Advertisement

Efficacy of SMS Text Message Interventions for Smoking Cessation: A Meta-Analysis

Published:February 02, 2015DOI:https://doi.org/10.1016/j.jsat.2015.01.011

      Highlights

      • Smoking quit rates for the intervention groups were 35% higher than control groups.
      • Pooled results were consistent at 3- and 6-month follow-up lengths.
      • Texting alone was as effective as texting plus additional intervention modalities.
      • Messaging frequency can impact intervention effectiveness.
      • Intervention assessment messages are unobtrusive and do not impact efficacy.

      Abstract

      Background

      Mobile technology provides new opportunities for health promotion communication. The purpose of this study was to conduct a current and extensive meta-analytic review of SMS (short message service) text message-based interventions for individual smoking cessation.

      Methods

      Academic Search Complete, PsycINFO, PubMed, and Scopus were reviewed for articles meeting selection criteria: 1) randomized controlled trials, 2) measured smoking cessation, and 3) intervention primarily delivered through SMS text messaging. Three and 6 month follow-up of 7-day point prevalence or continuous abstinence was considered from studies meeting criteria. All analyses were conducted with intention-to-treat. Both fixed and random effects models were used to calculate the global outcome measure and confidence intervals.

      Results

      Thirteen studies were identified that met inclusion criteria. The studies were found to be homogeneous [Q12 = 12.47, p = 0.14]. Odds ratios based on the random effects models suggested that interventions generally increased quit rates compared to controls, 1.36 [95% CI = 1.23, 1.51]. Intervention efficacy was higher in studies with a 3 month follow-up compared to 6 month follow-up. Text plus programs (e.g., text messaging plus Web or in-person intervention modalities) performed only slightly better than text only programs. Pooled results also indicate message frequency schedule can affect quit rates, in which fixed schedules performed better than decreasing or variable schedules. The use of quit status assessment messages was not related to intervention efficacy.

      Conclusion

      Smoking quit rates for the text messaging intervention group were 36% higher compared to the control group quit rates. Results suggest that SMS text messaging may be a promising way to improve smoking cessation outcomes. This is significant given the relatively wide reach and low cost of text message interventions. Identifying the components that make interventions efficacious will help to increase the effectiveness of such interventions.

      Keywords

      1. Introduction

      Mobile phone technology for health promotion and disease prevention is a rapidly growing area of research. The development of mobile phone interventions (e.g., short-message service [SMS], multimedia message service [MMS], Internet, applications) is increasing with the widespread acceptance of cell phones. In 2014, an estimated 4.55 billion people worldwide will use a mobile phone, and 1.75 billion will use a smartphone (). The growth and acceptance of mobile communication provides researchers with opportunities for delivering innovative health behavior change interventions.
      Mobile phones have been utilized in numerous capacities in health research. They can be used to collect data or help people self-monitor (e.g., diet/exercise tracking), provide behavioral reminders (e.g., health service appointments, medication compliance), deliver medical test results, serve as boosters for in-person or Web applications, or as stand-alone behavior change programs (e.g., smoking cessation, decreasing alcohol consumption) (
      • Cadmus-Bertram L.
      • Wang J.B.
      • Patterson R.E.
      • Newman V.A.
      • Parker B.A.
      • Pierce J.P.
      Web-based self-monitoring for weight loss among overweight/obese women at increased risk for breast cancer: The HELP pilot study.
      ,
      • Foreman K.F.
      • Stockl K.M.
      • Le L.B.
      • Fisk E.
      • Shah S.M.
      • Lew H.C.
      • et al.
      Impact of a text messaging pilot program on patient medication adherence.
      ,
      • Free C.
      • Knight R.
      • Robertson S.
      • Whittakers R.
      • Edwards P.
      • Zhou W.
      • et al.
      Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
      ,
      • Walters S.T.
      • Ondersma S.J.
      • Ingersoll K.S.
      • Rodriguez M.
      • Lerch J.M.A.
      • Rossheim M.E.
      • et al.
      MAPIT: Development of a web-based intervention targeting substance abuse treatment in the criminal justice system.
      ). Mobile technology also provides researchers with the versatility to target multiple behaviors for health promotion and disease prevention.
      There are numerous benefits of mobile technologies for both researchers and mobile phone users who are interested in improving their health. For instance, mobile technology allows for direct interaction between practitioners and clients without the need for face-to-face contact (
      • Fjeldsoe B.S.
      • Marhsall A.L.
      • Miller Y.D.
      Behavior change interventions delivered by mobile telephone short-message service.
      ). Researchers benefit from the use of mobile phone interventions because of their capability to collect and process large amounts of data efficiently, tailor messages based on user characteristics, and send time-sensitive information (
      • Fjeldsoe B.S.
      • Marhsall A.L.
      • Miller Y.D.
      Behavior change interventions delivered by mobile telephone short-message service.
      ,
      • Whittaker R.
      • McRobbie H.
      • Bullen C.
      • Borland R.
      • Rodgers A.
      • Gu Y.
      Mobile phone-based interventions for smoking cessation.
      ). Mobile technology is also cost effective and easily scalable to large populations (
      • Irvine L.
      • Falconer D.W.
      • Jones C.
      • Ricketts I.W.
      • Williams B.
      • Crombie I.
      Can text messages reach the parts other process measures cannot reach: An evaluation of a behavior change intervention delivered by mobile phone?.
      ,
      • Whittaker R.
      • McRobbie H.
      • Bullen C.
      • Borland R.
      • Rodgers A.
      • Gu Y.
      Mobile phone-based interventions for smoking cessation.
      ). Mobile phone users benefit from the use of health technologies primarily because they are convenient and easy to use. Mobile technology can easily provide health information, distractions, triggers, and social support for behavior change. In addition, mobile communication provides a certain amount of anonymity and unobtrusive assessment regarding sensitive behaviors/attitudes (
      • Fogg B.J.
      Persuasive technology: Using computers to change what we think and do.
      ,
      • Irvine L.
      • Falconer D.W.
      • Jones C.
      • Ricketts I.W.
      • Williams B.
      • Crombie I.
      Can text messages reach the parts other process measures cannot reach: An evaluation of a behavior change intervention delivered by mobile phone?.
      ).
      Behavioral interventions delivered by mobile phone have targeted a broad array of health behaviors (e.g., diet, weight loss, smoking cessation, medication adherence, diabetes management) (
      • Cole-Lewis H.
      • Kershaw T.
      Text messaging as a tool for behavior change in disease prevention and management.
      ). However, health promotion interventions delivered specifically through SMS text messaging have mostly targeted smoking cessation. For instance,
      • Whittaker R.
      • McRobbie H.
      • Bullen C.
      • Borland R.
      • Rodgers A.
      • Gu Y.
      Mobile phone-based interventions for smoking cessation.
      conducted a meta-analytic review of mobile phone interventions (i.e., using any function or application sent via mobile phone) for smoking cessation. The review included five studies (N = 9,109) with 6 month follow-up data. Pooled results indicated that mobile phone interventions increased quit rates by 71% as compared to controls (RR = 1.71, 95% CI 1.47, 1.99). However, the authors indicated that future research is needed to identify specific intervention moderators that are associated with improved user outcomes. Another meta-analytic review conducted by
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      reviewed 19 SMS text message interventions (N = 5,137) for a variety of health promotion areas, including smoking cessation, diet, and weight loss. Pooled results of smoking cessation programs indicated that text messaging interventions were moderately effective, d = .447, at increasing quitting compared to control groups. Text message interventions for health promotion were generally more effective when they were tailored, personalized, and/or used a decreasing schedule of message frequency. These results suggest that it may be important to consider intervention moderators when delivering services in an efficient and cost-effective manner. Overall, these meta-analyses suggest that mobile phone interventions can be an effective option for individuals who are seeking to quit smoking; however additional research needs to be conducted to identify the most efficacious intervention moderators or components that lead to improved outcomes for smokers.

