A review of web based interventions for managing tobacco use.

Web based interventions (WBIs) have been developed for various health conditions. These include interventions for various psychoactive substance use disorders including tobacco and alcohol. Tobacco use has remained the single largest preventable cause of global mortality and morbidity for many years. It is responsible for around 6 million deaths annually world-wide. Ironically, most of the tobacco users reside in resource poor low and middle-income countries. The article reviews the existing literature on WBIs for management of tobacco use. The literature search was performed using MedLine, PubMed, PsycINFO, Embase and Cochrane Review for relevant English language articles published from 1998 up to 2013. There is limited support for effectiveness of WBIs for managing tobacco use among adolescents. Although most of the trials among adults found WBIs to be more effective at short term follow-up (a few days to weeks), the benefits failed to extend beyond 3 months in most of the studies. All but one interventions studied in a randomized controlled trial is for smoking forms.

INTRODUCTION

Since its commercialization in early 1990s internet usage has been on a constant rise.[1] Use of internet has impacted various aspects of human life. This includes use of internet in health care service delivery. Web based interventions (WBIs) have been developed for various health conditions. These include interventions for various psychoactive substance use disorders including tobacco and alcohol. Whereas a variety of online eHealth tools emerged to help people manage their health, relatively little was known about their effectiveness until the middle of the last decade.[2] However, there has been an increase in interest in assessing the effectiveness of these interventions over the past few years.

The public health impact of an intervention is a product of the program's efficacy and reach.[3] Thus, the impact is highest when an intervention is effective and has a wide reach. Web based health interventions have a potential to reach a large section of the population. This mode of delivery ensures that large numbers of individuals can be reached at lower costs than with face-to-face interventions.[4] Furthermore WBIs enables the end users to access large amounts of information at a pace and time of their own convenience.[5] Recent findings also support use of online support groups by individuals with lower income as well.[6] This finding is contrary to earlier evidence[7] and supports the public health utility of WBIs.

Tobacco use has remained the single largest preventable cause of global mortality and morbidity for many years.[8] It is responsible for around 6 million deaths annually world-wide. Ironically, most of the tobacco users reside in resource poor low and middle income countries (LMIC). There are limited health care facilities and a dearth of human resources to deliver the services in these countries. In spite of a high proportion of tobacco users’ willingness to quit (approximately 70%) only a minority are able to quit and maintain abstinence.[9] Moreover, the increasing demand of today's busy life leave little time, if any, for even those with resources to access the traditional models of health care service delivery. Hence, WBIs for tobacco use management are likely to be of potential use in various settings including high income and LMICs.

The article reviews the existing literature on WBIs for management of tobacco use.

METHODOLOGY

Literature search

The literature search was performed using MedLine, PubMed, PsycINFO, Embase and Cochrane Review for relevant English language articles published from 1998 up to 2013. Key search terms used in the search were: ([“Online Systems” OR “Internet” OR “Web” OR “Computer”] AND [“Smoking Cessation” OR “Tobacco cessation” OR “Nicotine”] AND [“Randomized Controlled Trial”]). Only publications focused on managing smoking cessation through WBIs were included.

Selection of studies

The studies utilizing solely WBIs that were fully automated and excluded those that required additional elements, such as having face-to-face components or being delivered through intranet or mobile phone.

Titles and abstracts of all potentially relevant articles were reviewed for possible inclusion. Articles were included if (1) the primary intervention was delivered and accessed via the Internet, (2) the intervention focused on curtailing tobacco consumption and (3) the study was a randomized controlled trial (RCT) of an tobacco-related screen, assessment, or intervention with at least a no-treatment control.

Trials using internet only for recruitment or to remind participants of appointments for treatment but not for delivering tobacco cessation intervention were excluded.

Data extraction and analysis

Both authors independently carried out data extraction. The main outcome measure of interest was smoking cessation (e.g., motivation to quit, point prevalence [PP] abstinence and/or prolonged abstinence). Where data was insufficient or not available in the published paper or by contacting authors, studies were excluded from the relevant analysis. Articles describing the study protocols and dissertations were also excluded from analysis.

RESULTS

A total of 281 potentially relevant records were identified. Out of these seven were reviews or meta-analysis. A total of 28 studies evaluating internet sites with/without co-interventions were included in this review. The characteristics of the studies and participants, results of quality assessment and key findings are described below [Figure 1].

