Effectiveness of computer-tailored Smoking Cessation Advice in Primary Care (ESCAPE): a Randomised Trial

Smoking remains a major public health problem; 23% of all deaths in middle age are attributed to smoking [1]. Mortality from tobacco in the first half of the 21^st century will be affected much more by the number of adult smokers who stop than by the number of adolescents who start [2]. Declines in smoking rates have been greater among the better educated, leading to a strong socio-economic gradient in smoking prevalence [3] contributing to health inequalities and accounting for over half of the difference in risk of premature death between the social classes [4]. Developing effective interventions to encourage more quit attempts, and to improve the success rate of self-quit attempts, is essential to reduce the numbers of people who smoke, and to reach the Department of Health's targets to reduce the prevalence of smoking in manual groups from 32% to 26% over the next eight years [5].

Smoking cessation services are currently high on the NHS agenda, with strong drive to deliver a variety of interventions in primary care. Interventions for smoking cessation can be characterised in two extremes. The treatment model of the NHS services is based on the clinical approach, offering intensive support and focusing on individual face-to-face treatment. As a result of government initiatives [6], the number of specialist clinics offering intensive face-to-face treatment for smoking cessation, which can produce relatively high abstinence rates, has increased. However, they are limited by low participation rates, less than 2% of smokers make use of these specialist services [7‐9]. At the other extreme are large-scale, public health interventions and policy initiatives (e.g. legislation to control smoking in public places). This type of intervention can reach large numbers of smokers, but effectiveness rates are low partly because such interventions lack the personal element that is central to the success rates of the clinical approach. Computer-based systems offer a method for generating highly tailored behavioural feedback letters [10], which can be cost-effectively produced on a large scale, with the potential to help a greater proportion of the smoking population in changing their behaviour. The majority of smokers prefer less intensive methods [11] and tailored self-help materials can bridge the gap between these two extremes, combining the behavioural intervention principles used in clinical interventions with the participation rates of public health interventions. Furthermore, as many as 70% of smokers have no serious intentions to attempt to quit in the next year [12, 13]. Proactive mailing and recruitment outside the general practice offering personalised service, as an alternative or addition to brief intervention delivered by GPs, could also increase the impact of legislation, by serving as a prompt to motivate smokers to make quit attempts or to seek more intensive help through NHS services.

A large number of studies of tailored feedback have been published demonstrating positive results. This trial uses a modified version of a computer-based system developed by two of the authors (HG and SS) to generate individually tailored feedback reports designed to encourage and help smokers to quit. The smoker first completes a paper-based Smoking Behaviour Questionnaire (SBQ). The SBQ data are scanned into a database, and the computer program generates a feedback letter that is highly tailored to the individual's responses.

The efficacy of our tailored feedback was evaluated as an adjunct to telephone counselling (via the national Quitline) in a randomised trial of 1508 participants [14]. Results showed that at the 6-month follow-up, for the majority who were smokers at baseline (n = 1164), this intervention significantly increased 1-month prolonged abstinence rates from 11.3% in the control group to 16.4% in the intervention group (p < 0.02). These findings are consistent with previous studies, which have demonstrated a positive effect of individually tailored printed self-help material on smoking cessation [15‐17], and suggest that these effects may increase with time [9], in contrast to clinic based treatments which have large effects immediately post treatment but decline over time [18, 19]. A Cochrane review [20] reported on a meta-analysis of seventeen trials using materials tailored to the characteristics of individual smokers. While part of the effect could be due to the additional contact or assessment required to obtain individual data, there is evidence that tailored materials increase quit rates over and above standard materials and untailored materials (OR 1.42, 95% CI 1.26 – 1.61).

Two exceptions to this are recent UK general practice based studies that failed to demonstrate an increase in cessation rates for tailored feedback compared with an untailored letter [21] or with standard self help literature [22]. Both were based on the transtheoretical model (TTM) [23] which has serious conceptual and measurement problems [24, 25]. The materials used in the study by Lennox and colleagues were also lacking in the depth of individual tailoring and in practical advice on how to stop smoking [26]. Our tailored intervention differs from that used in these two studies. Development of our intervention was informed by evidence from the smoking cessation literature, using specific concepts from different theoretical models of behaviour change that have been shown to be relevant to change. It aims to change the cognitive determinants of smoking and smoking cessation, by addressing beliefs and expectations, and attitudinal ambivalence towards quitting, by enhancing perceived self-efficacy and re-evaluating the social environment. Theoretically, this is consistent with a number of social cognition models that have been applied to understanding smoking cessation. Previous research [15] has suggested the importance of tailoring to specific theoretical constructs. Matching advice to stage of readiness is common in many of the existing computerised programs. However, the use of one model exclusively, e.g. the transtheoretical or "stages of change" model (TTM) [23], to drive the tailoring does not allow for individual variation from that pattern of change or in the sequence of movement along the continuum. Stronger tailoring programs can be produced by using multiple theories to inform the tailoring process [27]. In addition to this strong theoretical base, our system is based on empirical findings (e.g. on predictors of quit attempts and success). By identifying specific factors that influence behaviour and decisions, we offer information to encourage individuals to bring about the desired cognitive states or behaviour for successful quitting. Moreover, the feedback reports were developed in consultation with smoking cessation counsellors and include conventional wisdom (e.g. the importance of setting a quit date). Our tailored feedback reports are therefore more psychologically detailed than those used previously in UK studies. The format resembles a personal letter, and differs from the promotional leaflet style used in other studies.

