A digital behaviour change intervention to increase booking and attendance at Stop Smoking Services: the MyWay feasibility RCT

Article history

The research reported in this issue of the journal was funded by the PHR programme as project number 15/183/26. The contractual start date was in March 2018. The final report began editorial review in March 2020 and was accepted for publication in October 2020. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The PHR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report 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 2021. This work was produced by Fulton 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.

2021 Queen’s Printer and Controller of HMSO

Background

Although UK smoking rates have hit an all-time low, 6.1 million adults in the UK are current smokers. 1 The significant role of smoking as a cause of chronic illness and preventable death endures worldwide. 2 UK NHS Stop Smoking Services (SSS) offer free behavioural and pharmacological interventions proven to increase the likelihood of quitting compared with ‘going it alone’. 3 SSS are typically offered within general practice and pharmacy settings, although increasingly they are offered as part of a package of lifestyle services hosted in a variety of community settings (CSs), such as supermarkets, libraries and community centres. 4

The effectiveness of SSS is determined by smoking abstinence 4 weeks after the quit date set, confirmed by standard carbon monoxide (CO) testing. 5 Studies indicate that 4-week quit rates are a reliable predictor of long-term abstinence, with 6-month follow-up data revealing only a small increase in the accuracy of such predictions. 6

Service uptake by smokers is, however, in decline. 7 Potential explanations for this include reduced economic resources for promoting services, and that smokers now favour the use of electronic cigarettes (ECs) over behavioural support for cessation, although this remains unclear. 8 Nonetheless, a survey of stop smoking advisors found that the majority (95%) had encountered smokers who were also using ECs. 9 This suggests that smokers may recognise ECs as an adjunct rather than an alternative to behavioural support for smoking cessation, and the effectiveness of ECs as a stop-smoking aid is likely to be enhanced when coupled with SSS behavioural support. 10 The growing use of ECs need not impede SSS access if services promote a clear EC-friendly approach, with consistent training and policy. 11

Despite significant budget cuts, sustained investment in SSS is being called for by the All Party Parliamentary Group on Smoking and Health,12 to support the ambition of achieving a smoke-free population by 2030. 13 However, evidence suggests that SSS reach only 5–10% of current smokers. 14 Investment in promotion and interventions to increase uptake of SSS is warranted. The UK Government funded a series of mass media tobacco control campaigns between 1999 and 2010 to emphasise smoking harms and SSS support. These campaigns were suspended owing to significant costs associated with delivery in 2010, enabling analysis of their impact on SSS uptake. Although calls to stop-smoking telephone lines, smoke-free website visits and requests for cessation support packs decreased significantly, there was no effect on SSS attendance,15 suggesting that the campaigns were not effective in enhancing SSS uptake.

Digital interventions promoting SSS help-seeking by smokers offer a potentially cost-effective solution. Current data suggest that internet access and smartphone use are increasingly widespread, with an estimated 96% of 16- to 24-year-olds, 98% of 25- to 34-year-olds, 94% of 35- to 44-year-olds, 87% of 45- to 54-year-olds and 71% of 55- to 64-year-olds owning a smartphone. 16 However, questions remain regarding whether or not they increase health inequalities if users of digital technologies are principally from higher socioeconomic status (SES) groups. Interventions require careful testing to ensure that they do not promote a ‘digital divide’. This is pertinent given that smoking is most prevalent in more deprived communities. 17 However, a trial of ‘StopAdvisor’,18 a digital smoking cessation intervention, found that it was effective at achieving cessation for smokers from lower but not higher SES groups, suggesting that digital interventions can help to reduce rather than increase health inequalities.

The uptake of SSS from lower SES groups remains varied across geographical localities. 19 Several studies suggest that smokers from more deprived groups are less aware that SSS exist, or of the services that they offer. 20–24 Negative beliefs about SSS include uncertainties about whether or not they will help or if they will be stigmatising, guilt-inducing or judgemental. 20–23 Concerns from smokers about a lack of information and perceived availability of services have been highlighted, alongside work commitments and time constraints, which can impede attendance. 25 The authors conducted studies to explore these perceived barriers and facilitators in depth,26 and discovered similar findings, including beliefs that needing support to quit smoking is a sign of weakness27 and that smokers should not need help. Therefore, an understanding of the potential barriers to access, particularly in lower SES groups, is paramount, and should be addressed within the development and evaluation of interventions designed to increase SSS access.

These barriers to accessing SSS have not, to the authors’ knowledge, been addressed sufficiently within interventions or health promotion campaigns targeted at smokers. Research has shown that providing information about SSS may increase access; for example, when booklets explaining how and why services help smokers are given to those who have registered with SSS, this can increase attendance rates. 28 A more recent trial (Start2quit) provided letters to smokers with personalised risk information and the offer of SSS taster sessions, resulting in improved SSS uptake. 29

Given that public health budgets are under increasing pressure, and evidence suggests that smokers seeking help themselves are more successful at stopping than those referred by others,30 it will become increasingly important that SSS are promoted to smokers who otherwise may not have recognised the benefits of attending. This will also involve ensuring that there is informed decision-making and encouragement to attend, to reduce ‘did-not-attend’ (DNA) rates and the costs that these incur. Implementing inexpensive strategies to address these factors will maximise effectiveness in a resource-restricted public health context.

StopApp™ (Coventry University, Coventry, UK) is a brief, digital behaviour change intervention [web application (app)] developed with the aim of increasing booking and attendance at SSS, especially by those from more deprived groups. Based on evidence about the barriers to service access and designed using the Behaviour Change Wheel,31 it first aims to enhance motivation to attend SSS using a series of behaviour change techniques (BCTs)32 derived from a systematic assessment of the barriers and facilitators that smokers typically experience. 27 StopApp then supports instant appointment booking, providing choice regarding time and location via an application programming interface (API). This infrastructure is supported by pharmacy and general practice data management software (PharmOutcomes and Outcomes4Health, both owned by Pinnacle Healthcare Ltd, Teddington, UK) that is used extensively by UK pharmacists and general practices to manage SSS use data. For the purpose of this feasibility trial, we had access only to pharmacy-based SSS data software (PharmOutcomes) to test the trial methods. However, StopApp could be used to increase attendance at any SSS for which it is possible to embed an API link to the SSS data software system.

