[Skip to Content]
[Skip to Content Landing]
Figure.
CONSORT Flowchart
CONSORT Flowchart

Steps of recruitment, number who dropped out, and proportion of participants included in follow-up are presented.

Table 1.  
Baseline Characteristics of the Participants in the Intervention and Control Groups
Baseline Characteristics of the Participants in the Intervention and Control Groups
Table 2.  
Primary and Secondary Outcomes
Primary and Secondary Outcomes
1.
Gilljam  H.  Quitting smoking brings quick health benefits [in Swedish].  Lakartidningen. 2012;109:554-557.PubMedGoogle Scholar
2.
Sussman  S, Sun  P.  Youth tobacco use cessation: 2008 update.  Tob Induc Dis. 2009;5:3.PubMedGoogle ScholarCrossref
3.
Lim  SS, Vos  T, Flaxman  AD,  et al.  A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010.  Lancet. 2012;380(9859):2224-2260.PubMedGoogle ScholarCrossref
4.
Agardh  E, Moradi  T, Allebeck  P.  The contribution of risk factors to the burden of disease in Sweden: a comparison between Swedish and WHO data [in Swedish].  Lakartidningen. 2008;105(11):816-821.PubMedGoogle Scholar
5.
Shiffman  S, Brockwell  SE, Pillitteri  JL, Gitchell  JG.  Use of smoking-cessation treatments in the United States.  Am J Prev Med. 2008;34(2):102-111.PubMedGoogle ScholarCrossref
6.
Gollust  SE, Schroeder  SA, Warner  KE.  Helping smokers quit: understanding the barriers to utilization of smoking cessation services.  Milbank Q. 2008;86(4):601-627.PubMedGoogle ScholarCrossref
7.
Stanton  A, Grimshaw  G.  Tobacco cessation interventions for young people.  Cochrane Database Syst Rev. 2013;8:CD003289.PubMedGoogle Scholar
8.
Backinger  CL, Fagan  P, Matthews  E, Grana  R.  Adolescent and young adult tobacco prevention and cessation: current status and future directions.  Tob Control. 2003;12(suppl 4):IV46-IV53.PubMedGoogle ScholarCrossref
9.
Mason  M, Ola  B, Zaharakis  N, Zhang  J.  Text messaging interventions for adolescent and young adult substance use: a meta-analysis.  Prev Sci. 2015;16(2):181-188.PubMedGoogle ScholarCrossref
10.
OECD: Better Policies for Better Lives. Data lab: health data. http://www.oecd.org/statistics/datalab/health.htm. Published 2015. Accessed June 12, 2015.
11.
Free  C, Phillips  G, Watson  L,  et al.  The effectiveness of mobile-health technologies to improve health care service delivery processes: a systematic review and meta-analysis.  PLoS Med. 2013;10(1):e1001363.PubMedGoogle ScholarCrossref
12.
Spohr  SA, Nandy  R, Gandhiraj  D, Vemulapalli  A, Anne  S, Walters  ST.  Efficacy of SMS text message interventions for smoking cessation: a meta-analysis.  J Subst Abuse Treat. 2015;56:1-10.PubMedGoogle ScholarCrossref
13.
Whittaker  R, McRobbie  H, Bullen  C, Borland  R, Rodgers  A, Gu  Y.  Mobile phone-based interventions for smoking cessation.  Cochrane Database Syst Rev. 2012;11:CD006611.PubMedGoogle Scholar
14.
Müssener  U, Bendtsen  M, Karlsson  N, White  IR, McCambridge  J, Bendtsen  P.  SMS-based smoking cessation intervention among university students: study protocol for a randomised controlled trial (NEXit trial).  Trials. 2015;16:140.PubMedGoogle ScholarCrossref
15.
Free  C, Knight  R, Robertson  S,  et al.  Smoking cessation support delivered via mobile phone text messaging (txt2stop): a single-blind, randomised trial.  Lancet. 2011;378(9785):49-55.PubMedGoogle ScholarCrossref
16.
Rodgers  A, Corbett  T, Bramley  D,  et al.  Do u smoke after txt? results of a randomised trial of smoking cessation using mobile phone text messaging.  Tob Control. 2005;14(4):255-261.PubMedGoogle ScholarCrossref
17.
Ybarra  ML, Holtrop  JS, Prescott  TL, Rahbar  MH, Strong  D.  Pilot RCT results of stop my smoking USA: a text messaging-based smoking cessation program for young adults.  Nicotine Tob Res. 2013;15(8):1388-1399.PubMedGoogle ScholarCrossref
18.
Bock  B, Heron  K, Jennings  E,  et al.  A text message delivered smoking cessation intervention: the initial trial of TXT-2-Quit: randomized controlled trial.  JMIR Mhealth Uhealth. 2013;1(2):e17.PubMedGoogle ScholarCrossref
19.
Michie  S, Hyder  N, Walia  A, West  R.  Development of a taxonomy of behaviour change techniques used in individual behavioural support for smoking cessation.  Addict Behav. 2011;36(4):315-319.PubMedGoogle ScholarCrossref
20.
West  R, Hajek  P, Stead  L, Stapleton  J.  Outcome criteria in smoking cessation trials: proposal for a common standard.  Addiction. 2005;100(3):299-303.PubMedGoogle ScholarCrossref
21.
SRNT Subcommittee on Biochemical Verification.  Biochemical verification of tobacco use and cessation.  Nicotine Tob Res. 2002;4(2):149-159.PubMedGoogle ScholarCrossref
22.
Hughes  JR, Keely  JP, Niaura  RS, Ossip-Klein  DJ, Richmond  RL, Swan  GE.  Measures of abstinence in clinical trials: issues and recommendations.  Nicotine Tob Res. 2003;5(1):13-25.PubMedGoogle ScholarCrossref
23.
Haug  S, Schaub  MP, Venzin  V, Meyer  C, John  U.  Efficacy of a text message-based smoking cessation intervention for young people: a cluster randomized controlled trial.  J Med Internet Res. 2013;15(8):e171.PubMedGoogle ScholarCrossref
24.
White  IR, Horton  NJ, Carpenter  J, Pocock  SJ.  Strategy for intention to treat analysis in randomised trials with missing outcome data.  BMJ. 2011;342:d40.PubMedGoogle ScholarCrossref
25.
Heatherton  TF, Kozlowski  LT, Frecker  RC, Fagerström  KO.  The Fagerström test for nicotine dependence: a revision of the Fagerström Tolerance Questionnaire.  Br J Addict. 1991;86(9):1119-1127.PubMedGoogle ScholarCrossref
26.
Jackson  D, White  IR, Leese  M.  How much can we learn about missing data? an exploration of a clinical trial in psychiatry.  J R Stat Soc Ser A Stat Soc. 2010;173(3):593-612.PubMedGoogle ScholarCrossref
27.
Jackson  D, White  IR, Mason  D, Sutton  S.  A general method for handling missing binary outcome data in randomized controlled trials.  Addiction. 2014;109(12):1986-1993.PubMedGoogle ScholarCrossref
28.
Stead  LF, Hartmann-Boyce  J, Perera  R, Lancaster  T.  Telephone counselling for smoking cessation.  Cochrane Database Syst Rev. 2013;8:CD002850.PubMedGoogle Scholar
29.
Stead  LF, Lancaster  T.  Group behaviour therapy programmes for smoking cessation.  Cochrane Database Syst Rev. 2005;2(2):CD001007.PubMedGoogle Scholar
30.
Lancaster  T, Stead  LF.  Individual behavioural counselling for smoking cessation.  Cochrane Database Syst Rev. 2005;(2):CD001292.PubMedGoogle Scholar
31.
Guerriero  C, Cairns  J, Roberts  I, Rodgers  A, Whittaker  R, Free  C.  The cost-effectiveness of smoking cessation support delivered by mobile phone text messaging: Txt2stop.  Eur J Health Econ. 2013;14(5):789-797.PubMedGoogle ScholarCrossref
32.
World Health Organization. Tobacco. http://www.who.int/mediacentre/factsheets/fs339/en/. Updated July 6, 2015. Accessed January 21, 2016.
Original Investigation
Less Is More
March 2016

