This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Happy Ending (HE) is an intense 1-year smoking cessation program delivered via the Internet and cell phone. HE consists of more than 400 contacts by email, Web pages, interactive voice response, and short message service technology. HE includes a craving helpline and a relapse prevention system, providing just-in-time therapy. All the components of the program are fully automated.
The objectives were to describe the rationale for the design of HE, to assess the 12-month efficacy of HE in a sample of smokers willing to attempt to quit without the use of nicotine replacement therapy, and to explore the potential effect of HE on coping planning and self-efficacy (prior to quitting) and whether coping planning and self-efficacy mediate treatment effect.
A two-arm randomized controlled trial was used. Subjects were recruited via Internet advertisements and randomly assigned to condition. Inclusion criteria were willingness to quit on a prescribed day without using nicotine replacement and being aged 18 years or older. The intervention group received HE, and the control group received a 44-page self-help booklet. Abstinence was defined as “not even a puff of smoke, for the last seven days” and was assessed by means of Internet surveys or telephone interviews 1, 3, 6, and 12 months postcessation. The main outcome was repeated point abstinence (ie, abstinence at all four time points). Coping planning and self-efficacy were measured at baseline and at the end of the preparation phase (ie, after 2 weeks of treatment, but prior to cessation day).
A total of 290 participants received either the HE intervention (n=144) or the control booklet (n=146). Using intent-to-treat analysis, participants in the intervention group reported clinically and statistically significantly higher repeated point abstinence rates than control participants (20% versus 7%, odds ratio [OR] = 3.43, 95% CI = 1.60-7.34,
This 12-month trial documents a long-term treatment effect of a fully automated smoking cessation intervention without the use of nicotine replacement therapy. The study adds to the promise of using digital media in supporting behavior change.
Two reviews [
The psychological processes that quitters experience are different across various time points and follow a certain chronology [
HE starts with a 14-day preparation phase. Every morning, the client receives an email containing a hyperlink. By activating the link, the smoker has access to that particular day’s website. See
The participant is educated about his or her psychological profile and responses, both as a person and as a smoker. Consequently, smokers will be more aware of, and will learn about, such things as their smoking behavior and nicotine dependence, reasons for previous failures to quit, motivational basis for quitting, general and task-specific self-efficacy, problems that people often experience when quitting, and stress and weight regulation. One of the most important predictors of the outcome of self-change processes is self-efficacy, or the extent to which the person is confident that he or she will succeed [
A crucial ingredient of the program is to educate the participants about the cognitive, affective, and behavioral reactions that smokers usually experience if a slip occurs (ie, if they smoke some cigarettes during the quit attempt). In HE, participants are told that the administrators expect that most of them will experience one or more slips [
Furthermore, we have applied principles from cognitive behavioral therapy [
In addition to the activities that take place on the websites, the participants stay in touch with HE via short message service (SMS) text messaging and interactive voice response (IVR). The purpose of this is twofold. First, it is important that the participants become used to communicating with HE via the cell phone because it plays a crucial role in the rest of the program. Second, the cell phone is used to support the other activities and processes that are initiated via the websites.
After the preparation phase comes a 30-day active quitting phase, which is initiated with the actual cessation attempt. Here, a number of activities are included to ensure that participants are actively involved in their own attempt to quit. Hence, there are numerous contact points every day between the participant and HE. Participants receive an email in the morning with a link to that day’s specific website. However, there are several differences between these websites and the ones in the preparation phase. First, the Web activities focus on the motivational conflict that many smokers will experience during the first smoke-free days. Along with the temptations and impulses to smoke, this motivational conflict implies that the effect of the expected consequences of smoking versus not smoking tends to change. In short, the positive short-term consequences of smoking (eg, feeling more relaxed, less irritable) tend to be inflated, while the long-term negative consequences of smoking (eg, health problems) seem to be deflated during the first days and weeks of a quit attempt [
The websites in the active quitting phase contain elements and activities collected from social cognitive learning theory [
An effective program should take into account the fact that a large proportion of quitters are likely to relapse. Relapses typically follow a pattern of intermittent episodes of smoking more often than they follow an abrupt resumption of smoking [
The quitter may experience close-call situations in which the ex-smoker is brought to the brink of smoking [
Finally, HE offers an 11-month follow-up phase. During this phase, the logging-off procedure continues daily for another 4 weeks, twice a week for another 2 weeks, and then once a week for the remaining follow-up period. Hence, the system will register slips and activate the relapse prevention system for the whole period. Furthermore, the participants have access to the craving helpline during the whole follow-up phase. Finally, the quitter receives a number of encouraging SMS and IVR messages during this phase.
