Nicotine replacement treatment, e-cigarettes and an online behavioural intervention to reduce relapse in recent ex-smokers: a multinational four-arm RCT

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 13/155/05. The contractual start date was in April 2016. The draft report began editorial review in January 2020 and was accepted for publication in July 2020. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

© Queen’s Printer and Controller of HMSO 2020. This work was produced by McRobbie et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2020 Queen’s Printer and Controller of HMSO

Original and curtailed trial design

This was an individually randomised factorial trial that included two interventions and usual care. The original plan was to recruit 1400 participants and follow them over 12 months. Owing to circumstances described below, the trial was curtailed with 236 participants randomised and the follow-up period was reduced to 6 months.

Changes to trial design/protocol (trial curtailment)

The trial encountered serious delays. The commencement of the trial was delayed as a result of the failure of the contracted IT consultants to deliver a version of the S3P programme, which, among other improvements and updates to the version trialled earlier, was to include data management capabilities (e.g. programing and scheduling of follow-up surveys, dynamic lists of scheduled calls for telephone interviewers, reminder e-mails). Unexpected technical issues and loss of staff with expertise were the primary reasons for this failure. The work had to be abandoned in April 2017, at which time we moved quickly to make necessary adaptations to the earlier version to the extent these were possible and to program the data management aspects in REDCap^® version 7.0.19 (Vanderbilt University Medical Centre, Nashville, TN, USA).

Integrating these two systems presented additional technical challenges. The integrated system was finally implemented in Australia in August 2017. UK regulations, however, did not allow the use of a system based abroad, and the integrated system had to be fully replicated and installed on a UK server. Experts responsible for implementing the Australian installation were not allowed direct access to it. The work proved difficult for local IT personnel and took a further 6 months to complete. During these delays, the English stop smoking services were subjected to substantial changes, which emerged as the next major problem for the study because of effects on client recruitment. Service management moved from the NHS to local councils that commissioned private providers. The stop smoking service throughput shrunk by > 60% and few services remained that were able to contribute to such studies.

Recruitment in Australia also proved difficult. The original plan was to recruit participants solely from the quitline in the state of Victoria. However, the number of participants it referred was smaller than anticipated; therefore, to help boost numbers we sought to recruit more widely and the quitline from Tasmania agreed to take part. Despite this, referrals remained low. To boost recruitment in England, recruitment methods were expanded to include smokers quitting through the Stoptober campaign. In Australia, ex-smokers were approached via targeted Facebook (Facebook, Inc., Menlo Park, CA, USA; www.facebook.com) adverts and through St Vincent’s Hospital Melbourne, Fitzroy, VIC, Australia, recruiting patients who were discharged following a period of abstinence in hospital. The Australian site also expanded its recruitment window from 21–45 days post quit date to 7–100 days post quit date. Despite these measures, the recruitment rate remained slow in both countries. After consultation with funders, the decision was made to curtail the study. Recruitment was stopped in January 2019 and the decision was made to complete follow-up at 6 months rather than 12 months. Saliva sample collection, which had originally been planned to validate self-reports of abstinence at 12 months, and cost-effectiveness analyses could not be carried out.

See Appendix 1, Tables 25 and 26, for details of changes to the protocol.

Inclusion/exclusion criteria

Recruitment and setting

The original plan was to recruit 1400 participants (700 participants in England and 700 participants in Australia). In total, 236 participants were randomised between September 2017 and January 2019. Two participants withdrew and requested that their data not be used, resulting in a total of 234 participants (131 participants in Australia and 103 participants in England).

In England, participants were originally recruited from six stop smoking services: Tower Hamlets, City of London, Leicester, Medway, Birkenhead and Durham. Later, those who had quit using the Stoptober 2018 campaign were also invited.

Participants in Australia were originally recruited from quitlines in Victoria and Tasmania. From March 2018, participants were recruited via Facebook advertising and in the final months participants were recruited from among patients discharged from St Vincent’s Hospital Melbourne.

Interventions were delivered remotely (i.e. online and/or by telephone/e-mail/text messages) and participants did not have any face-to-face contact. Follow-up was completed online or by telephone calls conducted by researchers at the Health and Lifestyle Research Unit (London, UK) and Cancer Council Victoria (Melbourne, VIC, Australia).

Study procedures

Potential participants from stop smoking services, Stoptober, quitlines and St Vincent’s Hospital Melbourne were informed of the study by service staff and, if interested, were given written information and referred to the research team. Potential participants from Facebook visited the study Facebook page and the associated study website where they self-referred.

Potential participants were contacted by e-mail as soon as possible after referral. Those relapsing back to smoking before being reached were not invited to take part.

Participants were screened for eligibility online or by telephone, depending on their preference. Informed consent was also provided either online or verbally. Consenting participants completed a baseline survey (see Measures) and were randomised into one of four study arms (see Randomisation). Following this, the allocated treatment was initiated (see Interventions).

All participants were contacted by telephone approximately 1 week later to check that all was well, including that they had begun receiving text messages or received their chosen product. Any difficulties with the intervention were discussed and resolved.

Interventions

Participants who, at randomisation, were using stop smoking medications obtained from their service (e.g. NRT or varenicline) were encouraged to complete the course regardless of treatment allocation. In this report, we refer to these medications as ‘base medication’, unless they were similar fast-acting nicotine products as those we offered in the nicotine product intervention. ‘Base medication’ is defined as nicotine patches, varenicline or bupropion.

Common to all interventions was the provision of text messages for up to 6 months post quit date. These text messages included reinforcers of milestones, general motivational messages and some general hints. Non-S3P messages were non-interactive, untailored and of lower frequency than the text messages provided in the S3P condition, which are described in more detail below.

Usual-care arm

Participants received a text messaging programme without any personalisation or reference to the specific strategies focused on in the S3P intervention. There was a maximum of 33 messages sent over 6 months and participants could request to stop the messages at any time. The usual-care static text messages were not expected to have a clinically significant relapse prevention effect and, therefore, this arm is considered for the purpose of analysis to have received no intervention. All non-S3P arms received these messages in addition to any post-treatment care provided by stop smoking services that delivered the initial stop smoking interventions. Such interventions include an invitation to contact services if experiencing difficulties or lapses, although these are very rarely acted on and would be present across randomisation groups.

Nicotine product arm

Participants were given a choice of oral nicotine products: Nicorette^® 1-mg nicotine mouth spray (Johnson & Johnson, Brunswick, NJ, USA), NiQuitin^® 4-mg nicotine minis (Omega Pharma Ltd, London, UK) or branded Nicabate^® 4-mg nicotine minis (GlaxoSmithKline Australia Pty Ltd, Melbourne, VIC, Australia). In Australia, additional choice of 1.5-mg nicotine minis was offered as well. A refillable e-cigarette (Innokin^®, Endura T18E, Innokin Technology, Shenzen, China) with a choice of menthol- or tobacco-flavoured liquid, containing 11 mg/ml of nicotine, was also offered. Participants chose one product to use as a coping strategy if they found themselves at risk of relapse. The products were mailed to participants along with instructions on their use.

The initial supply comprised an e-cigarette starter kit (including the refillable device, a USB charger, a spare battery, a pack of five coils and four 10-ml bottles of e-liquid, with a total cost of £32.75), two bottles of nicotine mouth spray (with a total cost of £17.95) or six tubs of minis (with a total cost of £19.58).

In Australia, nicotine e-liquid is available only on prescription, therefore, if a participant requested an e-cigarette, a local physician was asked by the study team to provide a prescription before the starter kit was sent to the participant.

Participants were offered further supply via an e-mail/text approximately 4 weeks later. Those taking up the offer received four bottles of e-liquid, two bottles of mouth spray or six tubs of minis. Participants were able to switch to a different product for their second supply.

In England, participants who enquired about further supplies were asked to buy the products themselves. To enable extended e-cigarette use in Australia, a note was sent with the second supply enabling participants to purchase further e-liquid and coils at a website managed by the study team via appropriate prescriptions.

The participant received a telephone call approximately 1 week after randomisation to check product receipt and to discuss product use as a strategy for coping with present or anticipated temptations to smoke.

Participants also received static text messages as per the usual-care study arm (see Usual-care arm).

S3P arm

Participants were asked to complete a web-based assessment [QuitCoach, www.quitcoach.org.au (accessed September 2020)], which generates a three- or four-page letter of personalised advice. The advice includes a list of priority activities, with the prioritisation based on assessment responses. The letter could be viewed on screen in HTML or PDF format and printed if desired. It could be retrieved from the study website for later reference until such time as a new assessment was commenced. New assessments were not mandatory. Participants received reminder e-mails, but could complete the assessments at any desired time by returning to the study website. The resultant advice would then reflect their current situation and relevant progress since their previous assessment. In addition, participants could use a structured tool (Problem Planner) for generating ‘if–then’ statements (i.e. implementation intentions17). Such self-generated statements link problematic situations (e.g. after dinner) with desired behavioural responses (e.g. staying inside and playing with the kids) and are designed to ensure that the appropriate response is triggered whenever the problematic situation is encountered. The Problem Planner could also be modified as frequently as desired.