Although increasing numbers of digital interventions for smoking cessation support exist,33,34 to our knowledge, StopApp is the first digital intervention to target increased uptake of SSS. If effective, StopApp could provide an important contribution to maximise SSS uptake. Given the novel digital approach that StopApp represents, and the challenges of recruiting smokers into research studies,35 it is unclear from the current evidence base whether or not it is possible to recruit the required sample size for a full randomised controlled trial (RCT). In addition, given concerns around the potential for digital interventions to contribute to health inequalities, the extent to which a trial of StopApp would reach and engage smokers from a range of sociodemographic backgrounds, including those most at risk of poor health outcomes from smoking, needs to be established. Therefore, a feasibility RCT was deemed necessary in order to assess whether or not a future, definitive RCT is possible.

Aims and objectives

The aim was to establish the viability of a future, larger, multicentre RCT of the StopApp intervention. The specific objectives are outlined below.

Design

This was a two-arm, 1 : 1 allocation, double-blinded, parallel-group, individual participant randomised feasibility RCT of StopApp (intervention) compared with usual promotion of and provision of contact details for SSS via an online leaflet (control; see Report Supplementary Material 1). The study included a nested qualitative process evaluation with participants and staff involved in participant recruitment across the three settings to assess the acceptability of the research processes, randomisation, measures and the intervention.

Study population for feasibility trial

Eligible participants were current smokers (of any tobacco amount and type) aged ≥ 16 years who were either (1) registered with one of six participating GP surgeries in Warwickshire, (2) accessing participating Warwickshire-based local community services (e.g. children’s centres, libraries, well-being hubs) or (3) viewing open-access advertising for the study on Warwickshire-based social media (SM) platforms [e.g. Twitter feed (Twitter, Inc., San Francisco, CA, USA; www.twitter.com) of local public health department; Facebook page (Facebook, Inc., Menlo Park, CA, USA; www.facebook.com) of ‘What’s on in Warwickshire’]. Participants were not excluded if they had previously attended SSS. Eligible participants had to be able to comprehend written English, have access to the internet and have a mobile phone to complete study measures, view intervention or control content, and be able to receive short message service (SMS) reminders. As SSS appointments were available only in Warwickshire, participants had to live or work in Warwickshire.

Study population for process interviews

Process interview participants comprised (1) smokers who had taken part in the main study who were willing to participate in an interview, and (2) professional staff from the pharmacies offering SSS appointments and two of the recruitment settings, including GPs and practice managers and staff from local community services.

Study protocol and intervention

The study protocol was approved by the NHS West Midlands – Edgbaston Research Ethics Committee (reference 18/WM/0170) and was later published. 36

Intervention

StopApp is a brief web app intervention designed to be used on a single occasion to address the psychological and practical barriers smokers may experience when considering help from SSS [Figure 1; see Report Supplementary Material 2]. It is intended to improve the reach and effectiveness of standard SSS promotion that local authorities (LAs) and SSS may use to recruit smokers. The process involved in the development of the StopApp, in terms of its theory and evidence base, and patient and public involvement (PPI) in content and functionality design, has been described elsewhere. 26,27,37 Figure 2 illustrates the participant flow and message tailoring content that StopApp recipients experienced. Initially, StopApp users are given persuasive messages applying BCTs tailored to past experiences of quit attempts and SSS use. They are given clear information about what to expect from the services and the benefits of SSS use; for example, the intervention provides key messages about the services and the multiple strategies for quitting that are available, the benefits of attending and the non-judgemental, friendly support provided. Some message content is derived from genuine feedback from SSS users. Subsequently, users are asked to select barriers to SSS use that they may be experiencing, and further persuasive content employing BCTs is tailored to personal perceptions. At all times, the opportunity to ‘book now’ is provided and anyone who does not yet feel ready but remains interested can request a reminder to come back to StopApp in the subsequent month. Navigation and user experience were designed using behavioural science insights to ensure that the intervention was intuitive, attractive and easy to use. To operationalise the ‘book now’ facility, StopApp was connected to the online outcomes reporting systems owned by Pinnacle Healthcare Ltd and used by SSS in pharmacies (PharmOutcomes) and GP surgeries (Outcomes4Health) in Warwickshire and other UK LAs. These secure systems support appointment booking and record SSS outcome data such as quit dates and CO-verified 4-week quit rates for service commissioners. StopApp was programmed to ‘talk to’ these systems and access live data about which services had appointment availability. StopApp users could then choose a location and could see appointment availability over the following 2 weeks. Participants selected an appointment time and received confirmation (date and location) via their mobile phone. In future, through collaboration with other providers of similar data management systems, StopApp could be adapted to link to a range of systems used by other types of service provider such as ‘QuitManager’ (Bionical Solutions Ltd, Willington, UK).

FIGURE 1.

Image of StopApp as viewed on a mobile phone.

FIGURE 2.

User flow and message tailoring within StopApp.

Process evaluation interviews

Participants were asked at the end of the baseline survey whether or not they would be willing to participate in an interview about their trial experience. Willing participants provided contact details. To ensure that a cross-section of views was captured, a sample of approximately 20 participants was targeted, with participants from both arms and different demographic groups represented. The intention was to oversample participants from more deprived backgrounds, as measured by their IMD scores and quintiles. This was to ensure that any factors potentially associated with deprivation (which may have influenced engagement with the intervention or the research process) were fully explored.

Staff from CSs, GP practices and pharmacies who took part were also invited to participate in telephone interviews about their trial involvement. Selected participants and staff were invited to participate in a telephone interview via a link that was sent by e-mail. The link took them back to eNgage, where they could view information about participating in the interviews and complete online consent statements. Participants who gave consent were then contacted by telephone to arrange a suitable time for the interview to take place. In some instances, this took place immediately. The interviews explored participants’ experiences of taking part in the trial, barriers to and facilitators of the process, suggestions for improvement, beliefs about the value of the research methods and the overall impact that it had on them. Interviews were conducted by KK and LS, were audio-recorded and transcribed verbatim, and qualitative data were subject to thematic analysis45 by EAF with input from KK and LS. A pragmatic, hybrid approach was used to conduct the analysis in that we were concerned with identifying positive and negative experiences with various predetermined processes within the study, but were open to identifying themes and ideas which summarised the data that had not been predetermined. All researchers involved were experienced in applying this thematic analysis approach to qualitative data. EAF, KK and LS analysed a 10% subset of the interview transcripts independently and met to consider their level of agreement and discrepancies. The transcripts were read in their entirety to gain familiarity with the data and then re-read to highlight examples of ‘meaning’, such as ideas, comments and descriptions, to form codes. Notes regarding these initial codes were made in the margin, for reflection and discussion with the other two researchers. Agreements about discrepancies were reached, and EAF applied agreed principles in all subsequent analyses. A complete list of codes, with example quotes from the text, from all the interviews by subgroup (e.g. all the codes for participants only) were reviewed and grouped together by commonality to create a list of overarching themes (based on the research questions) to describe the data. If themes could be grouped further to create higher-order themes, this was conducted. The final themes were reviewed a final time to ensure that codes were included in the most appropriate theme, and theme names were assigned.