Effectiveness of Short Message Service Text-Based Smoking Cessation Intervention Among University StudentsA Randomized Clinical Trial

Author Affiliations
  • 1Department of Medical and Health Sciences, Linköping University, Linköping, Sweden
  • 2Department of Computer and Information Science, Linköping University, Linköping, Sweden
  • 3MRC Biostatistics Unit, Cambridge Institute of Public Health, Cambridge, England
  • 4Department of Health Sciences, University of York, Heslington, England
  • 5Department of Medical Specialist and Department of Medicine and Health Sciences, Linköping University, Motala, Sweden
 

Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.

JAMA Intern Med. 2016;176(3):321-328. doi:10.1001/jamainternmed.2015.8260
Abstract

Importance  Smoking is globally the most important preventable cause of ill health and death. Mobile telephone interventions and, in particular, short message service (SMS) text messaging, have the potential to overcome access barriers to traditional health services, not least among young people.

Objective  To determine the effectiveness of a text-based smoking cessation intervention among young people.

Design, Setting, and Participants  A single-blind, 2-arm, randomized clinical trial (Nicotine Exit [NEXit]) was conducted from October 23, 2014, to April 17, 2015; data analysis was performed from April 23, 2014, to May 22, 2015. Participants included daily or weekly smokers willing to set a quit date within 1 month of enrollment. The study used email to invite all college and university students throughout Sweden to participate.

Interventions  The NEXit core program is initiated with a 1- to 4-week motivational phase during which participants can choose to set a stop date. The intervention group then received 157 text messages based on components of effective smoking cessation interventions for 12 weeks. The control group received 1 text every 2 weeks thanking them for participating in the study, with delayed access to the intervention.

Main Outcomes and Measures  The primary outcomes were self-reported prolonged abstinence (not having smoked >5 cigarettes over the past 8 weeks) and 4-week point prevalence of complete smoking cessation shortly after the completion of the intervention (approximately 4 months after the quit date).