In summary, compared with most other digital smoking cessation programs [
We previously investigated the same digital multimedia smoking cessation intervention using a similar design in an earlier 12-month RCT [
First, in the previous trial [
Second, the previous trial failed to document the mediation effect of the program on relevant psychological variables. Technically, a complete mediation effect was found [
We tested the hypothesis that a digital, fully automated smoking cessation intervention would produce an increased 12-month abstinence rate compared with a control condition of a self-help booklet. Furthermore, we expected the digital intervention to increase precessation levels of coping planning and self-efficacy. Finally, we expected the hypothesized increase in precessation coping planning and self-efficacy to partially mediate the treatment effect.
This was a two-arm randomized controlled trial. Subjects were randomized to either receive HE (intervention), or a 44-page self-help booklet (control), described in further detail below. The trial was registered and approved by the Regional Ethics Committee, Norway, South-East (project number: 2.2005.353).
Subjects were recruited by means of online banner advertisements in Norwegian regional newspapers from February 6 to 10, 2006. Banners were displayed 947,059 times, resulting in 2595 hits, which gave a hit rate of 0.3%. When clicking on a banner, potential subjects were routed to a website containing study information, an informed consent, and a baseline questionnaire. During the informed consent process, participants were informed that they would be arbitrarily split into groups that would receive different tools for smoking cessation. It was specified that the various tools did not include any form of medication and that participation in the study did not require attendance at face-to-face meetings or consultations. However, no information was provided whatsoever about the intervention conditions. Inclusion criteria were 1) willingness to quit on March 6, 2006, 2) at least 18 years old, 3) currently smoking five cigarettes or more on a daily basis, 4) willingness to quit without using NRT, 5) owning a mobile phone, 6) a Norwegian-registered phone number and postal address, and 7) having daily access to the Internet and email.
There were 427 unique registrations, 23 of which did not fulfill the inclusion criteria. Another 82 subjects were excluded because of missing values, and 19 subjects were excluded because they were suspected to know each other, based, for example, on sharing or having the same family name, postal address, email, IP address, or worksite. This was done to reduce the risk of communication across experimental conditions. Finally, seven subjects were excluded randomly because the required number of participants was 296 (according to a power analysis).
The control group received a 44-page self-help booklet issued by the Norwegian Directorate for Health and Social Affairs. The booklet contains general cessation information, a quit calendar, a 10-day quit log, the phone number of the national quitline, and links to relevant and open online tobacco cessation resources. The booklet recommends 10 days of preparation prior to quitting, in which readers are encouraged to map their smoking habits in the quit log. Additionally, for each of the 10 preparation days, the booklet suggests an exercise aimed at raising awareness about personal smoking habits. The 48-day quit calendar is composed of small, encouraging daily messages about improvements in health and well-being after quitting (eg, “Your risk of cardiovasculardisease is reduced.” and “Does food taste better to you now?”).
The treatment group received the digital multimedia intervention HE, described above. See
Overview of potential contact points between HE and user during the entire intervention period a
Component of HE | Week 1-2 | Week 3-6 | Week 7-8 b | Week 9-10 | Week 11-15 | Week 16-54 | ||||||||||||||||||||||||||||||||||||
∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | |||||||||||||||||||||||||||||
Web page | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ||||||||||||||||||||||||||||
Text message | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | |||||||||||
Log-on call | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ||||||||||||||||||||||||||||
Log-off call | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ | ∙ |
a The seven columns within the week correspond to the number of days in a week. Each dot represents one intended contact.
b The number of messages per day was gradually reduced from 3 to 1 over the span of these 2 weeks.
Based on computer-generated random digits, 296 subjects were randomly allocated to either the HE intervention or the booklet control condition. Stratified block randomization was applied to ensure equal numbers of both males and females in each group. Randomization was performed by the experimenter. The names and identities of the subjects, however, were concealed to the experimenter during randomization. After randomization, subjects received an email informing them which tool they would be provided with and when and how they would receive it. Subjects in the HE group were told that the intervention would begin on February 20, 2006, but that the designated quit date was March 6, 2006. Subjects in the control group were told about the booklet and were encouraged to read the booklet thoroughly before the designated quit date and to use it actively throughout their quit attempt. Information on the type of treatment provided to the other group was withheld for subjects in both experimental conditions.