The tailored advice contained separate strategies for those using and those not using a nicotine product, and provided general advice about countering more stable residual beliefs about the value of smoking. In addition, the advice provided suggestions for monitoring the ongoing benefits of having quit smoking and taking appropriate rewards for reaching milestones.

The web-based intervention was augmented by a series of interactive text messages [QuitTxt, www.quitcoach.org.au/QuitTextInformation.aspx (accessed September 2020)] that were provided for up to 6 months post quit date. These text messages further encouraged use of ‘if–then’ statements, provided motivational messages and gave some more generic advice. Unlike the non-S3P text messages, these text messages were tailored to baseline survey responses and to responses from the most recent QuitCoach assessment, particularly to measures of difficulty staying abstinent (e.g. evidence of a recent slip-up or frequent ongoing temptations to smoke). This primarily determined whether participants received the standard stream of 55 messages or an augmented stream of 72 messages. In addition, there was the potential to send two special streams of messages:

1. The ‘OFFMEDS’ stream was a timed or user-triggered module designed to be sent either when people reported going off their base medication (i.e. bupropion, varenicline or nicotine patches) or at around the time they indicated at baseline they would do so. Going off base medication was reported via either a QuitCoach assessment or an incoming short message service command (‘OFFMEDS’). The OFFMEDS stream consisted of 17 messages and ran concurrently with the underlying standard or augmented stream.

2. A ‘RELAPSE’ stream was sent immediately on notification of a relapse, following either an incoming ‘RELAPSE’ command or a QuitCoach assessment. The aim of this stream (12 messages) was to encourage immediate recommitment to the quit attempt. Once completed, the underlying messages resumed with the augmented stream.

The text messaging program was able to respond to a variety of requests for additional help from the user. The commands were listed on a one-page PDF instruction form that was e-mailed to the user immediately following randomisation. In addition to ‘OFFMEDS’ and ‘RELAPSE’, these took the form of ‘emergency help’ messages in response to commands such as ‘SOCIAL’ or ‘STRESS’. These were designed to provide help in different forms of tempting situations via an immediate message suggesting a strategy and a follow-up message 30-minutes later to encourage reflection. Any participant-requested additional messages were on top of the above frequencies.

Two interventions arm

Participants received both interventions described above (i.e. the S3P and a nicotine product). The S3P advice and text messages were modified to include references to nicotine products as a relapse prevention strategy.

Follow-up

Follow-up was originally planned for 3, 6 and 12 months post quit date to record smoking status and other measures (see Measures); however, because of the trial curtailment the 12-month follow-up was not completed. Participants had the option to complete the follow-ups over the telephone or online. Participants in England received £10 for completing each of these questionnaires and participants in Australia received AU$20. The final follow-ups took place between March 2018 and July 2019.

Measures

At baseline we collected the following:

• demographic details, smoking characteristics (e.g. the Heaviness of Smoking Index from when they were smoking),18 previous quit attempts, previous use of stop smoking medications and e-cigarettes, and medical history (e.g. screening for depression, measures of perceived stress and affect)

• information regarding the current quit attempt (e.g. type of support/medicines used), frequency and strength of cravings, extent of slip-ups (if any), plans on how long to continue use of base medication, self-efficacy for maintenance, perceived challenges and number of smokers in social network

• quality-of-life data [measured by the EuroQol-5 Dimensions (EQ-5D)]19

• health Service Use Questionnaire (HSUQ) data.

At all follow-up contacts we collected the following:

• self-reported smoking status and cigarette consumption

• lapse/relapse details for those relapsing (including number of lapses, when, where, reasons for relapse, how many cigarettes at first lapse, how soon after first cigarette was full relapse)

• strategies used by the participant to prevent relapse

• details regarding cravings to smoke

• details regarding use of any, including non-allocated, smoking cessation/relapse prevention treatments

• details regarding use of allocated interventions and ratings of the allocated interventions, including helpfulness (3-month follow-up only)

• details of any adverse events (AEs).

We also collected data regarding the use of the S3P, including the number of assessments completed, the number of text messages received, which message streams were received, and if participants requested ‘emergency help’ messages or stopped the text messages.

At 12-month follow-up we had planned to readminister the EQ-5D and HSUQ and collect saliva samples for cotinine analysis; however, because of the issues described above, follow-up was terminated at 6 months and so this could not be done. The baseline EQ-5D and HSUQ data were therefore not analysed as there were no accompanying follow-up data.

The curtailed schedule of measurements is shown in Table 1.

Adverse events and serious adverse events

Participants were asked in the follow-up surveys about any AEs and serious adverse events (SAEs) experienced. In those allocated to the nicotine product arms [i.e. nicotine product study (NIC) arm and NIC + S3P arm], the following AEs were deemed to be related and expected in the study protocol: nausea, throat/mouth irritation and sleep disturbance.

Substudies

Data management

Original primary outcome

The original primary outcome was relapse rate at 12 months post quit date. Relapse was defined as self-report of smoking on at least 7 consecutive days reported at any follow-up time point, or any cigarettes smoked (even just a puff) in the last month, biochemically validated.

Original secondary outcomes

Original secondary outcomes were sustained abstinence using different criteria to the primary outcome and different assumptions about missing cases; point prevalence and shorter-term period prevalence outcomes; sustained reduction in cigarette consumption; evaluations of likely mechanisms of effect, focusing on strategies that were encouraged and participant perceptions of effect (e.g. participant ratings and data from EMA/qualitative substudies); dose–response effects (testing whether or not the dose of the interventions, or extent of compliance, is associated with relapse); cost-effectiveness of the different strategies; the effects of intervention components (e.g. on relapse rates, participant ratings, etc.) by country and on people from different socioeconomic groups, ethnic groups, gender and prior smoking habits (including those who stopped smoking using different forms of medication); and rates of AEs/SAEs reported in people who use a study nicotine product compared with those who do not, and by type of product used.

Owing to the issues described in Changes to trial design/protocol (trial curtailment), the study outcomes were curtailed. All revised outcomes were prespecified and agreed by the independent DMEC prior to data download and analysis.

Curtailed primary outcome

The curtailed primary outcome was relapse rate at 6 months post quit date. Relapse was defined as self-report of smoking on at least 7 consecutive days reported at either follow-up time point, or any cigarettes smoked (even just a puff) in the last month at 6 months. Participants lost to follow-up were included as smokers.

Sensitivity analyses for primary outcome

To assess the robustness of the results, a series of sensitivity analyses were conducted, for example multiple imputation by chained equation27 and complete-case analysis (where we excluded cases with missing outcomes). To build the multiple imputation model, we explored differences in baseline measures between participants with complete and missing outcome measures.

Per-protocol analysis was conducted to exclude those participants who never accessed/used their allocated intervention. Initiation was coded as having tried the allocated product in the NIC and NIC + S3P arms, having read at least one text for the usual-care arm, and having completed at least one assessment for the S3P and NIC + S3P arms. Missing data were counted as not initiated treatment.

Curtailed secondary outcomes

The curtailed secondary outcomes were as follows:

1. abstinence from smoking using alternative definitions: sustained abstinence (Russell Standard, defined as self-report of not smoking more than five cigarettes since 2 weeks post quit date) and point prevalence abstinence at 3 and 6 months [defined as self-report of no smoking (not even a puff) in the past 7 days]

2. adherence to and ratings of the interventions by participants, and coping strategies used

3. characteristics of and reactions to lapses across the four groups explored in the EMA/qualitative substudies

4. rates of AEs/SAEs reported in people who use a study nicotine product compared with those who do not.

Statistical software

All analyses were carried out using Stata software.

Public and patient involvement

Members of the UK Centre for Tobacco & Alcohol Studies Public Engagement Panel and a member of the New Nicotine Alliance charity (London, UK) provided feedback on the study design and participant documents, and helped to inform the decision about which e-cigarette products to use. The Trial Steering Committee (TSC) included members of the public and contributed to decisions made regarding trial progress and curtailment.

Trial committees

In England, a TSC and DMEC were convened every 6–12 months. A Trial Management Group also met at regular intervals throughout the study. Appendix 1, Table 27, lists the trial committee members.

Quality control and quality assurance

In England, a risk assessment was carried out in conjunction with the study sponsor and Barts Clinical Trials Unit (CTU), now King’s CTU, which was used as a basis for the study monitoring plan. During the recruitment phase, a monitor from the co-ordinating site carried out 6-monthly monitoring at the English site. The Barts CTU was responsible for oversight of the monitoring process and overall audit of the trial.

