Telephone counselling for smoking cessation

William Matkin; José M Ordóñez‐Mena; Jamie Hartmann‐Boyce

doi:10.1002/14651858.CD002850.pub4

. 2019 May 2;2019(5):CD002850. doi: 10.1002/14651858.CD002850.pub4

Telephone counselling for smoking cessation

William Matkin ¹, José M Ordóñez‐Mena ^2,^✉, Jamie Hartmann‐Boyce ²

Editor: Cochrane Tobacco Addiction Group

PMCID: PMC6496404 PMID: 31045250

Abstract

Background

Telephone services can provide information and support for smokers. Counselling may be provided proactively or offered reactively to callers to smoking cessation helplines.

Objectives

To evaluate the effect of telephone support to help smokers quit, including proactive or reactive counselling, or the provision of other information to smokers calling a helpline.

Search methods

We searched the Cochrane Tobacco Addiction Group Specialised Register, clinicaltrials.gov, and the ICTRP for studies of telephone counselling, using search terms including 'hotlines' or 'quitline' or 'helpline'. Date of the most recent search: May 2018.

Selection criteria

Randomised or quasi‐randomised controlled trials which offered proactive or reactive telephone counselling to smokers to assist smoking cessation.

Data collection and analysis

We used standard methodological procedures expected by Cochrane. We pooled studies using a random‐effects model and assessed statistical heterogeneity amongst subgroups of clinically comparable studies using the I² statistic. In trials including smokers who did not call a quitline, we used meta‐regression to investigate moderation of the effect of telephone counselling by the planned number of calls in the intervention, trial selection of participants that were motivated to quit, and the baseline support provided together with telephone counselling (either self‐help only, brief face‐to‐face intervention, pharmacotherapy, or financial incentives).

Main results

We identified 104 trials including 111,653 participants that met the inclusion criteria. Participants were mostly adult smokers from the general population, but some studies included teenagers, pregnant women, and people with long‐term or mental health conditions. Most trials (58.7%) were at high risk of bias, while 30.8% were at unclear risk, and only 11.5% were at low risk of bias for all domains assessed. Most studies (100/104) assessed proactive telephone counselling, as opposed to reactive forms.

Among trials including smokers who contacted helplines (32,484 participants), quit rates were higher for smokers receiving multiple sessions of proactive counselling (risk ratio (RR) 1.38, 95% confidence interval (CI) 1.19 to 1.61; 14 trials, 32,484 participants; I² = 72%) compared with a control condition providing self‐help materials or brief counselling in a single call. Due to the substantial unexplained heterogeneity between studies, we downgraded the certainty of the evidence to moderate.

In studies that recruited smokers who did not call a helpline, the provision of telephone counselling increased quit rates (RR 1.25, 95% CI 1.15 to 1.35; 65 trials, 41,233 participants; I² = 52%). Due to the substantial unexplained heterogeneity between studies, we downgraded the certainty of the evidence to moderate. In subgroup analysis, we found no evidence that the effect of telephone counselling depended upon whether or not other interventions were provided (P = 0.21), no evidence that more intensive support was more effective than less intensive (P = 0.43), or that the effect of telephone support depended upon whether or not people were actively trying to quit smoking (P = 0.32). However, in meta‐regression, telephone counselling was associated with greater effectiveness when provided as an adjunct to self‐help written support (P < 0.01), or to a brief intervention from a health professional (P = 0.02); telephone counselling was less effective when provided as an adjunct to more intensive counselling. Further, telephone support was more effective for people who were motivated to try to quit smoking (P = 0.02). The findings from three additional trials of smokers who had not proactively called a helpline but were offered telephone counselling, found quit rates were higher in those offered three to five telephone calls compared to those offered just one call (RR 1.27, 95% CI 1.12 to 1.44; 2602 participants; I² = 0%).

Authors' conclusions

There is moderate‐certainty evidence that proactive telephone counselling aids smokers who seek help from quitlines, and moderate‐certainty evidence that proactive telephone counselling increases quit rates in smokers in other settings. There is currently insufficient evidence to assess potential variations in effect from differences in the number of contacts, type or timing of telephone counselling, or when telephone counselling is provided as an adjunct to other smoking cessation therapies. Evidence was inconclusive on the effect of reactive telephone counselling, due to a limited number studies, which reflects the difficulty of studying this intervention.

Plain language summary

Does telephone counselling help people stop smoking?

Background

There are a number of interventions available to help people stop smoking. One of them is using telephone calls to give smokers information, advice, and help to stop smoking. People can use these services by calling quitlines or by signing up to get calls from counsellors. We wanted to find out whether telephone counselling can help people quit smoking. Our most recent search for evidence was in May 2018.

Study characteristics

We found 104 studies (including 111,653 participants) testing the effect of any type of telephone counselling. The participants were mostly adult smokers from the general population, but some studies also looked at teenagers, pregnant women, and people with long‐term or mental health conditions.

Some studies included participants who had called helplines that provide smoking counselling (quitlines). Other studies included people who had not called quitlines, but received calls from counsellors or other healthcare providers.

Some studies provided telephone counselling alone, but many others provided telephone counselling along with minimal support such as self‐help leaflets, or more active support such as face‐to‐face counselling, or with stop‐smoking medication. The number of calls offered ranged from a single call to 12 calls. Some studies only recruited people trying to stop smoking, while others offered support even to those not actively trying to stop.

Studies needed to compare groups whose participants had similar characteristics at the start of the study, to investigate whether the participants had stopped smoking for at least six months, and ideally would test whether people had quit with blood or urine tests.

We judged few studies to be well designed and conducted. Most had at least one issue that could have affected the results.

Key results

In people who had called helplines, providing additional telephone counselling increased their chances of stopping smoking from 7% to 10%. In people who had not called a helpline, but received telephone calls from counsellors or other healthcare providers, their chances of stopping smoking increased from 11% to 14%. In studies which directly compared more versus fewer calls, people who were offered more calls (three to five) tended to be more likely to quit than those who received only one call. Telephone counselling appears to increase the chances of stopping smoking, whether or not people are motivated to quit or are receiving other stop‐smoking support.

Certainty of evidence

The overall certainty of the evidence was moderate, meaning that further research is likely to have an important impact on our conclusions.

Summary of findings

Background

Description of the condition

Worldwide, tobacco smoking kills more than seven million people each year, shortening life by an average of 10 to 11 years in people who smoke their whole lives (Doll 2004; Pirie 2013; WHO 2018). This increase in mortality is primarily due to cardiovascular disease, lung cancer, and chronic obstructive pulmonary disease, but smoking is also causally associated with other cancer‐ and non‐cancer‐related health conditions (USDHHS 2014; WHO 2018). Smoking cessation reverses much of the damage, and most smokers who are aware of these health hazards wish to quit (Doll 2004; Mons 2015; Müezzinler 2015; Ordóñez‐Mena 2016; WHO 2018).

Description of the intervention

Telephone counselling is a behavioural intervention to help people stop smoking. It can act as a stand‐alone intervention or one that runs alongside other interventions (e.g. an added component to face‐to‐face counselling). It can be proactive (the counsellor initiates contact) or reactive (the smoker initiates contact) (Lichtenstein 1996). In many areas, reactive services are available through helplines or quitlines, which may be specific to smoking, as for example the California Smokers' Helpline (Zhu 2000a) and the Quitlines in Australia (Borland 2001), the UK (Owen 2000), Sweden (Lindqvist 2013), and Denmark (Skov‐Ettrup 2016), or they may be embedded in broader health information services such as the Cancer Information Service in the USA (La Porta 2007). Quitlines may provide a regional or national service. They are often advertised in conjunction with population‐wide campaigns such as No‐Smoking Days. Helplines may also be provided on a smaller scale for a specific project or population. In some services, people may be enrolled in a formal smoking cessation programme, with further proactive calls from counsellors. Telephone counselling may also be provided as part of an integrated smoking cessation support service (e.g. Glasgow 1991). Access to hotlines or the opportunity to register to receive calls from a counsellor may also be offered as part of a cessation programme, including pharmacotherapy.

How the intervention might work

Behavioural and pharmacological interventions help people to quit smoking. Behavioural support can be given in individual counselling sessions (Lancaster 2017) or in group therapy (Stead 2017), where clients can share problems and derive support from one another. Standard self‐help materials have at best a small effect in helping people to quit, while those tailored to the characteristics of individuals are more likely to be effective (Livingstone‐Banks 2019a). Telephone counselling may supplement face‐to‐face support, or be a substitute for it as an adjunct to self‐help interventions or pharmacotherapy. Counselling may be helpful in planning a quit attempt, and in helping to prevent relapse during the initial period of abstinence (Livingstone‐Banks 2019b). Although intensive face‐to‐face interventions increase quit rates, there are difficulties with scalability. Telephone counselling may be a more feasible way to provide individual counselling to large populations. In addition, telephone contact can be timed to maximise the level of support around a planned quit date, and can be scheduled in response to the needs of the recipient. In a proactive approach the counsellor initiates one or more calls to provide support in making a quit attempt or avoiding relapse. Reactive counselling, in contrast, is available on demand to people calling specific services; quitlines, helplines or hotlines. These services take calls from people who smoke, or their friends and family (Zhu 2006). These telephone services may offer information, recorded messages, personal counselling or a mixture of components (Anderson 2007; Ossip‐Klein 2003).