Sample size

Based on RCT data investigating the recruitment of smokers to SSS via letters from their GPs,46 we would need to enrol 980 smokers to detect a 7% difference in attendance at SSS between control and intervention arms in a definitive trial. Based on this estimate, any trial would need to recruit 1.8 participants per day in total to achieve the required recruitment in 18 months. We planned to recruit for 3 months in a single LA provider of SSS, requiring one-sixth of the sample and providing a target of 162 participants in this time frame.

As a result of the challenges experienced during recruitment, outlined in Challenges with recruitment and deviation from planned protocol, a revised target of 120 participants was agreed with the NIHR Public Health Research (PHR) programme manager. This target adhered to recommendations from Teare et al. 47 about having a minimum of 60 participants per arm of a feasibility trial with a dichotomous primary outcome measure. 47 Given that this feasibility study was conducted in a single LA area and any future definitive RCT would be delivered as a multisite trial, a target of 120 participants still represents a conservative target for what may be possible in a national multicentre trial. With a sample size of 120 participants, the recruitment rate of smokers for a full RCT was calculated as estimable with a precision (95% confidence interval width) of ± 5%.

Data analysis

Data from Qualtrics were downloaded and analysed using the Statistical Product and Service Solutions (SPSS v25). The overall recruitment rate was calculated as the total number of participants recruited divided by the number of days recruitment was live, to provide a recruit per day figure for one LA. To enable comparison across recruitment settings, the same calculation was made for each of the different settings, and for GP surgeries the rate was also calculated as the proportion of those recruited from those identified as smokers and invited to participate, expressed as a percentage. Cost per baseline recruit was calculated by summing the total costs spent to promote and advertise in each setting divided by the number of recruits achieved there. Attrition rates across all three settings were calculated as the proportion of those lost to the 2-month follow-up, expressed as a percentage. Attrition at baseline (i.e. consented but no completion of measures) and at intervention access was also calculated in this way. Owing to low frequencies of data in some of the ethnicity categories, which would impair statistical analysis, the variable was reduced to four categories: ‘white British’, ‘white other’, ‘all mixed background’ and ‘African/Asian background’.

An analysis of variance (ANOVA) test was conducted for continuous data and the Kruskal–Wallis test was conducted for ordinal data, to assess any differential recruitment between the three settings. Statistical assessments were conducted at the 5% level of significance or 95% confidence intervals. We also asked participants in the process evaluation how they felt about the randomisation process and their experience of the study methods. To assess the acceptability of the StopApp intervention, we explored engagement using Matomo analytics and asked participants randomised to the StopApp arm for their views during the process evaluation interviews.

To assess the acceptability of measures, including outcome measures and measures needed to evaluate cost-effectiveness in a potential future definitive RCT, we considered the proportion of missing data across study measures. Missing value analysis identified data missing at > 5% on any single measure. Those interviewed were also asked about the acceptability of the questionnaires that they completed.

To assess the health equity of the study methods, we report on the spread of age, sex, ethnicity and SES represented in baseline recruits; those who completed follow-up and those who did not; and the differential attrition by sociodemographic characteristics of the sample.

To assess the feasibility of accessing SSS data on booking, attendance, quit dates set and 4-week quit rates, we report on the challenges experienced around data access. We also report rates of booking, attendance and quit rates from both the self-report data and the objective data, and have calculated the event rate of booking in the intervention arm in support of a future trial sample size calculation.

The key costs incurred in delivering the intervention and the usual-care data are examined in Chapter 5.

Patient and public involvement group

A PPI group of 11 smokers from a range of SES backgrounds was invited to support the delivery of the MyWay trial. A number from this group (n = 3) had previously been involved in the co-design and end-user testing of StopApp and were instrumental in contributing to several waves of improvements made to the content, functionality and user experience. We recruited new members to this group (n = 8) during preparation for our funding application, and we consulted and engaged with the wider group regularly on the design of this feasibility study. With the help of a NIHR Research Design Service PPI grant of £500, we held a workshop with and sent follow-up e-mail communications to the group to consult on study design. Payment via vouchers for all contributions was provided. In addition to several suggestions about locations and ways to recruit (e.g. through bus stop and public transport-based advertising), and recruitment messages, attendees also suggested that the follow-up measures should be collected at 2 months post baseline to encourage responding; any longer was deemed problematic, and would lead to a loss of interest and disengagement. The PPI group also read and approved the lay summary of the project for the funding application. A further three meetings were carried out following the commencement of the trial to (1) review the study materials and protocols to ensure that they were as inclusive and acceptable as possible, (2) discuss recruitment issues with suggestions for improvements and (3) review the acceptability of trial methods in the format of ‘think aloud’ sessions (members participated in the sessions as dummy participants). We provided training in research processes and met with members at regular intervals either face to face or via online methods, to consult and gain their input. Two representatives from the group also attended and contributed to the Study Steering Committee meetings. Detailed minutes and notes were taken by a co-facilitator throughout the PPI group sessions to record ideas and suggestions.

Primary objective: recruitment and attrition rates across three settings (general practitioner, community setting, social media) at baseline, intervention access and follow-up

A total of 838 participants signed up on the study website, consented to take part in the study and completed the baseline measures. Of these, 715 (all accessing the study via SM) were excluded because they did not meet the eligibility criterion of living or working in Warwickshire (see Challenges with recruitment and deviation from planned protocol for details). A total of 123 eligible participants were, therefore, recruited over a period of 116 days and included in the baseline analyses (overall recruitment rate of 1.06 participants per day). Sixty-one participants were recruited via SM (0.53 per day), 36 from CSs (0.31 per day) and 26 from GP practices (0.22 per day). Using text messages and letters sent by post to recruit people from GP practices, we were able to identify that only 1.62% of those who received an invitation were recruited. All recruits were obtained via text message. Five participants were not computer randomised within the trial and were instead randomised manually by our statistician. Seven participants were not randomised. Interrogation of web analytics suggest that these seven participants completed the baseline measures but closed their browser before being assigned.