Approvals

The study was sponsored by the Queen Mary University of London Joint Management Research Office, London, UK, and the Cancer Council Victoria Melbourne, VIC, Australia.

Ethics approval was obtained from the National Research Ethics Service Committee London – Camden and Islington on 15 November 2016 (reference 16/LO/1771) in England, and the Cancer Council Victoria Human Research Ethics Committee (HREC) (project number HREC 1606) on 15 August 2016 and the St Vincent’s Hospital Melbourne HREC (reference number HREC 092/18) on 6 August 2018 in Australia.

In Australia, the study was also notified to the Therapeutic Goods Administration (Clinical Trial Notification number CT-2016-CTN-02901-1).

Participant flow

Figure 1 shows the flow of participants through the trial.

FIGURE 1.

Main study flow diagram: relapse prevention trial. a, Participant did not want their data to be used in the study.

Sample characteristics

Table 2 shows participant characteristics. The sample comprised largely middle-aged smokers classified as ‘medium’ on the Heaviness of Smoking Index. Less than half were in full employment and 28% reported a history of mental illness.

At the time of recruitment, 49.1% of all participants reported using a base medication (69.9% and 32.8% of participants in England and Australia, respectively).

Overall, the follow-up rate was 88.5% at 6 months. This rate was similar in Australia (87.8%) and England (89.3%). Follow-up rates were also similar across arms (usual care, 90%; NIC, 90%; S3P, 86%; NIC + S3P, 88%).

Table 3 shows the missing data across primary and secondary outcomes, and Table 4 provides data on the differences in baseline characteristics between participants who provided primary outcome data and those who did not.

Relapse rates

Relapse to smoking was somewhat higher in usual care than in the individual or combined interventions, but the combined intervention did not show any higher efficacy compared with single interventions. The overall difference was not significant (p = 0.11) (Table 5).

The results of the sensitivity analyses of the primary outcome tallied with the main analyses (Table 6).

To provide more descriptive information, we looked at the relative risks for reducing relapse in each of the arms using the usual-care arm as the reference. These were 0.75 (95% CI 0.53 to 1.06), 0.70 (95% CI 0.48 to 1.00) and 0.83 (95% CI 0.60 to 1.14) for the NIC, S3P and NIC + S3P arms, respectively.

Sustained and 7-day point prevalence abstinence rates

Table 7 shows sustained abstinence rates, as defined by the Russell Standard (i.e. allows up to five lapses), and point prevalence abstinence rates (i.e. no smoking over the past 7 days). There were no significant differences between the study arms.

Treatment adherence and use of allocated treatments

Nicotine product selection by arm and by country is shown in Tables 8 and 9, respectively. More participants chose e-cigarette than NRT (63.4% vs. 36.6%) [chi-square(1) = 8.04; p = 0.005].

In Australia, 71.0% of participants chose e-cigarettes and 29.0% chose NRT [χ²(1) = 10.9; p = 0.001]. In England, there was no significant difference in the number of participants choosing e-cigarette and the number of participants choosing NRT (54.0% vs. 46.0%) [χ²(1) = 0.3; p = 0.57)].

Of those who chose NRT, nicotine minis and mouth spray were selected with similar frequency within NIC arms and countries.

Table 10 shows the use of and adherence to the study nicotine products. At 3 months, > 50% of participants in the NIC and NIC + S3P study arms reported using their product at least occasionally. At 6 months, this applied to 40% of participants in these arms. Among participants who chose e-cigarette, 26.8% were using them daily at 6 months, whereas 17.1% of participants who used NRTs were using it daily at 6 months.

Most participants (n = 66, 58.9%) requested and were sent a second supply of nicotine product.

With regard to continued use of any oral nicotine products in the non-NIC arms, at 6 months the proportions were 16 participants (26.7%) and 13 participants (22.8%) in the usual-care and S3P arms, respectively. In the usual-care arm, 15 (25.0%) participants were using oral NRT and 2 (3.3%) participants were using e-cigarettes (one participant was using both oral NRT and e-cigarette). In the S3P arm, 12 (21.1%) participants were using oral NRT and 1 (1.8%) participant was using e-cigarettes.

Table 11 shows the use of and adherence to the S3P. Most participants allocated to the S3P intervention completed one QuitCoach assessment only.

At 3 months, most responders in the S3P arms reported having read the advice generated by the intervention, with the majority responding ‘yes, but only quickly’. Only one-third of participants in the S3P arms reported having used the Problem Planner (see Table 11).

Table 12 shows the use of the text messages. Most responders reported reading all/most of the text messages sent. Roughly the same small proportion of participants (12% maximum) requested to stop the text messages across the study arms, although this was least in the NIC arm (5.2%).

A feature of the S3P arms was the interactive element of the text messages. Only one participant (in the S3P arm) used the interactive text message commands, sending the following: ‘STRESS’ (four times), ‘SOCIAL’ (twice), ‘MISSING SMOKING’ (twice) and ‘TEMPTATION’ (twice).

Ratings of the S3P intervention

Participant ratings of the S3P intervention at 3 and 6 months are shown in Tables 13 and 14.

At 3 months, the most common rating for the S3P was ‘somewhat useful’, with a median rating of 2 [interquartile range (IQR) 2–3] in both the S3P arm and the NIC + S3P arm. At 6 months, ratings seemed to improve in the S3P arm (median rating 1, IQR 1–3), while remaining the same in the NIC + S3P arm (median rating 2, IQR 2–3). Most participants considered the intervention useful, but only one-quarter of responders said that they would visit QuitCoach again.

Table 15 shows the ratings of the text messages. Despite different contents and intensity of the text messages, and the tailoring and interactive features of messages in the S3P condition, ratings were similar across the four study arms.

Use of S3P recommended strategies

Behavioural strategies specifically recommended in the S3P intervention were used across all four study arms (Tables 16–18). The S3P arms did not consistently report more use of the recommended strategies than the non-S3P arms.

The S3P intervention included a suggestion for participants to make a list of reasons for quitting and advised participants to consult their list when tempted to smoke. Somewhat surprisingly, participants in the usual-care and NIC arms reported doing this to a similar extent as those who received the S3P intervention (see Table 17).

The S3P intervention also encouraged participants to give themselves rewards for achieving milestones at 3 and 6 months. This was also carried out as frequently in non-S3P study arms as in S3P study arms (see Table 18).

Adverse events

At 3 months, the proportion of participants reporting AEs was higher in the usual-care arm than in the intervention arms (25% in the usual-care arm vs. 12.1%, 15.8% and 10.2% for the NIC, S3P and NIC + S3P arms, respectively) (Table 19).

At 6 months, the numbers of participants reporting AEs were similar across all study arms (Table 20).

Severe AEs were uncommon (reported by 6.0% of participants at 3 months and 3.9% of participants at 6 months) (see Tables 19 and 20).

There were eight SAEs reported across six participants [arm surgery and an incarcerated umbilical hernia in the usual-care arm; anorexia and a car accident in the S3P arm; and gallstones, a mastectomy with breast reconstruction (which later resulting in emergency breast reconstruction removal) and tarsometatarsal fusion in the NIC + S3P arm]. None of the events was deemed related to study procedures.

Participation in the qualitative and ecological momentary assessment substudies

The qualitative and EMA substudies were originally planned to involve only a small subsample of a large trial, with identical sampling across study arms. In the curtailed study, a much larger proportion of the sample took part in qualitative interviews (n = 94, 40%) and about one-third of participants (n = 79) took part in the EMA study. Of these participants, 48 participated in both substudies.

Relapse rates were lower among those participating in the substudies (qualitative substudy, 1.7%; EMA substudy, 10.4%; both substudies, 16.5%) than among those who did not participate (51.3%).

The results of these substudies are presented in Chapters 5 and 6.

Participant characteristics

Table 21 gives the characteristics of the qualitative sample across the trial arms. Approximately 60% were from Australia, just over half (55%) were female and 30% were aged between 51 and 60 years. Just over one-third of the sample had completed only secondary education, 54% were not working and 60% were in receipt of benefits.

In the below sections, when giving illustrative quotes, we list whether participants were abstinent, lapsed or relapsed (according to the definitions given earlier), their trial arm, whether they were interviewed at the 3- or 6-month follow-up, their gender, their age (in years) and their country of origin.

Summary of key findings

Analysis identified the following themes (see italicised phrases) as most persistent across participants’ accounts:

• Adversity in participants’ lives, which was interweaved with their smoking and relapse experiences.

• Many participants reported using multiple study and non-study relapse prevention strategies to cope with urges to smoke, there was differential engagement with study interventions and additional benefit was derived for some participants from the study surveys and the EMA substudy.