Figure 1

Flow chart depicting the data extraction procedure

CHARACTERISTICS OF INCLUDED STUDIES

Recruitment

Fifteen studies were solely from USA.[101112131415161718192021222324] One study each from Denmark[25] and Republic of Ireland,[26] two each from Norway[2728] and Switzerland;[2930] and four from Netherlands[31323334] were found. The studies by Muñoz et al. recruited from 74[35] to 68[36] countries, whereas one study was based in both USA and Canada.[37]

Overall the studies revealed data from more than 40,000 participants with sample sizes ranging from 61[25] to nearly 12,000.[29] The participants were mostly smokers motivated to quit smoking, who chose the Internet as a tool for smoking cessation support. Only one study focused on intervention directed toward smokeless tobacco (ST) users.[37] Thirteen studies recruited adults and four studies recruited adolescents or college students.[10141522] There were more female than male participants overall.

The subjects were recruited mainly from the web with the participants finding the WBI through online browsing or through search engines.[23456789101314151617182425262728293031323435363738] Other recruitment strategies included recruitment through the non-internet based advertising (such as television commercials, radio and newspaper announcements and flyers displayed in the schools and clinics at each respective site),[121522] a combination of non-internet based advertising and personal referrals from teachers or coaches,[14] a combination of web based and non-web based advertisements,[21323334] from attendees of lung cancer screening set up,[11] members of cancer research institute network,[1926] members of health care organization.[20]

Some studies reported use of monetary incentives to encourage adherence.[1014222432333435]

Selection bias

Risk of selection bias also varied across studies. Design of most of the studies resulted in low selection bias.[1013171819202127282932333536] However, some studies failed to provide details of the randomization process.[1112141522242634]

Nature of interventions

The nature of WBIs studied across these studies varied. These varied from included low intensity interventions such as list of web sites on smoking cessation with brief description of each site[11] to extensive tailored cumulative variants of an WBI.[36]

The intensity and rigorousness of interventions also varied. Some interventions used weekly visits to study web site.[10]

Some interventions included only E-mail reminders to the participants.[10] Other interventions used even more intense reminders in form of E-mail, web pages, interactive voice response, and short message service technology.[2728]

Although majority of the interventions used a fixed intervention module, some used a tailored approach.[20242932333436]

Four studies used pharmacotherapy along with WBI.[1228] In a study by Japuntich et al.[12] used Buprenorphine-SR (only for active arm). Three studies allowed use of NRT along with WBI.[192628] Studies by Strecher et al.[1926] assessed efficacy of WBI to support NRT assisted quit attempts. Brendryen and Kraft[28] offered NRT free of charge to both active and control arms of the trial. A study done by Swan et al.[20] offered 12 week free supply of varenicline to all participants.

Characteristics of study participants

The characteristics of participants also varied across the studies. Some studies included only active smokers.[101112131517192021222426272833343536] Others included both current as well as past smokers.[2930] Stoddard et al.[18] included active smokers and recent quitters.

Some studies included willingness to quit in near future (5 days to 30 days) as one of the inclusion criteria.[13181921262728333536] Smit et al.[34] in their study included subjects who were willing to quit within next 6 months and Te Poel et al.[32] included smokers who were willing to quit in next 1 year.

Overall more than 50% of the study participants were females across all studies. Almost all studies had preponderance of female participants. The highest proportion of female subjects was 75.4% in control group and 70.4% in intervention group in study by An et al.[10] The lowest proportion of female subjects was 41.3% in the study by Muñoz et al.[35]

Type of tobacco products

Almost all studies focused on cigarette smokers. However, study by Te Poel et al.[32] specifically mentioned inclusion of smokers of cigarettes and/or loose-cut tobacco. Only one study included users of ST forms.[37]

Outcomes studied

Twenty-one studies assessed smoking status at follow-up lasting at least 6 months after the start of the intervention.[1011121314161719202223242527283233343536] The longest follow-up was of 18 months.[24] Most studies reported intermittent assessments also. Six studies followed participants for <6 months.[182126293031] Only one study assessed cessation for ST and made assessments at end of 3 and 6 months.[37]

Multiple definitions of abstinence were employed for attributing outcome. Mostly 7-day abstinence was the main or secondary outcome measure while 30-day and 28-day continuous abstinence rates were also used in some studies. The longest abstinence duration assessed was that of continued abstinence for 12 months[33] and the shortest duration assessed was 24-h PP of abstinence.[34] Biochemical markers (CO levels) to confirm abstinence were also used by some studies.[101215] One study used cotinine assessments to validate reports of abstinence in a subsample.[33]

Nearly all the studies had used intent-to-treat (ITT) analysis.