      1.1 Intervention moderators

      Despite the growing body of literature on text message-based interventions for smoking cessation, relatively little is known about which intervention moderators are most important in helping people quit smoking (
      • Ybarra M.L.
      • Summers J.
      • Bagci A.T.
      • Emri S.
      Design considerations in developing a text messaging program aimed at smoking cessation.
      ). Drawing from the broader literature, there are several possible moderators of text message interventions, including variations in content, scheduling, and the availability of other on-demand features. Each of these moderators may contribute to intervention efficacy (
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      ,
      • Ybarra M.L.
      • Summers J.
      • Bagci A.T.
      • Emri S.
      Design considerations in developing a text messaging program aimed at smoking cessation.
      ).
      Smoking cessation programs vary by intervention type (i.e., SMS only or SMS plus). Some programs only include messages sent via SMS, while others combine SMS with in-person or Web support programs. For example,
      • Free C.
      • Knight R.
      • Robertson S.
      • Whittakers R.
      • Edwards P.
      • Zhou W.
      • et al.
      Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
      utilized an SMS text message only program, while
      • Bock B.
      • Heron K.
      • Jennings E.
      • Morrow K.
      • Cobb V.
      • Magee J.
      • et al.
      A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
      gave smokers a 30-minute individual smoking cessation counseling session prior to randomizing participants to receive the text message intervention or non-smoking related messages.
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      found no difference in intervention efficacy between SMS only and SMS plus programs for health promotion behaviors. However, intervention type should be assessed individually for different behaviors that may require different levels of support for change to occur.
      Smoking cessation programs vary by message frequency. For instance, many interventions adapt message frequency based on the quitting curve (i.e., high intensity messages at quit attempts, followed by a gradual reduction) (
      • Free C.
      • Knight R.
      • Robertson S.
      • Whittakers R.
      • Edwards P.
      • Zhou W.
      • et al.
      Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
      ,
      • Naughton F.
      • Prevost T.
      • Gilbert H.
      • Sutton S.
      Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
      ,
      • Rodgers A.
      • Corbett T.
      • Bramley D.
      • Riddell T.
      • Wills M.
      • Lin R.B.
      • et al.
      Do u smoke after txt? Results of a randomized trial of smoking cessation using mobile phone text messaging.
      ). For example,
      • Naughton F.
      • Prevost T.
      • Gilbert H.
      • Sutton S.
      Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
      utilized a decreasing message frequency schedule with the highest messages occurring around the person's quit date. Other programs utilized a fixed message schedule which delivered a relatively consistent number of messages per day across the intervention duration (
      • Haug S.
      • Schaub M.
      • Venzin V.
      • Meyer C.
      • John U.
      • Eysenbach G.
      Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
      ,
      • Naughton F.
      • Jamison J.
      • Boase S.
      • Sloan M.
      • Gilbert H.
      • Prevost A.T.
      • et al.
      Randomized controlled trial to assess the short-term effectiveness of tailored web- and text-based facilitation of smoking cessation in primary care (iQuit in Practice).
      ,
      • Shi H.
      • Jiang X.
      • Yu C.
      • Zhang Y.
      Use of mobile phone text messaging to deliver an individualized smoking behaviour intervention in Chinese adolescents.
      ). Still other interventions use a dynamic message track depending on which stage the client is in currently (
      • Bock B.
      • Heron K.
      • Jennings E.
      • Morrow K.
      • Cobb V.
      • Magee J.
      • et al.
      A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
      ,
      • Borland R.
      • Balmford J.
      • Benda P.
      Population-level effects of automated smoking cessation help programs: A randomized controlled trial.
      ,
      • Haug S.
      • Schaub M.
      • Venzin V.
      • Meyer C.
      • John U.
      • Eysenbach G.
      Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
      ). In sum, message frequency should be assessed to determine which schedule is most appropriate for smokers who want to quit.
      Smoking cessation interventions vary by message direction and initiation. Some programs use only unidirectional messaging initiated by the researcher while others use bidirectional messaging to assess real-time quit status from smokers. This serves as a data collection tool that can be used to provide tailored messages based on a person's current stage of change (e.g., contemplation, action, maintenance) (
      • Prochaska J.O.
      • DiClemente C.C.
      Transtheoretical therapy: Toward a more integrative model of change.
      ) or other factor. Individualized message tracks (i.e., tailoring message content based on smoking status or stage of change) are a particular strength of technology-based interventions; therefore, the majority of smoking cessation interventions utilize bidirectional messaging. For example,
      • Haug S.
      • Schaub M.
      • Venzin V.
      • Meyer C.
      • John U.
      • Eysenbach G.
      Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
      utilized weekly assessment messages such as “Have you recently smoked cigarettes?” and “How many cigarettes did you smoke this week?” as a way to tailor messages and gauge participant progress. Smoking cessation programs also vary based on who can initiate messages. Researcher-initiated messages often contain program intervention messages and assessment questions. User-initiated messages might contain requests for additional support or services (e.g., crave functions). Thus, it is important to assess message direction and initiation to determine their impact on overall intervention efficacy.
      Smoking cessation interventions also vary depending on the format of message tracks. Fixed message tracks have one track for participants to follow; users cannot change the course of the intervention. Dynamic message tracks can change messages based on user assessments such as quit status or stage of change. For example,
      • Free C.
      • Knight R.
      • Robertson S.
      • Whittakers R.
      • Edwards P.
      • Zhou W.
      • et al.
      Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
      utilized a fixed intervention track while
      • Naughton F.
      • Prevost T.
      • Gilbert H.
      • Sutton S.
      Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
      provided different message tracks to those in the smoking or non-smoking groups.
      • Bock B.
      • Heron K.
      • Jennings E.
      • Morrow K.
      • Cobb V.
      • Magee J.
      • et al.
      A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
      created four message tracks based on the participant's stage of change, “not ready”, “prepare”, “quit”, and “relapse”. Thus, the use of fixed vs. dynamic message tracks may be an important predictor of intervention efficacy as dynamic tracks inherently provide more tailoring than fixed tracks.
      Smoking cessation programs vary on the extent of message tailoring. Message tailoring utilizes personal characteristics (e.g., stage of change, coping, self-efficacy) to customize message content to a specific individual (
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      ,
      • Petty R.E.
      • Barden J.
      • Wheeler S.C.
      The elaboration likelihood model of persuasion: Developing health promotions for sustained behavioral change.
      ). It is generally accepted that message tailoring increases the relevance and salience of message content (
      • Petty R.E.
      • Barden J.
      • Wheeler S.C.
      The elaboration likelihood model of persuasion: Developing health promotions for sustained behavioral change.
      ). For instance,
      • Ybarra M.
      • Bagci A.T.
      • Korchmaros J.
      • Emri S.
      A text messaging-based smoking cessation program for adult smokers: Randomized controlled trial.
      created different content paths based on the participant's quit status. Likewise,
      • Haug S.
      • Schaub M.
      • Venzin V.
      • Meyer C.
      • John U.
      • Eysenbach G.
      Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
      tailored messages to participant demographic and smoking-related variables. Relatedly, targeted messages are customized based on shared characteristics of a population subgroup, such as gender, ethnicity, or location but not specific to one individual (
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      ,
      • Petty R.E.
      • Barden J.
      • Wheeler S.C.
      The elaboration likelihood model of persuasion: Developing health promotions for sustained behavioral change.
      ). Message targeting increases the relevance of messages when content is specific to a certain population. For example,
      • Whittaker R.
      • Dorey E.
      • Bramley D.
      • Bullen C.
      • Elley R.C.
      • Maddison R.
      • et al.
      A theory-based video messaging mobile phone intervention for smoking cessation: Randomized control trial.
      provided indigenous Maori smokers with population-relevant messages, although the messages were not tailored to individuals. Thus, the extent of message tailoring is important to assess for program efficacy.
      Finally, some smoking cessation SMS text message interventions provide on-demand messaging that allows participants to text a keyword (e.g., crave, help) in emergency situations to receive additional support (
      • Bock B.
      • Heron K.
      • Jennings E.
      • Morrow K.
      • Cobb V.
      • Magee J.
      • et al.
      A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
      ,
      • Borland R.
      • Balmford J.
      • Benda P.
      Population-level effects of automated smoking cessation help programs: A randomized controlled trial.
      ,
      • Haug S.
      • Schaub M.
      • Venzin V.
      • Meyer C.
      • John U.
      • Eysenbach G.
      Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
      ,
      • Naughton F.
      • Prevost T.
      • Gilbert H.
      • Sutton S.
      Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
      ,
      • Ybarra M.L.
      • Summers Holtrop J.
      • Prescott T.L.
      • Rahbar M.H.
      • Strong D.
      Pilot RCT results of Stop My Smoking USA: A text messaging-based smoking cessation program for young adults.
      ). Participants receive immediate motivational support messages to help them cope with the craving or relapse. Some text message interventions also allow participants to connect with others for social support and encouragement (
      • Bock B.
      • Heron K.
      • Jennings E.
      • Morrow K.
      • Cobb V.
      • Magee J.
      • et al.
      A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
      ,
      • Free C.
      • Whittaker R.
      • Knight R.
      • Abramsky T.
      • Rodgers A.
      • Roberts I.G.
      Txt2stop: A pilot randomised controlled trial of mobile phone-based smoking cessation support.
      ,
      • Free C.
      • Knight R.
      • Robertson S.
      • Whittakers R.
      • Edwards P.
      • Zhou W.
      • et al.
      Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
      ,
      • Rodgers A.
      • Corbett T.
      • Bramley D.
      • Riddell T.
      • Wills M.
      • Lin R.B.
      • et al.
      Do u smoke after txt? Results of a randomized trial of smoking cessation using mobile phone text messaging.
      ,
      • Ybarra M.L.
      • Summers Holtrop J.
      • Prescott T.L.
      • Rahbar M.H.
      • Strong D.
      Pilot RCT results of Stop My Smoking USA: A text messaging-based smoking cessation program for young adults.
      ). For example,
      • Rodgers A.
      • Corbett T.
      • Bramley D.
      • Riddell T.
      • Wills M.
      • Lin R.B.
      • et al.
      Do u smoke after txt? Results of a randomized trial of smoking cessation using mobile phone text messaging.
      designed a “quit buddy” component that connected participants with other people who had similar characteristics and quit days.
      In sum, text message interventions provide numerous benefits to researchers and participants. Previous review articles and meta-analyses suggest that text message interventions can be an effective behavioral change intervention for a variety of behaviors, and in particular smoking cessation. Yet relatively little is known about which program components are most useful in an SMS text message intervention for smoking cessation. Given the accelerated advancements in technology, it is important to assess the efficacy of text message interventions for smoking cessation and their intervention components as technology changes and opportunities grow. This study provides the most extensive review of text message interventions for smoking cessation to date. This meta-analysis extends previous reviews with broad inclusion criteria to capture additional relevant studies compared to previous reviews that used more stringent inclusion criteria. We also extend the literature by examining the impact of intervention moderators specific to smoking cessation programs. This meta-analysis assesses the efficacy of smoking cessation text message interventions, with specific attention to intervention moderators that may affect quit rates.