Of the 123 participants recruited, 116 were randomised (59 to the StopApp arm, and 57 to the control arm). A total of 60 participants completed the follow-up questionnaire (60/123; 48.8%) (n = 32, 54.2%, in the StopApp arm and n = 28, 49.1%, in the control arm) at 2 months post baseline (see Figure 3). This revealed an attrition rate of 51.2%. Loss to follow-up by site was as follows: SM (n = 24; 39.3%), CSs (n = 18; 50.0%) and GP (n = 21; 80.8%).

The demographic characteristics of participants are presented in Table 1. The distribution of participants by age, sex, ethnicity, SES and health-related quality of life between the two groups appeared balanced.

Secondary objectives

Process interview results

Eleven process evaluation interviews were conducted, involving five trial participants and six members of staff involved in supporting the study.

Thoughts about encouraging smokers to take part in the study

• The PH staff member reported a reluctance in some smokers (PH, line 22) as they become defensive when discussing smoking (PH, line 25), making it difficult for staff to approach the issue (PH, line 26). She felt that ‘her’ smokers did not want to give up or were heavy smokers, which provided an additional challenge (PH, lines 58–63).

• She felt that volunteers, in particular, found it uncomfortable (PH, lines 28–31). A lack of computer skills and age were also barriers (PH, lines 34–39), with participants primarily motivated by incentives rather than anything else (PH, line 45). She suggested the study was therefore more suited to a GP environment (PH, lines 50–56).

• Pharmacy staff also reiterated the ‘difficulty of the patient group’:Pharmacy32, line 97

  The onus is on the patient really.

Findings from the patient and public involvement group sessions during the study

Two meetings took place with the PPI group: one during data collection and one after data collection. The first meeting took place after recruitment had ended (June 2019) and was designed to review the user experience of participation. In order to do this, PPI members were asked to take part in the study as ‘dummy’ participants and to comment on the process throughout as part of a ‘think aloud’ exercise. There was also time to discuss some of the changes and issues within the study such as recruitment and the provision of vouchers. The second meeting took place at the end of the study (November 2019), and provided an opportunity to share with the group findings from the study, points of learning and future directions. The members had a final opportunity to comment on the research process and thoughts about a more targeted approach to SM in a full trial to include the use of external agencies offering services that enable targeted recruitment of smokers with the use of bespoke algorithms. A total of three PPI members attended both meetings and brief example notes were taken (see example PPI group meeting notes on the Journal Library website; URL: www.journalslibrary.nihr.ac.uk/programmes/phr/1518326/#/; accessed March 2021).

Introduction

The aim of the economic component of the study was to assess the feasibility of methods to collect cost and outcome data to inform the parameters of a future definitive RCT designed to assess the cost-effectiveness of StopApp. As StopApp is a web-based behavioural change intervention to improve motivation and opportunities to access SSS, the additional resources required to deliver the app should be compared with usual care. This is important because any additional benefits associated with an intervention must be assessed in terms of any additional costs attributed. 50

The economic costs associated with smoking are considerable, and these costs affect the health-care sector, society and families. In England, the estimated costs to the NHS associated with smoking were £2.6B in 2015, with these costs including approximately 520,000 smoking-attributable hospital admissions among those aged ≥ 35 years. 51 In addition, there were broader costs for society estimated to be > £9B, associated with a reduction in years of disability-free life, increases in absences from work and early retirement, increased morbidity for those of working age and impacts on mortality. 52 Moreover, smoking contributes to health inequalities, as those from lower socioeconomic groups are more likely to take up smoking, and are less likely to quit successfully. 53

A detailed description of the intervention is reported in Chapter 2, Intervention. Briefly, the intervention involved a behaviour change intervention to improve motivation and opportunities to access SSS. StopApp had been developed prior to this study, with input from smokers and ex-smokers from a range of socioeconomic backgrounds. For the feasibility RCT, participants were randomised to receive either StopApp or standard promotion and referral to SSS. Those who were randomised to the intervention arm, and accessed StopApp, received tailored content to address negative perceptions and could instantly book an SSS appointment or receive a reminder to reaccess the intervention at a later date. The web app is linked to an online system that supports service providers to record and report SSS outcome data. Those randomised to the control arm received access to a web page with standard information and details about how to access SSS.

As fully online RCTs are relatively new, there is ongoing discussion about the best ways to recruit participants to such studies and the relative costs involved. 54,55 In addition to assessing the feasibility of methods to collect cost and outcome data to inform the design of a future RCT, this study component also aimed to inform decisions around recruitment methods by analysing the costs and outcomes associated with different recruitment approaches. Within the feasibility study, recruitment took place in three settings (GP practices, CSs and online). The economic component collected the costs associated with different recruitment settings to assess the resource use associated with recruitment via different methods to inform future discussions about the best methods to recruit to the definitive RCT.

Methods

The objective of the economic analysis was to assess methods for collecting the cost and outcome data for the StopApp arm and the control arm (usual care). We explored methods to measure costs from the perspective of the public sector (including the NHS, LAs and other public sector agencies) in line with the National Institute for Health and Care Excellence (NICE) recommendations for public health interventions. 56 The study also attempted to assess private costs incurred by participants such as productivity costs (e.g. time off work to attend the intervention sessions) and costs associated with travel and the purchase of stop smoking aids.

Results

Health economic discussion

The economic component of the study aimed to design processes and capture data on the costs and outcomes associated with the intervention and comparator programmes, to inform the methods for a future economic evaluation alongside a definitive trial. On the whole, processes were successfully put in place to measure the costs associated with delivering the intervention and the control. A range of outcome data were successfully collected, including the use of the ICECAP-A instrument. The cost data collected were comprehensive and, in particular, the private costs borne by the participants were well captured. The findings of this feasibility trial suggest that, in a definitive trial, it would be important to capture a broad range of costs as participants reported that they had incurred a range of private costs that should be considered for any future trial.

