Retraining automatic action tendencies for smoking using mobile phone-based approach-avoidance bias training: A study protocol for a randomized controlled study

Tobacco smoking remains a major public health problem. Smoking is considered the largest preventable risk factor for morbidity and mortality worldwide, with smoking-related deaths estimated at over 5 million people per year [1]. In addition, smoking is strongly implicated in poorer quality of life [2] and imposes a considerably high economic burden on the healthcare system. For instance, it is estimated that smoking-related diseases are responsible for 1.5–6.8% of national health system expenditure [3].

Despite growing evidence about the hazardous effects attributed to smoking, prevalence rates are still high. In Germany, approximately 25–30% of the adult population smoke [4], mostly on a daily basis. Although many smokers want to stop smoking and quit attempts are frequent, relapse rates are upsettingly high [5, 6]. Even following effective treatment, only about one out of four smokers will achieve sustained (≥ 6 months) abstinence [7‐9]. Thus, there is an urgent need for developing more effective treatment options and/or improving the efficacy of the available smoking cessation interventions.

While most common interventions for smoking cessation rely on reflective reasoning (i.e. through education on smoking-related health risks, motivational interviewing, and pro–con debates), prominent theories of drug addiction [10, 11] and recent research [12, 13] insist on the notion that more automatic, hard-to-control impulsive processes are strongly implicated in the development and maintenance of addictive behaviors. According to dual-process theories, addiction arises from an imbalance between strengthened impulsive processes at the expanse of weakened reflective processes [10, 11]. Specifically, it is proposed that drug-related cues acquire motivational properties as they predict the availability of drugs, predispose drug taking and/or are associated with rewarding drug effects [14]. Hence, as addiction progresses, information processing is biased in favor of drug-related cues, resulting in various cognitive biases. For instance, research indicates that smokers automatically allocate their attention to smoking-related cues (attentional bias [15, 16]), display implicit positive attitudes toward smoking (association bias; for a review, see [17]), and automatically approach smoking-related cues (approach bias [18, 19]). While all of these biases have been linked to craving, drug seeking and drug use, the approach bias might most strongly relate to actual drug taking, since it incorporates biased information processing with actual motor movement.

Recent research shows that it is not only possible to measure cognitive biases using computer-based tasks, but that such tasks also have merit in modifying existing cognitive biases to promote healthier behavior or abstinence from drug use [12]. In this instance, the Approach-Avoidance Task (AAT) [20] has proven particularly valuable in both measuring and modifying a drug-related approach bias [18, 21]. During the task, different pictures are consecutively presented on a computer screen and participants are instructed to ignore image content and to pull or push a joystick attached to the computer depending on a content-irrelevant feature of the task (i.e., pull all images rotated to the left and push all images rotated to the right). In addition, the AAT incorporates a zoom feature: upon a pull movement, the picture becomes bigger, whereas upon a push movement the picture shrinks, creating a visual sense of approach versus (vs) avoidance. An approach bias is inferred from faster pulling than pushing a picture (e.g., a smoking-related picture). Using the AAT, an approach bias has been associated with different substances, including alcohol [22], cannabis [23], heroin [24], and recently nicotine [18]. Furthermore, an approach bias for cigarette cues has been reported in current smokers, but not in ex-smokers or never-smokers [19]. Lately, the AAT has been adapted to a training variant by changing the contingency between picture content and arm movements, namely by presenting all drug-related pictures in push-away format and all neutral pictures in pull-closer format, thus training automatic avoidance in response to drug-related stimuli. Recently, we applied four sessions of AAT-training as an add-on to a brief smoking cessation intervention in a sample of smokers hospitalized in a psychiatric ward [21]. Compared to sham training where smoking-related pictures had to be pulled and pushed equally often, the AAT-training led to a larger reduction in nicotine consumption at 3-month follow up. In another study conducted by Baird and colleagues [25], smokers motivated to quit were assigned to four sessions of AAT-training or sham training and were asked to make a self-guided quit attempt upon completion of the final training session. Results indicated that the reduction in approach bias was related to the number of days abstinent following the quit attempt. Hence, emerging evidence suggests that cognitive bias modification (CBM) by means of the AAT-training might be a useful intervention to reduce smoking or promote abstinence, but effects tend to be small in size or are even mixed sometimes [26]. Several reasons may account for these heterogeneous findings, with adherence to the demanding training protocols and generalization of training effects among the most significant. For example, although the optimal number of training sessions is still unknown (but see [27], for a systematic investigation on the optimal number of training sessions in the context of alcohol-avoidance training), it seems that training should be administered multiple times to produce stable and long-term improvements. Thus, participants are required to keep multiple training appointments, which begs the question of whether individuals are able to maintain adherence to a daily or weekly training prescription. Indeed, preliminary evidence suggests that high dropout rates are an issue of concern in CBM studies (e.g., [21, 28, 29]). The second reason for rather small training effects concerns the ecological validity of the training tasks. For instance, training takes place in a laboratory context that participants have never before associated with craving, drug taking, or beneficial drug effects. Increasing evidence from the literature on cue exposure and from extinction learning suggest that generalization is inhibited if it takes place in only one context [30‐32], leading to reduced treatment effects or increased relapse rates. Thus, CBM efficiency and effectiveness could be improved if the optimal parameters for bias modification are met, including the number and context of training sessions.

