Relapse Prevention Through Social Networking Groups

This research project consists of two parts: (1) Formative study on the use of social networking services for relapse prevention; and (2) A RCT to test the effectiveness of social networking services for relapse prevention 7.1 Phase 1 - Formative study 7.1.1 Study design and objectives Literature review and exploratory data analysis on the existing dataset of ICSCs will be conducted to better define smoking relapse and estimate the relapse rate in the natural setting. In addition, an extensive review of the effectiveness of other similar or relevant intervention studies, including our previous studies and preliminary results of using short messaging services at HKU will be conducted. A consultation group formed by the ICSCs and smoking cessation research team of HKU will be responsible to develop the intervention package through the social networking service, including the content of delivered messages, channels, format, time to be sent, intensity, moderator's protocol, manpower and other logistic arrangement. This intervention package will then be tested in some cognitive interview to assess the relevance, appropriateness, potential effectiveness and acceptance level of the SMS.

7.1.2 Selection of subjects We will invite about 6 quitters (without relapsing for at least one month) and 6 relapsers (smoke 5 cigarettes for 3 days after quitting) from the ICSCs by (i) gender (M, F) and (ii) age group (18-39, 40-59, 60+) to comment on the relevance, appropriateness, potential effectiveness and acceptance level of the intervention package.

7.1.3 Procedures for the cognitive interview The sites for conducting the cognitive interview of the intervention package contents will be the eight ICSCs. A face-to-face semi-qualitative interview will be conducted.

7.1.4 Instruments A semi-structured interview guide and instrument will be used to collect their opinions.

7.1.5 Sample size 12 participants (6 quitters and 6 relapsers) will be recruited. Phase 2 - Randomized controlled trial

1. Study design and objectives: This is a randomized controlled trial (RCT) to compare the relapse rate at the end of the first and third month between participants who join the social network (group A) and those who do not (group B). Research hypothesis: The relapse rate at 1-month and 3-month of the intervention arm (Group A) is lower than the control arm (Group B)

2. Target subjects

Inclusion criteria:

- Clients who have received the community-based smoking cessation counseling provided by ICSC

- Reported tobacco abstinence in the past 7 days at the 8-week follow-up from the first enrollment

- Aged 18 or above

- Able to communicate in Cantonese and read Chinese

- Have a local network mobile phone with mobile internet access

Exclusion criteria:

- Those with unstable physical or psychological conditions as advised by the doctor or counselor in charge

- Smokers with diagnosed alcohol dependence or the score of the Alcohol Use Disorders Identification Test (AUDIT) is 8 or higher

- Smokers who have difficulties (either physical or cognitive condition) to communicate

- Currently participating in other smoking cessation programs

3. Procedures:

Smokers who finished the 8-week treatment by TWGHs will be followed up by the staff of TWGHs through telephone or face-to-face visit. Those who meet the inclusion criteria will be introduced the RCT and invited to consent the participation of the RCT through oral or written consent. The social networking service for the intervention group will be available within 3 to 4 days after randomization.

Group A: Intervention arm with relapse prevention intervention in the social group via the social networking service (Potential channel: Whatsapp or Facebook). In order to sustain the abstinence among the participants in the social groups, the intervention content is about how to prevent smoking lapse and relapse, how to deal with high-risk situations and facilitate mutual support among participants. Moderators of the social groups will send messages of brief advice, give feedback to participants' enquiries and facilitate conversations in the platform. We propose the moderator to spend about 1 to 2 hours a day for the social group conversation in a flexible time schedule, and the duration will be at least one month. More details about the intervention will be confirmed at the preparation stage.

Group B: Control arm without any other interventions Counselors, from ICSCs and HKU, who have experience in smoking cessation and familiar with the operation of the social networking service, will be recruited as moderators of the social media groups to deliver the social networking services.

1- and 3-month follow up (after the random allocation) will be conducted by a trained interviewer (blinded to the group assignment) via telephone interview. Subjects who have reported abstinence from smoking at the follow-up will undergo a non-biochemical validation to validate their quitting status. They will be enquired a few questions about their quitting experience and impression, and their family members will be invited to report the smokers' smoking and quitting status thru telephone interview. Also, they will be invited by the interviewer to have biochemical validation tests (saliva cotinine test and exhaled CO test).

4 Randomization and Blinding: Simple randomization method by sequentially numbered, opaque sealed envelopes (SNOSE) will be conducted to ensure both the TWGHs counselors and the participants are masked from the allocation. The counselors who are responsible for the subject recruitment will be masked from the randomization. For each participant who is eligible for the study checked by the counselors, the counselor will call the site randomizer to allocate group identity for the participant by opening one sealed envelope according to the serial number sequentially and inform the counselor for further intervention. Assessors in the follow-ups will not be involved in the randomization process and be totally blinded to the group assignment (single blindness).

5 Instruments: The required baseline information related to demographics and background information will be confirmed at the preparation stage. ICSCs will collect these information from the smokers during the enrollment of the first visit. A structured questionnaire at 1- and 3- month from enrollment of ICSCs will be developed by adopting or modifying international and/or locally validated instruments. The questionnaire gathers information including smoking and quitting history, satisfaction and perceived-usefulness of the social networking services, risk perceptions of smoking, intention to quit smoking (stage of readiness to quit), antecedent factors of the Theory of Planned Behavior model (behavioural beliefs, outcome evaluations, normative beliefs, motivation to comply), pros and cons of smoking (decisional balance), self-efficacy to resist smoking, other lifestyle risk factors (drinking and physical activities), demographic information such as age, gender, marital status, education level and employment status.

6 Sample size determination: Since this is a pilot study to test the feasibility and generate preliminary estimates of effectiveness of the interactive messaging service for smoking cessation, the sample size would be 50 for each arm (i.e. total sample size = 100). Taking reference from the relapse rate (based on the 7-day point prevalence) of ICSCs of TWGH from 2009-2010, 251 out of the 763 quitters at 8 weeks reported that they smoked again at 6 months. In other words, the relapse rate (number of continuing smoking at 6 months among self-reported quitters at the 8-week follow-up) was 33%, assuming no new quitters. We assumed a similar result would occur for the control group of this RCT of relapse prevention (Group B). The effect size is set conservatively at 1:1.5 for the contrast between group A and B. In other words, the estimated relapse rates for A and B are 22% and 33%, respectively. As we only have 100 subjects in total, the power for detecting this difference using the Fisher's exact test is 26%, given alpha is 0.05. That means the study has low probability to reject the alternative hypothesis if it is false. The power would reach to 51.4% when the estimated effect size is set at 1:2. Alternatively, if the total sample size can be increased to 500, the power can reach to 85% for the 1.5 effect size.

7 Methods of statistical analysis Data will be entered into SPSS for Windows (version 20). Descriptive statistics such as frequency, percentage, and mean will be used to summarize the outcomes and other variables. Chi-square tests and t-tests will be used to compare outcome variables between subgroups. Fisher's exact test will be used to compare the quit rates between group A and group B. The intention-to-treat (ITT) analysis will be used such that those lost to contact and refused cases at the follow-ups will be treated as failure to achieve any cessation outcome.

8 Process evaluation To understand the perception and effectiveness of the social networking services, all the participants in the intervention group will also be invited to give feedback through the social networking platform. Furthermore, six participants from the intervention group will be selected (3 relapsers and 3 non-relapsers) to participate in an in-depth interview after the 3-month follow-up. Questions such as their satisfaction, perceptions on the advantages and disadvantages of the intervention, and any other suggestions for improvement will be explored and discussed.