However, some challenges were also encountered. In particular, difficulties were encountered in terms of capturing data on health-care resource use (including information on prescriptions) for the control arm. As part of the study, we aimed to collect data on the private costs borne by participants and their families that were related to smoking or trying to quit. Again, data were less complete for those in the control arm. We recommend that, for the future trial, attention is paid to increasing completion rates for the control arm in relation to resource use and private costs. In addition, for a definitive trial, it would be more efficient to collect data on attendance at SSS, missed SSS appointments, and so on, using data directly obtained from SSS. Similarly, primary care resource use could be obtained through accessing primary care patient data (via Outcomes4Health, which is also owned by the data controller Pinnacle Healthcare Ltd). We would therefore recommend that appropriate DSAs are put in place as part of the trial development to facilitate this.

The outcome data were collected successfully. For a full trial, further outcome data might be collected directly from SSS to increase understanding of the potential impacts of the intervention. Collecting a range of outcome data would permit different types of economic evaluation to be conducted in a future definitive trial. Firstly, a cost-effectiveness analysis could be conducted, using the primary clinical outcomes included in the trial, including cost per SSS attendance and cost per quit attempt. In addition, collection of EQ-5D data would allow a cost–utility analysis to be undertaken as recommended by NICE. Finally, collection of data on more general well-being (via ICECAP-A) would enable a broader analysis to be undertaken, which would focus on outcomes beyond health.

The feasibility trial has provided valuable information about the costs and benefits associated with different methods of recruitment. It would seem that online methods are cheaper, based on the cost per recruit. However, there would need to be further consideration of the quality of engagement with the trial and the impacts on follow-up.

In a definitive trial the costs associated with attempting to reduce cigarette consumption and stop smoking would need to be offset against the longer-term savings associated with reductions in smoking-related morbidity and mortality. As is the case with many public health preventative programmes, longer-term costs and outcomes would need to be considered. 65 Hence, it will be necessary to use a decision-analytic model to extrapolate costs and benefits beyond the follow-up period. There is a range of evidence that could guide the development of the model, including the cohort model used to inform NICE guidance on interventions for smoking cessation66 and other more recent work in this area,67,68 and adaptation of an existing model will be explored. The development of a model will require data from meta-analyses on long-term abstinence, relapse and lifetime abstinence69 and utility estimates for smokers and smoking-related diseases. 70–72 The aim of the decision-analytic model will be to provide information on the potential longer-term costs/savings and health benefits associated with the StopApp intervention.

Recruitment, attrition and equity of methods

The primary objective of this study was to estimate recruitment and attrition rates of participants for a future definitive RCT of StopApp across three settings (GP surgeries, CSs and online) at baseline, intervention access, and 2 months’ follow-up. Our data show that it was possible to recruit 123 smokers over 116 days in a single LA area through an overall recruitment rate of 1.06 smokers per day. Over 12 months it should therefore be possible to recruit 387 smokers within a single LA SSS provider, and the number of LAs could be increased to achieve a required full-trial target. For example, recruitment via four LAs with a similar population size might be expected to achieve baseline recruits in excess of 1500 participants. We also found that withdrawal of consent prior to randomisation did not occur. All participants who signed up to the study online and completed consent procedures also completed baseline measures in full.

Recruitment rates varied across the three recruitment settings. SM recruitment outperformed the other two settings, with 61 eligible recruits, compared with 36 recruits in CSs and 26 recruits in GP practices. There was, however, a significant problem with fraudulent sign-ups from people via SM who were not eligible to take part as a result of not living or working in Warwickshire. Changes to the protocol to prevent fraudulent activity are needed for any future definitive trial, and are discussed in Key costs that would be incurred in delivering the intervention and usual care, including a comparison of did-not-attend rates between each arm of the trial.

Social media recruitment was also the most cost-efficient method of recruitment at £50.20 (including e-voucher payment) per recruit, compared with £184.42 per recruit in CSs and £123.39 per recruit in GP practices. SM also achieved the lowest rates of attrition at the 2-month follow-up, with 61% of participants completing follow-up measures compared with only 50% of participants in CSs and 27% of GP recruits. These rates of attrition are relatively high. Researchers73 who recruited smokers to a ‘taster session’ of smoking cessation support via a tailored letter from their GP achieved a 77% follow-up rate. Other studies, however, have tended to use more intensive methods to pursue follow-up data, including phoning participants and pursuing only the primary outcome measures for those most resistant to completing follow-up measures. 46,73 We only sent two auto-generated e-mails, 2 weeks apart, with a request to complete measures online via a web link and the offer of a £5 e-voucher for doing so, and achieved follow-up with 61% of participants in the SM setting. In a future definitive trial, we could aim to increase follow-up to a target of 80% of participants by employing more intensive follow-up procedures as other studies have done.

Even with the very unintrusive follow-up methods employed in the current study, the numbers achieved suggest that a full-scale trial would be feasible if using SM as the main driver of recruitment. If using SM alone, in six LA areas over 18 months with a similar population size to Warwickshire, it may be possible to achieve as many as 1830 baseline recruits and 1116 participants’ self-report data at follow-up (based on 61% follow-up achieved in the present study), or more with more intensive follow-up methods. The same calculation, applied to CS data alone, suggests a baseline of 1080 recruits and 540 participants at follow-up. The most problematic recruitment setting was the GP surgery setting. Follow-up rates were very poor and, based on the data obtained in this feasibility study, to recruit 1000 smokers at baseline we would need to invite smokers from 228 GP surgeries. Gilbert et al. 46 had to expand the number of GP settings that they recruited from during their RCT of an invite letter to a taster SSS session, but they ultimately achieved more than 4000 baseline recruits through 99 GP surgeries.

It seems likely that the relative success of the SM recruitment method is due, in part, to the fact that we ran the entire trial online, and, to our knowledge, we were one of the first to attempt this for a smoking cessation-related trial. Those who came across study information online and signed up were therefore likely to be smokers who were familiar with and comfortable in online spaces. Participants recruited via SM and the CSs were also more likely to have an interest in stopping smoking before deciding to take part. This will potentially have affected participants in both the intervention group and the control group and reduced the difference in SSS booking between groups. Engaging in the study will have been more effortful for those who found out about it in CSs (e.g. saw a poster or bus campaign message). We also know that no-one who received a paper-based letter from their GP signed up to the study, and a preference for texts was discussed in the process interviews. All sign-ups through GP surgeries came from those who were sent text messages containing a link to the study website. Because StopApp is a digital intervention designed to reach smokers and motivate those interested in quitting to use the convenient SSS booking system it provides, it is not surprising that those who could achieve a rapid and seamless engagement process by simply clicking on a link signed up in greater numbers. Furthermore, it seems likely that the differential attrition rate across settings was, to some extent, due to the different motivational drivers that may affect smokers in each setting. Smokers who signed up to the study following a message from their GP may have been more reluctant to participate in the 2-month follow-up if they had not been prompted to book and attend a SSS appointment than those who had the greater perceived anonymity of participation via other routes. No outcomes from the study were shared with GPs, but GPs will remain aware of the smoking status of their patients in the future; therefore, social desirability effects may have played a role in the low follow-up from smokers via GP recruitment.