Results  A total of 1590 participants, mainly between 21 and 30 years of age, were randomized into the study; 827 (573 [69.3%] women) were allocated to the intervention group and 763 (522 [68.4%] women) were included in the control group. Primary outcome data were available for 783 (94.7%) of the intervention group and 719 (94.2%) of the control group. At baseline, participants were smoking a median (range) of 63 (1-238) and 70 (2-280) cigarettes per week, respectively. Eight-week prolonged abstinence was reported by 203 participants (25.9%) in the intervention group and 105 (14.6%) in the control group; 4-week point prevalence of complete cessation was reported by 161 (20.6%) and 102 (14.2%) participants, respectively, a mean (SD) of 3.9 (0.37) months after the quit date. The adjusted odds ratios (95% CIs) for these findings were 2.05 (1.57-2.67) and 1.56 (1.19-2.05), respectively.

Conclusions and Relevance  With the limitation of assessing only the short-term effect of the intervention, the effects observed in this trial are comparable with those for traditional smoking cessation interventions. The simple NEXit intervention has the potential to improve the uptake of effective smoking cessation interventions.

Trial Registration  isrctn.org Identifier: ISRCTN75766527

Introduction

Smoking is responsible for more than 60 diseases and is globally the most important preventable cause of ill health and death. For every death related to smoking, more than 20 additional individuals will have at least 1 serious smoking-related illness.1

Tobacco is responsible for approximately 9.6% of the total disease burden in Sweden.2 Approximately 6000 people die every year in Sweden of diseases associated with smoking.1 Thousands of young people in Sweden start smoking each year. Tobacco use increases with age, and the earlier one starts smoking, the higher the risks of becoming addicted to nicotine and developing illnesses due to smoking.3 According to national surveys,4 the prevalence of daily smoking among young people between ages 16 and 29 years has been stable in the past 5 years in Sweden at approximately 11% to 13% among women and 7% to 10% among men. Most smokers start in their teens, and, over the course of a year, most young smokers want to quit or cut down.5,6 There is limited evidence on effective smoking cessation interventions among young people, and the younger population is less likely to seek treatment compared with older adults.2,7,8 Consequently, there is a need to increase interventions and develop attractive, effective programs that are capable of reaching underserved populations, such as young smokers.8,9 Apart from having the lowest smoking prevalence, Sweden is typical of other high-income countries in these respects.10

The rapid increase in the number of people owning a mobile telephone has led to new applications in self-management of chronic diseases and behavioral change interventions, particularly for SMS (short message service) text message–based services.9,11-13 Furthermore, mobile telephone interventions could overcome access barriers to traditional health services. Research on smoking cessation interventions using text messaging on mobile telephones is scarce and shows a mixed picture of effectiveness; most trials are pilot studies or are underpowered. A Cochrane review13 from 2012 included 5 trials with more than 9000 participants, and a significant long-term quit rate compared with a control program was seen (relative risk [RR], 1.71; 95% CI, 1.47-1.99). The review included a large trial that found a significant 6-month smoking cessation benefit of the intervention (RR, 2.20; 95% CI, 1.80-2.68). However, this trial and most others included all age groups. In addition, not all steps were automated, because eligibility and baseline data were collected by telephone. In a more recent review,12a total of 13 studies were included in a meta-analysis of the efficacy of SMS text message–based smoking cessation interventions. The studies included in the meta-analysis were homogeneous, and odds ratios (ORs) suggested that text-based interventions generally increased quit rates compared with control interventions (OR, 1.35; 95% CI, 1.23-1.48). The meta-analysis also considered the content and structure of the interventions, and the pooled results indicated that more advanced technical interventions in addition to text messaging, including dynamic messaging that tracks pending answers to specific questions, tailoring messages to the individual participant, specific assessment messages requesting a response, and the provision of peer-to-peer support, did not significantly affect the effectiveness of text-based interventions.12 Therefore, an important research issue to be addressed is whether a simple, text-based smoking intervention, offering only one-way communication, is sufficient to accomplish smoking cessation. In this trial, we report on the effects of a simple SMS text message–based smoking cessation intervention targeting young adults in Sweden.

Methods
Study Design and Participants

Nicotine Exit (NEXit) was a single-blind, 2-arm, randomized clinical trial of an SMS text–based messaging smoking cessation intervention in which participants were randomized to an immediate-intervention or a delayed-intervention (control) group. The study was undertaken simultaneously in 25 student health care centers at all universities and colleges in Sweden except one university (Luleå University of Technology), which participated in a pilot study undertaken to refine trial procedures. Recruitment of participants was completed over a 3-week period (October 23 to November 13, 2014).14 The study was approved by the Regional Ethical Committee in Linköping, Sweden. The complete trial protocol is included in Supplement 1.

