Using text messaging to prevent relapse to smoking: intervention development, practicability and client reactions
Correspondence to: Sarah Snuggs, Wolfson Institute of Preventive Medicine, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, 55 Philpot Street, London E1 2JH, UK. E-mail: firstname.lastname@example.org
The NHS Stop Smoking Service (NHS-SSS) helps approximately half its clients to quit for 4 weeks. However, most initially successful quitters relapse within 6 months. Short message service (SMS) texting has been shown to facilitate stopping smoking. We describe the development, implementation and subsequent evaluation, in terms of practicability and client response, of an SMS text-based relapse prevention intervention (RPI) delivered within routine community and specialist National Health Service (NHS) Stop Smoking Service (SSS) provision in four Primary Care Trusts.
Text messages aimed at motivation to remain abstinent, preventing careless lapses and continuing the full course of medicine for smoking cessation were developed and sent weekly to clients' mobile phones for 12 weeks and fortnightly for 6 months. They were asked to respond to some of the texts and contact the NHS SSS if they lapsed. They were also offered free nicotine mini lozenges to be sent via the mail on three occasions.
202 clients who had been abstinent for 4 weeks.
Feasibility of introducing RPI into routine care; response to interactive messages and requests for the medication; rating of the helpfulness of RPI; self-reported and carbon monoxide (CO)-validated smoking status for up to 26 weeks.
A text-based RPI was easy to implement within the NHS SSS provided by specialist advisers, but enrolment of clients from services provided by a network of pharmacists was difficult because client contact details were often lacking. Where records of the number of people invited to RPI were available, 94% of eligible participants enrolled. The RPI was well received by both SSS clients and staff, with 70% (n = 63) of clients who completed follow-up considering the intervention helpful. Eighty-five per cent (n = 172) of clients responded to at least one of the nine interactive text messages. Sixty-four clients (32% of the total, 47% of those we managed to contact) reported continuous abstinence at 6 months. Eighteen (9%) clients who relapsed to smoking used the RPI to re-engage with the NHS SSS and 10 (5%) successfully re-established abstinence.
In smokers attending National Health Service Stop Smoking Services who are abstinent 4 weeks after their quit date, a relapse prevention intervention based on SMS text messaging was well received, and can be implemented economically and rapidly. A controlled trial is needed to establish whether it has a significant impact on relapse.
Current treatment offered by the National Health Service (NHS) Stop Smoking Service (SSS) consists of advice and medication provided by pharmacists and practice nurses (‘level 2’ service) and more intensive treatment provided by smoking cessation specialists, usually employed for the purpose (‘level 3’ service). The provision of intensive support generally follows the withdrawal-orientated (Maudsley) treatment model , comprising multi-session behavioural support, typically delivered weekly for 4 weeks after a quit day, combined with medication that is usually provided for up to 3 months. Although clients can maintain some contact with the advisers for the duration of medication use, both approaches focus on the acute withdrawal period and are time-limited.
More than 700 000 people use the service every year, approximately half of whom achieve short-term abstinence . However, more than 70% of these short-term quitters return to smoking within a year [3, 4]. As the health benefits of stopping smoking are only realized with long-term abstinence, relapse reduces the public health benefit of investment in smoking cessation.
The Cochrane Review of relapse prevention literature identified 47 randomized trials of behavioural treatments, with no single trial or group of trials showing an effect . Most trials evaluated Marlatt & Gordon's ‘skills-based’ approach, which includes identifying relapse situations and coping strategies . The interventions were often brief and one-off; many trials mixed smoking cessation and relapse prevention interventions, with only 36 randomizing abstaining smokers, and most trials posed significant methodological problems. The Cochrane Review identified the need for good quality research in this key area as the primary priority in the field of smoking cessation.
The lack of proven relapse prevention interventions has not stopped NHS SSS attempting to provide some relapse prevention help for their clients. A survey of current practice reported that most (58%) services were providing some kind of relapse prevention intervention (RPI) . ‘Common-sense’ approaches are usually used, such as an offer of regular drop-in sessions. Anecdotally, some clients find these helpful, but the sessions are usually poorly attended and their efficacy is unknown.