Why it is important to do this review

Telephone counselling services have the potential to provide access to information for large numbers of people. Some services have reported reaching substantial proportions of the target population (Ossip‐Klein 1991; Platt 1997). They have the potential to reach underserved populations such as ethnic minorities (Zhu 2000a) or younger people (Chan 2008; Gilbert 2005).

Telephone counselling is a widely used intervention for smoking cessation, and is often supported through public funds. It is therefore important to evaluate its effects, as well as different variations of telephone counselling that may impact on quit rates.

Objectives

To evaluate the effect of telephone support to help smokers quit, including proactive or reactive counselling, or the provision of other information to smokers calling a helpline.

We tried to address the following questions:

Do telephone calls from a counsellor increase quit rates, compared to other brief non‐pharmacological interventions or to no intervention?
Do telephone calls from a counsellor offered together with other interventions increase quit rates, compared to other interventions alone?
Does an increase in the number of telephone contacts increase quit rates?
Do differences in counselling protocol related to the type or timing of support lead to differences in quit rates?
Does the availability of a reactive helpline increase quit rates?

Methods

Criteria for considering studies for this review

Types of studies

Randomised or quasi‐randomised controlled trials (RCTs), with the unit of allocation being one of the following: the individual smoker; counsellor; group; intervention site; or geographical area.

Types of participants

Individuals who were current smokers at the time of inclusion in the trial. We included trials with a mixture of current smokers and recent quitters if the recent quitters were only a small proportion of the entire study population. The definition of recent quitters was that used by the trial recruitment protocols, or by the participants themselves. We excluded trials that exclusively recruited quitters or were focused on telephone counselling as an intervention for relapse, as they fall within the scope of a separate Cochrane Review on preventing relapse (Livingstone‐Banks 2019b).

We included trials recruiting exclusively teenagers or pregnant women, but we considered them as a potential source of heterogeneity in meta‐analyses. There are separate Cochrane Reviews for these population groups (Chamberlain 2017; Fanshawe 2017).

Types of interventions

Provision of proactive or reactive telephone counselling to assist smoking cessation, to any population. We excluded studies if the contribution of the telephone component could not be evaluated independently of another therapy. We included studies that compared a combination of telephone counselling and self‐help materials versus no telephone counselling, as the effect of self‐help materials alone is limited (Livingstone‐Banks 2019a). We also included studies in which the effect of telephone counselling as an adjunct to another smoking cessation treatment was evaluated, e.g. print‐based self‐help, brief face‐to‐face intervention, pharmacotherapy, or incentives. We also included studies that compared different modalities or strategies of telephone counselling, and different theories of behavioural change.

Types of outcome measures

Long‐term smoking cessation (i.e. at least six months after the start of intervention). We excluded trials with shorter follow‐up. We used the strictest definition of smoking cessation available in a trial and biochemically‐validated abstinence data whenever available.

Search methods for identification of studies

We identified studies from the Cochrane Tobacco Addiction Group Specialised Register (until May 3, 2018), World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) and ClinicalTrials.gov (until July 30, 2018) using the MeSH term 'hotlines' or free‐text terms telephone* OR phone* OR quitline* OR helpline. See Appendix 1 for the full search strategy. At the time of the search the Register included the results of searches of the Cochrane Central Register of Controlled trials (CENTRAL), issue 1, 2018; MEDLINE (via OVID) to update 20180404; Embase (via OVID) to week 201814; PsycINFO (via OVID) to update 20180212. See the Cochrane Tobacco Addiction Group website for full search strategies and a list of other resources searched.

Data collection and analysis

We identified trials where at least one of the arms included telephone contact. For this update, two review authors (WM and JMOM) extracted data from included studies, compared extraction and discussed disagreements with a third review author (JHB). We recorded the following information in the Characteristics of included studies tables:

The country and setting of the trial
The method of recruitment to the study
The method of randomisation and allocation concealment
Details of participants, including whether they were selected according to motivation to quit, and their age, gender and average baseline cigarette consumption
Description of intervention and control, including the schedule of telephone contacts
Definition of smoking abstinence used for the primary outcome, including timing of longest follow‐up and whether quit status was based on recent behaviour (point prevalence abstinence, e.g. in the past seven days) or on abstinence for an extended period since a quit date or a previous follow‐up (continuous or sustained abstinence)
Description of method of any biochemical validation or other method used to confirm self‐reported quitting
Description of numbers lost to follow‐up by treatment condition

In the Characteristics of excluded studies tables, we describe key trials that did not meet all inclusion criteria.

Assessment of risk of bias

We assessed each included study for risks of bias in the following domains: random sequence generation (allocation bias), allocation concealment (allocation bias), blinding of outcome assessor (detection bias), and incomplete outcome data (attrition bias). Following standard Cochrane guidance, we rated items in the 'Risk of bias' tables as being at low, unclear, or high risk of bias.

We judged studies to be at low risk of bias for sequence generation if an acceptable method for generating a randomisation sequence was described, at unclear risk if the study was described as 'random' but with no further information, and at high risk if it did not use a true randomisation method, e.g. alternation by calendar week. We judged the quality of the allocation concealment to be at low risk if the group to which a participant was to be allocated remained unknown to investigators and participant until enrolment was complete. We did not assess studies for performance bias, because blinding of participants and study personnel is not possible due to the nature of the intervention. We rated the quality of the blinding of outcome assessment at low risk if a biochemical verification of abstinence was carried out in the entire study population or in a random subsample, or if smoking status was measured by self‐report but the intervention and control arms received similar amounts of person‐to‐person contact, and at high risk otherwise. We judged attrition bias to be at low risk if the percentage of loss to follow‐up was less than 50% and if the difference between arms in percentage loss to follow‐up was less than 20%.

Choice of outcome and treatment of missing data

The primary outcome was the number of quitters at the longest follow‐up, using the strictest measure of abstinence reported. We preferred sustained and biochemically‐validated abstinence to point prevalence or to self‐reported quitting. If a less strict definition of quitting seemed more appropriate for showing an effect of the intervention on recovery from lapses or relapses, we planned a sensitivity analysis.

Where possible and appropriate, we used as denominators the number randomised to each condition, with losses to follow‐up assumed to be continuing smokers. We noted any exceptions in the 'Risk of bias' table for a study. Population‐based studies typically have relatively high loss to follow‐up, because of change of address or disconnected telephones. Non‐response might be independent of both treatment condition and smoking status, although possibly associated with other variables such as age or socioeconomic status. Dropout might be related to smoking status but not to treatment condition. Imputing as smoking all those missing, irrespective of, for example, whether they could not be contacted, or declined to respond, may not be appropriate. For individual studies it is possible to use analysis methods such as generalised estimating equations (GEE) for imputing missing data (Hall 2001). We noted whether studies that explored alternative assumptions about missing data reported any impact on the conclusions. When proportions lost are similar across conditions, and trial arms are balanced, the choice of denominator does not alter the relative effect, although the percentage quit and the absolute difference between conditions will be conservative.

Data synthesis

We summarised individual study results as a risk ratio (RR), calculated as: (number of quitters in intervention group/number randomised to intervention group)/(number of quitters in control group/number randomised to control group). Where appropriate, we performed meta‐analysis using a Mantel‐Haenszel random‐effects model to estimate a pooled risk ratio with a 95% confidence interval (CI). When trials had more than one arm with a less intensive intervention we used only the most similar intervention without a telephone component as the control group in the primary analysis. We considered pooling of study results if both the intervention and control arms were sufficiently similar across trials. We assessed statistical heterogeneity between trials using the I² statistic which describes the percentage of total variation between studies that is due to heterogeneity rather than chance (Higgins 2003). We used threshold values of 50% and 70% as suggesting moderate and substantial heterogeneity respectively.