Drop our rates

Proportion of the subjects completing the trial varied across the studies. It ranged from a high of 100%[11] to a low of <50%.[131718293233343536] Five studies had a retention rate of more than 80%.[1012192728] and six studies had a retention rate ranging from 50% to 80% respectively.[14152021222426] the study with longest follow-up of 18 months had a follow-up rate of 68.2%.[24]

WBI TRIALS AMONG ADOLESCENTS AND COLLEGE STUDENTS

A total of four studies recruited adolescents or college students [Table 1].[10141522]

Table 1

Trials of WBIs for management of tobacco use (arranged alphabetically by author name)

One cessation induction study[10] in college students compared the WBI (named RealU) with one-off untailored e-mail. It found significantly higher PP abstinence at 7 months for the WBI (RealU 59.1% vs. one-off untailored e-mail 38.5% [relative risk (RR) = 1.54, 95% of confidence interval (CI): 1.28-1.85]). However, there was no difference between groups (overall 6%) for prolonged abstinence at 7 months.

Patten et al.[15] studied adolescent smoking cessation and did not detect any difference in abstinence among the study groups. The 30-day, point-prevalence smoking abstinence rates for brief office intervention (BOI) and stomp out smokes (SOS) intervention were comparable (12% vs. 6% at week 24 and 13% vs. 6% at week 36 for BOI and SOS, respectively). Although the SOS group had a significantly greater reduction in average number of days smoked than BOI (P = 0.006).

Another study among adolescents compared American Lung Association's Not on Tobacco program (NOT) with a Web-based adjunct (NOT Plus). The study utilized hierarchical linear modeling and detected a significant effect of NOT Plus to the comparator group for smoking cessation.[14] Among the Web-based adjunct users, there was a significant association of use of the web site with smoking cessation at end-of-program (P < 0.05). However, it was not observed at 3 months.

Another study among high school students utilized an Internet-based, virtual reality world intervention comparing it with motivational interviewing (MI) by a counselor.[22] Intervention participants had significantly higher 7-day abstinence rates than controls (35% vs. 22%). There was no difference among the groups in smoking abstinence at 12 month follow-up (RR 0.93, 95% CI: 0.60-1.44), although the number of quit attempts were significantly different between the groups (P < 0.05).

WBI TRIALS AMONG ADULTS

There were a total of 24 trials among adult populations.[111216202123242527283133343536]

Comparison of WBI to non-WBIs or no interventions at all

A study among 171 smokers during lung cancer screening test found effect of self-help materials similar to a recommended written list of Internet resources. The 7-day PP quit rates were comparable between the two groups (5% vs. 10%; RR 0.45, 95% CI: 0.14-1.40) at 1-year follow-up.[11]

Another study using Comprehensive Health Enhancement Support System for Smoking Cessation and Relapse Prevention as an adjunct to counseling and bupropion (N = 284, 140 WBI, 144 control) failed to find a significant improvement in abstinence rates at 6 months (RR 1.27, 95% CI: 0.70-2.31).[12]

One short-term follow-up (90 days) study in 351 participants detected a significant effect of WBI compared with no intervention at all (24.1% vs. 8.2%; RR 2.46, 95% CI: 1.16-5.21).[21]

The Happy Endings trials recruited 290 participants (144 intervention, 146 control) via Internet advertisements.[2728] In the first of two studies, the intervention was delivered as 1-year smoking cessation program through the Internet and cell phone while controls received a self-help booklet.[27] The second study offered nicotine replacement therapy (NRT) to both groups.[28] WBI group was significantly more effective on sustained abstinence at 12 months compared with self-help controls without adjunct NRT (20% vs. 7%, 7%; RR 2.94, 95% CI: 1.49-5.81) as well as with adjunct NRT (22.3% vs. 13.1%, RR 1.71, 95% CI: 1.10-2.66).

The study by Oenema et al.[31] showed that the WBI was not more effective than providing no intervention at all for self-reported smoking status at 1 month (RR = 1.28, 95% CI: 0.52-3.13).

Another cessation induction study evaluated the effectiveness of Online Transtheoretical Model tailored communications and MI with the adjunct of Health Risk Intervention.[16] This study reported statistically insignificant difference in PP abstinence at 6 months between the two groups (21% vs. 35%).

In a study among general practitioners, no significant additional effect of referral to group-based (Odds ratio [OR]: 1.05; 95% CI: 0.6-1.8) or internet-based smoking cessation programs (OR: 0.91; 95% CI: 0.6-1.4) was noted among smokers (n = 760) at 1-year follow-up.[25]

Swan et al.[20] compared WBI (n = 401) to proactive telephone counseling (PTC) (n = 402), or as an adjunct (n = 399) with varenicline in all groups. The study failed to find any significant difference in 7-day point prevalent abstinence at 6 month (30.7% vs. 34.3% vs. 33.8%; RR 0.94, 95% CI: 0.79-1.13). Although the PTC group was found to be significantly more effective with regards to PP abstinence than the WBI group at 3 months follow-up (OR = 1.48, 95% CI: = 1.12, 1.96).