There are several leads in the existing literature and clinical experience that warrant proper evaluation. The main lead includes simply extending the initial support. Maintained contact can provide an ongoing opportunity to offer motivational support, encourage extended and ‘emergency’ use of medications, advise on developing coping strategies and overcoming lapses, and offer an opportunity to re-engage in face-to-face support if needed. As mentioned above, however, the uptake of an offer to attend continuing sessions is normally very low.
Recently, short message service (SMS) messaging (i.e. sending automated messages to clients' mobiles phones to help promote and maintain abstinence) was shown to be effective in helping smokers to quit [7-10]. This approach also seems eminently suitable to maintain contact with clients and to provide an RPI. Previous studies have evaluated the use of SMS messages to help people stop smoking, although some included relapse prevention messages. This pilot focused on relapse prevention by including only clients who had quit smoking for 4 weeks.
This paper describes the development, implementation and subsequent evaluation, in terms of practicability and client response, of a text-based RPI delivered within a routine NHS SSS service.
Methods and Measures
NHS SSS clients who achieved carbon monoxide (CO)-validated 4-week abstinence were offered the RPI. Clients were identified from within the level 2 and level 3 services in City and Hackney and Tower Hamlets Primary Care Trusts (PCTs) in London and from level 3 services in Leeds and East Kent NHS SSS. Advisers were asked to offer the service to 4-week quitters and to inform the project team if any clients did not want to receive it.
Some clients were offered the RPI as an ‘opt-in’ service, whereby they were required to request the messages, while others (usually those in group clinics) were simply informed that they would be receiving messages to help prevent relapse after the end of their initial treatment unless they chose to opt out.
Development of the text-based RPI
To develop a coherent RPI using brief text messages the project team, consisting of stop-smoking specialists, drafted a pool of 40 messages. The messages were based around two principal themes: extended motivational support (e.g. ‘Hi Jo, a 20 a day smoker will have saved nearly £600 by now—stay smokefree and you will be healthier and wealthier!’), and encouragement of continued use of medication (e.g. ‘Hi John, you should keep using your Champix for up to 12 weeks. Give us a call if you have any questions about it’.). The pool was adjusted and finalized after consultations with an advisory group and a panel of service users.
There were 17 messages in the final version, of which nine were interactive (e.g. ‘Hi Alison, just checking to see if you are still smokefree? Please text back Yes or No’). The messages were similar to those used by other SMS smoking cessation interventions. All clients received all messages unless they relapsed, but the messages were personalized to address the recipients by their first name and, where appropriate, included other personalized details such as medication choice. The messages were sent weekly for 12 weeks following 4 weeks' abstinence, and then fortnightly for up to 6 months. The messages were aimed at maintaining motivation to remain abstinent, preventing absent-minded and indulgent lapses, not stopping medication prematurely and offering additional emergency medication where needed. The nine interactive messages required reporting of current smoking status, ratings of self-efficacy and whether a further supply of medication should be mailed to clients (a supply of 20 nicotine mini lozenges was offered at 6, 12 and 18 weeks post-quit date). Every response to the interactive messages received a reply. If clients reported abstinence, they were sent a congratulatory message. If clients reported a lapse or expressed low confidence in maintaining abstinence, they were contacted by telephone and provided with support tailored to their individual circumstances. Where appropriate, they were encouraged to re-attend the local NHS SSS.
Clients were sent text messages using PageOne Messaging and Business Services. With the use of the SMS plug-in provided by PageOne, messages could be sent directly to clients from an Excel spreadsheet at a cost of 4 pence (stirling) per message.