Investigations of heterogeneity: meta‐regression and subgroup analyses

We ran a meta‐regression in R version 3.5.0 (R 2018) using the meta and metafor packages (Schwarzer 2007; Viechtbauer 2010) to test the moderation of the effect of telephone counselling on smoking cessation by telephone counselling intensity (continuous: defined as maximum number of calls offered as part of the telephone counselling intervention), trial selection of participants that were motivated to quit (binary: yes/no), and type of baseline support offered in both the intervention and control arms (categorical: print‐based self‐help, brief face‐to‐face counselling, pharmacotherapy, or financial incentives). In meta‐regression, the effect size was summarised using the trial‐specific (natural) logarithm of the RR with its standard errors as weights. We fitted these trial characteristics separately in univariate models, and also together in a multivariate model to adjust for each moderator simultaneously.

We did not combine proactive and reactive approaches to counselling, so studies that provided access to a telephone helpline but did not call participants form a separate category. In earlier versions of this review we noted heterogeneity between studies of proactive telephone counselling, which was not explained by using subgroups based on the amount of support given for the control group. Lichtenstein 2002a has suggested that studies recruiting smokers who call quitlines should be considered separately. These studies share the characteristics that participants were actively seeking support at the time of their call, and that telephone counselling was the primary intervention. We therefore distinguish between trials in quitline callers and trials in smokers not calling a quitline.

We expected differences between the relative effect of telephone support, depending on whether it was being tested as the main intervention to aid cessation, or as an extra part of a multicomponent cessation programme. We therefore conducted a priori defined subgroup analysis to distinguish those studies in which telephone counselling was the most intensive component of a minimal contact intervention (print‐based self‐help was provided), from studies in which telephone counselling was assessed as an adjunct to a brief face‐to‐face counselling, or to pharmacotherapy. Where results of studies differed within the broad groupings described above we considered the following possible explanations: the difference between the intensity of the counselling based on the number of calls (two sessions or fewer, three to six sessions, and seven sessions or more), the counselling strategy used, and the characteristics of the participants, in particular their motivation to quit or stage of change at baseline.

'Summary of findings' table

Following standard Cochrane methodology, we created 'Summary of findings' tables for the two main subgroups of participants, i.e. callers to a quitline and those not calling a quitline. Both 'Summary of findings' tables include data for the same primary outcome of long‐term smoking cessation. Also following standard Cochrane methodology, we used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the certainty of the body of evidence for each outcome, and to draw conclusions about the certainty of evidence within the text of the review.

Results

Description of studies

One hundred and four studies met the criteria for inclusion in the review, with a total of 111,653 participants, and a median trial size of 735. Only seven studies had fewer than 100 participants (Brown 1992; Cossette 2011; Duffy 2006; Ebbert 2007; McClure 2011; Osinubi 2003; Vander Weg 2016), whilst seven studies, all involving callers to quitlines, had more than 3000 (Hollis 2007; Joyce 2008; Rabius 2004; Rabius 2007; Sherman 2017; Zhu 1996; Zhu 2002).

The most recent search resulted in 511 studies to screen (Figure 1). After title, and abstract, and then full‐text screening, we found 30 new studies to include in this update, plus 15 new ongoing studies.

Study flow diagram for most recent update

Most trials were conducted in North America (72). Nine were in Australia (Borland 2001; Borland 2003; Borland 2008; Brown 1992; Girgis 2011; MacLeod 2003; Tzelepis 2011a; Young 2008; Zwar 2015), six in Canada (Chouinard 2005; Cossette 2011; Reid 1999a; Reid 2007; Reid 2018; Smith 2004), three in Spain (Miguez 2002; Miguez 2008; Ramon 2013), three in the UK (Aveyard 2003; Gilbert 2006; Ferguson 2012), two in Hong Kong (Abdullah 2005; Chan 2015), two in Germany (Flöter 2009; Metz 2007), two in Sweden (Lindqvist 2013; Nohlert 2014), one in Norway (Hanssen 2009), one in Malaysia (Blebil 2014), one in the Netherlands (Schuck 2014), one in Denmark (Skov‐Ettrup 2016), and one in China (Wu 2017). Participants were predominantly older adults with an average age typically in the 40s. One study recruited teenagers (Lipkus 2004), one high school students (Peterson 2016), one young people aged 18 to 24 (Sims 2013), and three recruited older people, aged over 50 (Rimer 1994), over 60 (Ossip‐Klein 1997), or over 65 (Joyce 2008). Four recruited pregnant women (Cummins 2016b; McBride 1999b; McBride 2004; Stotts 2002) and a further five recruited only women (McBride 1999a; McClure 2005; Flöter 2009; Solomon 2000; Solomon 2005). Four predominantly recruited men (Abdullah 2005; An 2006; Osinubi 2003; Sorensen 2007a). One was culturally tailored for Chinese, Korean and Vietnamese smokers (Zhu 2012) and one recruited Arabic smokers in Australia (Girgis 2011).

Most of the studies were trials of proactive calls from a counsellor, or from an automated interactive voice response system (IVR) (Reid 2007 IVR with counsellor follow‐up in case of need, Velicer 2006 IVR only). Only five assessed interventions that did not involve a counsellor contacting a participant (McFall 1993; Orleans 1998; Ossip‐Klein 1991; Sood 2009; Thompson 1993). Sixteen studies recruited participants who had phoned a quitline, but evaluated the addition of further proactive contacts (Borland 2001; Borland 2003; Bricker 2014; Cummins 2016b; Ferguson 2012; Gilbert 2006; Hollis 2007; Lindqvist 2013; Nohlert 2014; Rabius 2004; Rabius 2007; Sims 2013; Smith 2004; Zhu 1996; Zhu 2002; Zhu 2012). One study recruited proactively to quitline counselling (Tzelepis 2011a). Sixteen studies recruited participants in healthcare settings and referred them to services provided by quitlines, involving proactive counselling for those following through referral (Bastian 2012; Blebil 2014; Borland 2008; Brunette 2017; Collins 2018; Cummins 2016a; Duffy 2006; Ebbert 2007; Piper 2016; Ramon 2013; Rogers 2016; Schlam 2016; Sherman 2017; Warner 2016; Wu 2017; Zwar 2015). Ellerbeck 2009 repeatedly mailed primary care patients an offer of free pharmacotherapy and tested two levels of disease management, including proactive calls or no contact. One study offered either a proactive or reactive service as covered benefit (Joyce 2008). Additional details are in the Characteristics of included studies tables.

The number, duration and content of the telephone calls was variable. The potential number of calls ranged from one to 12 and in some studies was flexible. The duration of the calls also varied; 10 to 20 minutes was common, although the initial call might be longer. The call schedule could be spaced over weeks or months. Amongst studies that did not recruit participants on the basis of their willingness to make a quit attempt, the content was typically individualised to enhance motivation in those undecided about quitting, or to support a quit attempt where appropriate. Counselling was most commonly provided by professional counsellors or trained healthcare professionals. One trial used trained postgraduate students (Aveyard 2003). Three trials used trained peer counsellors, in one case survivors of childhood cancer (Emmons 2005), and in the other two women ex‐smokers (Solomon 2000; Solomon 2005).

Many trials reported both short‐term point prevalence abstinence (seven‐day or 24‐hour) and sustained abstinence, at one or more longer follow‐ups. Long‐term sustained abstinence was available for 53 of the 104 trials (51%). For the remainder, the outcome was based on point prevalence abstinence at the longest follow‐up. Length of longest follow‐up ranged from six months from start of intervention (35 trials), to seven years (Peterson 2016).

We grouped trials into three broad categories: trials of interventions for smokers who contacted a helpline; trials assessing the effect of providing access to a helpline; and trials that offered support proactively in other settings. Finally there are 10 trials that do not fit into any of these categories, so are considered individually (Bastian 2012; Collins 2018; Halpin 2006; Klemperer 2017; Reid 2018; Smith 2013; Sumner 2016; Vander Weg 2016; Warner 2016; Wu 2017).

Trials of interventions for people calling helplines

Nineteen trials recruited people who had phoned helplines/quitlines. We distinguished between trials where the intervention involved further proactive contact by the counsellor, and those that tested different interventions at the initial call. Fourteen studies tested proactive calls back to people who had initiated the contact with the quitline. The number of calls varied, with four studies comparing more than one schedule (Hollis 2007; Rabius 2007; Smith 2004; Zhu 1996). There were small differences in the support for the control group. In one trial, all participants had brief counselling during their initial call (Borland 2001); in three, some control group participants received some counselling (Borland 2003; Gilbert 2006; Zhu 2002). Ferguson 2012 was a factorial trial comparing proactive counselling to standard support, which included further contact by email, letter or text, and the offer of proactive calls. Participants were also randomised to an offer of nicotine replacement therapy (NRT). In the others the control group received self‐help materials (Cummins 2016b; Hollis 2007; Nohlert 2014; Rabius 2004; Rabius 2007; Sims 2013; Smith 2004; Zhu 1996; Zhu 2012).