      2. Materials and methods

      2.1 Search strategy

      The search strategy consisted of three steps. First, a comprehensive search of electronic article databases was conducted; databases included Academic Search Complete, PsycINFO, Health Source, Psychology and Behavioral Sciences Collection, PubMed, Scopus, and Web of Knowledge. The key search words utilized were smok* (e.g., smoke, smoking, smoker), tobacco, and SMS, text messaging, texting, mhealth, phone, mobile, or text messag*, and randomized/randomised controlled trial, controlled trial, or RCT. No date limit was included on the electronic search; all articles identified by October 2014 were considered for inclusion. Second, a review of the reference sections of previous reviews (
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      ,
      • Whittaker R.
      • McRobbie H.
      • Bullen C.
      • Borland R.
      • Rodgers A.
      • Gu Y.
      Mobile phone-based interventions for smoking cessation.
      ) and included articles was conducted to identify articles that might have been missed in the database search. Third, a review of the grey literature was conducted searching several Web sites and publically available databases, RAND Publications, the Grey Literature Report of the New York Academy of Medicine, The Community Guide, the National Association of County and City Health Officials (NACCHO) Model Practices Database, and WorldCAT Dissertations and Theses.
      Studies were included in the meta-analytic review if they met the following conditions: 1) targeted smoking cessation; 2) randomized participants to intervention and control/comparison groups; 3) delivered the main intervention primarily through text messaging; 4) included a follow-up measure of smoking abstinence, and 5) published in English in a scientific peer-reviewed journal. This search strategy resulted in 377 unique articles. The first and fourth authors reviewed all studies for inclusion criteria separately. The first round of coding resulted in a reliability of α = 0.86. Authors then met to resolve any discrepancies until total agreement was reached. After careful examination of article titles and abstracts, 36 articles were downloaded and fully reviewed for inclusionary criteria. The resulting review revealed 13 articles that met inclusionary criteria.
      Articles were discarded for various reasons: One hundred thirty-two (35.0%) articles did not target smoking cessation; 124 (32.9%) articles did not utilize a text message-based intervention for smoking cessation; 45 (11.9%) articles were systematic reviews or meta-analyses; 28 (7.9%) articles did not include randomization and/or a control group; 14 (3.7%) articles were duplicate trial data reported elsewhere; 10 (2.7%) articles were smoking cessation interventions not delivered primarily through text messaging or were multi-media interventions in which the effect of text messaging could not be isolated; 9 (6.3%) articles were commentaries or theoretical articles; 1 (0.3%) article provided smoking cessation messages to both the experimental and control groups; and 1 (0.3%) article did not report follow-up smoking abstinence rates.

      2.2 Article coding

      The first author extracted information about study population characteristics (e.g., age, gender) and location. Potential intervention moderators were coded by the first and fifth authors separately. The first round of coding resulted in a reliability of α = 0.87. Authors then met to clarify intervention moderator definitions and discussed discrepancies until total agreement was reached. Potential moderators included intervention type (i.e., text only or text plus additional intervention), message frequency (i.e., decreasing schedule, fixed schedule, and variable schedule), message direction and initiation, assessment of quit status messages, message track (i.e., fixed or dynamic), message content tailoring (i.e., targeted, tailored), availability of on-demand messages (i.e., help or crave functions for immediate messages), inclusion of social or peer-to-peer support functions, and the provision of nicotine replacement therapy.

      2.3 Outcome measure

      Smoking cessation studies typically utilize outcome measures such as 7-day point prevalence, continuous 6-month abstinence, and biologically verified saliva cotinine levels. In order to maintain consistency in the effect size calculations, 7-day point prevalence was selected as the primary outcome measure as 11 out of 13 studies reported these results. Two studies only reported continuous 6-month abstinence. Seven-day point prevalence abstinence was collected at multiple follow-up periods for many of the studies, for instance at 3 and 6 months. The primary model and moderator analysis included the longest follow-up period available. Four studies reported 3-month follow-up data (
      • Naughton F.
      • Prevost T.
      • Gilbert H.
      • Sutton S.
      Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
      ,
      • Shi H.
      • Jiang X.
      • Yu C.
      • Zhang Y.
      Use of mobile phone text messaging to deliver an individualized smoking behaviour intervention in Chinese adolescents.
      ,
      • Ybarra M.
      • Bagci A.T.
      • Korchmaros J.
      • Emri S.
      A text messaging-based smoking cessation program for adult smokers: Randomized controlled trial.
      ,
      • Ybarra M.L.
      • Summers Holtrop J.
      • Prescott T.L.
      • Rahbar M.H.
      • Strong D.
      Pilot RCT results of Stop My Smoking USA: A text messaging-based smoking cessation program for young adults.
      ) and nine studies reported 6-month follow-up data (
      • Abroms L.C.
      • Boal A.L.
      • Simmens S.J.
      • Mendel J.A.
      • Windsor R.A.
      A randomized trial of Text2Quit: A text messaging program for smoking cessation.
      ,
      • Bock B.
      • Heron K.
      • Jennings E.
      • Morrow K.
      • Cobb V.
      • Magee J.
      • et al.
      A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
      ,
      • Borland R.
      • Balmford J.
      • Benda P.
      Population-level effects of automated smoking cessation help programs: A randomized controlled trial.
      ,
      • Free C.
      • Whittaker R.
      • Knight R.
      • Abramsky T.
      • Rodgers A.
      • Roberts I.G.
      Txt2stop: A pilot randomised controlled trial of mobile phone-based smoking cessation support.
      ,
      • Free C.
      • Knight R.
      • Robertson S.
      • Whittakers R.
      • Edwards P.
      • Zhou W.
      • et al.
      Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
      ,
      • Haug S.
      • Schaub M.
      • Venzin V.
      • Meyer C.
      • John U.
      • Eysenbach G.
      Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
      ,
      • Naughton F.
      • Jamison J.
      • Boase S.
      • Sloan M.
      • Gilbert H.
      • Prevost A.T.
      • et al.
      Randomized controlled trial to assess the short-term effectiveness of tailored web- and text-based facilitation of smoking cessation in primary care (iQuit in Practice).
      ,
      • Rodgers A.
      • Corbett T.
      • Bramley D.
      • Riddell T.
      • Wills M.
      • Lin R.B.
      • et al.
      Do u smoke after txt? Results of a randomized trial of smoking cessation using mobile phone text messaging.
      ,
      • Whittaker R.
      • Dorey E.
      • Bramley D.
      • Bullen C.
      • Elley R.C.
      • Maddison R.
      • et al.
      A theory-based video messaging mobile phone intervention for smoking cessation: Randomized control trial.
      ).

