What kind of smoking identity following quitting would elevate smokers relapse risk?

INTRODUCTION

Most quit attempts by smokers end in relapse back to smoking. Relapse rates for unaided and aided attempts to quit smoking are high, particularly within the first few weeks of abstinence (1). Of smokers who reach 4-week abstinence, approximately 70% relapse within a year (2, 3). Additionally, upwards of thirty quit attempts required before a smoker successfully quits smoking (4). By identifying the challenges and supporting factors for successful abstinence, the rate of long-term abstinence may be improved (5). One key factor that has been hypothesized as contributing to smoking persistence is smoking identity (6). Understanding how smoker identity can contribute to the success/failure of behaviour change during a quit attempt could help improve interventions to prevent relapse during this vulnerable period (7).

Comprehensive theories of health behaviour change such as PRIME theory (8) and CEOS theory (9) both postulate an important role for self-identity in influencing behaviour. PRIME theory defines identity as the mental representations that we form of ourselves and the thoughts and feelings we attach to these (8). Identity is an important source of our motives, providing us with the ability for self-regulation, boundaries and stability of behaviour. Identity is composed of labels (i.e., the categories we use to define ourselves), attributes (i.e., the characteristics that arise from labels) and rules (i.e., the behaviours we permit ourselves). According to Tombor et al. (10), it is through these elements that a person’s identity directs behaviour and generates motives that may override an impulse.

CEOS theory focuses more on the role of self-identity as a manifestation of people’s values and expectations and it is this which influences the value we give specific behaviours (9). Both theories predict that changing an established identity to one where the act of smoking and its associated experiences are less valued should reduce the desire to return to an old behaviour and thus provide a valuable focus for behaviour change intervention (8, 9). Research has supported the influential role a person’s identity plays in their actions and attitudes towards smoking. A positive image of oneself as a smoker predicts continuation of smoking and reduces susceptibility to take on evidence of health risks for smoking behaviour (11, 12). Cessation attempts are less desirable and less likely to have occurred in smokers who maintain a positive smoker identity (13, 14). Research has shown that a non-smoker identity is strongly associated with smoking cessation (15).

During smoking cessation, there can be a shift from a smoker identity to a non-smoker identity (12, 16). Where this happens, both theories predict that the adoption of a ‘non-smoker’ identity should provide an important protection against smoking relapse. Identifying as a non-smoker would allow for a revision of one’s values (9) along with motives and evaluation (8) allowing for rules to be generated to override motives associated with smoking behaviour.

Failure to adopt a non-smoker identity has been linked to relapse. Following a quit attempt, many smokers report developing a sense of lost social identity (17). This disparity between current behaviour and identity as a ‘smoker’ has been highlighted as resulting in relapse in order to realign identity with behavioural outcomes. Further, research has indicated that in the short-term, smokers who continue to retain smoking behaviour as an important part of their identity are less likely to remain abstinent after three months than those who adopt an abstainer self-concept (18).

Despite the pivotal role that identity plays in smoking behaviour, only a handful of studies have explored the potential long-term protective benefits for relapse prevention that the adoption of a non-smoker identity could provide (10, 19). Research reviewing the effectiveness of behaviour change techniques (BCT) in English stop smoking services (19) found that BCT focused on strengthening an ex-smoker identity were associated with higher quit rates at 1-month post treatment. Another English study examined non-smoker identity beyond 1-month post-quitting and relationship to smoking abstinence and found that majority of participants who had quit smoking identified themselves as non-smokers at the 3-month and 6-month follow up assessments, and those who did so were more likely to have remained abstinent (10). Taken together, both of these studies provide some support for the role of identity in promoting continued and sustained positive behaviour change, with strong suggestion that non-smoker identity can have a protective effect against smoking relapse.