Given the novelty of our intervention and online trial methods, it is difficult to make direct comparison with recruitment to other smoking-related trials. Recruitment via GP settings does, however, provide some notable points for comparison. First, based on the number of adult smokers in England, which currently stands at nearly 15%,1 and the fact that there was a combined list size of 52,608 in the six GP surgeries from which we recruited, we should have expected to invite close to 7890 smokers to the study. Instead, only 1602 smokers were identified on participating GP lists by the CRN that supported the study and sent either a letter or a text message invitation. We therefore invited an average of 267 smokers per participating surgery, compared with an average of 1075 smokers per surgery in previous research. 73 It is not clear why so few smokers were identified. Of those ineligible, 76 people were ineligible because their first name was not listed first alphabetically out of the known smokers in the household (as per the criterion applied by the CRN). It is possible that the data that these surgeries held on the smoking status of patients were simply inaccurate and left many smokers unidentified in the searches of patient lists. Second, of the 1602 smokers who were invited to participate, only 1.62% responded to that invitation and completed baseline measures. Other studies have achieved recruitment at closer to 11%,74 although recruitment has been as low as 0.27% in other smoking studies. 75 It is possible that the very brief information about the study that was included in the text messages was not sufficient to entice individuals to find out more. Participants in the Gilbert et al. 46 study received a much more detailed, personalised invitation that contained health risk information tailored to their own current health status, age, sex and smoking habits. Furthermore, participants in another trial76 were invited to attend SSS in correspondence sent by their GP that sought to update their medical records on smoking, which may have encouraged and normalised participation, leading to a higher recruitment rate. It is possible, therefore, that recruitment via GP practices in a future definitive RCT may be improved by a combination of better identification of smokers on GP patient lists and changes to the brief text message invitation.

Differences in some participant demographic characteristics were noted between the recruitment settings, with slightly older participants consenting via GP recruitment and younger participants consenting from CSs. CS participants also reported smoking the most cigarettes, with GP recruits smoking the fewest. GP participants also reported the greatest intentions and likelihood of stopping smoking, with CS participants reporting the least. The fact that the samples recruited from each setting vary slightly is not surprising, but needs to be considered in any decision about the recruitment methods in a future definitive RCT.

Attrition at the point of intervention access was measured using analytics data from Matomo, which indicated that StopApp received 60 home page views from the 59 participants randomised to receive this content, indicating that one participant viewed it more than once. Of the 59 participants, we can see that nine participants (15%) proceeded beyond the home page, interacting further with the site. We also know that five participants (56% of those who engaged with it) booked a SSS appointment using StopApp. Of the 60 home page views, a total of 51 exits occurred. It remains unclear whether the 51 exits from the home page occurred without any reading of the content or scrolling, and whether or not the exits were intentional. During the PPI group think-aloud session, it became apparent that one member of the group had been randomised to the intervention arm, but had been unprepared for being presented with the home page for StopApp and clicked the back button on the assumption that they had missed some introductory text regarding what it was. This inadvertently took them out of eNgage, and therefore the study, and they were unable to return to StopApp; therefore, it is possible that some of the exits may have been unintentional and we may have lost some people who would have otherwise booked an appointment at SSS via StopApp. Consequently, as we cannot be clear whether some attrition at intervention access was accidental or intentional, our event rate for booking at SSS may be under-reported. A future definitive trial may benefit from the provision of information at the end of the baseline measures to brief participants about what they are about to view, including clear messages to avoid clicking the back button, if the facility to do this cannot be removed. The research team has also received intensive training in Matomo analytics since the end of the trial so that we can make better use of the sophisticated participant tracking options available in the software for both trial arms in a planned future definitive trial.

We observed recruitment and attrition rates across socioeconomic groups, ethnicity, age and sex. The only issue identified was a small imbalance in sex compared with the general population profile, with a greater number of females recruited. However, this is not an unusual finding as it has been experienced in other smoking trials (e.g. 70% females77 and 59.4% females78). One way to address this may be to use a targeted approach in SM recruitment to monitor imbalances and respond accordingly. This might include the possibility of utilising algorithmic approaches within SM advertising that enable reach to particular groups, at particular times and in particular contexts; for example, this may include recruiting from particular workplace settings where there are known to be higher numbers of male smokers.

The demographic and smoking status profiles of those who completed follow-up and those who did not were broadly similar, suggesting that the study did not incur an attrition bias. There was, however, some indication from reviewing demographic data in Table 2 that those from black, Asian and mixed ethnic groups were lost to follow-up in greater numbers than those from white ethnic groups. Given the need for more intensive follow-up methods in general, this could be addressed by pursuing these follow-up data, in particular from minority groups, to avoid a lack of representative follow-up data in a future trial. In addition, some exploratory data analysis identified that a lack of internet access was associated with being from a mixed, black or Asian ethnic background. Recent government data1 suggest that overall internet use by ethnicity is actually lowest for white people (90.5% have access) and Asian ethnic groups (92.4% have access), compared with 92.8% of black people, and 95.1% of ‘other including mixed’ ethnicity groups. Interestingly, the data showed that internet access was lowest across ethnic groups and regions (i.e. among Asian people and people living in the West Midlands), compared with other UK regions, where this study was conducted. This suggests that our sample was not necessarily representative of internet use among ethnic minorities across the general UK population; therefore, the likelihood is that a definitive RCT of the StopApp intervention would reach a sufficiently broad sample of the population.