      2.4 Effect size extraction and calculation

      Odds ratios (OR) and log-OR with 95% confidence intervals were calculated to determine the efficacy of text message interventions for smoking cessation, as well as potential intervention moderators. Effect sizes were calculated with intention-to-treat analyses. When more than one intervention existed in the same study, the most potent text messaging intervention was utilized for analyses. Effect sizes were also calculated for each reported follow-up period to determine intervention efficacy over time.

      2.5 Meta-analytic approach

      A fixed effects model with inverse-variance weighting scheme was used to obtain an overall effect size. In this approach, it is assumed that all studies share a true effect size with minimal variation. A random effects model was also used, incorporating both between-study and within-study variability. Compared to fixed effects models, confidence intervals for random effects models tend to be wider, making them a more conservative estimate. The criterion for utilizing a fixed or random effects model can be determined based on the test of heterogeneity. If the heterogeneity between studies is not statistically significant, then the fixed effects model is a statistically valid model, but the conclusions from the meta-analysis should not be extended beyond the set of studies in consideration. Therefore, both fixed and random effects models are presented given that either of the two models is appropriate in this analysis. A publication bias analysis was conducted by generating and evaluating the symmetry of included studies in a funnel plot. A statistical test for publication bias was also performed as described by
      • Egger M.
      • Davey Smith G.
      • Schneider M.
      • Minder C.
      Bias in meta-analysis detected by a simple, graphical test.
      . The Q statistic was calculated to determine the statistical significance of heterogeneity between studies. Natural log odds ratio was utilized to calculate the QB statistic for determining the statistical difference between effect sizes of intervention moderators. All analyses were conducted using packages “RMeta” and “Meta” in R-Studio with R version 3.0.1.

      3. Results

      3.1 Characteristics of individual studies

      The search strategy resulted in 13 articles for inclusion in the meta-analysis. Text message interventions for smoking cessation were conducted in seven countries: four in the United Kingdom, three in the United States, two in New Zealand, and one each in Australia, China, Switzerland, and Turkey. The 13 articles resulted in a cumulative sample of N = 13,626. Participants were primarily adult smokers interested in quitting, six studies only recruited participants 18 and over, four studies recruited participants 15 and over, and three studies targeted adolescents and young adults (range 16–25). Mean participant ages ranged from 15 to 93 years with a mean weighted average of 35.1 years. Gender was evenly distributed among most studies; however, five studies reported disproportionate rates (
      • Abroms L.C.
      • Boal A.L.
      • Simmens S.J.
      • Mendel J.A.
      • Windsor R.A.
      A randomized trial of Text2Quit: A text messaging program for smoking cessation.
      ,
      • Borland R.
      • Balmford J.
      • Benda P.
      Population-level effects of automated smoking cessation help programs: A randomized controlled trial.
      ,
      • Free C.
      • Whittaker R.
      • Knight R.
      • Abramsky T.
      • Rodgers A.
      • Roberts I.G.
      Txt2stop: A pilot randomised controlled trial of mobile phone-based smoking cessation support.
      ,
      • Naughton F.
      • Prevost T.
      • Gilbert H.
      • Sutton S.
      Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
      ,
      • Shi H.
      • Jiang X.
      • Yu C.
      • Zhang Y.
      Use of mobile phone text messaging to deliver an individualized smoking behaviour intervention in Chinese adolescents.
      ).
      Studies utilized various types of control conditions with varying intensities for comparison to text messaging-based interventions for smoking cessation. Only one study utilized an assessment only control group (
      • Haug S.
      • Schaub M.
      • Venzin V.
      • Meyer C.
      • John U.
      • Eysenbach G.
      Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
      ). Three studies only provided the control group with a self-help pamphlet for smoking cessation (
      • Naughton F.
      • Prevost T.
      • Gilbert H.
      • Sutton S.
      Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
      ,
      • Shi H.
      • Jiang X.
      • Yu C.
      • Zhang Y.
      Use of mobile phone text messaging to deliver an individualized smoking behaviour intervention in Chinese adolescents.
      ,
      • Ybarra M.
      • Bagci A.T.
      • Korchmaros J.
      • Emri S.
      A text messaging-based smoking cessation program for adult smokers: Randomized controlled trial.
      ,
      • Ybarra M.L.
      • Summers J.
      • Bagci A.T.
      • Emri S.
      Design considerations in developing a text messaging program aimed at smoking cessation.
      ) while one study provided brief information on publicly available Web- and phone-based assistance for smoking cessation available in Australia (
      • Borland R.
      • Balmford J.
      • Benda P.
      Population-level effects of automated smoking cessation help programs: A randomized controlled trial.
      ). Three studies only provided control participants with biweekly generic study text messages describing the importance of participation and reminders for follow-up appointments (
      • Free C.
      • Whittaker R.
      • Knight R.
      • Abramsky T.
      • Rodgers A.
      • Roberts I.G.
      Txt2stop: A pilot randomised controlled trial of mobile phone-based smoking cessation support.
      ,
      • Free C.
      • Knight R.
      • Robertson S.
      • Whittakers R.
      • Edwards P.
      • Zhou W.
      • et al.
      Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
      ,
      • Whittaker R.
      • Dorey E.
      • Bramley D.
      • Bullen C.
      • Elley R.C.
      • Maddison R.
      • et al.
      A theory-based video messaging mobile phone intervention for smoking cessation: Randomized control trial.
      ). Two studies provided control group participants with general information regarding publically available quitlines or Web sites for smoking cessation followed by generic study text messages regarding participation and follow-up appointments (
      • Abroms L.C.
      • Boal A.L.
      • Simmens S.J.
      • Mendel J.A.
      • Windsor R.A.
      A randomized trial of Text2Quit: A text messaging program for smoking cessation.
      ,
      • Rodgers A.
      • Corbett T.
      • Bramley D.
      • Riddell T.
      • Wills M.
      • Lin R.B.
      • et al.
      Do u smoke after txt? Results of a randomized trial of smoking cessation using mobile phone text messaging.
      ). All participants in the
      • Bock B.
      • Heron K.
      • Jennings E.
      • Morrow K.
      • Cobb V.
      • Magee J.
      • et al.
      A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
      study received a 30-minute counseling session and a quit smoking guide prior to randomization, the control group then received 8 weeks of daily motivational text messages not related to smoking cessation.
      • Ybarra M.L.
      • Summers Holtrop J.
      • Prescott T.L.
      • Rahbar M.H.
      • Strong D.
      Pilot RCT results of Stop My Smoking USA: A text messaging-based smoking cessation program for young adults.
      also provided control participants with general health text messages but in a frequency similar to that received by the intervention group. Finally,
      • Naughton F.
      • Jamison J.
      • Boase S.
      • Sloan M.
      • Gilbert H.
      • Prevost A.T.
      • et al.
      Randomized controlled trial to assess the short-term effectiveness of tailored web- and text-based facilitation of smoking cessation in primary care (iQuit in Practice).
      provided all participants regardless of randomization with routine smoking cessation advice by a healthcare provider and options for pharmacotherapy with the option for follow-up visits, but the control group did not receive any additional assistance.
      A summary of the included studies and intervention moderator coding is provided in Table A.1. Approximately half (k = 7) of the text message interventions were SMS text plus programs combining text messaging with additional intervention modalities (e.g., individual counseling session, tailored self-help pamphlet, and video messaging), while 46% (k = 6) of programs reported using an SMS text only intervention. All text messaging interventions were automatically initiated by the researcher (with the exception of on-demand crave messages). All but one study (k = 12) offered bidirectional messaging that allowed for participants to respond to assessment messages and request craving support.
      Sixty-two percent of studies (k = 8) utilized a tailored decreasing message frequency distribution based on the quitting curve, in which participants received the greatest number of messages around the quit attempt followed by a gradual reduction until the end of the program. Most of these programs (k = 5) also offered a preparatory phase of messages leading up to a quit attempt. Twenty-three percent (k = 3) of studies reported utilizing a fixed message frequency schedule which remained relatively constant across the intervention duration. The final 15% (k = 2) of studies reported a variable message frequency schedule, where participants could change between message tracks at any time.
      Approximately half of studies (k = 7) included some sort of assessment messaging to evaluate quit status. The majority of studies (k = 8) utilized a dynamic message track that adjusted to an individual's quit status or current stage of change, while 38% percent of studies (k = 5) utilized a fixed message track in which participants were on the same fixed schedule of messaging. All studies implemented some sort of message tailoring or targeting; 62% (k = 8) used tailored messages only, 8% (k = 1) used targeted messages only, and 30% (k = 4) used both tailoring and targeting. The majority of studies (k = 11) provided participants with on-demand messaging for craving or relapse support. Thirty-eight percent of studies (k = 5) included a social or peer-to-peer support function for participants to communicate with each other. Finally, half of studies (k = 7) promoted the use of nicotine replacement therapy.