A major challenge of interpreting the findings of past research is the lack of a consistent definition and/or operationalization of identity construct. It is unclear whether an ex-smoker identity differs from a non-smoker identity as the terms are often used interchangeably. Our research aimed to address this shortcoming to further understand the relationship between post-quit smoker identities and subsequent risk of relapse. The sample was taken from a non-treatment seeking smoker population who participated in a longitudinal cohort-based survey allowing for assessment of relapse past the initial stage of cessation and ensuring results will be generalizable to a larger population. Unlike past research, the survey question used four instead of two categories to assess smoking identity, allowing a more nuanced measurement of identity transition. Whilst research supports a post-quit transition of a smoker identity to non-smoker identity (16), a significant minority of long-term ex-smokers have been found to retain both a smoker identity and residual attraction towards smoking (7). Research has found smokers who retained a residual attraction to smoking could be more at risk of relapse (20, 21). Meijer, Vangeli, Gebhardt and van Laar (22) conducted a qualitative, longitudinal investigation that found those who identified as a ‘smoker who does not smoke’ found abstinence from smoking more challenging to maintain than those who identified as a non-smoker. Continued residual attraction to smoking and maintenance of a smoker identity in some former smokers suggests it is not a straightforward shift in identity from a smoker to non-smoker, but a more nuanced shift of attraction and identification throughout the abstinence process (23). To account for this, smoker identity was distinguished into those who saw themselves as smokers trying to quit smoking with no clear rule versus those who had set a clear rule for stopping smoking. A distinction between non-smoker versus ex-smoker identity was also made, the latter being those who still held some residual image of themselves as a smoker versus the former who did not want to be associated with past smoking identity and saw themselves now completely as a non-smoker.

Past research has established abstinence self-efficacy, post-quit urges to smoke, enjoyment of smoking and perceived importance of smoking as strong key predictors of both attempts at quitting and relapse to smoking in the short-term (21, 24, 25). Our study included these variables in analysis to further examine whether the predictive effect of post-quit smoker identities is over and above that of these other predictors of relapse.

Based on the theories, existing research in this area and our conceptualisation of smoking-related identity post quitting, it was hypothesised that following smoking cessation, smokers who perceived themselves as either a smoker who had chosen to no longer smoke, an ex-smoker or a non-smoker would be less likely to relapse than those who perceived themselves as a smoker trying to quit. Further, it was hypothesized that smokers who adopted a non-smoker identity following smoking cessation would be less likely to relapse compared to both those who saw themselves as smokers who had chosen to no longer smoke and as ex-smokers. No prediction was made for the latter two groups. As a secondary aim, differences in identity-abstinence relationships were explored across UK and Australia.

METHODS

RESULTS

DISCUSSION

The present study examined the association between smoker identity and risk of relapse among adult ex-smokers in Australia and the UK. Confirming our hypothesis, smoker identity significantly predicted smoking relapse, even after controlling for sociodemographic variables, quit duration, survey mode and year of recruitment. Former smokers who identified themselves as a smoker who was trying to quit had the highest rates of relapse and were significantly more likely to relapse when compared to those who identified as either a smoker who had chosen to no longer smoke, a non-smoker or an ex-smoker.

Contrary to the hypothesis, risk of relapse was not significantly different between subgroups identifying as either a smoker who had chosen to no longer smoke, a non-smoker or an ex-smoker although the percent relapsed among the smoker subgroup was somewhat higher than that of the non-smoker and ex-smoker subgroups. Also, of note is that the identity of a smoker was the most influential predictor of relapse, over and above other known predictors, such as abstinence self-efficacy which was also found to be significantly associated with relapse risk. The predictive relationship between smoker identity and risk of relapse did not significantly differ by country.

