Relapse dynamics during smoking cessation: recurrent abstinence violation effects and lapse-relapse progression

Abstract

Smoking cessation is a process that unfolds over time and is characterized by intermittent lapses. We used parametric recurrent event survival analyses to better understand the dynamic relationship between a set of Abstinence Violation Effect (AVE, Marlatt & Gordon, 1985) responses to lapsing and subsequent lapse-relapse progression. Participants were 203 smokers who achieved abstinence and responded to items assessing three core components of the AVE (internal attribution, abstinence self-efficacy and guilt) following a total of 1,001 lapse episodes in near real time. Neither self-blame, self-efficacy, nor guilt following participants' first lapse predicted relapse. Controlling for responses to their first lapse, responses to each additional lapse did prospectively predict lapse progression, such that drops in self-efficacy were associated with accelerated progression to a subsequent lapse (HR = 1.09, CI = 1.02-1.15), while increases in internal attributions of blame actually protected against lapsing (HR = 0.98, CI = 0.97-0.99). Treatment with nicotine patches slowed recurrent lapse progression (HR = 0.58, CI = 0.48-0.70), but this effect dissipated over multiple lapses, and was moderated by elevated ratings of postlapse guilt (HR = 1.08, CI = 1.01-1.18), which predicted accelerated progression within the active patch group, while protecting against lapse in the placebo group. Results highlight the dynamic nature of lapse responses during smoking cessation, indicating that self-efficacy predicts progression from one lapse to the next, while attributions of self-blame and guilt influenced progression in unexpected ways.

Figure 1

Lapses distributed by the sequence they occurred (Left Axis), along with median hours of abstinence preceding each lapse (Right Axis).

Figure 2

Model-based predicted median hours of prospective abstinence preceding each lapse, plotted as a function of Active versus Placebo NRT patch assignment.

Figure 3

Additional hours of prospective abstinence time, plotted across each 1-unit change in post-lapse self-efficacy.

Figure 4

Additional hours of prospective abstinence time across each 1-unit change in post-lapse internal attribution of blame, plotted as a function of abstinence duration (days) preceding the lapse.

Figure 5

Additional hours of prospective abstinence time across each 1-unit change in post-lapse guilt, plotted separately for those assigned to Active versus Placebo NRT patch.