      3.2 Intervention efficacy

      Studies were not significantly heterogeneous (Q12 = 12.47, p = .41). Therefore, we report results for both fixed and random effects models. The fixed effects model indicated that smoking quit rates for the text messaging intervention groups were 37% higher compared to the quit rates for controls (OR = 1.37, 95% CI = 1.25, 1.50) (Fig. A.1.). The random effects model similarly reported a 36% increase in smoking cessation rates for text message interventions when compared to control (OR = 1.36, 95% CI = 1.23, 1.51). These effect sizes suggest that text messaging interventions are an effective means for reducing smoking. Due to the similarity in results between models, only the random effects model will be discussed in text (see Table A.2. for all results). An informal assessment of the funnel plot indicates that the included studies were symmetrical and the test assessing funnel plot symmetry was not significant (p = .12) also indicating that there was no statistically significant publication bias in our meta-analysis.

      3.3 Intervention moderators' efficacy

      Intervention moderators were analyzed to identify differences between group effect sizes (see Table A.2.). First, we assessed the efficacy of text message interventions at 3 and 6 month follow-up. There were no significant differences between intervention efficacy over time (QB = 0.46, df = 1, p = .49); however, studies with a 3 month follow-up period showed a slightly higher efficacy compared to studies with 6 month follow-up data. These results indicate that intervention efficacy is consistent across 3 and 6 month follow-up periods but short-term intervention efficacy is slightly better.
      Second, we compared smoking cessation interventions that provided only text messages (k = 6) to those that provided text messaging plus additional modalities (k = 7). There were no significant differences between intervention types (QB = 1.66, df = 1, p = .20), but text plus programs indicated a slightly higher pooled effect size. These results indicate that smoking cessation programs that include additional intervention modalities (i.e., individual counseling session, tailored self-help pamphlet, and video messaging) were not significantly more effective than those that only utilized text messaging.
      Third, we examined the differences between messaging frequencies. There was no statistically significant difference between the three message frequencies: decreasing schedule (k = 8), fixed schedule (k = 3), and a variable schedule (k = 2), (QB = 0.86, df = 2, p = .65). However, interventions that utilized a fixed message schedule had the highest significant effect size (OR = 1.57, 95% CI = 1.14, 2.17) compared to a decreasing (OR = 1.34, 95% CI = 1.17, 1.54) or varied schedule (OR = 2.13, 95% CI = 0.44, 10.26). These results may suggest that using a consistent number of daily messages throughout the intervention period may increase intervention efficacy and prove to be the most effective messaging strategy.
      Fourth, we compared text message interventions that implemented a fixed message track (k = 5) to those that used a dynamic message track (k = 8). We found no significant difference based on choice of message track (QB = 1.03, df = 1, p = .31). These results indicate that there is no particular benefit of utilizing a fixed or dynamic message track in terms of intervention efficacy.
      Fifth, we examined the difference in effect sizes between studies that used message tailoring (k = 8), targeting (k = 1), or the use of both techniques (k = 4). While there was no significant difference between message content tailoring and targeting, (QB = 1.15, df = 2, p = .56), all studies utilized some type of message content tailoring. Pooled results indicated the use of message tailoring (OR = 1.51, 95% CI = 1.17, 1.94) and the combined use of message targeting and tailoring (OR = 1.39, 95% CI = 1.25, 1.56) techniques were equally effective for increasing smoking cessation when compared to control groups. As only one study included only message targeting, there are no pooled results for this group, and it is difficult to determine the isolated effect of message targeting on smoking cessation.
      Sixth, we compared programs that offered on-demand messaging services for additional support (k = 11) to those that did not provide crave/relapse support (k = 2). There was no significant difference between programs that offered on-demand messages and those that did not (QB = 0.11, df = 1, p = .74). While pooled results of studies not offering on demand messaging had a higher efficacy than studies that offered on demand messaging, this result was not statistically significant (OR = 1.50, 95% CI = 0.91, 1.55). These results may indicate that offering on-demand support messages can increase the intervention efficacy. However, information was not available regarding the actual use of the on-demand service from the included studies.
      Seventh, we compared interventions that included assessment messages (k = 7) to those that did not assess for quit status or stage of change (k = 6). There was no statistically significant difference between interventions that utilized assessment messages versus those that did not (QB = 0.69, df = 1, p = .41). These results indicate that the use of assessment messages did not affect intervention efficacy.
      Eighth, we compared text message interventions that provided a peer-to-peer support component (k = 5) to those that did not (k = 8). We did not find a statistically significant difference between studies that included a social support component and those that did not (QB = 0.31, df = 1, p = .58). These results suggest that a social support function does not significantly increase intervention efficacy. However, data regarding the actual use of the social support function was not available from all included studies.
      Finally, we compared studies that promoted the use of NRT (k = 7) to those that did not (k = 6). While we did not find a significant difference between programs that allowed the use of NRT, (QB = 0.75, df = 1, p = .39), programs that did not utilize NRT (OR = 1.57, 95% CI = 1.14, 2.16) displayed a slightly higher effect size than those that allowed NRT use (OR = 1.34, 95% CI = 1.16, 1.53). These results indicate that text message interventions may be more effective without the use of NRT; however, information regarding the actual use of NRT is not available from all of the included studies.