Consistent with past research (10, 18), results show that smoker identity is significantly associated with outcome of a quit attempt. This extends on findings from Tombor et al. (14) that the establishment of a non-smoker identity following a recent quit attempt was associated with successful smoking abstinence. While reasons for this association were not explored, it is consistent with the theorized proposal that formation of a smoker identity no longer associated with smoking provides a basis for long-term behaviour change through establishment of absolute non-smoking rules (8, 9). Extending on research which found an association between smoker identity and relapse in the early periods of quitting (10), our results suggest that post-quit smoker identity continues to play an important role for long-term relapse (i.e., beyond 12 months). However, it may have been the implicit commitment to not smoking again that is more important given the low relapse rate in the group who still identified as smokers but were committed to not smoking again. This suggestion is supported by research highlighting the commitment to quit smoking as a key predictor of smoking relapse (28, 29). The absence of direct measurement of level of commitment limits the conclusions that can be drawn, but future research could seek to further understand commitment as a function of identity and the potential relapse protection it offers.

Smoker identity has been recognised as more complex than a binary smoker vs. non-smoker construct (6) and this study attempted to expand on previous research by capturing smoker identity in more detail. A key finding was that those who were not as committed to a non-smoker identity, labelling themselves as a smoker trying to quit, were most likely to relapse. Tombor et al (6) suggested the degree to which a specific smoker identity is internalised drives motivation to engage or refrain from smoking. The label of ‘a smoker who is trying to quit’ acknowledges that the individual is a smoker refraining from a behaviour to which they still closely identify. The degree of commitment to quitting smoking does not reflect that of an individual who labels themselves as a non-smoker, dissociating from smoking behaviour completely.

The results suggest the way former smokers define their identity can impact the success of quit attempts. PRIME theory encapsulates this as rules undertaken by a smoker (8) and resulting motives that drive behaviour. Rules defined by attempting not to smoke may not provide high enough motivation levels to resist smoking in all circumstances. However, those with rules that define no smoking behaviour at all, boundaries are clearly established and power to prevent smoking relapse is motivated. Identification as a smoker trying to quit may form a label that is less committed to cessation and an absence of rules providing motivation to prevent relapse.

Past research has found that individuals who identify as a ‘smoker who does not smoke’ find quitting to be challenging and that some element of non-smoking needs to be incorporated into identity to reach stable abstinence (30). However, our finding seems to contradict this given the low relapse rate observed among former smokers who described themselves as a smoker but committed to not smoking again which suggests that identity label is less important than level of commitment to a non-smoking lifestyle and having the non-smoking rules established in maintaining abstinence. Nevertheless, the important role of identity self-label cannot be entirely ruled out since an alternative explanation might be that this subgroup of former smokers who still described themselves as smokers represented mainly those in the early stages of quitting who would eventually transition to an ‘ex-’ or ‘non-smoker’ identity after further months of abstinence which would lower their relapse risk. This possibility appears to be supported by the association found of shorter quit duration at baseline for the two ‘smoker’ groups compared to the ‘ex-’ and ‘non-smoker’ groups (see Appendix Table 1A).

Quit duration was identified as an additional independent predictor of relapse in the analysis, supporting research that has established this as a key predictor in success of long-term quit attempts (10, 31). Studies have found that a non-smoker identity is more prevalent and identification with smoking decreases amongst quitters after longer abstinence length (32). Transition from a smoker to a non-smoker identity, and formation of rules that provide motive to successfully abstain from smoking occurs over time (12, 16) and high relapse rates before successful abstinence may be due to these not yet being established. Many determinants of relapse change as a function of length of abstinence (21, 33). Future research should seek to more clearly define how smoker identity changes as a function of length of abstinence and the resulting impact on success of a quit attempt. It may also be useful to explore how self-identity changes in the early days of a quit attempt, we only asked those quit for at least 1 month.

In addition, abstinence self-efficacy was identified as an independent predictor of relapse supporting research which has found individuals with higher confidence in their ability to quit smoking are more likely to achieve a successful quit attempt (21, 34). This could be linked with the conceptualisation of an individual’s post-quit identity and the rules that were established. The commitment an individual has towards the quit attempt and the rules that are established as a result could be directly influenced by their confidence in the success of their quit attempt. Future research should look at how these predictors interact and influence the success of a quit attempt.

The lack of country differences in the predictive relationship between smoker identity and relapse risk could be limited to the two countries studied. It would be interesting to explore it in countries with quite different cultural and tobacco control traditions.