Five trials compared different interventions at the time a participant called the helpline. Sood 2009 compared counselling at the initial call to mailed self‐help materials only. Orleans 1998 and Thompson 1993 compared different counselling interventions provided during the initial call; Orleans 1998 compared counselling and materials targeted at African‐American smokers to standard advice and materials, and Thompson 1993 compared a counselling approach based on the 'stage of change' model to the provision of more general information. Two trials offered two different modalities of proactive telephone counselling. Bricker 2014 compared Acceptance and Commitment Therapy (ACT) counselling with Cognitive Behavioural Therapy (CBT), with both arms receiving a two‐week supply of NRT. Lindqvist 2013 compared Motivational Interviewing (MI) telephone counselling with standard telephone counselling.

Trials providing access to a helpline

Three studies assessed the impact of offering reactive counselling by providing access to a helpline/quitline/hotline, compared to not being provided access to a telephone helpline/quitline/hotline. One randomised counties to hotline access or not, and followed up smokers who were planning to stop and had registered for a smokers' self‐help project (Ossip‐Klein 1991). One compared a referral to a quitline to the usual care of a GP practice (Zwar 2015). One combined newsletter mailings and hotline access compared to no follow‐up support for smokers who had registered for a self‐help televised cessation programme (McFall 1993). Three studies compared referral to a quitline to proactive telephone counselling (Rogers 2016; Sherman 2017; Skov‐Ettrup 2016). Joyce 2008 compared four different levels of benefit for Medicare beneficiaries aged 65 or older. The most intensive intervention offered a choice of accessing either a reactive hotline or multisession proactive counselling, along with self‐help materials and coverage of nicotine patch with a small co‐payment. Other arms offered coverage of brief provider counselling with or without coverage of pharmacotherapy, and usual care.

Trials of proactive counselling, not initiated by calls to quitlines

There were 65 trials in this category that we judged to have sufficient features in common to consider pooling their results. There were some differences in the intensity of the telephone component, the amount of cessation support that was common to both the control and intervention groups, and the populations recruited.

Studies with minimal intervention controls

There were 33 studies in this subgroup. In 26 studies proactive telephone counselling calls were the only form of personal contact in the cessation intervention. The control groups generally received mailed self‐help materials, but Graham 2011 provided access to a cessation website. In six studies in healthcare settings the telephone intervention was an adjunct to usual care that involved at most a brief smoking intervention (Duffy 2006; Hanssen 2009; Holmes‐Rovner 2008; Rigotti 2006; Stotts 2002; Young 2008). In two further studies that recruited participants through healthcare systems, advice and support were part of usual care but not all participants had clinic visits; the telephone counselling was delivered independently of any clinic visit rather than being an adjunct to a specific episode of care (An 2006; Lipkus 1999). Pharmacotherapy was not systematically offered to all intervention participants in any of the above trials, but in two there was greater use of pharmacotherapy by intervention participants (An 2006; McClure 2005). An 2006 encouraged the use of NRT or bupropion for intervention group participants making a quit attempt and this increased their use, although pharmacotherapy was available to all participants as part of their usual care. In McClure 2005 all participants could enrol in the Free & Clear phone‐based support programme, which could also provide access to pharmacotherapy; this was used more by intervention than control groups.

Studies with brief intervention/counselling controls

Thirteen trials incorporated what we judged to be more substantial face‐to‐face advice for all participants, but without systematic use of pharmacotherapy (Borland 2008; Brown 1992; Brunette 2017; Chouinard 2005; Cossette 2011; Ebbert 2007; Flöter 2009; McBride 2004; Metz 2007; Ockene 1991; Osinubi 2003; Ramon 2013; Reid 2007). The support common to all participants included: a single information session and the provision of a self‐help manual (Brown 1992); usual prenatal care including provider advice and self‐help materials (McBride 2004); assessment, advice or brief counselling from a physician (relevant arms of Borland 2008; Ockene 1991; Ramon 2013) or hygienist/dentist (Ebbert 2007); advice from an occupational physician to consult a personal physician (Osinubi 2003); inpatient nurse counselling (Brunette 2017; Chouinard 2005; Cossette 2011; Reid 2007); or multisession group counselling (Flöter 2009; Metz 2007).

Studies of counselling added to pharmacotherapy

Eighteen trials provided telephone counselling as an adjunct to pharmacotherapy. In 15 trials there was a systematic offer or provision of NRT (Bastian 2013; Blebil 2014; Cummins 2016a; Fiore 2004; Fraser 2014; Hughes 2010; Lando 1997; MacLeod 2003; NCT00534404; Ockene 1991; Reid 1999a; Schlam 2016; Solomon 2000; Solomon 2005; Velicer 2006). Swan 2010 provided varenicline. Boyle 2007 recruited health maintenance organisation (HMO) members who were filling a prescription for any cessation medication, and Ellerbeck 2009 offered free medication four times over two years. The support common to all participants in other trials included: access to a website (Swan 2010); physician advice and an offer of free nicotine gum (relevant arms of Ockene 1991); provision of free nicotine patch after a primary care visit (Fiore 2004); three sessions of physician advice and free nicotine patch (Reid 1999a); a single 90‐minute session, a free prescription for nicotine patch and access to a helpline (Lando 1997); or provision of free nicotine patches (two‐week supply only) but no face‐to‐face contact (Bastian 2013; Blebil 2014; Fraser 2014; MacLeod 2003; Solomon 2000; Solomon 2005; Velicer 2006). Cummins 2016a provided up to six weeks of NRT with the number of weeks depending on how many cigarettes smoked per day , whereas NCT00534404 and Schlam 2016 provided up to eight weeks' free supply of NRT. Velicer 2006 provided nicotine patches to participants meeting criteria for readiness to make a quit attempt; 86% received some during the study.

Studies of counselling added to incentives

One study provided telephone counselling as an adjunct to incentives. Thomas 2016 compared the effect of adding telephone counselling as an adjunct to entry into a 'Quit and win' contest.

Telephone counselling intensity

The number of calls and the period over which they were delivered in this group of 65 studies was very varied. We provide a summary in the following table.

Maximum no. of calls	Within 4 weeks	Within 3 months	Within 6 months	Over longer period/other
Single call	Fiore 2004; Miguez 2008	‐	‐	‐
2 calls	Ellerbeck 2009 (moderate intensity arm, up to 2 after each offer of pharmacotherapy); Lando 1992; Lichtenstein 2000; Lichtenstein 2008; Lipkus 1999	Ossip‐Klein 1997; Stotts 2002 (in late pregnancy)	Rimer 1994	‐
3 calls	Ebbert 2007; Flöter 2009	Abdullah 2005; Curry 1995; Lipkus 2004; McBride 1999a; Ockene 1991; Reid 1999a	Aveyard 2003; Prochaska 2001	McBride 1999b (part, during pregnancy)
4 calls	Blebil 2014; Young 2008	Lando 1997; Reid 2007 (average 2 automated and 2 counsellor)	McClure 2005; Prochaska 1993; Ramon 2013; Tzelepis 2011a (if not setting TQD)	Chan 2015; Orleans 1991
5 calls	Fraser 2014	Graham 2011; Hughes 2010; MacLeod 2003; Metz 2007; NCT00534404; Osinubi 2003; Swan 2010		Rigotti 2006 (4 in pregnancy and 1 postpartum)
6 calls	Brown 1992; Tzelepis 2011a (if setting TQD)	Bastian 2013; Borland 2008; Chouinard 2005; Cossette 2011; Ellerbeck 2009 (high intensity arm, up to 6 after each offer of pharmacotherapy); Girgis 2011; Holmes‐Rovner 2008; Miguez 2002; Sorensen 2007a	Emmons 2005	McBride 1999b (part) and McBride 2004 (3 during pregnancy and 3 postpartum)
7 or more	Brunette 2017	An 2006; Schuck 2014; Solomon 2000	Boyle 2007 (up to 9, average 5); Duffy 2006 (9 to 11); (Hanssen 2009 (9); (Velicer 2006 (up to 10 automated calls); (Solomon 2005 (up to 12); McClure 2011 (up to 12, also covering depression and physical activity); Schlam 2016 (up to 8)	Cummins 2016a (up to 10); Peterson 2016 (up to 10)