Eligible participants were students who were daily or weekly smokers and were willing to set a quit date for smoking cessation within the 4 weeks following enrollment.14 They indicated their interest by either responding to an email invitation or by sending an SMS text message to a dedicated telephone number, after which they received an email on how to register for the trial. Participants provided informed consent by clicking on a link in the email invitation; they were then referred to a baseline assessment page. The participants did not receive financial compensation. Follow-up was performed from March 6 to April 17, 2015.14

Invitations to participate were emailed over a 1-week period. Individuals consenting and randomized to the intervention could set a quit date 1 to 4 weeks after being exposed to preparatory content before the 12-week core program (see Interventions subsection below). Quit dates were thus set between 1 and 7 weeks after the initial invitation, and follow-up invitations were emailed 19 weeks after the initial recruitment email. The follow-up assessment was undertaken a mean (SD) of 3.9 (0.37) months after the quit date and thus assessed the short-term effect of the intervention. Because the study used a delayed-intervention (control) group design, ethical reasons limited the delay time.

Randomization and Blinding

After answering the baseline questionnaire and confirming their telephone numbers, participants were immediately randomized to the intervention group or the delayed-intervention group (Figure). Randomization was fully computerized and automated, used no blocks or strata, and allocated each participant a number 1 or 2 with equal probability using Java’s built-in random number generator (java.util.Random; https://docs.oracle.com/javase/7/docs/api/java/util/Random.html). Participants in both groups were aware that they were participating in a trial and that they had been randomized to the intervention or control group. After the 4-month follow-up period, the control group received access to the intervention.

Interventions

We developed the messages in our intervention based on existing evidence-based practice, including components derived from expert guidance and official smoking cessation manuals recommended in Sweden (http://dok.slso.sll.se/CES/FHG/Tobak/Rapporter/srl-behandlingsupplagg-och-rutiner.2013_4.2014.pdf). We included key elements from previous text-based interventions and Internet-based interventions (http://www.viss.nu/Global/Bilagor/Tobaksavvanjning_0709.pdf).13,15-18 The intervention included elements such as making a public declaration about quitting (ie, telling friends about the quit attempt), asking friends and relatives for support, using problem-solving tips and distraction techniques, and the option to text for more help if craving to smoke or smoking (eMethods in Supplement 2).19

The NEXit core program lasts 12 weeks and is preceded by a 1- to 4-week motivational phase during which participants can choose to set a stop date. If the participants did not set a stop date within 4 weeks, they had agreed at the outset to try to stop at this time. The 12-week core program consists of 157 text messages, with the option to request extra messages when having cravings to smoke, relapse, or concerns about weight gain. The participants received 4 to 5 text messages per day in the first week, followed by a decreasing number of messages throughout the 12-week intervention.14 Both the intervention and control groups received text messages every 2 weeks thanking them for participating in the study.

Outcomes

The first primary outcome followed the Russell standard definition of prolonged abstinence since the quit date20 (restricting abstinence to the last 8 weeks of the 12-week intervention, thus allowing a grace period and applying the usual threshold of not smoking >5 cigarettes during the 8-week period). We calculated follow-up duration as the time since the quit date, departing from the Russell standard definition in the absence of biochemical verification since that approach was designed for studies with face-to-fact contact. The timescales for follow-up were constrained by university term dates. The second primary outcome was 4-week point prevalence of not having smoked a single cigarette at the time of follow-up (ie, immediately after the intervention period). This time was chosen to capture delayed effects of the intervention as suggested by the Society for Research on Nicotine and Tobacco21 and used in previous studies.11,15,16,22,23

There were 4 secondary outcomes. These outcomes included (1) self-reported, 7-day point prevalence of smoking abstinence (defined as not smoking any cigarettes in the past 7 days)15,16,22,23; (2) mean number of quit attempts since taking part in the study23; (3) number of uses of other smoking cessation services (eg, prescribed medication including nicotine replacement, counseling, using telephone quitlines, or any other forms of professional help) since the first invitation to the study15; and (4) the number of cigarettes smoked weekly among participants still smoking at the time of follow-up.23

For the 4-month follow-up, a link to an electronic follow-up questionnaire was emailed to all participants. Initially, 2 reminders were sent to nonresponders 1 week apart. To minimize attrition, nonresponders received additional email reminders every other day for 6 days (ie, 3 emails). If still not responding, these participants received a text message every other day for 6 days (ie, 3 texts) with only 2 questions capturing the 2 primary outcome measures. Those still not responding were telephoned a maximum of 10 times per participant, again assessing only the 2 primary outcomes. To aid in follow-up, participants were told that they would be entered into a lottery for 1 of 2 iPads (Apple Corp) after answering the follow-up questionnaire.

Sample Size

Based on previous studies,15,16we expected an absolute difference of 5% in cessation rates between the intervention and control groups (with 10% quitting in the intervention group and 5% in the control group).

To achieve 80% power with a significance level of P ≤ .05 (2-sided) and correction for continuity, a sample size of 474 participants is needed in each group. If there is 30% attrition in the follow-up measurement, the number needed in each group is 677 and the total required sample size is 1354. The allowance for attrition was deliberately conservative. All trial procedures were automated and implemented simultaneously for all participants (ie, intervention and control). A minority of university students in Sweden are smokers, and not all of the smokers are willing to participate in research. On the basis of a pilot study in one university not included in this study, we estimated the total number of invitations needed to recruit the required sample size. The sample size achieved was greater than required, indicating a slightly higher participation rate in the trial compared with the pilot study (0.9% vs 0.7%).