      4. Discussion

      The purpose of this meta-analysis was to examine the efficacy of text message interventions for smoking cessation. We were particularly interested whether different intervention moderators affected subsequent quit rates. The results of this meta-analysis indicate that text message interventions are an effective means for reducing smoking. The overall random effects model indicates that text message interventions for smoking cessation increase the odds of quitting by 1.36 compared to quitting in control groups. These findings are similar to results found in other meta-analytic reviews of text messaging and mobile technology for health behavior change and smoking cessation (
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      ,
      • Whittaker R.
      • McRobbie H.
      • Bullen C.
      • Borland R.
      • Rodgers A.
      • Gu Y.
      Mobile phone-based interventions for smoking cessation.
      ). Text message interventions also display comparable quit rates with other types of smoking cessation interventions: telephone quitlines can increase the odds of quitting by 1.6, counseling and behavior therapies, such as social support and practical counseling, increases the odds of quitting by 1.3 and 1.5, respectively, while NRT and medications can increase the odds of smoking cessation by 1.5 to 3.1 over placebo (
      • Fiore M.
      • Jaen C.
      • Baker T.
      • Bailey W.
      • Benowitz N.
      • Curry S.
      • et al.
      Treating tobacco use and dependence: Clinical practice guideline.
      ). Our results indicate that text message interventions are as effective as other smoking cessation interventions, with the added benefit of being delivered at a presumably lower cost.
      Our moderator analyses provided important insights regarding text messaging program components and their effect on smoking cessation. It is important to note that moderators are not randomly distributed across studies and thus are always confounded by the presence or absence of other moderators. Our results indicate that smoking cessation rates were generally consistent at 3- and 6-month follow-ups. However, pooled results of studies with 3 month follow-up data indicate a slightly better efficacy of short-term smoking cessation. Our results are consistent with those of
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      in which text plus interventions were not significantly better than text only interventions. This is an important finding for the development of future text message interventions for smoking cessation. Researchers and program developers may not need to include additional intervention modalities in order to obtain the same efficacy from text messages alone.
      Based on a limited number of studies, we found some evidence that message frequencies utilizing a fixed message frequency may be the most effective strategy for smoking cessation. While this result was not significantly greater than the decreasing or variable message frequencies, it does provide an opportunity for future investigation. Our results also suggest that there may not be a benefit of utilizing a fixed or dynamic message track. This is an interesting finding considering that dynamic message tracks inherently provide more tailored information by varying the participants' messaging based on quit status or stage of change. However, the tailored messaging tracks did not significantly improve intervention efficacy when compared to fixed messaging tracks in this sample of studies.
      In this meta-analysis, all studies utilized some type of content tailoring or targeting making it difficult to determine an effect between message tailoring and no tailoring. However, studies that utilized tailoring or a combination of tailoring and targeting exhibited similar efficacy for smoking cessation. These results are similar to those found by
      • Head K.J.
      • Noar S.M.
      • Iannarino N.T.
      • Grant Harrington N.
      Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
      in which the combination of targeting and tailoring techniques were the most efficacious, followed by message tailoring, while message targeting was not found to be effective. This method of content tailoring is particularly beneficial for use in text message interventions because it has the ability to provide tailored information on multiple levels (e.g., demographics, psychosocial variables).
      Our results found that on-demand support messages did not increase quit rates. However, information was not available regarding the actual use of these on-demand services in the included articles. Therefore, results should be interpreted with caution, and further research is necessary to determine the exact relationship between on-demand support and quit rates. We also found that the use of assessment messages did not affect intervention efficacy and appeared to be unobtrusive in the quitting process. Assessment messages are an important component of text message interventions in order to obtain real-time participant updates and assess treatment fidelity. While the use of assessment messages did not increase quit rates, they also did not appear to negatively impact intervention efficacy through increased participant burden.
      • Irvine L.
      • Falconer D.W.
      • Jones C.
      • Ricketts I.W.
      • Williams B.
      • Crombie I.
      Can text messages reach the parts other process measures cannot reach: An evaluation of a behavior change intervention delivered by mobile phone?.
      also found process measures collected during a brief text message alcohol intervention, which were unobtrusive and cost-effective.
      Our findings also suggest the inclusion of a social support communication function did not significantly increase quit rates. However, these results should be interpreted with caution given that data regarding the actual use of the social support function were not available in the included studies. Further research into the importance of social support communication between smokers may provide promising avenues for future interventions. Finally, our results indicate that the promotion of NRT use did not significantly increase the quit rates among smokers. This could present a potential confounding variable when assessing text message intervention efficacy, however studies that did not encourage the use of NRT actually had slightly higher effect sizes than those that encouraged its use. Again, however, data were not available regarding the actual use of NRT in all of the included studies. Further research is necessary to determine to effectiveness of text message programs with the use of NRT. Treatment combinations (e.g., behavioral interventions with medications) have generally been found to be more effective than with one alone (
      • Fiore M.
      • Jaen C.
      • Baker T.
      • Bailey W.
      • Benowitz N.
      • Curry S.
      • et al.
      Treating tobacco use and dependence: Clinical practice guideline.
      ).
      Our results provide additional support for this increasingly popular area of using technology to increase smoking cessation. Due to the growing capabilities of mobile technology, text message interventions have the capacity to provide users with numerous program options, such as social support from peers, tailored message tracks and content, and on-demand support. This analysis of intervention components will inform future intervention design regarding which components are most effective at increasing quit rates.

      4.1 Future research

      Future research should focus on adaptive, tailored programming for text message interventions. Ecological momentary assessment (EMA) may provide a promising avenue for assessing smoking relapse risk and priming a text message program to intervene with additional support at the most appropriate time. EMA can be particularly useful for assessing cue-induced craving (
      • Watkins K.L.
      • Regan S.D.
      • Nguyen N.
      • Businelle M.S.
      • Kendzor D.E.
      • Lam C.
      • et al.
      Advancing cessation research by integrating EMA and geospatial methodologies: Associations between tobacco retail outlets and real-time smoking urges during a quit attempt.
      ). When EMA is used in conjunction with a behavioral intervention, the craving or relapse can be intervened upon and prevented in real-time.
      • Watkins K.L.
      • Regan S.D.
      • Nguyen N.
      • Businelle M.S.
      • Kendzor D.E.
      • Lam C.
      • et al.
      Advancing cessation research by integrating EMA and geospatial methodologies: Associations between tobacco retail outlets and real-time smoking urges during a quit attempt.
      utilized EMA to assess smoking urges and geo-spatial location via smartphones. Their findings suggest that closer proximity to tobacco outlets increased urges to smoke. By utilizing this kind of information as well as various other triggers, location, and psychosocial characteristics, text message programs can provide increasingly tailored support for smoking cessation.

      4.2 Limitations

      There were some limitations to this meta-analytic review. First, this review only included 13 studies, a small sample for a meta-analysis. For the moderator analyses, the sample sizes were even smaller when grouped by intervention moderator; hence it was difficult to detect some effects. It will be important to assess these intervention moderators again with a larger sample of text message interventions. Second, we mainly relied on information collected from the published article. This resulted in items being coded as not occurring if the component was not specifically mentioned in the article. For example, if NRT was not mentioned in the article it was coded as ‘not promoting’ the use of NRT. It is possible that some studies could have been misclassified if the article did not specifically reference the component. Data regarding the actual usage of the optional intervention components (e.g., crave messages, social support, NRT use) must be interpreted given that actual usage rates were not available for all of the included articles. Finally, it is further speculated that the explanation for non-significant results in the QB analysis was due to studies being too homogeneous. The test for heterogeneity as not significant and the similarity between studies could have led to non-significant results regarding the differences in effect sizes between moderator groups.

      5. Conclusions

      Our results add to a growing body of literature indicating that smoking cessation interventions via mobile phone are effective. The moderator analyses provide information for future research and text message intervention development. Moderators such as these will be important when developing new text message interventions that include the most efficacious design components. The benefits of mobile technology include the ease of use, cost-effective intervention delivery, the ability to tailor message content and timing to individual characteristics, and sending and receiving time-sensitive information. However, our results must be interpreted with caution given that the relatively small sample size and future research should assess a larger pool of text message interventions for smoking cessation.

      Conflict of interest

      None reported.

      Acknowledgements

      The authors want to thank Brittany Marshall for her support and thoughts during this project.

      Appendix A.