• The acceptability, use and impact of study interventions varied across study arms. For example, in relation to the NIC arms, the products offered had different perceived strengths and weaknesses and there were concerns around e-cigarette safety, particularly in Australia. In addition, there was a desire to taper nicotine content over time. The S3P advice and strategies were useful when accessed, but difficulties with access were commonly reported and a preference was expressed for an easier method of use, specifically a mobile phone application (app) rather than a web-based program. In addition, text messages were a helpful reinforcement and could be improved if their content and timing were tailored to individual circumstances, but there were some reports that text messages triggered urges to smoke.

• The battle to overcome craving played a key role in relapse.

• Participants had differential responses to lapses and relapse.

It should be noted that the trial was complex and involved a number of different processes, including baseline, 3- and 6-month surveys; differential text messages according to study arm and user interactions; choice of product in the NIC arm and the option to switch product if desired; postal delivery of product/leaflet; 1-week follow-up call (all arms); an option for further supplies, if desired; prompts to use the S3P, which varied according to participants’ assessment responses online; and the EMA substudy. In addition, some participants continued to use their base medications [i.e. those used in their quit attempt (e.g. varenicline and nicotine patches)] during the study period. Furthermore, especially when the inclusion criteria were extended in Australia to include participants who had quit for 1 week, there was potential for the participants to confuse elements of any support they had received for their quit attempt with the relapse prevention interventions. Some participants struggled to distinguish the research elements (such as the surveys and EMA substudy) from the study interventions, and at times to disentangle their acute cessation treatment from the relapse prevention interventions. These reported instances should be borne in mind when reading further.

Adversity in participants’ lives

A very strong theme running through the majority of the interviews, which provided an important context, was the difficult circumstances of participants’ lives. Many participants discussed lapses or relapses in the context of people close to them dying of diseases (often smoking related) and, on occasion, suicides. In addition, participants mentioned mental health issues (e.g. depression and bipolar disorder) and their use of smoking to combat loneliness, stigma and blame. Occasionally, participants referred to relationship problems, such as divorce. Frequently, participants referred to their own physical health with descriptions of smoking-related diseases, such as heart and vascular disease and lung disease, and on a few occasions struggling with abstaining from other substances, such as alcohol or marijuana. Many had been smoking for a long time and sometimes at very high daily cigarette consumption:

I was smoking 80 ciggies a day, look where I am a year later I’m on five or six.

Relapsed/S3P/6-month follow-up/female/51 years/England

In addition, participants reported having been trying to quit for a long time (e.g. ‘it’s 25 years I’ve been trying to quit’) and talked about how they had lost periods of their lives because of smoking:

I’ve lost 15 years to smoking.

Abstinent/NIC + S3P/3-month follow-up/female/31 years/Australia

Multiple study and non-study relapse prevention strategies

The majority of participants reported using different motivational, behavioural and pharmacological strategies for different situations and at different time points in the study period:

The vaporiser was useful for that earlier part and the Nicabates now I find are pretty useful.

Abstinent/NIC + S3P/3-month follow-up/female/59 years/Australia

Some tools were study interventions, others were non-study methods and many participants used a combination of strategies, including both study and non-study methods, simultaneously. It was commonly recognised that different strategies were likely to work for different people:

What works for one mightn’t work for another.

Abstinent/NIC/6-month follow-up/male/62 years/Australia

Not everyone fits the same mould obviously I know that but erm there’s not a one thing fits all you know, their approach to their recovery is different and different things will work obviously, I can’t say one thing works even for me type-thing, you know I can’t say anything constructive on that really, I’m still learning myself.

Relapsed/usual care/6-month follow-up/female/33 years/Australia

In terms of motivational strategies for relapse prevention, participants drew attention to the importance of having a clear understanding of their reasons for quitting, planning in advance and setting goals, self-positioning as a non-smoker or ex-smoker and having a strong sense of willpower, commitment and determination. Willpower was frequently perceived as the most important element, regardless of support strategies:

And I don’t think anybody’s in a position to, the support is great but I don’t think you can, I don’t think one can stop for anybody, if that makes sense, it’s a personal thing.

Abstinent/NIC/3-month follow-up/male/57 years/England

In relation to behavioural strategies, the importance of rewarding oneself for abstinence and changing routines (e.g. tidying up, going shopping or not drinking coffee in the morning, or at all) were highlighted. Distraction as a strategy was commonly reported in the form of walks, going to the gym, keeping busy, drinking water, using non-nicotine-containing gum and mints, breathing techniques and meditation:

Like every morning that was my favourite cigarette, that was the one I enjoyed, the rest of them during the day were just habit, nicotine addiction, but the only one that I really really enjoyed was my first one with my cup of coffee so now I don’t have my cup of coffee first thing when I wake up I wait.

Abstinent/NIC/3-month follow-up/male/54 years/England

Some participants mentioned noting down urges and feelings (e.g. stress) in a booklet, or recording days quit and money saved on a calendar. Some used the strategy of confronting risky situations for relapse (e.g. socialising with alcohol), whereas others felt that it was important to avoid such situations, particularly at the beginning of quitting. Indeed, reducing alcohol consumption, or stopping altogether, was mentioned by some as crucial to them staying quit:

Originally I didn’t go out to places where smoking or alcohol was going to be.

Relapsed/NIC + S3P/6-month follow-up/male/62 years/Australia

Smoke-free environments and a sense of social stigma towards smoking were mentioned as important elements in staying quit.

Several participants were still using non-study medications, such as nicotine patches or varenicline, when they joined, which they carried on using throughout the study period, sometimes in addition to the study interventions:

I just went on the Champix and it blocks your cravings and I found that a lot better, [ . . . ] and I have your e-cig [e-cigarette] and the e-ciggie [e-cigarette] is always there in case you do fall, and I think that’s a good idea.

Abstinent/NIC + S3P/3-month follow-up/female/47 years/England

I had the lozenges as part of my support from the doctors, NHS quit smoking service, she gave me loads of those so I’ve had them left over.

Lapsed/NIC + S3P/3-month follow-up/male/29 years/England

A sense of feeling supported was highlighted as important in staying quit:

That positive feeling you know that they’re with you and they want you to succeed.

Abstinent/NIC + S3P/6-month follow-up/female/61 years/Australia

For example, the role of social support from partners, family and friends (or for some a lack of support) was highlighted by many as an important motivation to maintaining abstinence. For some, this was having people close to them who were quitting at the same time. For others, having people to offer support when needed provided enough motivation:

But addiction is pretty strong, so you need to have a wing man that can be very strict with you if they need to, someone that you really trust.

Lapsed/S3P/6-month follow-up/female/35 years/Australia

Others discussed more structured support from their general practitioner, quitlines and mobile phone apps as being helpful in addition to or instead of the study interventions:

They gave me what they called an emotional backpack.

Abstinent/usual care/3-month follow-up/female/39 years/Australia

I’ve been using an app since I stopped smoking it’s not the NHS one because I didn’t like the NHS one it’s another one called QuitNow which I found really motivating and I still do when I look at it I can see how many days it’s been since I stopped smoking and how many cigarettes I’ve avoided and how much lunch money I’ve saved erm yeah so that’s sort of enough of a motivation rather than having those text messages.

Abstinent/NIC/3-month follow-up/female/46 years/England

For several participants, it was difficult to distinguish between all the methods that they had used and state, definitively, the single strategy that worked for them:

It’s not just one of them, it’s all of them.

Abstinent/S3P/3-month follow-up/female/37 years/England

The strategies used were part of an ‘overall package’ or a ‘concerted effort’ to make use of all the tools available to them, without which many of those who were abstinent imagined they might have relapsed:

The texts, the QuitCoach and the nicotine substitute or in my case I had the e-cigarette down to no nicotine erm but it was just the comforting fact that I’ve spent 30 years with a cigarette in my hand and now I have something else in my hand so that is the whole package . . . I think if someone tried to give up or tried an e-cigarette or that Nicabate or whatever it is erm without other support I don’t think that it’d work.

Abstinent/NIC + S3P/6-month follow-up/male/62 years/Australia

Acceptability, use and impact of study interventions

In this section we discuss feasibility, acceptability, barriers to and facilitators of the interventions in the different arms of the trial. For the combined arm (i.e. NIC + S3P) we discuss any synergistic or other effects of having both interventions [note that comments on the individual interventions are discussed within the specific intervention sections (i.e. NIC or S3P)]. As text messages were sent in all arms, these are discussed separately, including the enhanced messages alongside the S3P.

Battling to overcome craving

Craving was a key concern for most interviewees. When asked about when they started to become aware of urges, most frequently it was in the context of quitting. Indeed, one participant stated that they only really became aware of these when they stopped smoking, as usually urges were indulged. Many participants described how it felt to have urges or cravings, with some having to be constantly on their toes:

Early on I would say I was getting quite a few, but yeah it was always quite reassuring, but it is a daily battle I guess, not so much now but earlier on it was an ongoing thing that you’d have to be on your toes and be aware of possible things that could be a trigger or erm yeah, without even knowing what’s a trigger for you.