Statistical Analysis

The data analysis conformed to the prespecified statistical analysis plan as published in the trial protocol.14 Following the intention-to-treat analysis strategy, all primary analyses included only participants with follow-up data in their groups as randomized, thus assuming absent data to be missing at random (MAR). However, subsequent sensitivity analyses included all randomized participants to explore different assumptions about the missing data.24

Continuous variables were summarized with descriptive statistics (number and mean [SD] for data with normal distribution or median [interquartile range] for nonnormally distributed data). Frequency counts and percentages of participants within each category were calculated for categorical data.

The binary outcomes of self-reported 8-week prolonged abstinence, 4-week point prevalence of complete smoking cessation, and 7-day point prevalence of smoking abstinence were analyzed by logistic regression, and the results are presented as ORs (95% CIs). The number of quit attempts and number of uses of other smoking cessation services were analyzed by negative binomial regression, and the results are presented as ratios of means (95% CIs). The number of cigarettes smoked weekly was analyzed by logarithmic transformation and linear regression, and the results are presented as the ratio of geometric means (95% CIs).

All regression analyses were adjusted for the following baseline variables: sex, years of smoking, mean number of cigarettes smoked weekly, severity of dependence as measured by the Fagerström Nicotine Dependence Scale,25 and amount of snus used at baseline. Effect modification analyses were performed for the 2 primary outcomes and the following potential effect modifiers that were measured at baseline: sex, mean number of cigarettes smoked weekly, amount of snus used weekly, and severity of dependence as measured by the Fagerström Nicotine Dependence Scale. Each effect modification analysis was performed by comparing adjusted logistic regression models excluding and including the interaction parameter using the likelihood ratio test. All tests were 2-sided with a 5% level of significance. A reviewer requested an additional, nonprespecified, effect modification analysis for frequency of smoking at baseline (daily vs weekly smoking); this analysis included the randomized group, the dichotomous measures of frequency of smoking at baseline and their interaction, and the adjustment variables listed above.

Sensitivity Analysis

A sensitivity analysis explored the effects of departures from the MAR assumption in the main analysis.24,26,27 As suggested by Jackson et al,27 we quantified departures from the MAR assumption by the informative missing OR (IMOR). We assumed that the IMOR was the same in each randomized group and we varied the IMOR over the range 0.5 to 1.0. If 10% of the observed data are on abstinence from cigarettes, this range implies that the abstinence factor is responsible for 5% to 10% of the missing data. We also set the IMOR to be 0 (missing = smoking, the Russell standard).20 Furthermore, we used data on the number of follow-up emails, texts, and telephone calls needed before an individual responded to explore the plausibility of the MAR assumption: first by exploring the association between quitting and the number of follow-up attempts needed, and then by fitting the repeated-attempts model of Jackson et al,26,27 which allowed us to estimate the degree of departure from MAR and to adjust for departure from MAR. Data analysis was performed from April 27 to May 22, 2015; SPSS, version 23 (IBM Corp) and Stata, version 13 (StataCorp) were used to conduct the analyses.

Results

A total of 1590 participants were randomly assigned: 827 (52.0%) to the NEXit intervention group (573 [69.3%] women) and 763 (48.0%) to the control group (522 [68.4%] women) (Figure). A summary of the participants’ characteristics at baseline is given in Table 1. There were no significant differences in any of the sociodemographic characteristics or smoking variables.

Outcome Analyses

The primary outcome analysis was done on a total of 783 (94.7%) randomized participants in the intervention group and 719 (94.2%) in the control group. The number of participants who achieved 8 weeks of prolonged abstinence at the 4-month follow-up (having smoked ≤5 cigarettes during this time) was 203 (25.9%) in the intervention group and 105 (14.6%) in the control group (adjusted OR, 2.05; 95% CI, 1.57-2.67). The 4-week point prevalence of complete smoking cessation occurred in 161 (20.6%) vs 102 (14.2%) participants (adjusted OR, 1.56; 95% CI, 1.19-2.05) (Table 2). No evidence of effect modification was shown for either primary outcome (P values between 0.13 and 0.58).

Secondary outcome data were available only for participants completing the follow-up questionnaire by email, and included 557 (67.4%) of the participants in the intervention group and 429 (56.2%) in the control group (χ21 = 20.86; P < .001). There were 377 (67.7%) smokers remaining in the intervention group and 361 (84.1%) in the control group (Table 2).

Sensitivity Analysis

Sensitivity analyses explored the effects of departures from the MAR assumption on the primary outcomes. Varying the IMOR, or using missing = smoking, changed the estimated ORs by less than 0.01. The repeated-attempts model suggested that the IMOR was 0.3 to 0.6 in the control arm and 0.7 to 1.4 in the intervention arm and gave ORs (95% CIs) of 1.82 (1.39-2.38) for 8 weeks of prolonged abstinence and 1.44 (0.09-1.91) for 4-week point prevalence of complete smoking cessation. Further details are provided in the eTable in Supplement 2.