      Table A.1Summary of SMS text messaging interventions.
      ArticleNOutcome measureFollow-up lengthIntervention typeMessage content typeMessage trackMessage frequencyAverage number of messages sentOn demand messagingAssessment messagingPeer-to-peer supportUse of NRT
      • Abroms L.C.
      • Boal A.L.
      • Simmens S.J.
      • Mendel J.A.
      • Windsor R.A.
      A randomized trial of Text2Quit: A text messaging program for smoking cessation.
      5037-day point prevalence6 monthsText plus

      Emails and supporting Web site
      Tailored

      Quit date, reasons for quitting, money saved, and use of NRT
      Dynamic

      Messaging track dependent on quit success
      Decreasing schedule

      5 on quit day

      2 per day for 1 week

      3 per week for 8 weeks

      1 per week for 4 weeks
      45 over 3 monthsYesYesNoYes
      • Bock B.
      • Heron K.
      • Jennings E.
      • Morrow K.
      • Cobb V.
      • Magee J.
      • et al.
      A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
      607-day point prevalence6 monthsText plus

      30 minute counseling session
      Tailored

      Stage of change matched
      Dynamic

      Messaging tracks dependent on current stage of change
      Variable schedule

      Not ready —1 per day for 14 days

      Prepare—2 per day for 14 days

      Quit—4 per day for 2 weeks then 2 per day for 4 weeks
      Unable to calculate

      Dependent on message track, variable over 8 weeks
      YesYesYesNo
      • Borland R.
      • Balmford J.
      • Benda P.
      Population-level effects of automated smoking cessation help programs: A randomized controlled trial.
      35307-day point prevalence6 monthsText onlyBoth

      Stage of change matched

      and gender
      Dynamic

      Messaging track dependent on current stage of change
      Variable schedule

      Participants could choose 3 different message frequencies
      Unable to calculate

      Dependent on message track, variable over 6 months
      YesYesNoYes
      • Free C.
      • Whittaker R.
      • Knight R.
      • Abramsky T.
      • Rodgers A.
      • Roberts I.G.
      Txt2stop: A pilot randomised controlled trial of mobile phone-based smoking cessation support.
      2007-day point prevalence6 monthsText OnlyBoth

      Demographic information and concerns about quitting
      FixedDecreasing schedule

      1 per day until quit

      5 per day for 5 weeks

      3 per week for 26 weeks
      225 over 7–8 monthsYesNoYesNo
      • Free C.
      • Knight R.
      • Robertson S.
      • Whittakers R.
      • Edwards P.
      • Zhou W.
      • et al.
      Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
      58007-day point prevalence6 monthsText OnlyBoth

      Demographic information and concerns about quitting
      FixedDecreasing schedule

      5 per day for 5 weeks

      3 per week for 26 weeks
      225 over 7–8 monthsYesNoYesYes
      • Haug S.
      • Schaub M.
      • Venzin V.
      • Meyer C.
      • John U.
      • Eysenbach G.
      Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
      7557-day point prevalence6 monthsText onlyBoth

      Demographic information and stage of change matched
      Dynamic

      Messaging track dependent on current stage of change
      Fixed schedule

      1–3 daily messages
      58 over 3 monthsNoYesNoNo
      • Naughton F.
      • Prevost T.
      • Gilbert H.
      • Sutton S.
      Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
      2077-day point prevalence3 monthsText plus

      Tailored self-help pamphlet
      Tailored

      26 characteristics: motivation to quit, confidence, nicotine dependence, beliefs about harms
      Dynamic

      Messaging track dependent on quit success
      Decreasing schedule

      1–3 messages daily

      Highest during first 4 weeks then frequency reduced
      80 over 11 weeksYesYesNoYes
      • Naughton F.
      • Jamison J.
      • Boase S.
      • Sloan M.
      • Gilbert H.
      • Prevost A.T.
      • et al.
      Randomized controlled trial to assess the short-term effectiveness of tailored web- and text-based facilitation of smoking cessation in primary care (iQuit in Practice).
      602Continuous abstinence6 monthsText plus

      Smoking cessation advice from primary care provider and tailored self-help pamphlet
      Tailored

      24 characteristics: motivation to quit, determination, nicotine dependence, reasons for quitting
      Dynamic

      Messaging track dependent on quit success
      Fixed schedule

      0–2 daily messages
      108 over 3 monthsYesYesNoYes
      • Rodgers A.
      • Corbett T.
      • Bramley D.
      • Riddell T.
      • Wills M.
      • Lin R.B.
      • et al.
      Do u smoke after txt? Results of a randomized trial of smoking cessation using mobile phone text messaging.
      17057-day point prevalence6 monthsText onlyTailored

      Preferences, smoking history, barriers to cessation
      FixedDecreasing schedule

      5 per day for 1 week pre-quit

      5 per day for 4 weeks

      3 per week for 4.5 months
      199 over 6 monthsYesNoYesYes
      • Shi H.
      • Jiang X.
      • Yu C.
      • Zhang Y.
      Use of mobile phone text messaging to deliver an individualized smoking behaviour intervention in Chinese adolescents.
      1797-day point prevalence3 monthsText plus

      Supporting Web site and online chatting
      Tailored

      Stage of change matched
      FixedFixed schedule

      1 message per day
      84 over 3 monthsYesNoNoNo
      • Whittaker R.
      • Dorey E.
      • Bramley D.
      • Bullen C.
      • Elley R.C.
      • Maddison R.
      • et al.
      A theory-based video messaging mobile phone intervention for smoking cessation: Randomized control trial.
      2267-day point prevalence6 monthsText plus

      Supporting Web site
      Targeted

      Role model for video messages, message timing
      FixedDecreasing schedule

      1 per day for 1 week pre-quit

      3 per day for 5 weeks

      3–4 per week for 2 weeks

      1–2 per week for 20 weeks
      136 over 6 monthsYesNoNoYes
      • Ybarra M.
      • Bagci A.T.
      • Korchmaros J.
      • Emri S.
      A text messaging-based smoking cessation program for adult smokers: Randomized controlled trial.
      1517-day point prevalence3 monthsText OnlyTailored

      Stage of change matched
      Dynamic

      Messaging track dependent on quit success
      Decreasing schedule

      5 per day during 2 week pre-quit

      9 quit day

      1 less each day for the first week

      2 per day for 2 weeks

      1 per day for 1 week
      119 over 6 weeksNoNoNoNo
      • Ybarra M.L.
      • Summers Holtrop J.
      • Prescott T.L.
      • Rahbar M.H.
      • Strong D.
      Pilot RCT results of Stop My Smoking USA: A text messaging-based smoking cessation program for young adults.
      211Continuous abstinence3 monthsText plus

      Supporting Web site
      Tailored

      Stage of change matched
      Dynamic

      Messaging track dependent on quit success
      Decreasing schedule

      4 per day during 2 week pre-quit

      9 quit day

      1 less each day for the first week

      2 per day for 2 weeks

      1 per day for 1 week
      146 over 6 weeksYesYesYesNo
      Table A.2Effect sizes by categorical moderating variables.
      Fixed effectsRandom effects
      kOR95% CIOR95% CIQQB
      Combined131.371.25–1.501.361.23–1.51NS
      Follow-up periodNSNS
       3 month41.581.06–2.361.571.05–2.35
       6 month91.361.24–1.491.351.18–1.49
      InterventionNSNS
       Text only61.331.21–1.471.321.17–1.48
       Text plus71.591.26–2.011.581.25–1.99
      Message frequencyNSNS
       Decreasing schedule81.361.23–1.501.341.17–1.54
       Fixed schedule31.571.14–2.171.571.14–2.17
      Variable schedule21.320.99–1.762.130.44–10.26
      Message trackNSNS
       Fixed51.331.19–1.481.281.08–1.52
       Dynamic81.491.25–1.781.481.24–1.77
      Message typeNSNS
       Targeted11.000.53–1.86
       Tailored81.351.15–1.591.511.17–1.94
       Targeted/Tailored41.391.25–1.561.391.25–1.56
      Messages on demandNSNS
       Yes111.371.24–1.501.361.21–1.52
       No21.480.97–2.261.500.91–2.48
      Assessment messagesNSNS
       Yes71.471.23–1.751.461.22–1.75
       No61.341.20–1.491.301.09–1.55
      Social/Peer-to-PeerNSNS
       Yes51.351.21–1.501.321.07–1.63
       No81.431.20–1.701.431.20–1.70
      NRTNSNS
       Yes71.351.23–1.491.341.16–1.53
       No61.601.16–2.191.571.14–2.16
      k = number of studies, OR = odds ratio, CI = confidence interval, Q = measure of heterogeneity, QB = measure of effect size difference. NS = non-significant.
      Figure thumbnail gr1
      Fig. A.1Efficacy of text message-based interventions for smoking cessation.