Interventions for callers to quitlines ‐ effect of additional proactive calls for smoking cessation
Patient or population: callers to quitlines  Intervention: additional proactive calls
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Additional proactive calls
Smoking cessation  Self‐reported abstinence (majority)  Follow‐up: 6+ months	Study population		RR 1.38   (1.19 to 1.61)	32,484  (14 studies)	⊕⊕⊕⊝  moderate^b,c	‐
	72 per 1000	100 per 1000  (85 to 116)
	Low
	50 per 1000^a	69 per 1000 (59 to 81)
	High
	150 per 1000^a	208 per 1000 (178 to 242)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Proactive telephone counselling for smokers not calling quitlines
Patient or population: smokers not calling quitlines  Intervention: proactive telephone counselling
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Control	Proactive telephone counselling
Smoking cessation  Self‐reported abstinence (majority)  Follow‐up: 6+ months	Study population		RR 1.25   (1.15 to 1.35)	41,233  (65 studies)	⊕⊕⊕⊝  moderate^a,b
	110 per 1000^a	137 per 1000  (127 to 149)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Date	Event	Description
30 October 2018	New citation required but conclusions have not changed	30 new studies added, no major change to conclusions. Additional authors: WM and JMOM
3 May 2018	New search has been performed	Searches updated

Date	Event	Description
23 October 2013	Amended	Information added to ongoing studies tables.
20 June 2013	New search has been performed	Searches updated
20 June 2013	New citation required but conclusions have not changed	Twelve new studies added, no major change to conclusions. Additional author JHB.
12 May 2009	New search has been performed	Updated for issue 3, 2009. Nineteen new studies, no change to conclusions, strengthened evidence of effect overall and for some subgroups.
4 August 2008	Amended	Converted to new review format.
11 April 2006	New citation required but conclusions have not changed	Updated for Issue 3, 2006. Twenty two new studies, studies of relapse prevention now excluded. Comparisons reorganised, additional subgroup analyses.
14 October 2002	New citation required but conclusions have not changed	Updated for Issue 1, 2003. Four new trials, of which 3 contribute to meta‐analysis. No major changes to conclusions

Parameter	β	(95% CI)	RRC^a	(95% CI)	P value
Intercept	−0.17	(−0.45 to 0.12)	0.85	(0.64 to 1.13)	0.25
Baseline Support	‐	‐	‐	‐	‐
Pharmacotherapy	0	Reference	1.00	Reference
Self‐help	0.30	(0.10 to 0.51)	1.35	(1.10 to 1.67)	< 0.01
Brief intervention	0.31	(0.05 to 0.58)	1.37	(1.05 to 1.79)	0.02
Incentives	−0.12	(−0.75 to 0.51)	0.89	(0.47 to 1.67)	0.71
Per each additional call	0.02	(−0.02 to 0.05)	1.02	(0.98 to 1.05)	0.29
Selected for motivation	‐	‐	‐	‐	‐
No	0.00	Reference	1.00	Reference	‐
Yes	0.23	(0.04 to 0.42)	1.26	(1.04 to 1.52)	0.02

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cessation at longest follow‐up	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
1.1 Seven versus three phone calls	1	1908	Risk Ratio (M‐H, Random, 95% CI)	1.44 [1.09, 1.89]
1.2 Five versus three phone calls	1	3669	Risk Ratio (M‐H, Random, 95% CI)	1.28 [1.00, 1.64]
1.3 Seven versus five phone calls	1	3939	Risk Ratio (M‐H, Random, 95% CI)	1.12 [0.93, 1.36]
1.4 Five versus two phone calls	1	2874	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.89, 1.23]
1.5 Five versus one phone call	1	2189	Risk Ratio (M‐H, Random, 95% CI)	1.32 [1.01, 1.74]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cessation at longest follow‐up	14	32484	Risk Ratio (M‐H, Random, 95% CI)	1.37 [1.20, 1.57]
1.1 Two sessions or fewer	2	3867	Risk Ratio (M‐H, Random, 95% CI)	1.22 [1.02, 1.46]
1.2 Three to six sessions	11	22612	Risk Ratio (M‐H, Random, 95% CI)	1.38 [1.17, 1.63]
1.3 Seven sessions or more	4	6005	Risk Ratio (M‐H, Random, 95% CI)	1.49 [0.98, 2.25]

Outcome or subgroup title	No. of studies	Statistical method	Effect size
1 Cessation at longest follow‐up	5	Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
1.1 Reactive counselling vs self‐help materials	1	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
1.2 Tailored counselling versus standard counselling	1	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
1.3 Stage‐based counselling versus general information	1	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
1.4 ACT + NRT versus CBT + NRT	1	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
1.5 Motivational Interviewing TC versus standard TC	1	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cessation at longest follow‐up	65	41233	Risk Ratio (M‐H, Random, 95% CI)	1.25 [1.15, 1.35]
1.1 Adjunct to self‐help or minimal intervention	35	22917	Risk Ratio (M‐H, Random, 95% CI)	1.35 [1.16, 1.57]
1.2 Adjunct to brief intervention or counselling	12	4234	Risk Ratio (M‐H, Random, 95% CI)	1.30 [1.12, 1.50]
1.3 Adjunct to pharmacotherapy	18	12865	Risk Ratio (M‐H, Random, 95% CI)	1.14 [1.03, 1.26]
1.4 Adjunct to incentives for smoking cessation	1	1217	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.67, 1.65]

Methods	Setting: Parents of children in a birth cohort study, Hong Kong  Recruitment: Active; by mail, current smokers, not selected for motivation
Participants	903 current smokers with young children (49 recent quitters not included here); 84% M, > 50% aged 36 to 45, 91% smoked ≤ 20/day
Interventions	1. Single mailing of stage‐matched S‐H (either preparation/action or contemplation/precontemplation)  2. As 1, plus 20 to 30 mins of TC at time of enrolment by trained nurse counsellor. Hotline number, further counselling at 1 month and 3 months. Average duration of counselling 38 mins over 3 contacts
Outcomes	Validated abstinence at 6 m (7‐day PP).  Validation: CO < 9 ppm or urine cotinine < 100 mmol/mol
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised, method not described
Allocation concealment (selection bias)	Low risk	Numbered sealed opaque envelopes
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Quote: "Independent interviewer...was unaware of subjects' group allocation... All respondents who reported they were not smoking during the preceding 7 days were invited to attend the research centre for biochemical validation."
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Losses to follow‐up 11% intervention/ 4% control. Included as continuing smokers

Methods	Setting: 5 Veterans Administration medical centres, USA  Recruitment: By mail, planning to quit in next 30 days
Participants	821 smokers interested in quitting (excludes 16 deaths, 1 withdrawal); 91% M, av. age 57, av. cigs/day 26. 26% had > 7‐day abstinence in previous year, 44% ever use of bupropion, 82% ever use NRT
Interventions	1. Mailed S‐H and standard care; opportunity for intervention during routine health care and referral to individual or group cessation programmes. NRT and bupropion avail on formulary  2. As 1, plus proactive TC, modified California helpline protocol, 7 calls over 2 m, relapse‐sensitive schedule. NRT and bupropion available, could be mailed directly after screening and primary provider approval for bupropion
Outcomes	Abstinence at 12 m (sustained > 6 m, 7‐day PP also reported)  Validation: none
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised, method not described
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Self‐reported outcomes from participants not blinded to treatment condition. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Losses to follow‐up included as smokers, 16 deaths excluded

Methods	Setting: 65 general practices, UK  Recruitment: Active; volunteers from random selection of smoking patients, not selected for motivation  Randomisation: Centralised, minimisation to balance SoC, addiction and SES
Participants	2471 smokers (2058 in relevant arms); > 80% in precontemplation or contemplation, 10% to 14% in preparation, 46% M, av. age 41, av. cigs/day 20
Interventions	1. Standard S‐H materials, single mailing  2. S‐H manual based on Transtheoretical model, expert‐system letter tailored on baseline questionnaire. Further questionnaires at 3 and 6 m for additional letters (approx 50% received 3 letters)  3. As 2, plus proactive TC after receipt of each questionnaire (max 3 calls). Designed as reminders, scripted, delivered by trained postgraduate students
Outcomes	Abstinence at 12 m (sustained for 6 m)  Validation: saliva cotinine < 14.2 ng/ml
Notes	We included arms 3 vs 2 in the analyses. Sensitivity analysis comparing arms 3 vs 2+1. 66% received 1st phone call, 36% 2nd, 31% 3rd
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Centralised randomisation procedure, with minimisation to balance SoC, addiction and SES
Allocation concealment (selection bias)	Low risk	Centralised
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	12 m PP "was confirmed with salivary cotinine, so that we had unconfirmed and confirmed prevalence of quitting." Confirmed figures used in analysis
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Loss to follow‐up 24% in group 1, 31% in 2 and 3. All included as smokers. Sensitivity analysis allowing for differential dropout did not change findings