Routine data required for standard NHS SSS service monitoring were collected , including client demographics and smoking history. Clients were contacted at 6 months post-quit date to assess smoking status and to collect feedback regarding the RPI. Sustained abstinence rates were measured in accordance with Russell Standard  and NHS SSS monitoring guidance , i.e. on an intention-to-treat basis, with clients not available for follow-up counted as smokers. Follow-up contact was made initially via the telephone; those who did not respond to these calls were also approached by post, text message and/or e-mail. All self-reported abstainers were asked to complete a CO breath test [a CO reading of <10 parts per million (p.p.m.) was required for a client to be classified as abstinent], either by attending the clinic (for which they were paid travel expenses), or if convenient an adviser would visit their work-place/home. Point prevalence abstinence (no smoking over the past 4 weeks) and time to first lapse and relapse were also measured (calculated by time from 4 weeks post-quit date to the first cigarette) at the 6-month follow-up. The ‘NHS Stop Smoking Service Client Satisfaction Survey’  and an RPI-specific feedback questionnaire were used to collect client feedback. The former included 24 questions on satisfaction with the service, including questions about staff, waiting times, clinic structure and convenience of appointment times and locations. The RPI questionnaire included questions on the helpfulness of the interactive messages, frequency of the messages, the offer of free nicotine replacement therapy (NRT) and the helpfulness of the RPI overall. Clients were asked to rate these elements on a scale of 1 to 5, where 5 was extremely helpful and 1 was not at all helpful. Follow-up calls were conducted by researchers other than clients' counsellors to minimize misreporting.
Discussions were also held with service managers and advisers regarding the practicability of the RPI and whether they perceived its implementation to be practicable for the service.
A total of 202 clients used the RPI; 18 (9%) were from level 2 services run by pharmacists offering less intensive support, and the remaining 184 (91%) were from level 3 services run by specialist staff providing a more intensive treatment. All clients set a quit day between September 2010 and July 2011. The clients' characteristics are shown in Table 1.
Table 1. Client characteristics.
|Age (mean ± SD)||43 (12.1)|
|Female n (%)||102 (50.5%)|
|Lower-income SESa||88 (43.6%)|
|Intermediate occupation||21 (10.4%)|
|Full-time student||11 (5.4%)|
|None of these||3 (1.5%)|
|None (%)||9 (4.5%)|
|NRT (%)||57 (28.2%)|
|Bupropion (%)||1 (0.5%)|
|Varenicline (%)||132 (65.3%)|
|Varenicline and NRT (%)||3 (1.5%)|
|Level 2||18 (8.9%)|
|Level 3||184 (91.1%)|
|Tower Hamlets||105 (52%)|
|City and Hackney||82 (40.1%)|
|East Kent||3 (1.5%)|
Initially, we expected the RPI to be useful primarily to clients of the level 2 service, as this typically provides a less intensive intervention. This approach, however, encountered problems. Service records often did not include client contact details, it proved difficult to ensure that the extensive network of level 2 advisers (mainly pharmacists) informed their clients about the service, and there was a general problem in that clients' data were often inputted late into the databases. Good records were kept by the level 3 services, and these were consequently the main source of recent quitters offered the RPI.
Due to a variation in clinic procedures, some clients participated on an opt-out basis while others were asked to opt in. Of the 178 clients who were offered the RPI on an opt-out basis, none opted out. Nine of 21 (43%) people took up the offer when they were able to opt in. The remaining 15 (7% of the sample) clients opted in through services where there is no information as to how many clients overall were offered the service.
The PageOne texting service was simple to use and functioned well. It did not include automatic responding to interactive messages, which had implications for staff time, because every incoming text needed to be replied to individually. This was always conducted as soon as possible, but sometimes meant that clients would wait for more than a day before receiving a response, particularly over weekends. Newer versions of SMS systems provide this option.
At 6 months, we managed to contact 137 of the 202 participants (68%). Of these, 77 (38% of the total) reported 7-day point prevalence abstinence, of whom 47 were CO-validated (23%). Sixty-four (32%) were abstinent continuously since their quit date (47% of those who we managed to contact), of whom 36 (18%) were CO-validated. Thirteen (7%) had had lapses, but were abstinent for at least 4 weeks at the time of follow-up (this includes eight clients who re-attended NHS SSS within the 6-month period). Table 2 presents a breakdown of abstinence rates across the 6-month period.