Data were collected by means of online questionnaires at baseline, precessation, and at 1, 3, 6, and 12 months after cessation. An email containing a link to the questionnaire was sent to the subjects at each data collection point. Two email reminders were sent to nonresponders. For all postcessation follow-ups, telephone interviews were conducted with subjects who had not responded after the second reminder. The telephone interviews were structured and standardized with no person-to-person counseling or face-to-face contact between experimenters and subjects at any point. Four attempts were made to contact nonresponders by telephone in both conditions at every data collection point.
Abstinence was defined as having been completely smoke-free for the past 7 days. Subjects with missing values on abstinence data were coded as smokers. Abstinence data were based on self-reports with no biochemical verification and were assessed at 1, 3, 6, and 12 months after cessation. The main outcome in this trial was repeated point abstinence, that is, abstinence on all four postcessation measuring points.
Nicotine dependence was assessed by the Fagerström Test for Nicotine Dependence (FTND) [
The present study intended to evaluate the effect of HE without the adjunct use of NRT. All eligible candidates for the study were informed about this and agreed to attempt quitting without using NRT. However, it is important to note that subjects received information and recommendations regarding NRT in both conditions. For technical reasons, it was not possible to modify this feature from the program or the booklet. Therefore, to be able to control for possible NRT use, the subjects were asked at 3 months whether they had used NRT to quit smoking.
An alpha level of .05 was chosen for all statistical tests and all tests were two-tailed. To check for differences between experimental conditions at baseline,
For repeated point abstinence at 12 months and for point abstinence at 1, 3, 6, and 12 months after cessation, the odds ratio (OR) with the 95% confidence interval (CI) and a chi-square test for experimental condition were carried out, respectively. Hierarchical logistic regression was applied [
The flow of participants is depicted in
At baseline, there were no variables on which treatment and control subjects differed significantly (
Flowchart of participants
Baseline sample characteristicsa
Characteristic | HE (n = 144) | Control (n = 146) |
Female, No. (%) | 72 (50) | 73 (50) |
Has a college degree, No. (%) | 70 (49) | 76 (52) |
Age (years) | 39.5 ± 11.0 | 39.7 ± 10.8 |
Nicotine Dependence (FTND) | 4.5 ± 2.3 | 4.6 ± 2.2 |
Cigarettes smoked per day | 16.6 ± 7.2 | 17.6 ± 7.0 |
Precessation self-efficacy | 5.1 ± 1.4 | 5.1 ± 1.3 |
Precessation coping planning | 2.3 ± 0.6 | 2.4 ± 0.7 |
a Numbers are mean ± SD except where noted.
Computerized logging routines revealed that subjects in the treatment condition, to a large extent, accomplished the actions intended in the program design (ie, in 5-6 out of 10 cases). See
Mean number of active client actions for three components of HE (n = 144)a
Active Client Action | Range | Mean | SD | % |
Log-on call | 0-42 | 26 | 16 | 62 |
Opening Web page | 0-44 | 26 | 13 | 59 |
Responding to log-off call | 0-102 | 53 | 37 | 52 |
a The log-off call was initiated by the program. Here, responding means answering either “yes” or “no” to the abstinence question. Theoretical range and observed ranges coincide with one exception: theoretical maximum for log-off calls is 104. The right-hand column shows the average percentage of actions completed.
In total, 57 subjects discontinued the intervention, of which 36 did so during the first 6 weeks. The reason for dropout was not recorded. These subjects were approached by Web and telephone interviews in exactly the same way as were program participants and subjects in the control group. At 1 month, 17 subjects (12%) reported that they found HE “not at all helpful,” 74 subjects (51%) found HE to be “helpful,” and 46 subjects (32%) reported HE to be “very helpful”; data were missing for 7 subjects (5%).