Since our previous study with the Quitline, [14] we have done extensive work to develop the intervention, consistent with the MRC Framework for the Development and Evaluation of Complex Interventions [28].

Smoking cessation literature has been criticised in the past for being written at a level beyond the literacy skills of many smokers [29], leading to higher levels of education associated with the use of self-help materials [16]. Results from our previous study [14] suggest that the tailored feedback may be as effective among more deprived smokers as in the less deprived. Extensive work on the development of the intervention, in consultation with literacy experts, has modified the questionnaire and tailored the format and content of the advice to different educational levels. Focus groups engaged users (i.e. smokers) in exploring the acceptability and suitability of the materials, and their suggestions incorporated into the modifications.

In addition, the acceptability and feasibility of the delivery of this modified feedback in general practice using a proactive recruitment strategy has been assessed in a pilot trial [30]. Questionnaires were sent to a random sample of 200 smokers identified from GP records from each of four practices. Results indicate that a 10% response rate can be achieved with one mailing. There are strategies that can be used to increase response rates [31], and by using some of these strategies (e.g. a reminder with a second copy of the questionnaire, a more attractive questionnaire with the use of colour) we estimate that the response rate could be increased to 15%.

The results of the pilot trial also indicate that a high proportion of the respondents are not planning to quit in the next six months [30]. It is essential that we develop methods to engage these smokers in cognitive change to encourage modification of attitudes and beliefs to promote cessation activity in this population group. To address this we have incorporated cognitive strategies and motivational interviewing techniques which are generally more appropriate for smokers not ready to quit [32], into the tailored interventions in varying degrees to suit the motivation of the smoker. In our previous study of callers to the Quitline [14], participants received a single feedback letter. However, this system can store data, to be used at a later date in combination with new information provided by the smoker. In the present study the data is being used in this way to expand the intervention by including a second assessment and letter. Multiple tailoring has been shown to be more effective than single tailoring in promoting cognitive change in smokers with low readiness to quit [33].

The aim of this study is to test the hypothesis that personally tailored feedback reports, based on an assessment of individual needs and tailored to levels of reading ability, sent to smokers identified from general practitioners lists with varying levels of motivation and readiness to quit, will increase quit rates and quitting activity over and above that found with standard self help and usual care received from the practice. The study objectives are 1) to compare the effectiveness of sending personalised computer tailored feedback reports to smokers with sending standard self-help materials 2) to explore the effectiveness of tailored feedback reports by socio-economic status to determine their effect in more deprived groups 3) to determine the characteristics of smokers who are prompted to change their behaviour after receiving tailored feedback reports.

In the UK there are 450 new smokers every day [42]. At this rate of recruitment to the smoking population, there is an urgent need to develop and deliver cost-effective interventions to the bulk of the population, rather than to the minority of dependent smokers who seek specialized help [43].

It is important to offer a range of interventions that appeal to different individual needs and preferences [44]. Computer tailored personal feedback, adapted to reading levels and motivation to quit, is a simple and inexpensive intervention which could be widely replicated and delivered cost effectively to a large proportion of the smoking population. The current portfolio of NHS services does not include a highly tailored approach for smokers who do not wish to attend clinics or have face-to-face counselling.

Reaching and changing the behaviour of smokers, particularly heavily dependent smokers and those from lower socio-economic groups is a challenging task, and the effect of any intervention likely to be small. However, contrary to common assumptions, the proportion of hardcore smokers (defined by cigarette consumption and time from waking to first cigarette), forms a lower proportion of all smokers than in 1996 [45]. Although, the prevalence of smoking by socio-economic group remains unchanged, there are no differences by socio-economic group or by cigarette consumption in intentions and desire to quit and in attempts to quit [45], suggesting that these harder to reach groups will be as receptive to help and encouragement as higher socio-economic groups and less dependent smokers. Furthermore, a modest success rate could have a large effect on public health given its recruitment potential, and make a valuable contribution to lowering smoking prevalence.

Brief advice from GPs has been shown to increase quit rates [46], and this method enables the standardised collection of relevant information from smokers by practice nurses or other health professionals, and could offer an efficient tool to integrate smoking cessation counselling into a busy primary care practice. The degree of uptake by GPs of this intervention is an important issue requiring research in its own right. However, feedback from general practices will inform the degree of uptake as well as reasons for agreeing or refusing to take part in the study, and also identify any potential barriers to the implementation of the intervention. The provision of this option to smokers in primary care could complement existing services, and work synergistically with other measures to produce more quitters and reduce the prevalence of smoking in the UK. This is also an intervention that would fit into the broader scope of tobacco control, by prompting more quit attempts, and increasing referrals to specialised services.