Table 2. Self-reported smoking status over time since quit date.a
|Abstinent n (%)||202 (100%)||103 (51%)||86 (42%)||67 (33%)||46 (23%)||77 (38%)|
|Smoking n (%)||0 (%)||2 (1%)||1 (1%)||4 (2%)||2 (1%)||61 (30)|
|No response n (%)||0 (%)||97 (48%)||115 (57%)||131 (65%)||154 (76%)||64 (32%)|
Eighteen clients (see Table 3) who relapsed to smoking used RPI to re-engage with the NHS SSS, and 10 (5%) of these re-established abstinence successfully.
Table 3. Smoking cessation outcomes at 6-month post-quit date.
|Continuously abstinent (not a single puff)|| |
|Self-report n (%)||64 (32%)|
|CO-validated n (%)||36 (18%)|
|Lapsed or relapsed but now abstinent (not smoked at all in the past 4 weeks)|| |
|Self-report n (%)||13 (6%)|
|CO-validated n (%)||11 (5%)|
|Relapsed and rebooked into NHS SSS||n = 41|
|(1) Rebooked before 6 months follow-up|| |
|Abstinent at the end of NHS treatment||10 (5%)|
|Smoking at the end of NHS treatment||8 (4%)|
|(2) Rebooked after 6 months follow-up, outcome unknown||4 (2%)|
|(3) Took NHS SSS details to rebook themselves, outcome unknown||19 (9%)|
Clients who returned to smoking and provided details of their relapse (n = 40) had relapsed on average 15.5 [standard deviation (SD) = 6.48] weeks after their quit day.
Response to interactive text messages
A total of 172 (85%) of clients responded to at least one of the nine interactive text messages. A total of 141 (70%) responded to at least one message specifically about their smoking status. Clients who reported to be continuously abstinent at 6-month follow-up responded to more text messages than those who were smoking (3.2 versus 1.9, F = 30.57, P < 0.001). This was also the case when using the 4-week point prevalence measure (3.1 versus 1.9 responses in abstainers and non-abstainers, respectively, F = 25.42, P < 0.001).
A total of 84 (42%) clients accepted at least one of the three offers of NRT ‘minis’. There was no significant association between abstinence at 6-month follow-up and whether clients requested NRT.
Three clients (1%) did not want to continue with receiving the messages several weeks into the texting schedule. One felt he no longer needed the contact and two texted their request without providing a reason.
Clients who provided feedback at 6-month follow-up (n = 90) gave a mean approval score of 3.9 (of a maximum score of 5) for the messages overall and 4.8 for the NHS SSS overall (see Table 4). Sixty-three (70%) gave an overall score for helpfulness of the messages of 4 or 5 on a 5-point scale.
Table 4. Client feedback scores.
|Ratings of text messages (where 1 = not at all helpful and 5 = extremely helpful)|
|How helpful did you find it being able to text back?||80||3.69||1.35||4||1||5|
|How helpful did you find the frequency of the messages?||87||3.95||0.86||4||1||5|
|How helpful did you find the free Niquitin minis offered?||61||3||1.7||3||1||5|
|How helpful did you find the messages overall?||90||3.93||1.11||4||1||5|
|Satisfaction from NHS SSS monitoring form (where 1 = very unsatisfied and 5 = very satisfied)|
|Overall, how satisfied are you with the support you have received to stop smoking?||82||4.83||4.66||5||2||5|
|How satisfied are you with how supportive staff have been?||74||4.91||0.29||5||4||5|
|How helpful has the information and advice that staff have given you been?||73||4.79||4.70||5||3||5|
|How helpful did you find having your carbon monoxide (CO) reading done at every visit?||75||4.47||1.09||5||1||5|
Of the 90 people who provided quantitative feedback, 36 provided other feedback. Most comments were positive, e.g. that it was ‘great to feel remembered’ and that the messages felt like ‘a pat on the back’. Some clients felt that the messages had had little impact on them, and three clients reported that the messages reminded them too much of smoking. Two clients who had relapsed but not informed the RPI project team commented that they felt too embarrassed to do so.
Staff feedback was also generally positive; some staff had been told by clients that they liked the messages and none felt that there was any negative impact of the RPI.
The text-messaging RPI proved practicable and acceptable to SSS clients and staff.