As shown in
Number of Web, phone, and total responses across conditions at specified time points; HE (n = 144) and control (n = 146)
Time | ------ Web ------ | ----- Phone ----- | ----- Total ----- | |||
HE | Control | HE | Control | HE | Control | |
Preparation | 132 | 131 | – | – | 132 | 131 |
1 month | 128 | 119 | 11 | 8 | 139 | 127 |
3 months | 119 | 110 | 16 | 21 | 135 | 131 |
6 months | 101 | 97 | 23 | 23 | 124 | 120 |
12 months | 101 | 89 | 30 | 34 | 131 | 123 |
The main finding from this trial was that participants in the intervention condition (n = 29, 20%) reported clinically and statistically significantly higher repeated point abstinence rates than control participants (n = 10, 7%) (OR = 3.43, 95% CI = 1.60-7.34, n = 290,
Despite agreeing to quit without using NRT, 34 subjects (24%) in the treatment condition and 14 subjects (10%) in the control condition reported NRT use. The proportion of NRT users was significantly higher in the treatment condition compared with the control condition (
Point abstinence and repeated point abstinence rates across conditions at specified time points
Time After Cessation | HE (n = 144), |
Control (n = 146), No. (%) | OR | 95% CI |
|
|
|
||||||
1 month | 60 (42) | 25 (17) | 3.46 | 2.01-5.95 | .001 | |
3 months | 51 (35) | 23 (16) | 2.93 | 1.67-5.14 | .001 | |
6 months | 42 (29) | 20 (14) | 2.59 | 1.43-4.69 | .002 | |
12 months | 47 (33) | 33 (23) | 1.66 | 0.99-2.79 | .07 | |
|
||||||
1 + 3 months | 43 (30) | 17 (12) | 3.23 | 1.74-6.00 | .001 | |
1 + 3 + 6 months | 34 (24) | 10 (7) | 4.24 | 1.99-8.89 | .001 | |
1 + 3 + 6 + 12 months | 29 (20) | 10 (7) | 3.43 | 1.60-7.34 | .002 |
a Point abstinence was based on 7-day point prevalence and intent-to-treat.
Pearson
Pearson
The between-group difference for both coping planning and self-efficacy was small, at only one-third of a standard deviation. Coping planning and self-efficacy were tested formally [
In summary, HE slightly increased the level of both coping planning and self-efficacy during the 2-week preparation phase of the program. The increase in self-efficacy could explain at least some of the initial success in gaining abstinence (ie, at 1 month after cessation).
A complete case analysis showed the repeated point abstinence rate at 12 months to be 25% (29/118 subjects) in the treatment group versus 9% (10/108 subjects) in the control group (
We also looked into what happened when subjects who did not use the intervention at some minimal level were excluded. Excluding subjects who performed fewer than five actions in each of the three categories of log-on calls, opening Web pages, and answering log-off calls resulted in an abstinence rate in the treatment condition of 26% (n= 111). Inclusion of only those who used the intervention at some minimum level and applying a complete case approach further increased the quit rate to 29% (n = 100).
This trial demonstrated the efficacy of the digitally delivered and fully automated HE smoking cessation intervention over a self-help booklet condition—without the combined use of NRT—in producing increased repeated point abstinence at 12 months. The ability of HE to increase precessation self-efficacy could explain some success in gaining early abstinence.
The fact that some quitters used NRT, even though they had promised not to do so, resulted in a somewhat inflated effect size. However, this could not seriously compromise conclusions because the main effect from the experimental condition is still clinically and statistically significant even after controlling for NRT use. Hence, the success of HE can be explained by the psychological support provided by the program. Exactly what mechanisms are at play to cause the treatment effect is not fully clear at this stage. We do know that HE instilled a somewhat higher level of precessation self-efficacy compared with the control condition and that this could explain at least some of the initial success in gaining abstinence.
In a previous trial on the same intervention and with a similar design [
This trial could not biochemically verify self-reported claims of abstinence due to the geographic spread of the sample, cost, and other practical concerns. However, false reporting is considered to be minimal when there is little or no personal contact between treatment provider and subjects [
In summary, this trial extends the public health significance of digital multimedia interventions for smoking cessation. It shows that psychological support can be effectively mediated through modern distance communication technology and that automated support as a stand-alone intervention is, in fact, sufficient for a significant effect on long-term behavior change.
This trial was made possible through cooperation and cofunding among the University of Oslo, Happy Ending AS, and the Norwegian Research Council. Results are owned by the University of Oslo. There are no contractual constraints from any of the sponsors regarding publication.
The second author has a financial interest in the intervention under scrutiny, as a shareholder of Happy Ending AS.
Fagerström Test for Nicotine Dependence
Happy Ending
interactive voice response
nicotine replacement therapy
randomized controlled trial
short message service