Post Hoc Analyses

There was no evidence of effect modification of frequency of smoking at baseline (daily vs weekly smoking) on self-reported abstinence (P = .79) or on self-reported, 4-week point prevalence of complete smoking cessation (P = .48). A total of 257 (31.1%) of the participants in the intervention group requested extra messages. Of these, 128 (49.8%) requested 1 extra message; the remainder made use of the function a mean (SD) of 3.4 (2.1) times (range, 2-13 times). Only 1 of the 6 outcomes (number of quit attempts) had a statistically significant difference between those who requested extra messages and those who did not request extra messages (mean [SD] quit attempts, 2.81 [2.87] vs 2.14 [2.29]; P = .01).

Discussion

The NEXit text-message intervention approximately doubled the rate of prolonged abstinence (allowing occasional lapses) at the 4-month follow-up, with a risk difference of 11.3% (number needed to treat, 9). The risk difference was somewhat smaller (6.4%) for the 4-week point prevalence of complete smoking cessation, increasing the number needed to treat to 16.

Sensitivity analyses allowing for different IMOR values for the 2 primary outcomes showed similar results, which remained statistically significant, thus providing strong support for an intervention effect. The intervention group also did better than the control group on all secondary outcomes in the main analyses, with an RR difference of 16.4% and number needed to treat of 6 for the 7-day complete smoking cessation outcome. Accounting for multiplicity of outcomes by applying a Bonferroni correction to the primary outcomes alone, to the secondary outcomes alone, or even to the combined sets made no difference to the study findings, all of which remained statistically significant.

A major strength of this study is the low rate of attrition for the primary outcomes. Large numbers of individuals were randomized using a fully automated system that cannot be subverted and does not involve initial personal contact, as used in previous studies.15,16 Limitations of the study include the reliance on self-reported smoking data in our approach to outcome evaluation. The Society for Research on Nicotine and Tobacco21 recommends that, in population-based studies with limited face-to-face contact, it is neither required nor desirable to use biochemical verification. Recruiting only young adults studying at universities is another limitation, and we do not suggest that the observed effects will be found in nonstudents or in other age groups, even though 2 large previous studies15,16 did not find any significant difference between age groups. Another limitation is that we do not have long-term follow-up outcomes; we hypothesized that if the intervention did not have any short-term effect, it would not have any long-term effect. Thus, the durability of these effects is unclear. We also did not incorporate a trial-based health economic evaluation. There were significant levels of attrition for the secondary outcomes, and this was significantly different (P < .001) between the groups. We thus recommend caution in the interpretation of these secondary outcomes, mindful of the risk of bias.

The structure and content of text-based smoking cessation interventions vary considerably. A recent meta-analysis12 of SMS text message–based interventions for smoking cessation including 13 studies found that, in most cases, less technically complex interventions performed as well as the more technically advanced interventions. The structure and content of the NEXit intervention differs somewhat from most of the previous text-based interventions; for example, the structure of our study was less complex without using personalized messages and it was shorter, lasting for 3 months as opposed to the 6-month program used in most previous trials.12 Despite these differences, our study showed no attenuation in risk difference compared with 2 larger studies15,16 of 6-month intervention that included more technically advanced features, such as personalized messages and peer-to-peer support. The first study15 reported risk differences for not smoking during the last 7 days of 5.9% and during the last 4 weeks of 6.3% compared with 16.9% and 6.4%, respectively, observed in the present study. The second study16 reported a risk difference of 15.3% for not smoking in the last 7 days, after 6 weeks of follow-up.

In a recent meta-analytic review12 of the effectiveness of SMS text-message interventions for smoking cessation, 13 studies yielded an OR (95% CI) of 1.35 (1.23-1.48) for not smoking in the past 7 days at the time of follow-up in the intervention group, with no heterogeneity. This finding compares with an unadjusted OR of 2.53 (1.85-3.47) for our equivalent secondary outcome measure, which is subject to attrition, and with an adjusted OR of 1.56 (1.19-2.05) for the more stringent short-term (4-week) primary outcome unlikely to be affected by attrition.

Considering that we assessed only the short-term effect of the intervention, the ORs for prolonged abstinence, 7-day, and 4-week point prevalence cessation outcomes are comparable to those of traditional smoking cessation interventions. With findings reported as RR (95% CI), these interventions include telephone quitlines (1.37 [1.26-1.50]),28 group behavior counseling (1.98 [1.60-2.46]),29 and individual behavior counseling (1.39 [1.24-1.57]).30

The nationwide proactive recruitment approach used in our study reached young adults who had low rates of use of any forms of smoking cessation support before participating in this trial. The previous number of quit attempts was low, with a median of 3 previous attempts during a median duration of smoking of 8 years. Our results thus add to the growing body of evidence for the reach and effectiveness of text-based smoking cessation interventions. The quit rates compare well with those of traditional smoking cessation interventions and cost less to enlarge; therefore, they should be highly cost-effective. The technical simplicity of the NEXit intervention also indicates that there might not be any need to complicate the structure of text-based interventions with sophisticated pathways that may not provide added benefit and would also increase the cost of developing and delivering such interventions. However, one important remaining issue to demonstrate is the reach of text-based interventions when offered to whole populations, since this study recruited participants to a research study rather than a direct offer of intervention.