The platform that was used to send text messages (PageOne) was simple and cheap, as is typical of these types of intervention. Provided that it is efficacious, such interventions would be much less expensive than alternatives such as multi-session face-to-face.
A minor problem was identified concerning automatic responding to interactive messages. The system also does not allow monitoring of whether clients rejoined SSS, so clients would continue to receive text messages not synchronized with their new treatment plan unless they informed the RPI team. One possible remedy would be to integrate RPI with SSS client management programmes, such as QuitManager . Such software is used widely within NHS SSS and has the capability of receiving SMS text messages and sending automated responses at no extra cost. RPI could be triggered when the 4-week smoking status is entered. QuitManager was not used for the RPI because none of the PCTs involved used it at the time and the project needed the flexibility of adapting messages and having manual control.
The main implementation problem we encountered concerned client contact records from level 2 services. Incomplete records and problems with obtaining data on 4-week quitters in other than quarterly intervals presented significant barriers to implementation. There were no problems in implementing the service in level 3 services. SSS practices vary widely across the country, but it is possible that services which rely largely on extensive networks of level 2 advisers may find RPI implementation more challenging than services employing specialist advisers.
Regarding the contents of RPI messages, there may be scope for adjusting the messages to encourage clients who experienced lapse or relapse to remain engaged with the service. Very few clients reported lapses. Feedback from clients suggested that people felt embarrassed and nervous about informing the system of their ‘failure’. This is typical of stop smoking interventions generally, in that people who fail to quit are reluctant to attend follow-up sessions. The messages could perhaps put greater emphasis on encouraging such contact. There may also be the potential to adapt the RPI to develop, for example, a mobile phone application (‘app’), providing more opportunity for interaction and personalization.
An encouraging finding was that two out of five clients accepted the offer of NRT. Initially, we did not think that many people would take up this offer, especially as in our sample the majority of clients used varenicline for their quit attempt. This is not typical of NHS SSS services (in 2011, 64% of clients used NRT and 26% used varenicline ), but is probably reflective of the fact that the majority of the RPI clients used level 3 services.
Clients who responded to interactive messages were mainly those who were doing well. The finding that those who responded to the messages were more likely to be abstinent at 6-month follow-up is likely to be primarily the result of successful clients being more likely to remain successful. However, it is also possible that, to some extent, such reporting may be reinforcing and contributing to an effort to remain successful.
Another encouraging finding was that at least 10 clients who lapsed were abstinent again at 6-month follow-up and RPI helped another 41 people who relapsed to get back in touch with the NHS SSS to restart treatment. A number of clients also reported that they found the RPI helpful in maintaining their abstinence. Such information, of course, does not provide proof of the efficacy of this approach. Only a controlled trial would be able to assess this.
In summary, the main limitation to this project concerned the incomplete client databases in some services, which restricted the offer of the RPI. This, however, also represents a finding relevant to considerations of both a full-scale trial and any practical implementation of the intervention. The strengths of the project include the fact that the RPI was implemented and evaluated within routine NHS SSS delivery, and the robustness of the follow-up procedures.
The observational data from this pilot project do not provide any indication of whether the intervention contributes to preventing relapse. However, the results suggest that the new text-messaging RPI is practicable and well received, and that it has a potential to be implemented with minimal expense throughout the NHS SSS. The next step would be to evaluate the impact of this RPI on relapse in a randomized controlled trial.
Declarations of interest
This research was carried out as part of a programme of inequalities projects implemented by the UK Centre for Tobacco Control Studies and funded by the Department of Health. S.S., K.M., F.S. and J.G. have no conflicts of interest to declare. H.M. and P.H. have received research funding from, and provided consultancy to, manufacturers of smoking cessation medications.
We would like to thank Christabel Balogun, City and Hackney PCT, Tower Hamlets PCT, Leeds NHS Stop Smoking Service and East Kent Stop Smoking Service for their help with this project. We would also like to thank Dr Shade Agboola, Janet Ferguson and Professor Ann McNeill for comments and support throughout the project. This project was funded by the UK Department of Health.