A similar design employed in Quit Using Internet and Telephone Treatment study comparing static WBI either to tailored WBI or as adjunct to PTC also failed to find any significant difference in 30-day single PP abstinence rates at 18 months.[24]

Quitting is Winning, a cognitive-behavioral program evaluated among Korean Americans (n = 1112), found no significant difference in 30-day smoking cessation rates between the Internet (11%) and booklet (13%) groups (ITT difference = −2%, 95% CI: = −6% to 2%).[23]

The stay quit for you study (n = 2031) reported two differently tailored web-based smoking relapse prevention programs (action planning [AP] and AP plus program) to be significantly more effective than the control group (participants with no intervention at all) for self-reported continued abstinence at 12 months follow-up (AP program [OR: 1.95, P = 0.005], AP + program [OR: 1.61, P = 0.049]).[33] The study also suggested that the number of completed program elements had a dose-response relationship to abstinence rates.

Another recent study reported tailored WBI (n = 552) to be more effective than no intervention at all (n = 571) at 6 weeks (24-h PP abstinence [OR: 1.85, 95% CI: 1.30-2.65], 7-day PP abstinence [OR: 2.17, 95% CI: 1.44-3.27] and prolonged abstinence [OR: 1.99, 95% CI: 1.28-3.09]). However, the difference failed to extend until end of 6 months.[34]

Comparison of different WBIs

Studies comparing tailored to untailored WBIs report ambiguous findings. While few studies have found that tailored WBI is equally efficacious to untailored ones[131718303536] others have reported tailored intervention to be better.[192632] Two studies reporting significantly effective results for tailored WBIs assessed outcomes as 24-h or 7-day PP at 6 months.[1932] One short term (12 weeks) study also reported better continuous abstinence rates for tailored WBIs.[26] A study by Muñoz et al.[35] found more complex intervention to be having significantly lower quit rates at 12 months.

WBIs for ST

The Chew Free trial is the only study conducted for ST cessation.[37] Participants were recruited online through MyLastDip program and provided two fully automated WBI as tailored (n = 857) or static (n = 859) text. The tailored WBI had significantly higher quit rates than static condition at 6 months assessed using complete case analysis (40.6% vs. 21.2%; P < 0.001) and ITT analysis (12.6% vs. 7.9%; P < 0.001).

Reviews and meta-analysis

We included 7 reviews and meta-analysis in the current review.[39404142434445]

The review by Strecher et al.[43] included 10 RCTs comparing tailored versus the general or targeted modalities. They reported of very few characteristic patterns amongst the studies, strong impact on smoking cessation by trials combining tailored materials with nicotine replacement therapy and a significant positive impact of tailored materials among pre-contemplators.

Another review identified 19 studies of computer and Internet-based interventions for smoking cessation published between 1995 and August 2004.[45] The authors reported of few patterns in terms of subject, design or intervention characteristics leading to positive outcomes. The mailed computer-generated feedback report intervention format was observed to be most consistently associated with improved outcomes.

Shahab & McEvan[42] conducted systematic review and meta-analysis of the literature (1990–2008) and included 11 RCTs. They concluded that although there was no overall effect of interactive compared with static WBIs, web-based-tailored-interactive smoking cessation interventions were effective compared with untailored booklet or e-mail interventions [rate ratio (RR) 1.8; 95% confidence interval (CI) 1.4–2.3] increasing 6-month abstinence by 17% (95% CI 12–21%) without any publication bias. Effective Interventions were those that were only aimed at smokers motivated to quit (RR 1.3, 95% CI 1.0–1.7) or were fully automated (RR 1.4, 95% CI 1.0–2.0).

One meta-analysis of 22 RCTs found that Web- or computer-based smoking cessation programs yielded an abstinence rate about 1.5 times higher than controls (RR, 1.44; 95% CI, 1.27%-1.64%).[44] Pooled analysis revealed significantly higher abstinence rate at 12-month follow-up in the intervention group (9.9%; 95% CI, 8.9%-10.9%) than the control group (5.7%; 95% CI, 5.1%-6.3%). Both stand alone or supplemental interventions were effective in adults but none was significantly effective in adolescents.