Methods	Setting: North Carolina, USA; Medical Center Recruitmenet: Mailed smokers an introductory letter from the Chief of Cardiology, Chief of Oncology, or a primary care physician (the Principal Investigator (PI)) informing them of the study and encouraging smoking cessation
Participants	471 smokers enrolled in Durham Veterans Affairs Medical Center, receiving treatment for chronic illnesses (i.e. cancer, cardiovascular disease, HTN, diabetes, COPD) and wanting to quit in the next 30 days; 91.5% M, av. age 59.2, av. cigs/day not reported
Interventions	1. Standard telephone counselling, a letter from a VA physician encouraging smoking cessation, NRT, if not contra‐indicated, a S‐H cessation kit, and up to 5 TC calls (every 3 ‐ 4 weeks, av. duration 20 minutes) 2. "Family‐supported telephone counseling, included all components of the standard TC arm plus an enhanced family‐supported intervention that included a support skills booklet and additional telephone counseling content focusing on social support skills [...] The main distinction between the two arms of this comparative effectiveness study was the family‐supported intervention that aimed to help increase positive interactions between the participant and their designated support person, to facilitate smoking cessation [...] Participants randomized to family support‐based intervention also received an 8‐page disease‐specific family support booklet."
Outcomes	Abstinence at 12 m (7‐day PP) Validation: available for only 50.5% of the participants
Notes	New for 2018 update Funding: "This material is based upon work supported by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, and Health Services Research and Development. IIR‐05‐202." Declarations of interest: "SCG serves as a consultant to Gilead Sciences and Watermark Research Partners. Although these relationships are not perceived to represent a conflict with the present work, it is included in the spirit of full disclosure. Presented in part at the Society of General Internal Medicine Annual Meeting, Phoenix AZ May 2011."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "blocked randomization, stratified by sex and disease type"
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of outcome assessment (detection bias)   All outcomes	High risk	The investigators mailed participants saliva‐sampling kits to measure cotinine, but the return rates for saliva samples were low at all follow‐ups. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Quote: "Follow‐up rates were 86% and 81.1% at 5 months and 12 months, respectively. Loss to follow‐up was similar in both arms."

Methods	Setting: North Carolina, USA; Academic setting Recruitment: Investigators asked lung cancer patients’ permission to contact members of their social networks who smoked and were aged 18 and older; these received a letter describing the study and providing them a toll‐free number to call to decline participation
Participants	496 relatives of lung cancer patients, 41.5% M, av. age 46.9, av. cigs/day 19.5. More than half the participants have made more than 3 quit attempts before inclusion in the study
Interventions	1. Tailored self‐directed materials and nicotine patches 2. As above, plus proactive TC, 6 x weekly sessions scheduled over an intervention period of 12 weeks, with an av. duration of each session of 30 mins. The mean number of sessions completed was 2.4. 81 participants in this group (33%) did not complete at least 1 session
Outcomes	Abstinence at 12 m (7‐day PP) Validation: saliva cotinine, not possible due to low return rates
Notes	New for 2018 update. Funding: "This work was supported by National Cancer Institute grant 5U01‐CA‐92622. This research was also supported in part by the Intramural Program of the National Human Genome Research Institute, National Institutes of Health." Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was blocked by patient, with entire social network units stratified by site and size of social network enrolled (one vs. two or more) assigned to the same condition."
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Self‐reported. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	High risk	Large withdrawal of participants after 12 months of follow‐up (> 50%), although similar across arms

Methods	Setting: Malaysia; Outpatient Quit Smoking Clinic based at 2 hospitals  Recruitment: Quote: "All individuals who attended the clinics during the period under review were invited to participate in the research."
Participants	231 outpatient smokers, 96.1% M, av. age 48.3, av. cigs/day 13.8
Interventions	1. Usual care, which included a combination of nicotine gum and CBT (4 counselling sessions during the 1st month, 2 counselling sessions during the 2nd month + 2 phone calls (av. duration 20 ‐ 30 mins), and 1 counselling session during the 3rd month plus 2 phone calls (av. duration 20 ‐ 30 mins)) 2. As above, + 1 extra weekly proactive call (av. duration 10 ‐ 15 mins) during the first month of the quit attempt
Outcomes	Abstinence at 6 months (4‐week PP) Validation: exhaled CO level < 7 ppm
Notes	New for 2018 update Funding: not reported Declarations of interest: none declared
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Even though "urn design was used to achieve balanced groups", participants walking in the study and being referred from outpatient clinics at the hospitals were more likely to be assigned to the intervention than those coming from other hospitals
Allocation concealment (selection bias)	Low risk	Quote: "the assignments of treatment within a sequence created by the urn design are not as predictable as those of other types of restricted randomisation processed"
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Biochemical validation of self‐reported outcome. The data were collected by another research member not connected with counselling and the data analysis
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Percentage lost to follow‐up was overall low (8%), although larger in the usual‐care (12%) than in the intervention group (4%).

Methods	Setting: Community, Australia  Recruitment: Callers to a quitline
Participants	998 smokers interested in quitting; 48% M, 37% aged 15 ‐ 29, 26% aged 30 ‐ 39, av. cigs/day 23
Interventions	1. Proactive call‐back TC following initial call to quitline: Multiple calls, first pre‐quit, quit, then according to need. Up to 6 m. Mailed materials  2. Control: Mailed materials  Both groups also received the standard motivational counselling in response to their first call
Outcomes	Self‐reported abstinence at 12 m (sustained for 9 m)  Validation: none
Notes	Average number of calls 2.8, 67% received 1 or more. 20% refused call‐back or wanted to initiate the calls, further 7% did not receive any
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised, method not described
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Self‐reported outcomes from participants not blinded to treatment condition. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Loss to follow‐up 37% intervention, 30% control. All participants included as smokers in the meta‐analysis

Methods	Setting: General practice, Australia  Recruitment: 45 participating GPs recruiting patients who smoked
Participants	1039 smokers, not selected for motivation but ˜80% had previously tried to quit; 45% M, av. age: 41, av. cigs/day 17
Interventions	1. Referral: Smokers with any interest in quitting referred by fax to Victorian Quitline. Proactive contact attempted with up to 2 pre‐quit and 4 post‐quit sessions typically using relapse‐sensitive schedule. Internet support available as an alternative (4.4% reported use)  2. In‐practice support, could include external referral if this was clinical preference  All participants given guideline‐based assessment of readiness to quit and offer of pharmacotherapy if appropriate
Outcomes	Self‐reported abstinence at 12 m (sustained ≥ 10 m)  Validation: none
Notes	TC as adjunct to face‐to‐face intervention.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Cluster‐randomised by GP (1:2 ratio). Computer allocation before GPs attended education session for their assigned intervention
Allocation concealment (selection bias)	Unclear risk	Initially concealed but 13 referral (30%) and 11 (42%) control GPs failed to recruit participants. Allocation not blind at time of recruitment of individual participants, so further selection bias possible. Measured characteristics at baseline were similar
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Quote: "Three‐ and 12‐month questionnaires were administered...by trained interviewers who were blind to treatment condition until after the outcome data were collected." However, reliant on self‐reported outcomes from participants not blinded to treatment condition. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	Low risk	33% lost in referral condition, 39% in control, all included as smokers in MA. Excluding losses does not affect MA

Methods	Setting: Health Maintenance Organisation, USA  Recruitment: Proactive recruitment of members filling a prescription for cessation medications (motivated)
Participants	1329 HMO members; 42% M, av. age 47, 66% smoked > pack/day
Interventions	All participants had filled a prescription. Almost 95% used; ˜51% only bupropion, 26% only NRT, remainder both  1. No further intervention  2. Proactive call to offer counselling, up to 9 calls, given choice of structured course or unstructured format
Outcomes	Abstinence at 12 m (repeated 7‐day PP at 3 and 12 m)  Validation: none
Notes	49% of intervention group reached, 36% of those declined, 31% of total accepted counselling. Average no of calls 5.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised, stratified by presence of chronic disease. Method not described
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Quote: "The follow‐up survey was conducted by the Data Collection Center within the Health Partners Research Foundation, using staff not involved in the intervention." However, reliant on self‐reported outcomes from participants not blinded to treatment condition. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	Low risk	˜33% lost to follow‐up, balanced across groups, included in MA as smokers

Methods	Setting: South Carolina, USA  Recruitment: Staff advertised the study to the quitline callers
Participants	121 uninsured callers to the South Carolina State Quitline who wanted to quit in the following 30 days
Interventions	1. Telephone counselling (CBT) + NRT 2. Telephone counselling (ACT) + NRT 5 weekly calls, 30‐min first session and 15‐min subsequently, were offered. All participants received standard 2‐week NRT (patch or gum) of choice
Outcomes	Self‐reported abstinence at 6 m (30‐day PP) Validation: none
Notes	New for 2018 update Funding: "This study was supported by the National Institutes of Health (T32MH082709 to RV, K23DA026517 to JLH, R21DA030646 to JB) and the Fred Hutchinson Cancer Research Center." Declarations of interest: not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomized study arm assignments were computer generated"
Allocation concealment (selection bias)	Low risk	Quote: "Randomized study arm assignments were [...] concealed"
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Abstinence not biochemically validated, but same level of personal contact in different study arms
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Proportion of individuals lost to follow‐up was greater in CBT than in ACT arm (39% and 27%, respectively) but less than 50% overall

