Commentary on Herd & Borland (2009) and Herd et al. (2009): Illuminating the course and dynamics of smoking cessation


The two papers in this issue from the ITC study of smoking and cessation open up important perspectives on the experience of smokers as they quit and endeavor to maintain abstinence over weeks, months, and years [1,2]. The ITC study lies in a crucial gap between two common approaches to research—a gap that has hampered our understanding of cessation and relapse. On the one hand, population-level epidemiological studies have provided data on large and representative samples, but have generally been cross-sectional and lacked detailed assessment of cessation-related processes. On the other hand, clinical studies have provided rich longitudinal data on cessation processes, but have been limited by small, unrepresentative clinical or convenience samples and short time-frames. The ITC study helps bridge this gap, providing us with psychological process measures in the context of a multi-year longitudinal population study. We need long-term longitudinal data on cessation processes in large representative samples if we are to make significant progress in understanding quitting. The paucity of detailed longitudinal studies of smoking and cessation represents a significant gap in our research strategy.

As much as the ITC study's long-term longitudinal data provide important perspective on cessation and relapse process, the authors recognize that the annual measurements in the ITC may be too widely spaced to capture key processes that play out over much shorter intervals. For example, some of our research [3–5] has shown that, at least early in the quit process, there are important dynamics at much shorter intervals, with relevant shifts in craving, mood, expectancies, and efficacy, and effects on smoking, occurring from one day to the next or even across hours within a single day. In their writing on ‘temporal design’ in longitudinal studies, Collins & Graham [6] caution that when measurements are spaced too far apart, they may yield misleading conclusions. Designs with more frequent measurement and finer temporal resolution are needed for a clear understanding of cessation and relapse process. How to provide fine temporal resolution that can capture dynamic effects, while allowing observation over long periods, is a challenge that has yet to be resolved.

While limited in fine temporal resolution, the ITC study analyses demonstrate the power and utility of a broad temporal scope. By looking at cessation dynamics over several years, the analyses provide a long-term perspective on the process of quitting smoking, which can be divided into stages. Most studies have focused, with good reason, on what I'll call the ‘acute’ phase of quitting—the first few weeks, when craving, symptoms, and mood changes are most prominent and most dynamic, and when the risk of relapse is highest. The ITC analyses help distinguish this from what we might call the ‘protracted’ phase of cessation, during which both the intensity and the volatility of symptoms, concerns, etc., are declining, but still relevant. Least well studied is an implied third phase, ‘resolution,’ reached when experiences such as craving and temptation are not only absent but also irrelevant, and abstinence is stable and well-established.

Research has not clearly delineated, much less characterized, these phases, but clinical experience and data such as the ITC analyses suggest how smokers' experience may progress. In the acute phase, the smoker experiences a seemingly constant struggle to abstain and cope with intense urges and temptations, as well as with disruptive withdrawal symptoms. In the protracted phase, the struggles become easier and more intermittent. Withdrawal symptoms are largely resolved, the ex-smoker has established some comfort and confidence with abstinence, and there may be long periods when there are no urges. The smoker may still miss smoking, and abstinence is still threatened by periodic temptations that arise suddenly, often as a result of exposure to smoking cues or unusual stress. The resolution phase is most easily observed in long-time, stable ex-smokers, for whom smoking, craving, and temptations are simply not an issue, even when cues are presented or stress looms. No effort is needed to ‘maintain abstinence’ nor is there any sense of struggle—the person has simply become a non-smoker (albeit one who used to smoke). At this stage, smoking is simply not on the table as a behavioral option, which may have important psychological consequences. Studies [7,8] have shown that craving is potentiated when smoking is perceived as an available option. When a person's self-definition transitions from being ‘a smoker struggling to keep from smoking’ to being an ‘ex-smoker who does not smoke,’ their vulnerability to craving—and smoking—may disappear. The aggregate ITC data did not show an obvious asymptote, but individuals seem to reach stable resolution at some point. Research on resolution would be very useful—particularly if it suggested ways to accelerate achievement of resolution.

The ITC study and analyses in this issue point the way towards the benefits of long-term longitudinal study of cessation process. More research like this is needed.

Declaration of interest

Saul Shiffman is a cofounder of invivodata, which provides electronic diaries for research; through his work with Pinney Associates, serves as consultant to GlaxoSmithKline Consumer Healthcare exclusively regarding matters relating to smoking cessation and also is developing new nicotine medications.