      References

        • Abroms L.C.
        • Boal A.L.
        • Simmens S.J.
        • Mendel J.A.
        • Windsor R.A.
        A randomized trial of Text2Quit: A text messaging program for smoking cessation.
        American Journal of Preventive Medicine. 2014; 47: 242-250
        • Bock B.
        • Heron K.
        • Jennings E.
        • Morrow K.
        • Cobb V.
        • Magee J.
        • et al.
        A text message delivered smoking cessation intervention: The initial trial of TXT-2 Quit: Randomized controlled trial.
        Journal of Medical Internet Research. 2013; 1: e17
        • Borland R.
        • Balmford J.
        • Benda P.
        Population-level effects of automated smoking cessation help programs: A randomized controlled trial.
        Addiction. 2012; 108: 618-628
        • Cadmus-Bertram L.
        • Wang J.B.
        • Patterson R.E.
        • Newman V.A.
        • Parker B.A.
        • Pierce J.P.
        Web-based self-monitoring for weight loss among overweight/obese women at increased risk for breast cancer: The HELP pilot study.
        Psycho-Oncology. 2013; 22: 1821-1828
        • Cole-Lewis H.
        • Kershaw T.
        Text messaging as a tool for behavior change in disease prevention and management.
        Epidemiologic Reviews. 2010; 32: 56-69
        • Egger M.
        • Davey Smith G.
        • Schneider M.
        • Minder C.
        Bias in meta-analysis detected by a simple, graphical test.
        British Medical Journal. 1997; 315: 629-634
        • eMarketer
        Smartphone users worldwide will total 1.75 billion in 2014.
        (Retrieved May 15, 2014)
        • Fiore M.
        • Jaen C.
        • Baker T.
        • Bailey W.
        • Benowitz N.
        • Curry S.
        • et al.
        Treating tobacco use and dependence: Clinical practice guideline.
        US Public Health and Human Services, Washington, DC2008
        • Fjeldsoe B.S.
        • Marhsall A.L.
        • Miller Y.D.
        Behavior change interventions delivered by mobile telephone short-message service.
        American Journal of Preventive Medicine. 2009; 36: 165-173
        • Fogg B.J.
        Persuasive technology: Using computers to change what we think and do.
        Morgan Kaufmann Publishers, San Francisco, CA2003
        • Foreman K.F.
        • Stockl K.M.
        • Le L.B.
        • Fisk E.
        • Shah S.M.
        • Lew H.C.
        • et al.
        Impact of a text messaging pilot program on patient medication adherence.
        Clinical Therapy. 2012; 34: 1084-1091
        • Free C.
        • Knight R.
        • Robertson S.
        • Whittakers R.
        • Edwards P.
        • Zhou W.
        • et al.
        Smoking cessation support delivered via mobile phone text messaging (txt2stop): A single-bling, randomized trial.
        Lancet. 2011; 378: 49-55
        • Free C.
        • Whittaker R.
        • Knight R.
        • Abramsky T.
        • Rodgers A.
        • Roberts I.G.
        Txt2stop: A pilot randomised controlled trial of mobile phone-based smoking cessation support.
        Tobacco Control. 2009; 18: 88-91
        • Haug S.
        • Schaub M.
        • Venzin V.
        • Meyer C.
        • John U.
        • Eysenbach G.
        Efficacy of a text message-based smoking cessation intervention for young people: A cluster randomized controlled trial.
        Journal of Medical Internet Research. 2013; 15: e171
        • Head K.J.
        • Noar S.M.
        • Iannarino N.T.
        • Grant Harrington N.
        Efficacy of text messaging-based interventions for health promotion: A meta-analysis.
        Social Science and Medicine. 2013; 97: 41-48
        • Irvine L.
        • Falconer D.W.
        • Jones C.
        • Ricketts I.W.
        • Williams B.
        • Crombie I.
        Can text messages reach the parts other process measures cannot reach: An evaluation of a behavior change intervention delivered by mobile phone?.
        PLoS ONE. 2012; 7: e52621
        • Naughton F.
        • Jamison J.
        • Boase S.
        • Sloan M.
        • Gilbert H.
        • Prevost A.T.
        • et al.
        Randomized controlled trial to assess the short-term effectiveness of tailored web- and text-based facilitation of smoking cessation in primary care (iQuit in Practice).
        Addiction. 2014; 109: 1184-1193
        • Naughton F.
        • Prevost T.
        • Gilbert H.
        • Sutton S.
        Randomized controlled trial evaluation of a tailored leaflet and SMS text message self-help intervention for pregnancy smokers (MiQuit).
        Nicotine & Tobacco Research: Official Journal of the Society for Research on Nicotine and Tobacco. 2012; 14: 569-577
        • Petty R.E.
        • Barden J.
        • Wheeler S.C.
        The elaboration likelihood model of persuasion: Developing health promotions for sustained behavioral change.
        in: DiClemente R.J. Crosby R.A. Kegler M.C. Emerging Theories in Health Promotion Practice and Research. 2nd ed. Jossey-Bass, San Francisco, CA2009
        • Prochaska J.O.
        • DiClemente C.C.
        Transtheoretical therapy: Toward a more integrative model of change.
        Psychotherapy: Theory, Research and Practice. 1982; 19: 276-288
        • Rodgers A.
        • Corbett T.
        • Bramley D.
        • Riddell T.
        • Wills M.
        • Lin R.B.
        • et al.
        Do u smoke after txt? Results of a randomized trial of smoking cessation using mobile phone text messaging.
        Tobacco Control. 2005; 14: 255-261
        • Shi H.
        • Jiang X.
        • Yu C.
        • Zhang Y.
        Use of mobile phone text messaging to deliver an individualized smoking behaviour intervention in Chinese adolescents.
        Journal of Telemedicine and Telecare. 2013; 19: 282-287
        • Walters S.T.
        • Ondersma S.J.
        • Ingersoll K.S.
        • Rodriguez M.
        • Lerch J.M.A.
        • Rossheim M.E.
        • et al.
        MAPIT: Development of a web-based intervention targeting substance abuse treatment in the criminal justice system.
        Journal of Substance Abuse Treatment. 2014; 46: 60-65
        • Watkins K.L.
        • Regan S.D.
        • Nguyen N.
        • Businelle M.S.
        • Kendzor D.E.
        • Lam C.
        • et al.
        Advancing cessation research by integrating EMA and geospatial methodologies: Associations between tobacco retail outlets and real-time smoking urges during a quit attempt.
        Nicotine & Tobacco Research: Official Journal of the Society for Research on Nicotine and Tobacco. 2014; 16: S93-S101
        • Whittaker R.
        • Dorey E.
        • Bramley D.
        • Bullen C.
        • Elley R.C.
        • Maddison R.
        • et al.
        A theory-based video messaging mobile phone intervention for smoking cessation: Randomized control trial.
        Journal of Medical Internet Research. 2011; 13: 61-72
        • Whittaker R.
        • McRobbie H.
        • Bullen C.
        • Borland R.
        • Rodgers A.
        • Gu Y.
        Mobile phone-based interventions for smoking cessation.
        Cochrane Database of Systematic Reviews. 2012; 11: CD006611
        • Ybarra M.
        • Bagci A.T.
        • Korchmaros J.
        • Emri S.
        A text messaging-based smoking cessation program for adult smokers: Randomized controlled trial.
        Journal of Medical Internet Research. 2012; 14: e47
        • Ybarra M.L.
        • Summers Holtrop J.
        • Prescott T.L.
        • Rahbar M.H.
        • Strong D.
        Pilot RCT results of Stop My Smoking USA: A text messaging-based smoking cessation program for young adults.
        Nicotine & Tobacco Research: Official Journal of the Society for Research on Nicotine and Tobacco. 2013; 15: 1388-1399
        • Ybarra M.L.
        • Summers J.
        • Bagci A.T.
        • Emri S.
        Design considerations in developing a text messaging program aimed at smoking cessation.
        Journal of Medical Internet Research. 2012; 14: e8