Acceptability of randomisation methods

Our data suggest that the randomisation methods used were feasible and acceptable to participants, as disclosed during the process evaluation interviews. We did identify an issue with some participants missing randomisation; however, this was likely to have been because participants closed the browser too early, not realising that there was something to see after the measures. For the full trial, careful guidance would be provided to participants and embedded monitoring and alerts would be used to ensure that randomisation was occurring for all participants, as this cannot adequately be managed by the research team or a clinical trials unit manually without time lapse and error.

Acceptability of the trial measures

The measures used were largely completed as expected, although some items contained consistent missing data or participants provided contradictory answers, suggesting that people were not willing to complete these measures or were finding them confusing or boring. Further work with our PPI group is required to revise these particular measures before including them in a full trial. Feedback on the measures was principally positive from the process interview findings, with agreement that item measures were not too onerous and were easy to comprehend and that there was a clear rationale for their inclusion.

Key costs that would be incurred in delivering the intervention and usual care, including a comparison of did-not-attend rates between each arm of the trial

Key costs incurred in delivering the intervention and usual care have been set out and discussed in Chapter 5. It is clear that the ongoing costs of delivering the intervention rather than usual care are negligible for commissioners and, therefore, the main costs to be considered are the resource use costs of those in the intervention arm versus those in the control arm. The provision of e-vouchers may have contributed to the high rate of participants who DNA their appointments. In a full definitive trial, the provision of e-vouchers needs some consideration to avoid attracting fraudulent behaviour. There is an indication in the data that resource use was higher among intervention arm participants, but there were also issues with the completeness of data from control arm participants on these measures that would need to be addressed for a full trial, in conjunction with our PPI group. Data on DNA rates were not sufficiently reliable nor available in the local Warwickshire LA for a comparison of associated costs to be carried out (see Feasibility of accessing Stop Smoking Services data). However, in supporting the feasibility trial, Pinnacle Healthcare Ltd did add a field to indicate whether an individual attended the appointment or not (i.e. DNA) to their routine process (as a result of supporting the MyWay trial), which we believe is a successful tangible impact on SSS attendance reporting. However, this approach did demand additional effort from staff completing the fields and puts the onus on these staff members to accurately report attendance.

Feasibility of accessing Stop Smoking Services data

This feasibility trial demonstrated that data on participant SSS use and smoking outcomes could be accessed for a future definitive trial; however, some amendments to the study set-up procedures and protocol are needed to ensure that there is accurate and complete data collection. For the data to be released to the research team, Pinnacle Healthcare Ltd required additional, direct consent from participants, collected by the pharmacist at the first appointment, rather than online in eNgage. We found that some participants who had signed up for the trial stated that they were not part of the trial when prompted by pharmacists and, therefore, were not invited to consent and, as a result, their data could not be accessed. Consequently, it will be necessary to consider more carefully how consent for access to SSS data is collected, when and by whom, to ensure the collection of accurate, objective smoking outcomes. We recognised that some participants may choose to book an appointment at SSS within a GP practice or a non-participating pharmacy (which could include participants in the StopApp arm who chose to book an appointment via alternative means to the web app, despite being prompted to do so via StopApp content). In the trial design that we were ultimately forced to adopt for this feasibility study, we were unable to collect data about these services and could only rely on self-reports of appointment bookings, which may be subject to bias. This could be addressed in the planning of a future definitive RCT of StopApp by working with Pinnacle Healthcare Ltd and the tobacco control commissioners in participating LAs or public health departments to set up a clear data-sharing protocol that meets the research needs. This could be organised with all relevant parties from the outset and alongside preparation of an application for trial funding.

Rate of Stop Smoking Services booking in the intervention and control groups to estimate the event rate of the primary outcome measure for a future trial and to support future trial sample size calculations

The booking event rate calculations based on the available objective and self-report data from participants have suggested that a future definitive RCT should be powered to detect an approximate 6% difference between the control and intervention arms. This is a conservative estimate of the likely difference between trial arms, since the true difference in booking rate between trial arms could be nearly three times higher. This more conservative estimate is in line with other smoking trials targeting attendance at SSS. 46,76 A sample size calculation based on this difference for a dichotomous outcome such as booking a SSS appointment with the significance level held at 0.05 and power at 0.95 suggests that we need an overall sample size of 1036 participants or 518 participants in each arm of a definitive trial to detect a difference of 6%. This sample size has been identified as being possible to achieve if using a well-targeted SM campaign with or without other recruitment settings also included.

Limitations, proposed improvements for a full trial and potential of StopApp

Based on our findings, a full trial would benefit from some amendments to the feasibility trial protocol. First, a more proactive approach to reducing attrition and improving follow-up data collection is warranted; for example, in many smoking trials follow-up data are collected via telephone calls with participants,46,73,79 whereas in this trial two e-mails were sent. This would also help to ensure representation of all demographic groups in follow-up data. Second, we demonstrated that use of SM was the most effective and cost-effective recruitment method. However, the approach to recruitment was not sufficiently targeted towards smokers and could benefit from employing specialist digital marketing agencies to recruit on our behalf to design bespoke algorithms to strategically target specific groups online for trial recruitment. Third, as discussed in Data collection at participating Stop Smoking Services, we experienced difficulties in collecting information from SSS, including issues with pharmacists obtaining consent. Prior to commencing a future definitive trial, there is a need to work closely with the data controller, Pinnacle Healthcare Ltd, LAs that commission SSS and are interested in participating in a trial, and the SSS providers using the software within pharmacies and GP surgeries to ensure that we can collect this information without attrition. For example, consent might be collected by the research team via text messaging at the point of SSS booking, or evidence of baseline consent could be provided to the data controller as proof for data release. Last, because of issues with the study management software (i.e. eNgage not being fully optimised for mobiles initially, and problems identified with hitting the back button in StopApp after randomisation), we would need to fully test digital trial procedures in each site of a multisite full RCT ahead of starting recruitment.