Methods	Setting: New Hampshire, USA; community mental health centres  Recruitment: Through flyers, clinician referral, and direct mail
Participants	661 medicaid beneficiaries with mental illness and low income (< USD 1317 a month) willing to initiate cessation treatment within 30 days, 36% M, av. age 45, av. cigs/day 17.3
Interventions	1. Usual care, a prescriber visit for smoking cessation (NRT or cessation medications, i.e. bupropion/varenicline) 2. As in 1, plus referral to New Hampshire Tobacco Helpline which provides an average of 3 manualised TC sessions 3. As in 1, plus TC (av. 9 sessions) CBT initiated by a CBT therapist
Outcomes	Abstinence at 12 m (7‐day PP) Validation: breath CO ≤ 4 ppm and urine cotinine < 100 ng/ml (or solely breath CO if using NRT)
Notes	New for 2018 update. Funding: "This research received financial support from the Centers for Medicare and Medicaid Services (Medicaid Incentives for the Prevention of Chronic Diseases grant 1B1CMS330880) and from the New Hampshire Department of Health and Human Services (NHDHHS)." Declarations of interest: "Dr. Brunette reports receipt of research funding from Alkermes. The other authors report no financial relationships with commercial interests."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Computer‐generated tables for each strata within each site were used for random assignment."
Allocation concealment (selection bias)	High risk	Quote: "We used equipoise randomization [...] that allowed participants to opt out of one of the cessation treatment conditions or allowed randomization to any of the three options. [...] Randomization strata were defined by conditions to which the participant was willing to be randomly assigned. Within the stratum, a participant was then randomly assigned with equal probability to the selected treatment condition options." Not a true randomisation method; participants can choose what intervention they do not want to be allocated to and this can lead to selection bias. This led to different numbers between arms, and significant baseline age differences
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Biochemical validation for only half of the participants in the trial (those receiving an incentive), and there are significant differences between those receiving and not receiving an incentive. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Proportion of participants lost to follow‐up was lower than 50% overall

PERMALINK

Telephone counselling for smoking cessation

William Matkin

José M Ordóñez‐Mena

Jamie Hartmann‐Boyce

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Search methods for identification of studies

Data collection and analysis

Assessment of risk of bias

Choice of outcome and treatment of missing data

Data synthesis

Investigations of heterogeneity: meta‐regression and subgroup analyses

'Summary of findings' table

Results

Description of studies

1.

Trials of interventions for people calling helplines

Trials providing access to a helpline

Trials of proactive counselling, not initiated by calls to quitlines

Studies with minimal intervention controls

Studies with brief intervention/counselling controls

Studies of counselling added to pharmacotherapy

Studies of counselling added to incentives

Telephone counselling intensity

Recruitment and motivation of participants

Studies comparing intense versus minimal telephone counselling

Other studies

Risk of bias in included studies

2.

Allocation

Blinding of outcome measurement (detection bias)

Incomplete outcome data

Effects of interventions

Summary of findings for the main comparison. Interventions for callers to quitlines ‐ effect of additional proactive calls for smoking cessation.

Summary of findings 2. Interventions for smokers not calling quitlines ‐ effect of proactive telephone counselling.

Trials of interventions for people calling helplines

Effect of additional proactive support

3.

1.1. Analysis.

Counselling intensity

2.1. Analysis.

3.1. Analysis.

Comparisons between different types of support at initial call

4.1. Analysis.

Trials providing access to a helpline

5.1. Analysis.

Trials of proactive counselling, not initiated by calls to quitlines

Overall effect of counselling

4.

6.1. Analysis.

Baseline support offered

Counselling intensity

7.1. Analysis.

8.1. Analysis.

Motivation

9.1. Analysis.

Results of a meta‐regression

Other studies

10.1. Analysis.

Methods	Setting: Hong‐Kong, China; community‐based Recruitment: Participants were approached by investigators at shopping malls or public areas in 16 out of the 18 districts in Hong Kong. Participants who expressed an interest in joining the contest were screened for eligibility and tested on their exhaled CO to ascertain their smoking status
Participants	1003 Hong Kong residents aged 18 or older, who smoked 1 or more cig/day in the past 6 months, 82% M; 38% 18 – 39 years, 49% 40 – 59 years, 13% 60+ years, 42% 1 – 10 cigs/day, 43% 11 – 20 cigs/day, 15% > 20 cigs/day
Interventions	1. S‐H booklet and the contact information of the smoking cessation services at the enrolment 2. As 1, plus 8 mobile phone text messages corresponding to the 8 pages of the S‐H booklet (not used in review) 3. As 1, plus 4 sessions (within 1 week, after 2, 6 and 12 m) of 5‐min smoking cessation telephone counselling provided by a trained nurse, using the AWARD Protocol
Outcomes	Abstinence at 12 m (7‐day PP) Validation: exhaled CO < 4 ppm and salivary cotinine level < 10 ng/ml
Notes	New for 2018 update Funding: "This work was funded by Hong Kong Council on Smoking and Health." Declarations of interest: "Prof. Tai‐hing Lam is the principal investigator of the FAMILY project, which was funded by the Hong Kong Jockey Club Charities Trust. All other authors do not have connection with the tobacco, alcohol, pharmaceutical or gaming industries, and nobody was substantially funded by these organizations."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomisation was used to ensure similar group sizes.
Allocation concealment (selection bias)	Low risk	Quote: "The randomization and allocation were conducted by the author who did not participate in subject recruitment to ensure allocation concealment"
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Abstinence biochemically validated
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Efforts were made to minimise loss to follow‐up: Quote: "at least seven call attempts at different times were made before participants were considered as loss to follow‐up." In the end follow‐up was comparable across arms. Reasons for losses to follow‐up are provided

Methods	Setting: Canada  Recruitment setting: Inpatients with cardiovascular disease (myocardial infarction, angina, congestive heart failure) or peripheral vascular disease, unselected by motivation
Participants	168 past‐month smokers; 27% M, av. age 56, 60 % in preparation or action SoC
Interventions	1. Counselling by research nurse (1 x 10 ‐ 60 mins, av. 40 mins, based on Transtheoretical Model, included component to enhance social support from a significant family member)  2. As 1, plus telephone follow‐up, 6 calls over 2 m post‐discharge  3. Usual care cessation advice (not used in review)
Outcomes	Abstinence at 6 m (sustained at 2 and 6m)  Validation: Urine cotinine or CO
Notes	TC as adjunct to face‐to‐face counselling. 75% received 6 calls
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Cluster‐randomised in groups of 3 ‐ 6 "to prevent contamination between groups", method not described
Allocation concealment (selection bias)	Low risk	Quote: "Individuals not familiar with the study were in charge of the randomization procedure which included inserting the information into envelopes that were sealed and would be opened by the investigator only at the time of recruitment."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Biochemical validation used
Incomplete outcome data (attrition bias)   All outcomes	Low risk	4 deaths (3 in Grp 1, 1 in Grp 2) and 3 not meeting follow‐up criteria excluded from MA denominators. Other losses to follow‐up included

Methods	Setting: North and West Philadelphia, PA, US; 4 paediatric clinics  Recruitment: Clinic providers referred smoking parents of children exposed to SHS to the cessation resources, including the current study
Participants	327 smoking parents from predominantly low‐income, racial‐ and ethnic‐minority families of children under the age of 11, 16.5% M, av. age 33, av. cigs/day 11.5
Interventions	1. Individual TC health education attention control (AC) intervention that focuses on improving family nutrition on a budget 2. Individual behavioural TC intervention that focuses on reducing child SHS exposure and parent smoking cessation The TC dosage (5 sessions over 12 weeks) was similar between arms
Outcomes	Abstinence at 12 m (7‐day PP) Validation: cotinine‐verified (cut‐off not reported)
Notes	New for 2018 update Funding: "Supported by Temple University. Funded by the National Cancer Institute, National Institutes of Health, grant CA158361. Funded by the National Institutes of Health (NIH)." Declarations of interest: "The authors have indicated they have no potential conflicts of interest to disclose."
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was seeded using values obtained from random.org."
Allocation concealment (selection bias)	Low risk	Quote: "The project biostatistician provided the allocations to the data collection team in opaque sealed security envelopes."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Cotinine‐verified smoking cessation
Incomplete outcome data (attrition bias)   All outcomes	Low risk	The percentage of participants lost to follow‐up is minimal (12%), although slightly different between intervention (17%) and control (8%) arms.