Lapsed/NIC + S3P/3-month follow-up/female/30 years/Australia

Some participants would often describe a conversation in the brain as ‘a battle’ that played out in different ways. However, one participant said that the voices changed to encouragement when trying not to relapse:

This constant sort of heckling in the brain about needing to satisfy particular sort of, yeah a particular craving for what you know is a cigarette, it’s not food it’s actually a cigarette.

Relapsed/usual care/3-month follow-up/male/52 years/Australia

Oh look I think a person you know has to be more strong-willed in their efforts to quit if you know what I mean, and not fall for the voice in the head [ . . . ] it seems that all of a sudden it has a grip on you. Craving gets really really strong sometimes it seems overpowering, you know, and you just buckle under pressure sometimes I think.

Lapsed/NIC/6-month follow-up/male/28 years/Australia

When asked how long urges to smoke lasted, answers varied from ‘a couple of minutes’ to ‘a good 10 minutes’. Participants described how sometimes they gave in to the urges, as they were completely overwhelming:

. . . when I was dying for a, like actually dying for a smoke.

Relapsed/usual care/6-month follow-up/male/54 years/Australia

To the extent that they would:

. . . go down the street and bum one off someone.

Relapsed/usual care/6-month follow-up/male/54 years/Australia

Sometimes participants differentiated among various types of urges:

. . . there’s so many different types of cravings and yeah urges to smoke.

Relapsed/usual care/3-month follow-up/female/47 years/Australia

For example, one participant differentiated physical urges from emotional urges when they were going through a difficult time:

It was just a pressure building and I just couldn’t cope.

Relapsed/NIC + S3P/6-month follow-up/female/69 years/Australia

It sort of like starts off as like a tightness in my chest type-thing, from there I’m like I need a smoke, my mind won’t stop thinking about it where it’s like erm obsessive thoughts about it almost and then if I can’t get it I get really agitated you know I get the tightness in the chest and things start happening.

Relapsed/usual care/6-month follow-up/female/33 years/Australia

The first wave I found it so easy, physical cravings were taken care of with patches and lozenges, I really really wanted to quit and still do, but yeah and it was really the second wave the cravings it was kind of sort of more emotional more than anything else.

Relapsed/NIC + S3P/6-month follow-up/female/69 years/Australia

So at that point I didn’t care about stopping any more, my sort of erm feeling of having almost like no resources was stronger than anything else at that time.

Lapsed/NIC/3-month follow-up/female/57 years/Australia

Some abstainers and lapsers reported that they were still getting urges, even after some time, but for others that battle lessened over time:

I’m always feeling tempted to smoke.

Lapsed/usual care/3-month follow-up/female/38 years/Australia

But yeah that went away over time and now it’s not even a thought that’s in my head like . . . It’s totally gone.

Lapsed/NIC + S3P/3-month follow-up/female/30 years/Australia

Differential responses to lapses and relapse

There were differences between participants in terms of how they interpreted lapses to smoking. In addition, regardless of how they interpreted lapses, frequently lapses and relapses were not defined as they are in academia, with a slip-up (i.e. smoking even a few puffs of a cigarette) being referred to as a ‘relapse’.

For some abstainers, a lapse would be regarded as a failure and likely to progress to a full-blown return to smoking. Some suggested that ‘you’re either a smoker or not’ and therefore a lapse was the equivalent of relapse and there was no distinction between the two. Lapsing was therefore to be avoided at all costs. In these instances, a lapse was grounds for having to start a new quit attempt:

I just keep telling myself a lot of the time that if I even have one cigarette I have to start again and that’s a very strong motivator for me.

Abstinent/usual care/3-month follow-up/female/39 years/Australia

I know that if you have one puff that can just get you back in [ . . . ] I would say the reason I’ve been more successful this time is because it happened to me before, that they sucked me back in.

Abstinent/NIC/3-month follow-up/female/54 years/Australia

The minute you have a ciggie it opens that up in your head and then once you’ve done that you’re just back to square one, so never having that first cigarette is the biggest thing and remember there’s no such thing as just one.

Abstinent/NIC/3-month follow-up/male/54 years/England

For these participants who believed a lapse was a personal failure, then the lapse could progress to relapse:

I had a cigarette and I stopped again [ . . . ] I had the temptation but I was like fighting with the temptation that I shouldn’t be smoking I shouldn’t be smoking, and I kept telling myself that, reminding myself, and then eventually I said one more won’t hurt me, and after a week I think I took another fag, and from there it was frequent.

Relapsed/usual care/6-month follow-up/male/50 years/England

Sometimes the physiological response to a lapse meant that more cigarettes would follow. A few alluded, to being ‘gripped’ by urges to smoke more after one puff, referring to it as a recurring ‘cycle’ that ‘sucked you in’ or was ‘easy to fall back in to’.

For others, there was a sense that learning the best strategies takes time and comes with experience (e.g. ‘slipping isn’t failing’). Among many, a lapse was viewed as a transitional learning experience and for these participants the progression to relapse was less likely. Many participants reported how they were more likely to experiment with alternative coping strategies in the future, which may lead to the learning of more effective coping responses in high-risk situations. Indeed, for these participants, relapse prevention appeared to be an iterative process, as they talked about learning from the situations and strategies used during past lapses to succeed. Abstainers, in particular, conceived of relapse as a dynamic process of learning strategies and techniques to cope with temptation and cravings over successive quit attempts, of expanding and honing their capabilities over time and making use of existing resources available to them to achieve this.

For these participants, a lapse was differentiated from relapse. More specifically, for these participants a lapse or a slip was not a failure, ‘it’s not falling off the wagon’, but rather an opportunity to experiment and learn about what is required for them and in their particular circumstances to stay quit:

Relapse for me, it doesn’t mean that I’ve failed, it means that I’ve tried and I’ll have to try harder [ . . . ] relapse prevention I think is about how to control my urges, it just makes me more confident if I manage to do it, so I also think it links to someone’s self-confidence, if you know what I mean, the more confident you are in thinking that you’ll achieve your goal they more successful you become.

Abstinent/NIC/3-month follow-up/female/38 years/England

Your success will probably be built on failures it doesn’t matter yeah [ . . . ] it’s just part of the process so actually it’s getting to know how your addiction works and getting to know how you overcome it [ . . . ] just go straight back to giving up.

Abstinent/NIC/3-month follow-up/male/59 years/England

For those who viewed a lapse as a learning experience, their successful coping was described as having been built on by learning from previous quit attempts. Lapses were positioned as part of the learning process to the extent that they help one know what to do to recover and return to abstinence:

Don’t beat yourself up because you might’ve slipped up once in the week so what, don’t beat yourself up about it, if you slip-up every day that’s a different thing, well that’s what I’m doing, I didn’t beat myself up about it, I didn’t go oh well I’ve blown it I’ve got back on it.

Lapsed/NIC/3-month follow-up/female/60 years/Australia

It’s just a setback, you haven’t failed, it’s just a setback and start again but you know what you’ve learnt in the last few months and just apply them again, you can do it, [ . . . ] it’s easier to go back to ciggies than it is to stay stopped, I’d just say you know just try and do it again and keep on going, use these cigarette things where they help you and you’ll do it, eventually you’ll do it.

Abstinent/NIC/3-month follow-up/male/54 years/England

Keep persisting and if you fall off the band wagon just get back on again and [laughs] yeah it’s like that saying if you fall off the horse just get back on and try again. I mean it took me you know like three attempts just to finally stop.

Abstinent/NIC + S3P/3-month follow-up/female/33 years/England

Some participants described how they had learnt which triggers lead to smoking, which situations to avoid, or which attitude or approach was required through previous lapsing experiences, which enabled them to have more ‘control’ over urges to smoke:

When I have done it in the past, it’s more about how I’ve approached, like if I go into it with oh it doesn’t really matter type-thing and I’m really blasé about it then I find I’m just like one more packet won’t hurt me, if I’m not strict with myself it turns back into full-blown relapse where I’m smoking all the time [ . . . ] I find that I’m a lot more strict with myself, yeah if I go into it very blasé I find that I lose control.

Relapsed/usual care/6-month follow-up/female/33 years/Australia

Many participants reported that lapses were particularly common in situations involving alcohol consumption. Alcohol and smoking were frequently described as going together ‘hand in hand’, and several participants reported difficulty in refraining from smoking while consuming alcohol:

It’s mostly just the nights out when there’s alcohol involved that’s when I relapse.

Lapsed/NIC + S3P/3-month follow-up/male/24 years/Australia

There was some discussion of ‘permissive’ lapses. 29 These were lapses that they decided to allow:

I’ve just thought that I’ll decide to have a smoke today.

Lapsed/S3P/6-month follow-up/female/35 years/Australia

Alternatively, ‘permissive’ lapses were lapses that participants felt were somehow under their control:

I felt that it was my own decision and I was kind of empowered.

