Objective: To test the efficacy of a self-determination theory-based intervention plus instant messaging to help smokers with cancer quit smoking.

Hypothesis to be tested: Subjects who are allowed to choose their quit schedule and receive regular instant messaging about smoking cessation will show a significantly higher biochemically validated quit rate than those who receive only brief advice to quit immediately.

Design and Subjects: An RCT will be conducted to 1448 smokers with cancer attending the outpatient clinics of five major acute care hospitals in Hong Kong for medical follow-up.

Instruments: A structured questionnaire will be used to assess subjects' smoking history and demographic and clinical characteristics. EQ-5D-5L will be used to measure subjects' quality of life.

Interventions: The intervention group will receive brief advice and will be invited to choose their own quit schedules (immediate or progressive) in the outpatient clinics. They will receive instant messaging about smoking cessation during the first 6-month follow-up period. The control group will receive brief advice to quit smoking immediately in the outpatient clinics, and will receive a placebo intervention during the first 6-month follow-up period. Subjects in both groups will receive leaflets on smoking cessation.

Setting:

The setting is outpatient clinics of five major acute care hospitals in different clusters in Hong Kong.

Intervention group

In outpatient clinics Subjects will first receive a brief intervention using the Ask, Warn, Advise, Refer, and Do-it-again (AWARD) model. Previous clinical trials of smoking cessation have provided evidence of the effectiveness of brief interventions using the AWARD model. The brief intervention will comprise the following steps: (1) ask about smoking history; (2) warn about the high morbidity and mortality risks associated with smoking; (3) advise to quit: subjects will be allowed to select their own quit schedules after discussing their situation with the research nurse [quit immediately (QI), or quit progressively (QP) with the ultimate goal of complete cessation]. The content Scripts for the QI and QP interventions will emphasize the health benefits of quitting; (4) refer smokers to a smoking cessation clinic or a smoking cessation hotline: 1833183; and (5) do it again: repeat the intervention and encourage smokers who fail to quit or relapse to try again during each telephone follow-up. The whole intervention will last approximately 5-8 minutes. This brief intervention will be cost-effective and more feasible for routine use in clinical practice by healthcare professionals after minimal training.

Subjects who opt to quit progressively will receive a smoking reduction leaflet, which contains reduction strategies and a suggested plan to reduce smoking. Subjects will also be given the option to consider a tailored quit schedule for themselves after the discussions with the research nurse. The research nurse will motivate subjects to reduce cigarette consumption and then quit at their own pace, but the whole process should not exceed 6 months.

At the end of the face-to-face session, subjects will be informed that over the next 6 months, the research nurse will help them to adhere to their schedules by sending WhatsApp/WeChat messages. Each subject will receive a specially designed leaflet for smokers with cancer entitled, 'It is never too late to quit smoking.' The content will highlight smoking risks, the benefits of quitting, and myths about quitting smoking. The advisory messages given to subjects will be standardized with reference to the specially designed leaflet.

Follow-up booster intervention During the first 3 months of the study, the research nurse will deliver WhatsApp/WeChat messages at least once per week. In addition, subjects will receive four independent 1-minute videos (one video will be sent in weeks 1, 5, 9, and 13) via WhatsApp/WeChat. The video content will focus on the health benefits of smoking cessation after a cancer diagnosis. Between the 3-month and 6-month follow-up assessments, infrequent messages (approximately one WhatsApp/WeChat message per month) will be sent by the research nurse to follow the subjects ' progress, respond to their questions, and maintain contact.

Control group

In outpatient clinics Subjects will receive a brief intervention using the AWARD model. However, all subjects will be advised to quit immediately. Control group subjects will receive the same leaflet for smokers with cancer as intervention group subjects.

Follow-up Subjects will receive a placebo intervention that follows the same WhatsApp/WeChat schedule as the intervention group, but the messages will contain only general health advice, such as to perform more physical activity and eat more fruit and vegetables. The control group will not receive any videos.

Training and quality assurance All research nurses will attend a training workshop conducted by the research team before the study starts to ensure that they have the necessary knowledge and skills to deliver smoking cessation advice using the AWARD model. Regular case conferences, quality checks using audiotaping, and audit procedures will be conducted to maintain the quality and uniformity of the interventions. Moreover, research nurses will be trained to screen for eligible cases and conduct baseline/follow-up surveys and biochemical validation.

Quality and data security control:

The proposed study will follow the quality assurance protocol developed for a previous project. The principal investigator (PI), co-investigators (Co-Is), project coordinator, and research nurses will organize meetings with the Chief of Service, nurse managers, and frontline nurses to explain the protocol and the study logistics, and to examine the physical facilities available in the outpatient clinics. The project coordinator will be present during the first week to monitor the subject recruitment and data collection processes. All completed questionnaires will be sealed in separate opaque envelopes that will be kept in secure lockers provided by the outpatient clinics. The project coordinator will personally collect the completed questionnaires weekly. All collected questionnaires will be stored in a secure locker and the keys will be kept in our department. The project coordinator and other research assistant(s) involved in this project will be responsible for entering the data into SPSS software, and the resulting dataset will be encrypted and stored on an assigned computer in our department. Only the PI, Co-Is, project coordinator, relevant research assistants, independent data monitoring committee, and institutional review board will have access to the data.