There have been multiple Cochrane reviews concerning with smoking cessation but only 2 focused on web based interventions.[3940] One review included 20 RCTs including trials with WBIs only.[40] It concluded that though the trials did not show consistent effects, few WBIs having tailored information or frequent automated contacts with the users can effectively assist smoking cessation. There was no benefit detected of including a mood management component or an asynchronous bulletin board. Another recent Cochrane effectiveness review concluded that computer and other electronic aids increase the likelihood of cessation (aid to cessation as well as cessation induction studies) compared with no intervention or generic self-help materials, but the effect is small (prolonged abstinence: relative risk = 1.32, 95% confidence interval 1.21 to 1.45).[39] It was also observed that the chances of sustaining abstinence increases noticeably after successfully negotiating the first month.

The review by Hutton et al.[41] included 21 RCTs with 31,481 smokers. The studies in adults were observed to be of moderate quality with retention rates ranging from 27% to 86%, ninety percent in college students and losses to follow-up in 13% to 47% among the adolescents. The authors concluded that the evidence supporting the use of WBIs for smoking cessation is insufficient to moderate in adults and insufficient in college students and adolescents.

Cost effectiveness of WBIs

Some of the RCTs assessing effectiveness of WBIs for managing tobacco use have also commented on the cost-effectiveness of the intervention.

Etter[29] reported that the cost of implementing WBI for management of tobacco sue for a reach of 8000 participants in computer tailored programs (with 600,000 visitors per year to the website) is comparable to the cost of running a small smoking cessation clinic which would treat about 50 smokers a month.

Rabius et al.[17] in their study have reported WBI for management of tobacco use to be cost effective. In this study, 4 days of programming at a cost of less than US $2000 allowed approximately 5000 additional users for services from the five tailored interactive service providers. The cost was much less than the cost of serving 1000 new clients with telephone counseling amounting to approximately US $100,000.

CONCLUSIONS

Recent years have seen a significant growth in number of WBIs in field of health care service delivery. The same is true for psychoactive substance use disorders. The WBIs for some substances of abuse (alcohol and tobacco) is more researched than others.

Limited number of specialized health care professional, busy schedules due to demands of profession and increasing penetration of internet to cities as well as villages make the WBIs an option worth exploring in LMIC settings. These interventions are expected to be cost effective due to limited recurring and maintenance cost, especially keeping in mind the large consumer base. Although a limited number of studies have commented on the cost-effectiveness of WBIs for management of tobacco use, the findings are suggestive of substantial lower cost of offering such services.[1729]

Accessibility of internet through the hand held devices has offered even cheaper alternatives to computers and laptops. The freedom to access the service at one's own convenient time and pace also makes these a lucrative area to invest and investigate.

Previous estimates of potential reach of WBIs for smoking cessation were criticized for being based on either national figures for Internet access or reported interest among non-representative samples.[39] However, even the newer studies, assessing a representative sample of smokers, have estimated that 40-46% was interested in using a WBI for smoking cessation.[4041] Thus in this era of internet usage the applicability of WBIs for managing tobacco use is encouraging.

However, the existing evidence on WBIs for management of tobacco use is limited. These are restricted mainly to the developed world. Most of the interventions are in English. None have been developed in other languages spoken in LMICs. In addition, all but one intervention studied in a RCT is for smoking forms. ST use constitutes a major fraction of the tobacco used in LMICs including India.

In addition, the quality of trials is also heterogeneous. Some of the studies are likely to have high/uncertain risk of selection bias. Some of the studies have relatively small follow-up duration. Only a few studies have included biochemical markers as objective measure for ongoing tobacco use. Only a handful of studies have assessed the WBIs in combination with pharmacotherapy (including NRT, bupropion-SR and varenicline).[1219202628]

There is limited support for effectiveness of WBIs for managing tobacco use among adolescents. Although most of the trials among adults found WBIs to be more effective at short term follow-up (a few days to weeks), the benefits failed to extend beyond 3 months in most of the studies. The notable exception to this were studies by Brendryen et al.[2728] and Prochaska et al.[16] Even among the studies that have reported superiority of WBIs, the effect size is small.[42]

Randomized trials of WBIs are also limited by weakness due to generalized factors applicable to internet use itself. People who choose to participate in the WBIs based trials may be unrepresentative of the people who use websites in real life thereby raising concern about whether the same type of result would be obtained by all users of the website. The qualitative value of WBI may also be hindered by the fact that having WBI at health care setting may make participant compelled to fill out rather than when used the same service at home.

The findings from the exiting RCTs can help develop more refined WBIs for managing tobacco use. Use of tailored materials among pre-contemplators; combining tailored materials with NRT, mailed computer-generated feedback report intervention format; and successful negotiation of the 1st month have been found to be effective strategies and approaches across the existing studies.