Lapsed/S3P/6-month follow-up/female/35 years/Australia

Permissive lapses were described as ‘disappointing’, ‘upsetting’ and ‘annoying’, although they had been ‘justifiable’, as some participants reported how they ‘still felt quite in control of that decision’ to smoke, as they were not buying cigarettes or not carrying cigarettes around all the time, and therefore did not consider themselves to be a smoker:

A little bit disappointed in myself I guess, but I still felt quite in control of that decision, so I guess I was able to justify to myself, you know like it’s OK I’ll just have one, and I wouldn’t consider myself to be a smoker again, I haven’t bought a packet of cigarettes, I don’t carry them with me all the time you know. It doesn’t matter if I just have one, because that’s not going to ruin my whole plan of quitting so I was still quite focussed on the end goal of not becoming a full-on smoker again.

Lapsed/S3P/6-month follow-up/female/35 years/Australia

Following the lapse, participants would want to ‘move on’ and go back to quitting. However, serial lapses were common among these participants. Alcohol was, again, frequently involved in the occurrence of these permissive lapses. Previous quit attempts had equipped participants with knowledge about how to cope with temptations to smoke while socialising and drinking alcohol:

There were times when I hadn’t smoked for 2 years, and you just drop your guard, you say save me some of that cigarette, and the next minute you buy 20 then, so it’s the old adage isn’t it, if you’re socialising with alcohol, or even sometimes just socialising, it’s funny, you could be in a good mood or sometimes if something really goes wrong or you’ve had a bad day you know, I think if you’re not as strong as I am now, when I was weaker if something, if I got bad news, your first thought would be I need a cigarette, whereas now I know that if I buy a packet of cigarettes I’m not helping myself.

Lapsed/NIC/3-month follow-up/male/44 years/England

Despite determination to stay quit, it was apparent that some participants felt that a lack of support and negative affect (such as stress, loneliness or boredom) or unexpected life events (such as a relationship ending or loss of a loved one) contributed to lapses, which in some cases were followed by prolonged periods of returning to smoking. Having been smokers for a long period of time contributed to this, as smoking had been the response to such occasions over a period of time and unlearning this was difficult. However, there were also a few reports of lapse progressing to relapse while relaxed on holiday, for example, which for some had also been partly because of the low cost of cigarettes.

Acceptability of the interventions

Initiation of allocated interventions was high, suggesting that both interventions are acceptable to recent ex-smokers. The qualitative study suggests that both interventions were perceived as sensible and helpful. Both interventions also obtained generally positive ratings. These are encouraging findings, although, as noted below, these need to be taken into account with the findings that NRT/e-cigarettes were more likely to be used than QuitCoach, which most people used only once. Some relapse prevention interventions examined in the past seem to have generated little enthusiasm and there was a concern that smokers who have recently quit smoking may feel that they have achieved their objective and have little interest in trying to prevent any theoretical lapses. The positive ratings could have been in part because some of the study participants were still in the early stages of smoking cessation rather than abstinent for at least 4 weeks, as originally planned. This, however, concerns only a minority of participants in Australia, and treatment ratings were positive at both trial sites. Overall, we can conclude that most participants found their allocated intervention of some utility, but there was variation.

Nicotine products

Regarding nicotine products, e-cigarettes were selected more often than NRT in Australia (71% vs. 29%, respectively), but the difference was not significant in England (54% vs. 46%, respectively). This was unexpected in the English site, as e-cigarettes are more popular among UK smokers than NRT. 12,13 Given that nicotine-containing e-cigarettes are banned from general sale in Australia, we expected that they would be a popular choice there, as this study provided a way to access these products. It is possible that smokers in England who attend stop smoking services are less interested in vaping, or are more affected by media scares concerning e-cigarettes than smokers in Australia, where there is no vaping and, therefore, media scares are less prominent. Nevertheless, concerns around the safety of vaping were more likely to be voiced by participants in Australia in the qualitative research.

The intervention consisted of recommending opportunistic use when at risk of relapse and to use as much as needed to prevent a return to smoking. At the 3-month follow-up, we found that 53% of participants used their products at least occasionally (31% used daily). At the 6-month follow-up, we found that 40% of participants used their products at least occasionally (23% used daily). Of those still using their study product at least occasionally, 74% were using an e-cigarette and 26% were using NRT at 3 months, and the equivalent values at 6 months were 76% and 23% (one person used both). This shows a higher extended use of e-cigarettes than of NRT, but by a smaller margin than noted in a recent randomised trial15 (the trial, however, included nicotine patches as opposed to the fast-acting NRT that was used here). Faster-acting NRT products generate higher rates of extended use than nicotine patches. 30,31

Regarding product safety, there were no SAEs related to trial medications. The most common AE reported was weight gain, which is a common side effect of quitting smoking. 32 Indeed, it is notable that those in the groups provided with nicotine products were less likely to report AEs, suggesting that a proportion of AEs are due to abstinence from nicotine. However, as the differences were not significant this possibility should be treated with caution.

Behavioural interventions

Behavioural interventions included a series of practical recommendations for implementing behavioural and cognitive coping strategies imparted via the online QuitCoach and Problem Planner. The intervention was further accompanied by an intensive texting schedule that was tailored to participants’ responses to an online assessment and included an option for interactive messaging.

Although the comprehensive package received overall positive ratings, both for the online and for the text messaging components, its utilisation in terms of site visits, assessments and use of interactive texting was low. Most participants completed only a single QuitCoach assessment, typically immediately after joining the study (i.e. on the same website visit), and only one-third of participants reported using the Problem Planner. Only one participant used the interactive feature of the text messaging, indicating that participants either did not see a need for it or were not aware of it. Future work may need to focus on encouraging re-engagement with such tools. Most participants indicated that they read the advice that the programme generated for them ‘quickly’. The qualitative study offers some insights into the reasons for this. Clients needed usernames and passwords to visit the site, which represented a significant barrier to some of them. The extensive text and long questionnaires may have also limited engagement with the program. Some of these features were because of the trial components, but future iterations of the program may look at simplifying access to the information, as well as looking at ways to present the content differently. The revised version of QuitCoach that was being programed for the study (but ultimately was not delivered by our IT consultants) would have provided the capacity to break the assessment and advice into smaller chunks. Use of a version with capacity to break the advice up and provide summary versions, as well as the full advice, may better meet the varying needs of participants.

We hoped that the detailed follow-up questionnaires would identify coping strategies that the behavioural intervention imparted (i.e. strategies used by participants in the S3P and NIC + S3P study arms but not by participants in the other two arms) and highlight advice that was most popular, suggesting components that were particularly practicable and useful. Of the six listed strategies, five were used by the majority of participants at both the 3- and 6-month follow-ups. The most popular advice was reminding oneself of reasons for quitting, used by some 90% of participants. Some 80% used distraction. The most demanding feature that was the focus of the intervention (i.e. putting in place a pre-prepared plan for resisting temptations) was used by approximately one-third of participants. However, unexpectedly, the proportion of those who reported using each coping strategy when tempted to smoke was practically identical in the non-S3P groups (i.e. usual care and NIC only). This included the use of a pre-prepared plan. More relevant than whether or not planning occurred, however, may be the quality of planning, something which is more difficult to quantify. Recognising that most ex-smokers engage in some form of planning,33 the intention of the S3P intervention was primarily to increase the quality of existing planning strategies and to help ensure that they were remembered and acted on at times they were needed. At the intermediate follow-ups (i.e. at 3 and 6 months), we could not ask participants in all study arms about whether or not they had formed strategies in terms of ‘if–then’ plans (implementation intentions). This would have made little sense to those in the non-S3P arms, but more importantly risked exposing these participants to a key idea underlying the S3P intervention. The EMA substudy also found no differences in coping strategies used by different study arms when faced with lapse situations. Overall, these findings could suggest that the S3P intervention had limited effects on the actual behavioural responses to temptations to smoke. However, given the trend for this intervention to reduce relapse, it may have supported people in another way (e.g. promoting recovery from a lapse) that we have not investigated in this report.

Process of relapse

Approximately half of the participants maintained sustained abstinence at 6 months. Of those taking part in the EMA substudy, which monitored lapses in real time, two-thirds experienced a lapse during the 3 weeks of monitoring. This included lapses in participants who were classified as having lapsed or relapsed already (n = 20) and also 33 participants who were, until the EMA monitoring, classified as sustained abstainers. A good proportion of lapses happened when craving levels were relatively low, and many lapses made participants feel like abandoning their quit attempt. There was no clear signal of any intervention effects on these variables.

