxide values of 10 parts per million

Secondary measures of efficacy were continuous abstinence during month 24–36 and 24–48. Continuous abstinence for months 24–48 was defined by participants being continuously abstinent during months 24–36, having a diary cigarette count of zero during weeks 37–48, and having exhaled carbon monoxide values of 10 ppm or less at month 48.

All participates were asked to complete a questionnaires that included age, sex, current employment status, education level, marital status, physical activity, alcohol use, number of smoking family member, number of cigarettes smoked, motivation to stop smoking, smoking pack years (average daily numbers of cigarettes smoked divided by 20 and multiplied by the smoking years). Comorbidities including chronic bronchitis or emphysema, asthma, other lung disease, diabetes, treatment for blood pressure, stroke, coronary heart disease (angina or heart attack), or other heart disease and conditions were recorded on patient’s baseline reports. Education was categorized into below high school, high school, or above high school levels. Alcohol drinking was defined as the consumption of at least 30 g of alcohol per week for one or more years. The height and weight were measured, and body mass index (BMI) was calculated (BMI = weight in kilograms divided by height in meters squared). BMI was categorized as underweight (≤18.5 kg/m²), normal (18.5 to ≤24.0 kg/m²) or overweight/obese (≥24.0 kg/m²) [24]. Mental health (depression and anxiety) was evaluated by the Hospital Anxiety and Depression Scale (HADS) [25]. Dyspnea was measured using the Medical Research Council (MRC) dyspnea scale [26] and FagerstrÖm Test for Nicotine Dependence (FTND) [27].

Participants in control group were treated by healthcare providers or general practitioners as usual manner. The content and number of usual care services were not standardized. Participants were follow-upped once every 2 months, and asked whether the symptoms aggravated, what medication they used, etc.

We did all analyses on the intention-to-treat population, which consisted of patients who took at least one month of our study. All participants who withdrew from the study were taken to be smokers thereafter.

SPSS for Windows version 11.5 (SPSS Inc.) was used for data analysis. Tests were considered as significant when P < 0.05. Baseline demographic and smoking characteristics as well as mean scores of MRC, and mental health were compared to assess potential differences between intervention group and control. T-test was used to analyze the differences in mean scores between the groups. A chi-squared test was used to evaluate the statistical significance of smoking cessation rates between the two groups. We studied the effect of age, sex, education level, household income, alcohol user, depression, anxiety, smoking pack-years, Nicotine addiction, dyspnea, smoking history of family members and smoking history of the responsible doctor or nurse of patients with multivariate logistic regression. These variables were added to the basic model including intervention group assignment and centre. Candidate variables (P < 0.05) were entered into the multiple logistic regression models analysis to analyze risk factors affecting quit smoking. Odds ratios (OR), confidence intervals (CI) and p values were calculated.

The two groups were drawn from a single district and were similar in annual averaged income, access to health services and main demography at baseline (Table 1).

Among 3562 smokers, 26.8% did not complete the follow-up visit, i.e. 46 subjects had moved away from the region, 39 declined to participate in the follow-up visit for miscellaneous reasons, 20 did not complete exhaled carbon monoxide (CO) level test, 23 were incomplete data, and 827 died (intervention: 370 vs control: 457; χ² = 28.76, P < 0.001). The total number of lost in behavioral intervention and control group were 437 and 518, respectively. Significant difference in lost rates was observed between behavioral intervention and control group (χ² = 13.94, P < 0.001). There were no statistically significant differences in the smoking history, age and sex between those who attended the follow up and non-responders. Finally, 2607subjects were analyzed in the study, 1377 in behavioral intervention group and 1230 in control group (Figure 1). In behavioral intervention group, the mean ages, median smoking years, median nicotine dependence scales and median MRC scales were 61.6 ±10.2 years, 37.8 ±11.7pack-years, 5.1 ±2.1 and 2.4 ±1.5, respectively. In the control group, the mean age of 61.5 ± 10.1 years, median smoking years of 37.6 ±11.5 pack-years, and median nicotine dependence scales of 5.0 ±1.9. No statistically significant differences in these indicators were observed between the two groups. Other descriptive characteristics for behavioral intervention and control group were shown in Table 2. Which showed no statistically significant differences between the intervention and control groups (all Ps > 0.05).

The rates of continuous abstinence from month 24 to the end of month 30 was higher in the intervention group than that in the control group ( 46.4% vs 3.4% p<0.001). 639 out of 1377 (46.4% ) participants receiving behavioral intervention remained abstinent compared with 42 out of 1230 (3.4%) receiving usual care. Rates of continuous abstinence were significantly higher with behavioral intervention than with usual care throughout the 24-month intervention phase and at the 48 month follow-up visit (Table 3). At the 48 month follow-up, more participants receiving behavioral intervention remained abstinent than those receiving usual care (44.3% vs 5.1% , p<0.001). Behavioral intervention was significantly better than usual care at months 36–48 during the follow-up phase (p<0.001).

The significant differences of continuous smoking abstinent rates were consistently observed for all subgroups (such as different ages, gender, education levels, household income, alcohol user and comorbidities etc.), i.e. higher smoking cessation rate in intervention group than that in control group (Table 4).

Independent predictors of successful smoking cessation included women, older age, no-drinkers, comorbidities, lower HADS-anxiety (HADS-A), lower HADS -depression (HADS-–D), less smoking pack years, lower FagerstrÖm score, and higher MRC score (Table 5). Patients were more likely to quit smoking if they had the family physicians/nurses nonsmoking or no family member smoking (Table 5). In other words, the family physicians/nurses and family members smoking were two strong predictors of failure to quit smoking. Other confounders might also play a role. The multiple logistic regression analysis showed significant effects for alcohol drinkers, comorbidities, HADS-A, HADS-D, smoking pack years, FagerstrÖm score, MRC score, family members smoking, and the family physicians/nurses smoking as predictors of smoking cessation by adjusting age and gender (Table 5). The family members smoking was the strongest predictor of sustained smoking (OR = 12.1).