If a proposed full trial establishes that StopApp is effective and cost-effective in increasing attendance at SSS, over and above usual SSS promotion activity, then it also meets the APEASE criteria (affordability, practicality, effectiveness/cost, acceptability, side effects/safety and equity),80 in that it can be instantly scaled up nationally owing to the widespread existing use of PharmOutcomes systems by SSS. It could also be incorporated into the GP-based version (Outcomes4Health), which is also owned by Pinnacle Healthcare Ltd, which uses an identical platform. StopApp would incur minimal additional costs to local commissioners, who will require only an updated map of services for their own locality; therefore, it offers a highly affordable addition to existing SSS provision, targeting smokers who might otherwise have avoided services as a result of mistaken beliefs about what they offer, their approach and how to access them. A future trial will compare cost savings associated with StopApp, including reduced administrative burden in arranging appointments and the possibility of fewer wasted sessions due to non-attendance or attendance by those who were referred by health professionals and are, therefore, less motivated to stop smoking. The seventh World Health Organization report on the global tobacco epidemic81 states that its principal focus is the need for greater comprehensive cessation support worldwide. One-third of the global population (2.4 billion) in 23 countries now have access to SSS, two billion more people than in 2007. Given the clear recommendations to offer continued cessation support, reaching the harder-to-reach entrenched smokers from more deprived backgrounds becomes ever more imperative.

Conclusions

This feasibility RCT has demonstrated that it is possible to recruit and retain sufficient smokers to assess the effectiveness and cost-effectiveness of StopApp at increasing booking for and attendance at SSS compared with usual SSS promotion. The trial methods were identified as acceptable and equitable. A number of challenges relating to the feasibility of a definitive RCT have, however, been identified. It is arguable that these could be addressed in the process of the planning of a future RCT, before a funding application is submitted. The acquisition of service use data will demand careful planning with all relevant data controllers and processors in order to ensure that there is complete objective data on all bookings, attendances/DNAs, quit dates set and 4-week quits achieved among participants. We foresee this resolution involving detailed discussions with all relevant parties (including Pinnacle Healthcare Ltd) and a funding application for the full RCT being submitted only when everyone is in agreement with the appropriate data access. A clear SM recruitment strategy is also required to successfully target smokers and reduce the likelihood of fraudulent activity, and a NIHR-preferred provider has been identified for targeted SM marketing to address this need. In addition, in line with methods in other smoking trials, greater efforts to pursue follow-up data from participants would be applied to minimise missing data.

Research recommendations

The following specific research recommendations based on the experience of delivering this study should be considered and appropriately incorporated in the planning of a future definitive RCT of StopApp, and may also be of relevance in the planning of similar digital or smoking-related feasibility and full RCTs:

• Agree data-sharing protocols with the data controller and all data processors, which, in the case of the current trial, includes Pinnacle Healthcare Ltd, LA commissioners and LA service providers. Where LAs wish to be involved as a study site, they could use commissioning contract arrangements to require data sharing for research purposes.

• Booking and attendance remain the primary outcome measure but a broader measurement approach to secondary outcome measures, including recording any positive changes in smoking behaviour, may encourage those who have not successfully booked an appointment and/or quit smoking to respond to follow-up measures about their smoking behaviour and outcomes.

• Obtain better access to objective service use data to reduce the reliance on self-report data on resource use by participants for health economic analyses.

• Apply more intensive follow-up procedures (e.g. telephone calls and requests to access the primary outcome items) to reduce attrition and to ensure that demographic representation is achieved at follow-up compared with baseline.

• Work with the PPI group ahead of trial data collection commencement to refine and streamline all digital trial processes and the experience for participants – address any identified glitches or poor experiences. In particular, check that the study management software being used is functioning seamlessly and monitor this regularly.

• Make sure follow-up communications with participants are very clear about who they are from to avoid being viewed as ‘spam’ and consider reducing an initial follow-up from 2 months to avoid participants ‘forgetting’ about involvement.

• Improve the quality of study and digital marketing by using a professional digital marketing company to support targeted marketing to potential recruits via SM.

• Engage a stakeholder and public involvement co-ordinator to support not just the PPI group activities but communicate regularly with all relevant stakeholders about progress, delivery and outcomes of the research.

Ethics approval

Ethics approval for the study was granted by the West Midlands – Edgbaston Research Ethics Committee (reference 18/WM/1070).

Contributions of authors

Emily A Fulton (https://orcid.org/0000-0002-5979-3207) (Assistant Professor Health Psychology) conceived the idea for StopApp and led the stage 1 application for funding to NIHR, co-led the research project and co-wrote this monograph.

Katie Newby (https://orcid.org/0000-0002-9348-0116) (Associate Professor Health Psychology) supported the set-up of the trial data management system and took responsibility for the information governance of the study.

Kayleigh Kwah (https://orcid.org/0000-0003-2307-1285) and Lauren Schumacher (https://orcid.org/0000-0003-3430-1816) (Research Assistants Psychology) conducted the day-to-day management of the study, liaised with all study sites and ran the process interviews.

Kajal Gokal (https://orcid.org/0000-0002-2020-1876) (Research Fellow Health Psychology) managed the ethics application and liaison with the CRN and GP practices.

Louise J Jackson (https://orcid.org/0000-0001-8492-0020) (Lecturer Health Economics) conducted the health economic assessment.

Felix Naughton (https://orcid.org/0000-0001-9790-2796) (Senior Lecturer Health Psychology) and Tim Coleman (https://orcid.org/0000-0002-7303-4805) (Professor Primary Care) advised on the trial design and supported study decision-making throughout.

Alun Owen (https://orcid.org/0000-0002-4358-7118) (Associate Professor; Statistician) advised on all statistical analysis and reporting of statistics.

Katherine E Brown (https://orcid.org/0000-0003-2472-5754) (Professor Health Psychology) led the stage 2 application for funding, co-led the research project and co-wrote this monograph.

All authors contributed to the drafting of this monograph and read and approved the final version.

Publications

Fulton EA, Brown KE, Kwah KL, Wild S. StopApp: using the behaviour change wheel to develop an app to increase uptake and attendance at NHS Stop Smoking Services. Healthcare 2016;4:31.

Fulton EA, Kwah KL, Wild S, Brown KE. Lost in translation: transforming behaviour change techniques into engaging digital content and design for the StopApp. Healthcare 2018;6:75.

Fulton E, Newby K, Gokal K, Kwah K, Schumacher L, Jackson LJ, et al. Tailored digital behaviour change intervention with e-referral system to increase attendance at NHS stop smoking services (the MyWay project): study protocol for a randomised controlled feasibility trial. BMJ Open 2019;9:e028721.

Kwah K, Fulton E, Brown K. Accessing National Health Service Stop Smoking Services in the UK: a COM-B analysis of barriers and facilitators perceived by smokers, ex-smokers and stop smoking advisors. Public Health 2019;171:123–30.

Data-sharing statement

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

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

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 PHR 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 PHR programme or the Department of Health and Social Care.