The EMA substudy found few significant differences between the usual-care group and the intervention group. A significant difference was observed in the experience of low-craving lapses. As noted earlier, it has previously been reported that users of high-dose nicotine patches are more likely to lapse under situations that involve low levels of craving. 26 Here, however, participants in the S3P group were significantly more likely than those in the usual-care group to report lapses during low craving intensity. This finding is difficult to explain, as participants in the S3P group were provided additional information about coping with craving and avoiding lapses, including the use of implementation intentions. However, it may be that those allocated to S3P were more likely to recover from a lapse, which is also something the content focuses on. This and other differences between groups need to be explored in future, using more adequately powered studies, before firm conclusions can be drawn.

Some lessons from the trial tribulations

The trial faced serious difficulties in integrating a proprietary online programme with the data collection requirements of a randomised controlled trial, which led to significant delays in recruitment. Future trials using such interventions need to consider the vulnerability of specific programs and apps to the demands of data management. Even more importantly, trials that involve international collaborations need to research relevant data protection and IT security requirements, as these can generate major hurdles.

Another issue that can serve as a warning to future projects was difficulties in recruitment. Recruiting recent quitters from quitlines proved much more challenging than expected, and we expected it to be hard. The main problems included (1) marked drop-offs in the use of quitline services before recruitment; (2) reluctance to enrol in a new intervention study after succeeding with support already provided; (3) unwillingness to participate in a trial that could result in being assigned to a pharmacotherapy intervention; and (4) greater difficulties experienced by quitline specialists in contacting callers on their last scheduled call. In England, over the past few years, the throughput of the stop smoking services has shrunk by > 60% from their inception, and the services have been relocated from NHS to local councils, which commission private providers to deliver stop smoking treatments. The reorganisation and the change of service focus meant that only a few services remained that could contribute to research. Future studies, at least those conducted in England, that involve smoking cessation may need to identify recruitment venues other than stop smoking services. Planning for recruitment in other UK countries may have also been a useful strategy. However, this planning needs to take place early, as gaining regulatory approvals takes time.

In Australia, we attempted to recruit other quitlines. However, we were unsuccessful, largely as a result of one arm of the study involving vaping products. Nevertheless, even if we had been able to recruit all quitlines, recruitment would have still been below what we needed. In Australia, we conclude that the required strategy of recruiting ex-smokers after successfully quitting for 1 month is not feasible, at least from services that had already provided a quite intensive programme of help.

In Australia, we tested two alternative methods of recruitment: (1) use of social media and (2) recruiting from hospitals where patients are not allowed to smoke as inpatients. Our conclusion is that social media are probably a useful recruitment source, but their use can be complicated when there are complex inclusion criteria for the trial, as there were in our case. Furthermore, social media are not likely to be a major approach for service delivery, although promoting effective programmes on social media is likely to draw some traffic and may be able to do this at a reasonable cost.

With regard to recruiting from hospital inpatients, overall, initial recruitment in Australia relied heavily on manual screening processes of patients’ smoking status and receiving referrals from other health professionals within the hospital. This process was later further automated, but is something to consider early on in the set up process.

Common to both countries was a greater than expected intention to use medications or vaping devices for extended periods (i.e. > 3 months), therefore rendering participants ineligible for this trial.

These issues highlight the need for pilot phases within large trials with clear stop/go criteria.

Study limitations

The key study limitation is the curtailed sample size that limited the power that the trial had to detect intervention effects. However, study outcomes can still contribute to future meta-analyses. Related to study curtailment, the follow-up period was shortened from 12 to 6 months. As the efficacy of smoking cessation interventions typically declines over time, the hint we detected that the single interventions may show some efficacy needs to be interpreted with caution. Another consequence of the trial curtailment is the lack of biochemical validation. This means that the abstinence rates are likely to have been overestimated overall and, in particular, in the intervention groups that received more input and may have felt a stronger obligation to report success. It is reassuring that the NIC + S3P group reported somewhat higher relapse than the NIC and S3P groups, but the numerical advantage of all intervention groups compared with usual care could still reflect different response expectations.

The trial curtailment also meant that the planned health economics analyses could not be carried out, as follow-up data collection for the EQ-5D and HSUQ was planned only at 12 months.

The trial was originally designed to test the interventions following formal stop smoking treatment from either the English stop smoking services or Australian quitlines. The addition of recruitment via Facebook, St Vincent’s Hospital Melbourne and Stoptober meant that 34% of participants had not received any formal stop smoking support. Furthermore, extending the eligibility window meant that some participants entered the trial after only 1 week quit in Australia, whereas others joined after 3 months quit. This variation reduces the focus on relapse happening from around 1 month, as originally planned. However, the notion that relapse prevention only begins at 4 weeks of abstinence is arbitrary. It can be argued that from the moment that a person stops smoking they require strategies that help prevent them from relapsing to smoking.

Another feature of the trial that warrants additional caution is in assessing relapse rates overall, in particular the interpretation of findings for the control group. The qualitative and EMA substudies were originally planned to involve only a small subsample of a large trial, with identical sampling across study arms. In the curtailed study, the qualitative aspects of the project are relatively more important. A much larger proportion of the sample took part in qualitative interviews, approximately 40%, and approximately one-third of participants took part in the EMA substudy. We were also unable to balance the proportions of lapsers, relapsers and abstainers within or between the study arms, and those participating in these substudies were less likely to have relapsed than those who did not. A potential issue that this created is that qualitative interviews suggested that EMA data collection represented a perceived intervention, and the interviews themselves could have had this impact too. The EMA substudy and qualitative interviews could have provided motivational input, perhaps more memorable than online instructions because of their interactive nature, and could have even have imparted or reinforced coping strategies by focusing on their details during surveys and EMA monitoring. Such input could, in theory, have contributed to the lack of differences in the use of coping strategies between study arms, as discussed above. It could have also boosted abstinence rates, especially in the control group, and therefore have diluted any intervention effects. However, given the self-selection for these substudies, it is not possible to make any definitive conclusions regarding these effects.

The overall 6-month follow-up rate was 88.5%, which is relatively high for trials of this type. Achieving higher follow-up rates is difficult, as smokers who are unsuccessful tend to feel embarrassed and avoid contact. It is reassuring that follow-up rates across the four study arms were similar, as were the follow-up rates in the two countries. Nevertheless, incomplete data represent a limitation.

Generalisability

The majority of participants (66%) received structured stop smoking support before joining the trial. Therefore, the results may not be generalisable to populations who quit unaided or without formal support. This is particularly true of participants in the nicotine product arms, as e-cigarettes are readily available in England and are widely used without any structured support as an effective tool to quit smoking. 34,35

Recommendations for research

The efficacy of fast-acting nicotine products provided for use in emergencies and accompanied by text message reminders for relapse prevention may deserve further exploration. Online training programmes have good intuitive validity and should also be further examined. An intermediate aim could be to develop delivery modes that ensure that participants recall the advice and act on it. Once this has been established, evaluating the intervention in a simplified package (e.g. as a mobile phone app) would be the natural next step. Another recommendation is that future trials should aim to initiate relapse prevention interventions around the time of quitting. Delaying the offer until later in a successful quit attempt risks a reduced likelihood of treatment uptake. That said, it should also be available for those who have quit and seek extra help as well, but this is likely to represent only a minority (largely self-quitters who find themselves in trouble). There may also be utility in testing these interventions in those who are hospitalised and unable to smoke. Our experience suggests that this will be best done if the relapse prevention interventions can be initiated while hospitalised, and so advice and/or medications can be provided before the person is discharged. Future studies should also consider using simpler designs, and plan for the challenges faced when integrating trial data management systems into existing programs or apps. Furthermore, international collaborations involving such systems will need to factor in data protection and IT security requirements.

Conclusion

Both interventions were well received and safe. The effects of the combined interventions did not surpass those of each intervention alone. There was a trend in favour of single interventions reducing relapse, but it did not reach statistical significance and there are reasons to interpret the trend with caution.

The EMA substudy showed that two-thirds of participants experienced a lapse, and around one-third of these lapses happened when craving levels were relatively low. For many, lapses made participants feel like abandoning their quit attempt.

Adherence to nicotine products was high, with e-cigarettes preferred to medicinal nicotine products. Behavioural advice was appreciated, but did not seem to affect behaviour. Nonetheless, the findings of this curtailed trial suggest that these interventions may have promise, but more likely as alternative interventions, as there is no evidence they work additively or synergistically.

Dedication

In March 2020, while this report was being revised, James Balmford died suddenly after a short illness. He made a large contribution to the study and is greatly missed.

Contributions of authors

Hayden J McRobbie (https://orcid.org/0000-0002-7777-1845) (Professor in Public Health Interventions) led on the original grant application, co-designed the trial, co-wrote the statistical analysis plan, trained staff, interpreted the study findings and co-led on the drafting of the report.

Anna Phillips-Waller (https://orcid.org/0000-0001-6856-6439) (Research Manager) managed the English trial site, co-wrote the statistical analysis plan, trained staff, contributed to data collection, interpreted the study findings and co-led on the drafting of the report.