Subject recruitment We will obtain ethical approval from the institutional review boards of the five hospitals. Potential subjects will be assessed for eligibility, and the purpose, design, procedures, and potential benefits and risks of the study will be explained to them. Informed written consent will then be sought. All subjects will be assured that their participation will be voluntary, with no prejudice attached to refusal, and that the information they provide will be kept confidential.

Data processing and analysis

Process evaluation will be conducted with subjects after the 12-month follow-up. Process evaluation can help to identify the strengths and limitations of a new care delivery model from the subjects' perspective. Moreover, process evaluation can help to identify the most important elements of the intervention and improve future implementation of the intervention. Most importantly, it provides information about optimizing the quality and efficacy of a newly developed model of care. Based on their smoking status, 40 subjects from the intervention group (20 quitters and 20 non-quitters) will be interviewed and a final sample will be determined using data saturation. A one-to-one audiotaped semi-structured in-depth interview (face-to-face) will be conducted with each subject.

Sample size G*Power will be used to estimate the sample size, following two previous RCTs. We predict that the proposed intervention will result in at least a 4% difference (7% vs 3%) in biochemically validated abstinence between the two groups at 6 months. The research team has reached a consensus that such changes constitute a minimally importance difference (i.e. are clinically significant) and thus warrant a change in patient management. To detect a significant difference between intervention and control group quit rates with a power of 90% and a significance level of 5%, at least 507 subjects will be needed per group. To allow a potential retention rate of 70% at the 12-month follow-up, 1448 subjects will be recruited.

Data Analysis The baseline characteristics of the two groups will first be compared using the chi-square test for categorical variables and analysis of variance for continuous variables. Intention-to-treat analysis will be used by imputing all non-responses at follow-up by baseline values (i.e. assuming failure or no change after the intervention), to yield more conservative effect size estimates.

The primary analysis will involve (1) main effect: intervention vs. control on biochemically validated quit rates at 12 months using chi-square test or Fisher's exact test.

The secondary analyses will include (2) main effect adjusting for baseline difference; (3) all secondary outcomes at 6 and 12 months; and (4) use of a generalized estimating equation model to calculate the adjusted ORs for validated and self-reported abstinence after adjusting for fixed clustering effects, baseline demographic and clinical characteristics that show a significant difference, and the within- subjects effect of the repeated measure outcomes (3, 6, and 12 months).

Sensitivity analyses (e.g. complete case or per protocol) for the primary and secondary outcomes will be conducted, depending on the observed pattern of missing data, using several methods of statistical imputation (e.g. multiple imputations/completed case/last observation carried forward) to assess the robustness of the findings. Subgroup analysis will be conducted to explore any interaction effects.

Our health economist will conduct a CEA using standard methods. The CEA will be populated using the empirical 12-month RCT data. An ingredients approach will be used to estimate the cost of the intervention program, including intervention materials (e.g. leaflet and videos), administration fees, and time taken to deliver the intervention. The health effectiveness outcomes will include the number of quitters at 6 and 12 months and QALY. EQ-5D-5L utility scores at baseline and follow-up assessments will be used to construct the QALY measure using the area under the receiver operating characteristic curve approach. Treatment costs for smoking-associated morbidities and relative risks will be extracted from the literature and the intervention costs derived from the RCT. Incremental cost-effectiveness ratios in the form of incremental cost per incremental QALY gained from the intervention, and incremental cost per one additional quitter, will be calculated. Both deterministic and probabilistic sensitivity analyses will be conducted to test the robustness of the model.

The process evaluation data analysis will begin immediately after each individual interview in accordance with the thematic analysis framework introduced by Braun and Clarke, using NVivo version 12 (QSR International Pty Ltd, 2018). The recordings will be fully transcribed verbatim, in Cantonese, to capture nuances of expression unique to the dialect, and selected quotations relevant to the themes will be later translated into English. During the coding process, two researchers will be responsible for analysing the narratives. The analyses will begin with an intensive examination of the transcripts to search for general constructs and themes. Special attention will be given to constructs that diverge from the major topics as framed by the guiding questions. The transcripts will first be coded using the open coding method. Details in the interview conversations will be closely examined to allow a large number of initial categories to emerge. As the number of codes increases, some closely related codes will be merged, resulting in a smaller, more manageable set of codes. Selective coding will then be used to code the transcripts using the established categories. Finally, a complete set of codes will be generated to facilitate comparisons and the development of themes and categories. Codes, categories, and themes generated will be compared with the established taxonomy for evaluating intervention quality related to behavioural change techniques for smoking cessation. The research team has considerable experience in conducting qualitative research and will strictly follow the CONSORT-EHEALTH statement.

Inclusion Criteria:

• smoked weekly in the last 6 months,
• diagnosed with any cancer for at least 6 months and at all cancer stages (I, II, III, or IV),
• aged 18 years or above and can communicate in Cantonese, and
• own a smartphone and able to use instant messaging (e.g. WhatsApp or WeChat)

Exclusion Criteria:

• individuals with unstable medical conditions, poor cognitive state, or mental illness as advised by the doctor in charge and noted on their medical records, and
• those participating in other smoking cessation programmes.

The relevant anonymized patient level data, full dataset, technical appendix, and statistical code are available on reasonable request. The approval from the Principal Investigator for the purpose of data use is required.

After the project is completed and the results of the project has been published

Request could be sent to Principal Investigator (williamli@cuhk.edu.hk)