One way to improve training adherence and provide varying training contexts could be to apply them via smartphone application (app). An increasing number of people owns smartphones (i.e., 78% in Germany). Moreover, it is estimated, that 3.1 billion people will own a smartphone by the year 2021 [33]. Thus, through smartphone-delivered AAT-training, participants can perform training more often and in the exact same environment in which they otherwise smoke cigarettes, thereby facilitating generalization of training effects.

This study investigates the effectiveness of an app-based AAT-training (app-AAT) for smokers motivated to quit smoking and its effects on the malleability of the smoking-related approach bias, nicotine consumption, and smoking behavior, including cessation. It features a randomized controlled design and multi-session training by means of a newly developed app-based AAT-training during a 2-week training period. We expect that the app-AAT would reduce smoking-related approach biases as measured using the assessment version of the standard AAT [18] and reduce the number of daily smoked cigarettes over and above the control condition.

This study makes use of a randomized controlled design to explore the efficacy of an app-based approach bias retraining in a sample of regular smokers motivated to quit smoking. We seek to investigate whether our newly developed app-AAT training will modify cognitive biases, reduce smoking behavior and contribute to quit attempts and smoking cessation. To do so, an approach bias retraining app will be added to “treatment as usual”, which comprises information about smoking, a self-help book for smoking cessation, and self-monitoring daily cigarette use.

Recent work on CBM in the context of addiction and unhealthy behavior shows promise in changing maladaptive cognitive biases and/or reducing harmful behavior [12, 55]. However, several challenges have to be addressed, including small-to-medium training effects, adherence to the training, and generalization to long-term, real-world behavior. Providing the opportunity for participants to perform training via mobile phones at home represents a promising innovation in the field of CBM research, not only because the majority of people possesses smartphones, but also because individuals tend to keep their mobile phones close to hand in nearly any everyday situation. Hence, next to their ubiquity, smartphones can provide novel ways to optimize training efficacy and efficiency and to increase the number of opportunities to engage in training. This proof-of-principle study is novel in combining existing approach bias retraining for smoking with a mobile-phone-based application as an adjunct to brief behavioral counseling for smoking cessation. Our results can inform future research and clinical practice, as CBM could provide a cost-effective and easy-to-access intervention that could be used as an add-on to more traditional treatments, to promote abstinence from smoking or other harmful activities.

At the time of submission, the trial had not started. Recruitment is scheduled to begin in November 2019. The expected duration of the study is 12 months. The final results will be published as soon as possible after the analysis is completed.