Text-messaging interventions could also be used in combination with initial face-to-face contacts within health care settings or via telephone quitlines, facilitating smokers to register for text-based interventions. This combination could change the way staff interact with and use new technology, saving time and potentially increasing the effectiveness of health behavior interventions endeavoring to have an impact at a population level. In addition, analysis of a previous text-based message intervention clearly showed that it was cost-effective.31 Although undertaken in a high-income country, this study is relevant for low- and middle-income countries where approximately 80% of the more than 1 billion smokers worldwide now live and where interventions such as the one evaluated here can be delivered beyond health care settings.32

Conclusions

The present study shows that it is possible to use a proactive recruitment strategy to enroll a large number of young adults in a research study. The effectiveness of this fairly low-technology intervention was comparable with previous, more sophisticated, tailored text-messaging interventions as well as traditional face-to-face-smoking cessation interventions. The results are promising not only for smoking cessation but also for other areas of disease management and health interventions because, by using a relatively simple technology, the development costs can be low and sophisticated monitoring of technical aspects is not required.

Back to top
Article Information

Corresponding Author: Preben Bendtsen, PhD, Department of Medical Specialist and Department of Medical and Health Sciences, Campus US, SE 581 83, Linköping University, Motala, Sweden (preben.bendtsen@liu.se).

Accepted for Publication: December 13, 2014.

Published Online: February 22, 2016. doi:10.1001/jamainternmed.2015.8260.

Author Contributions: Dr Müssener had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Müssener, M. Bendtsen, McCambridge, P. Bendtsen.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Müssener, P. Bendtsen.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: M. Bendtsen, Karlsson, White, McCambridge.

Obtained funding: P. Bendtsen.

Administrative, technical, or material support: Müssener, M. Bendtsen, P. Bendtsen.

Study supervision: Müssener, P. Bendtsen.

Conflict of Interest Disclosures: Mr M. Bendtsen and Dr P. Bendtsen own a private company that develops and distributes evidence-based lifestyle interventions to be used in health care settings. No other disclosures were reported.

Funding/Support: The study was funded by grant 521-2012-2865 from the Swedish Research Council. Dr White was supported by Unit Programme No. U105260558 from the Swedish Medical Research Council.

Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Gilljam  H.  Quitting smoking brings quick health benefits [in Swedish].  Lakartidningen. 2012;109:554-557.PubMedGoogle Scholar
2.
Sussman  S, Sun  P.  Youth tobacco use cessation: 2008 update.  Tob Induc Dis. 2009;5:3.PubMedGoogle ScholarCrossref
3.
Lim  SS, Vos  T, Flaxman  AD,  et al.  A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010.  Lancet. 2012;380(9859):2224-2260.PubMedGoogle ScholarCrossref
4.
Agardh  E, Moradi  T, Allebeck  P.  The contribution of risk factors to the burden of disease in Sweden: a comparison between Swedish and WHO data [in Swedish].  Lakartidningen. 2008;105(11):816-821.PubMedGoogle Scholar
5.
Shiffman  S, Brockwell  SE, Pillitteri  JL, Gitchell  JG.  Use of smoking-cessation treatments in the United States.  Am J Prev Med. 2008;34(2):102-111.PubMedGoogle ScholarCrossref
6.
Gollust  SE, Schroeder  SA, Warner  KE.  Helping smokers quit: understanding the barriers to utilization of smoking cessation services.  Milbank Q. 2008;86(4):601-627.PubMedGoogle ScholarCrossref
7.
Stanton  A, Grimshaw  G.  Tobacco cessation interventions for young people.  Cochrane Database Syst Rev. 2013;8:CD003289.PubMedGoogle Scholar
8.
Backinger  CL, Fagan  P, Matthews  E, Grana  R.  Adolescent and young adult tobacco prevention and cessation: current status and future directions.  Tob Control. 2003;12(suppl 4):IV46-IV53.PubMedGoogle ScholarCrossref
9.
Mason  M, Ola  B, Zaharakis  N, Zhang  J.  Text messaging interventions for adolescent and young adult substance use: a meta-analysis.  Prev Sci. 2015;16(2):181-188.PubMedGoogle ScholarCrossref
10.
OECD: Better Policies for Better Lives. Data lab: health data. http://www.oecd.org/statistics/datalab/health.htm. Published 2015. Accessed June 12, 2015.
11.
Free  C, Phillips  G, Watson  L,  et al.  The effectiveness of mobile-health technologies to improve health care service delivery processes: a systematic review and meta-analysis.  PLoS Med. 2013;10(1):e1001363.PubMedGoogle ScholarCrossref
12.
Spohr  SA, Nandy  R, Gandhiraj  D, Vemulapalli  A, Anne  S, Walters  ST.  Efficacy of SMS text message interventions for smoking cessation: a meta-analysis.  J Subst Abuse Treat. 2015;56:1-10.PubMedGoogle ScholarCrossref
13.
Whittaker  R, McRobbie  H, Bullen  C, Borland  R, Rodgers  A, Gu  Y.  Mobile phone-based interventions for smoking cessation.  Cochrane Database Syst Rev. 2012;11:CD006611.PubMedGoogle Scholar
14.
Müssener  U, Bendtsen  M, Karlsson  N, White  IR, McCambridge  J, Bendtsen  P.  SMS-based smoking cessation intervention among university students: study protocol for a randomised controlled trial (NEXit trial).  Trials. 2015;16:140.PubMedGoogle ScholarCrossref
15.
Free  C, Knight  R, Robertson  S,  et al.  Smoking cessation support delivered via mobile phone text messaging (txt2stop): a single-blind, randomised trial.  Lancet. 2011;378(9785):49-55.PubMedGoogle ScholarCrossref
16.
Rodgers  A, Corbett  T, Bramley  D,  et al.  Do u smoke after txt? results of a randomised trial of smoking cessation using mobile phone text messaging.  Tob Control. 2005;14(4):255-261.PubMedGoogle ScholarCrossref
17.
Ybarra  ML, Holtrop  JS, Prescott  TL, Rahbar  MH, Strong  D.  Pilot RCT results of stop my smoking USA: a text messaging-based smoking cessation program for young adults.  Nicotine Tob Res. 2013;15(8):1388-1399.PubMedGoogle ScholarCrossref
18.
Bock  B, Heron  K, Jennings  E,  et al.  A text message delivered smoking cessation intervention: the initial trial of TXT-2-Quit: randomized controlled trial.  JMIR Mhealth Uhealth. 2013;1(2):e17.PubMedGoogle ScholarCrossref
19.
Michie  S, Hyder  N, Walia  A, West  R.  Development of a taxonomy of behaviour change techniques used in individual behavioural support for smoking cessation.  Addict Behav. 2011;36(4):315-319.PubMedGoogle ScholarCrossref
20.
West  R, Hajek  P, Stead  L, Stapleton  J.  Outcome criteria in smoking cessation trials: proposal for a common standard.  Addiction. 2005;100(3):299-303.PubMedGoogle ScholarCrossref
21.
SRNT Subcommittee on Biochemical Verification.  Biochemical verification of tobacco use and cessation.  Nicotine Tob Res. 2002;4(2):149-159.PubMedGoogle ScholarCrossref
22.
Hughes  JR, Keely  JP, Niaura  RS, Ossip-Klein  DJ, Richmond  RL, Swan  GE.  Measures of abstinence in clinical trials: issues and recommendations.  Nicotine Tob Res. 2003;5(1):13-25.PubMedGoogle ScholarCrossref
23.
Haug  S, Schaub  MP, Venzin  V, Meyer  C, John  U.  Efficacy of a text message-based smoking cessation intervention for young people: a cluster randomized controlled trial.  J Med Internet Res. 2013;15(8):e171.PubMedGoogle ScholarCrossref
24.
White  IR, Horton  NJ, Carpenter  J, Pocock  SJ.  Strategy for intention to treat analysis in randomised trials with missing outcome data.  BMJ. 2011;342:d40.PubMedGoogle ScholarCrossref
25.
Heatherton  TF, Kozlowski  LT, Frecker  RC, Fagerström  KO.  The Fagerström test for nicotine dependence: a revision of the Fagerström Tolerance Questionnaire.  Br J Addict. 1991;86(9):1119-1127.PubMedGoogle ScholarCrossref
26.
Jackson  D, White  IR, Leese  M.  How much can we learn about missing data? an exploration of a clinical trial in psychiatry.  J R Stat Soc Ser A Stat Soc. 2010;173(3):593-612.PubMedGoogle ScholarCrossref
27.
Jackson  D, White  IR, Mason  D, Sutton  S.  A general method for handling missing binary outcome data in randomized controlled trials.  Addiction. 2014;109(12):1986-1993.PubMedGoogle ScholarCrossref
28.
Stead  LF, Hartmann-Boyce  J, Perera  R, Lancaster  T.  Telephone counselling for smoking cessation.  Cochrane Database Syst Rev. 2013;8:CD002850.PubMedGoogle Scholar
29.
Stead  LF, Lancaster  T.  Group behaviour therapy programmes for smoking cessation.  Cochrane Database Syst Rev. 2005;2(2):CD001007.PubMedGoogle Scholar
30.
Lancaster  T, Stead  LF.  Individual behavioural counselling for smoking cessation.  Cochrane Database Syst Rev. 2005;(2):CD001292.PubMedGoogle Scholar
31.
Guerriero  C, Cairns  J, Roberts  I, Rodgers  A, Whittaker  R, Free  C.  The cost-effectiveness of smoking cessation support delivered by mobile phone text messaging: Txt2stop.  Eur J Health Econ. 2013;14(5):789-797.PubMedGoogle ScholarCrossref
32.
World Health Organization. Tobacco. http://www.who.int/mediacentre/factsheets/fs339/en/. Updated July 6, 2015. Accessed January 21, 2016.
×