Methods	Setting: Specialised cardiac hospital, Canada Recruitment: All smokers who were hospitalised were asked to participate by the study nurse (not selected by motivation)
Participants	40 current daily smokers with cardiovascular disease, 60% M, av. age 57. Most in preparation stage Therapists: nurse specialised in smoking cessation
Interventions	All participants had 1 or more sessions with the study nurse during hospitalisation. Conditions differed after discharge 1. Intervention: 6 phone calls by study nurse at weeks 1, 2, 3, 4, 8, 12. If needed additional phone calls could be arranged between 3 and 6m post‐discharge. At week 3 appointment with the study nurse if requested by participant 2. Control: referral to a national quitline or a community centre for smoking cessation Pharmacotherapy: NRT, bupropion or varenicline were suggested during hospitalisation and follow‐up
Outcomes	Self‐reported abstinence at 6 m (7‐day PP) Validation: only for 1 participant
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not specified, but generated by a centre for randomised controlled trials
Allocation concealment (selection bias)	Unclear risk	Opaque sealed envelopes
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Self‐reported outcomes from participants not blinded to treatment condition. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	High loss to follow‐up, but missing data similar in both groups and analyses are ITT, participants lost to follow‐up considered smokers

Methods	Setting: California, USA; hospital‐based Recruitment: Quote: "Recruitment procedures differed between healthcare systems based on the personnel involved and the hospital’s reliance on electronic medical records (EMRs)". In 1 study site recruitment was part of the therapists' workflow, while in another academic site, research staff were involved instead
Participants	1270 hospitalised adult smokers who smoked 6 or more cigs/day, were interested in quitting, spoke English or Spanish, and were not pregnant, 56.7% M, av. age 49.9, av. cigs/day 14.6
Interventions	Factorial 2 x 2 design comparing TC vs no TC, and NRT vs no NRT 1. No TC (usual care) ± NRT patches. In general usual care consisted of providing smokers with the quitline number, but some hospitals may have also provided counselling or prescribed quitting aids 2. TCg ± NRT patches, with 10 calls scheduled, but on av. 3.6 completed. The av. number of calls in the usual care arm was 1.7
Outcomes	Abstinence at 6 m (7‐day PP) Validation: saliva cotinine < 10 ng/mL
Notes	New for 2018 update. Previously listed under ongoing studies as Cummins 2012 Funding: "This research was supported by a grant from the National Cancer Institute (CA159533)." Declarations of interest: none reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "randomly assigned by computer"
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Biochemically‐confirmed abstinence
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Similar loss to follow‐up across arms (˜33%)

Methods	Setting: California USA; pregnant women Recruitment: Callers to University of California San Diego California Smokers' Helpline
Participants	1173 pregnant (< 27 weeks) women, willing to quit within 1 month or recent quitters, av. age 26.3, av. cigs/day 11.2
Interventions	1. Self‐help American Cancer Society's Make Yours a Fresh Start Family fact sheets, and additional tips for quitting while pregnant  2. As 1, plus proactive TC specifically developed for pregnant smokers, including 9 x 30 ‐ 45‐min sessions on days 0, 1, 3, 7, 14, and 30 after quit date, at 32 weeks of gestation, and 2 and 4 weeks after delivery
Outcomes	Abstinence at 6 m post‐partum (180‐day abstinence)  Validation: saliva cotinine < 13 ng/mL
Notes	New for 2018 update. Previously listed under studies awaiting assignment as Zhu 2004 Funding: "This research was supported by the Tobacco‐Related Disease Research Program (Grant 8RT‐0103) and First 5 California (Contract CCFC‐6810) and by funds received from the California Department of Health Services Tobacco Control Section (Contract 00–90605)." Declarations of interest: none declared
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Random allocation to condition was done by computer using blocks of 20"
Allocation concealment (selection bias)	Low risk	Quote: "staff were blind to group assignment until the end of the intake, when the appropriate script was presented"
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Abstinence biochemically validated
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Higher % lost to follow‐up in intervention arm

Methods	Setting: ENT clinics at 4 hospitals, USA  Recruitment: Patients with head and neck cancer who screened positive for smoking, alcohol problem or depression, not selected for motivation
Participants	89 current smokers used in MA, out of 184 trial participants who also included 26 quit within last month and 21 within last 6 m . Demographics are for all participants; 84% M, av. age 57
Interventions	1. Proactive counselling; 9 ‐ 11 CBT‐based calls from trained nurses, linked to use of CBT workbook. Smokers with problem drinking or depression received counselling for these too  2. Enhanced usual care with assessment and referral
Outcomes	Abstinence at 6 m (sustained)  Validation: none
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised, method not described
Allocation concealment (selection bias)	Unclear risk	No details given. Smokers were a higher proportion of the intervention than control groups, and a higher proportion of those randomised than those who refused, raising possibility of selection bias
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Self‐reported outcomes from participants not blinded to treatment condition. Level of personal contact differed between arms.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	22 in total (including non‐smokers) lost to follow‐up, evenly distributed. Losses appear to have been included as smokers

Methods	Setting: 8 dental practices, USA  Recruitment: Patients screened by questionnaire at routine hygiene appointments, not selected for motivation
Participants	82 smokers (60 intervention, 22 control). No baseline data for controls
Interventions	1. Control: Brief counselling (10 mins) from hygienist, reinforced by dentist  2. As 1 plus faxed referral to quitline, proactive counselling, 45 mins baseline, 20 mins at 1 week and 2 weeks, further calls if requested
Outcomes	Abstinence at 6 m (7‐day PP)  Validation: none
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Cluster‐randomised by practice, method not described
Allocation concealment (selection bias)	High risk	Hygienists who recruited participants after screening not blind, large difference in numbers recruited, not possible to establish baseline similarity
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Self‐reported outcomes from participants not blinded to treatment condition. Level of personal contact differed between arms
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	No description of number lost at follow‐up

Methods	Settng: Primary care patients, 50 rural practices, Kansas, USA  Recruitment: Smokers not selected for motivation, but 67% of those eligible enrolled, only 8.7% in precontemplation stage of change
Participants	750 smokers of > 10 cigs/day, 41% M, av. age 47, av. cigs/day 24, 61% contemplation, 30% preparation
Interventions	All participants mailed an offer of free pharmacotherapy every 6 m, 4 times in total. Nicotine patch 21 mg for 6 weeks or bupropion SR (150 mg twice daily) for 7 weeks 1. Control. No other contact. 2. Moderate‐intensity disease management: up to 2 calls from counsellor in each cycle encouraging uptake of pharmacotherapy, newsletter mailings and periodic progress reports with counselling suggestions faxed to physician 3. High‐intensity disease management, up to 6 calls at approx 1, 3, 6, 9, 12 weeks from start of each cycle
Outcomes	Abstinence at 24 m (PP). Study also reported analysis based on combination of effects at all follow‐up points. Sustained abstinence not a suitable outcome since no quit date and repeated intervention Validation: attempted saliva cotinine (< 15 ng/ml) by mail at 12 and 24 m. Proxy report used at 24 m for non‐returners. Rate of validation similar across groups
Notes	For analysis on counselling intensity, classified on basis of average calls; moderate in 3 ‐ 6 sessions, high in 7+ subgroups
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "computer generated random‐number table" in blocks of 24
Allocation concealment (selection bias)	Low risk	Quote: "To conceal allocation, we placed these cards in sequentially numbered, opaque, sealed envelopes."
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Biochemical validation used
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Differential rates of loss to follow‐up (1: 22.0%; 2: 31.3%; 3: 31.1%). Participants lost to follow‐up counted as smokers but sensitivity analysis shows no significant difference in analysis outcome if excluding those lost to follow‐up

Methods	Setting: Childhood Cancer Survivors Study cohort, USA  Recruitment: Smokers contacted by telephone to assess eligibility and enrol, not selected for motivation
Participants	794 smokers (excludes 2 deaths in control); 53% M, av. age 31, av. cigs/day 12
Interventions	1. S‐H control. Mailed manual (Clearing the Air) and letter from study physician  2. Peer counselling. Up to 6 calls in 7‐m period, by trained cancer survivor. Motivational, tailored to SoC. Free NRT available. Individually‐tailored materials before 1st call and other materials during intervention
Outcomes	Abstinence at 12 m (7‐day PP)  Validation: none (warning that samples might be requested)
Notes	No data on average number of calls. Longer‐term follow‐up, assessed at 2 ‐ 4 years, reported in Emmons 2009. Not used in MA ‐ sustained rates not reported.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised, method not described
Allocation concealment (selection bias)	Unclear risk	No details given
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Bogus pipeline procedure used, no further details provided
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	19% lost in intervention vs 24% in control at 12 m. All included as smokers in MA. Excluding losses does not affect MA