Catherine El Zerbi (https://orcid.org/0000-0003-4041-2379) (Postdoctoral Research Associate) conducted the qualitative interviews, co-wrote the statistical analysis plan, interpreted the study findings and assisted with the drafting of the report.

Ann McNeill (https://orcid.org/0000-0002-6223-4000) (Professor of Tobacco Addiction) led the qualitative substudy, co-wrote the statistical analysis plan, interpreted the study findings and assisted with the drafting of the report.

Peter Hajek (https://orcid.org/0000-0001-9160-4296) (Professor of Clinical Psychology) contributed to trial design, co-wrote the statistical analysis plan, interpreted the study findings and co-led on the drafting of the report.

Francesca Pesola (https://orcid.org/0000-0002-2054-7930) (Research Fellow) co-wrote the statistical analysis plan, analysed the study data and assisted with the drafting of the report.

James Balmford (https://orcid.org/0000-0003-1423-4361) (Research Fellow) assisted with the development, refinement and maintenance of the S3P intervention and electronic data management system, co-wrote the statistical analysis plan, interpreted the study findings and assisted with the drafting of the report.

Stuart G Ferguson (https://orcid.org/0000-0001-7378-3497) (Associate Professor) led the EMA substudy, co-wrote the statistical analysis plan, interpreted the study findings and assisted with the drafting of the report.

Lin Li (https://orcid.org/0000-0002-4764-1679) (Postdoctoral Research Fellow) managed the Australian trial site, co-wrote the statistical analysis plan, trained staff, contributed to data collection, interpreted the study findings and assisted with the drafting of the report.

Sarah Lewis (https://orcid.org/0000-0001-5308-6619) (Professor of Medical Statistics) co-wrote the statistical analysis plan, provided statistical oversight and assisted with the drafting of the report.

Ryan J Courtney (https://orcid.org/0000-0003-1339-3221) (Senior Research Fellow) co-wrote the statistical analysis plan, interpreted the study findings and assisted with the drafting of the report.

Coral Gartner (https://orcid.org/0000-0002-6651-8035) (Associate Professor) assisted with the study set-up in Australia, interpreted the study findings and assisted with the drafting of the report.

Linda Bauld (https://orcid.org/0000-0001-7411-4260) (Bruce and John Usher Professor of Public Health) assisted with the drafting of the report.

Ron Borland (https://orcid.org/0000-0003-0059-178X) (Professor of Psychology, Health Behaviour) led the Australian study site, co-wrote the original grant application, co-designed the trial, co-wrote the statistical analysis plan, trained staff, interpreted the study findings and assisted with the drafting of the report.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to anonymised data may be granted following review.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care.

Qualitative substudy semistructured interview questions

Questions will differ depending on smoking/abstinence status, intervention condition and country. Below is a comprehensive list of the questions that will be tailored in accordance with the above criteria. Probes will be used throughout to elicit maximum information. We will have access to baseline and follow-up data on smoking history, lapses, relapses, etc., for the purposes of reference only.

• Preamble on format and approximate length of time (45 minutes maximum).

• Emphasise confidentiality.

• Emphasise that there are no right or wrong answers.

• Warm up conversation and test sound on audio-recorder with a question: how is it going? (Based on the response to this question, we will reallocate to lapse/relapse/abstainer.)

• Clarify when the participant’s last cigarette was, including just a puff of a cigarette.

Below, the questions marked with * refer to the intervention questions for appropriate arm of the study.

Take notes and return to unanswered and unclear responses at the end of interview.

Questions for lapsers (adapted in line with any responses to the warm up)

Lapsers: a puff in the last 2 weeks but not smoked daily for at least 7 consecutive days, including participants who have been serial lapsers.

• Have you had any cigarettes since (see REDCap for recruitment date), including just a puff of a cigarette?

• Can you tell me about any cigarettes you have smoked in the last couple of weeks? First, when did you last smoke? How many and how often in the last 2 weeks? Did you have just a few puffs or the whole cigarette? (If relapse – switch, if abstinent – switch.) If multiple: was this one period of smoking or were they separated by cigarettes resisted? Now thinking about the first cigarette (in the last bout of smoking), what triggered that cigarette?

• Did you do anything to try to resist? If yes, what?

• Had you been in a similar situation before and successfully resisted? If yes, what was different this time?

• How many cigarettes did you smoke before you recovered and resumed abstinence?

• What did you do to resume your quit attempt? What thoughts did you have?

• Had you tried previously to resume abstinence? How did it go?

• Explore relationship to use of interventions below as appropriate.*

• Explore relationship between intervention and identified triggers/situations as appropriate.*

• How did you feel after smoking it? (Relate to intervention questions below as appropriate.)

• Have you smoked any further cigarettes since that lapse? Can you remember the interval between your first lapse and second?

• Have you been supported or influenced by family members, friends and other smokers to quit smoking?

• Have you felt tempted to smoke further cigarettes since then? What strategies have you used to stop yourself smoking? (Relate to intervention questions below as appropriate.*)

• How has this lapse experience differed from any previous quit attempts and lapses. If so, why was it different?

Questions for relapsers (adapted in line with any responses to the warm up)

Relapsers: smoking daily for at least 7 consecutive days, including participants who were lapsing and then relapsed subsequently, as well as participants who were abstinent and then relapsed.

• Have you had any cigarettes since (see REDCap for recruitment date), including just a puff of a cigarette?

• When was the last time you smoked? (Within 2 weeks?)

• When did you first lapse? Can you say a bit about what you think might have led to this situation? Explore the situation leading up to the relapse, was it a lapse initially or a full-blown relapse in one go? [Use questions above, as appropriate, if initially lapses (particularly in relation to triggers, interval between lapses, and between lapses and relapse, plus any differences with previous quit attempts and full-blown relapses).]

• Could you talk a little bit about resuming smoking and how you feel about this? Have you previously tried to remain abstinent from smoking and how do you feel about not being able to remain abstinent? (Explore timing of relapse in relation to use of interventions.*)

• Prior to the relapse had you been using any effective coping strategies, for example when you had an urge for a cigarette? Had you been using any coping strategies and, if yes, which strategy did you find to be most effective? Why did you think this strategy was particularly helpful? Were there any strategies you found to be ineffective and, if yes, could you say why you think this strategy failed? (Relate to intervention questions below as appropriate.*)

• What, if anything, might have prevented you from relapsing? (Relate to intervention questions below as appropriate.*)

• How confident were you feeling about staying stopped in the run-up to your relapse?

• Have you been supported or influenced by family members, friends and other smokers to quit smoking?

• Will you be trying to quit again soon? (Explore interest in using the interventions if so, and explore if this is different from previous relapses and why.)

• How did you feel before relapsing? And now?

• How do you think you would feel if you successfully quit smoking?

Questions for abstainers (adapted in line with any responses to the warm up)

Abstainers: previously lapsed, or indeed relapsed, but subsequently abstinent in the last week.

Abstinence at 12 months: not a single puff in the last month at 12-month follow-up.

Continuous abstinence: not smoking a puff at each follow-up at 12 months.

• Just to check, have you smoked any cigarettes at all since the start of our study?

• What do you think lies behind your success on this quit attempt, and why? What are the main strategies/methods you have been using to help you stay quit? When did you start using the strategy/method?

• How often, if at all, have you had to make special efforts to resist smoking? Have you been supported or influenced by family members, friends and other smokers to quit smoking?

• Which, if any, strategies did you adopt? Which did you find to be the most successful?

• Explore use of interventions as below.* (Probe individual strategies and relate to intervention questions below as appropriate.)

• How do you feel since quitting smoking?

*Specific questions on interventions specific to the 3-month follow-up

Use knowledge of follow-up data to explore answers.

Final questions

• Do you have any thoughts as to how we could improve our materials/product selection/provision, etc.? When, by whom, where, etc.? If favourable, do you think our intervention should be given to all those who have stopped smoking after using the stop smoking services or quitline?

• Do you think getting this help at around 1 month quit was a good time, or would you have liked it at around the time you stopped?

• Strategies around confrontation and avoidance of situations identified by participants as high risk? Ask what their views are.

• Confidence about staying quit in the future and perceived barriers to staying quit.

Definitions

• Slip-up.

• Serial lapses.

• Full-blown relapse.

• Permissive lapses.

• Anticipating lapses.

Context

• Settings.

• Stress.

• Loneliness.

• Boredom.

• Alcohol.

• Adverse situations.

Interpretations

• Failure because of own ability.

• Failure because of context.

• Learning experiences.

• Unlearning previous patterns.

• A test of resolve.

• Confidence.

• Shame and guilt.

Responses

• Physiological sensations.

• Recovering and resuming.

• Learned techniques for coping with cravings.

• Persistence.

• Relapse.