Toward precision smoking cessation treatment II: Proximal effects of smoking cessation intervention components on putative mechanisms of action

Abstract

Background: Understanding how smoking cessation treatments exert their effects can inform treatment development and use. Factorial designs allow researchers to examine whether multiple intervention components affect hypothesized change mechanisms, and whether the affected mechanisms are related to cessation.

Methods: This is a secondary data analysis of smokers recruited during primary care visits (N=637, 55% women, 87% white) who were motivated to quit. Participants in this fractional factorial experiment were randomized to one level of each of six intervention factors: Prequit Nicotine Patch vs None, Prequit Nicotine Gum vs None, Preparation Counseling vs None, Intensive In-Person Counseling vs Minimal, Intensive Phone Counseling vs Minimal, and 16 vs 8 Weeks of Combination Nicotine Replacement (nicotine patch+nicotine gum). Data on putative mechanisms (e.g., medication use, withdrawal, self-efficacy) and smoking status were gathered using daily assessments and during follow-up assessment calls.

Results: Some intervention components influenced hypothesized mechanisms. Prequit Gum and Patch each reduced prequit smoking and enhanced prequit coping and self-efficacy. In-Person Counseling increased prequit motivation to quit, postquit self-efficacy, and postquit perceived intratreatment support. Withdrawal reduction and reduced prequit smoking produced the strongest effects on cessation. The significant effect of combining Prequit Gum and In-Person Counseling on 26-week abstinence was mediated by increased prequit self-efficacy.

Conclusions: This factorial experiment identified which putative treatment mechanisms were influenced by discrete intervention components and which mechanisms influenced cessation. Such information supports the combined use of prequit nicotine gum and intensive in-person counseling as cessation interventions that operate via increased prequit self-efficacy.

Keywords: Factorial design; Mechanisms; Smoking cessation; Treatment.

Conflict of interest statement

No conflict declared.

Copyright © 2016 Elsevier Ireland Ltd. All rights reserved.

Figures

Figure 1

The mediation model.

Figure 2

Timing of phases, interventions, and ecological momentary assessments.

Figure 3

Summary of the significant main and interaction effects of Preparation and Cessation phase intervention components on putative mechanisms. The straight lines indicate that the effects were significant after controlling for postquit smoking; the dashed lines indicate an effect that was no longer significant after controlling for postquit smoking. See Figure 4 for the nature of the interaction effects.

Figure 4

The graphic depiction of the significant interaction effects on putative mechanisms. a. The Preparation Counseling x Prequit Gum interaction predicting prequit cigarettes smoked per day. This figure depicts the mean number of cigarettes smoked per day for each treatment group and the 95% confidence intervals. b. The In-Person Cessation Counseling x Prequit Gum interaction predicting postquit talking with others to cope. This figure depicts the mean postquit reports of talking with others to cope with stress and urges for each treatment group and the 95% confidence intervals. c. The Preparation Counseling x In-Person Counseling interaction predicting prequit motivation. This figure depicts the mean prequit motivation scores and the 95% confidence intervals. d. The Prequit Gum x In-Person Counseling interaction predicting prequit self-efficacy. This figure depicts the mean prequit self-efficacy scores and the 95% confidence intervals.

Figure 5

Mean prequit self-efficacy mediates the effect of receiving both Prequit Gum and Intensive In-Person Counseling, relative to receiving either one without the other or neither of these interventions, on 26-week self-reported abstinence. A = the path between treatment and the mediator; B = the path from the mediator to the outcome; C = the direct effect of treatment on outcome, without taking into account the indirect effect; C’ = the direct effect of treatment on outcome when the indirect effect is included in the model.