Beyond BCTs and published materials: reliability coding of behaviours and application methods within smoking cessation

Abstract

Background: The behaviour-change-technique (BCT) taxonomy v1 (BCTTv1) is widely used for describing and identifying BCTs in interventions. This systematic review of smoking cessation interventions extends BCTTv1 coding by evaluating whether the behaviour targeted by coded BCTs (ie quitting, abstinence, medication adherence, intervention engagement) and the method of BCT application (BCT tailoring, active/passive engagement of BCT recipient) can be reliably determined. We also evaluated the usefulness of requesting additional intervention materials from study authors. Methods: Published articles and supplementary materials were examined by 2 independent coders for BCTs regarding targeted behaviour and method of delivery. First, second and last authors of all studies were contacted for additional intervention materials. Inter-coder reliability analyses were used to examine the data. Findings: 148 randomised controlled trials were identified. In total 2546 intervention BCTs were coded (Kappa=0.984, PABAK=0.999, %agreement =99.93 [Yes=2.24, No=0.07]). Inter-coder reliability for BCTs tailoring (PABAK=0.993, 2577 instances), for passive/active recipient (PABAK=0.991, 2547 instances) and for targeted behaviours [quitting, abstinence, medication adherence, or intervention engagement] (PABAK=0.996-0.999) were produced. Authors of 64/148 trials (43.24%) provided additional intervention material in which 930/1757 (52.93%) of the BCTs from these 64 intervention trials were identified. Discussion: Method of BCT application (tailoring, active/passive recipient) were coded with moderate to good reliability, whereas targeting specific smoking cessation behaviours produced excellent reliability coding. Contacting study authors for additional intervention materials adds substantial amounts of information to intervention content. Both methods are recommended for future systematic reviews