The State of the Science on Cancer Diagnosis as a “Teachable Moment” for Smoking Cessation: A Scoping Review

Abstract

Introduction

Theoretically, a cancer diagnosis has the potential to spur health behavior changes in physical activity, diet, substance use, medication adherence, and the like. The Teachable Moment heuristic is a parsimonious, transtheoretical framework for understanding the conditions under which behavior change might occur, with constructs that include affective, cognitive, and social factors. Application of the Teachable Moment to smoking cessation after cancer diagnosis might aid selection of predictors in observational studies and inform how to optimally design interventions to promote quit attempts and sustain abstinence, as many smoking cessation interventions for cancer survivors do not yield positive outcomes.

Aims and Methods

This scoping review of 47 studies that span nearly 20 years of literature examines the measurement of the Teachable Moment constructs and what empirical support they have in explaining cancer survivors’ smoking behavior.

Results

From this review, it appears the construct of affective response is more widely explored than risk perceptions, social role, and self-concept. Strong, negative affective responses (e.g., anxiety, general distress) may be a powerful contributor to continued smoking after a cancer diagnosis. Risk perceptions may also play a role in smoking behavior, such that never and former smokers espouse stronger perceptions of smoking-related risks than current smokers. Finally, due to a paucity of studies, the role of cancer survivors’ self-concept (e.g., identity as a “cancer survivor”) and changes in their social role (e.g., employee, athlete) are unclear contributors to their smoking behavior. In summary, the Teachable Moment holds promise in its application to smoking cessation after a cancer diagnosis, though more direct research is needed.

Conclusions

This scoping review of the scientific literature is the first formal test of the extent to which cancer diagnosis has been explored as a “teachable moment” for smoking cessation, with results that provide insight into issues of measurement precision and breadth as well as empirical support of the “teachable moment” heuristic.

A plethora of health behavior performance and change theories exist, with varying levels of foci that range from the individual to the community to even system- and environmental-level contributors. Across health behaviors (e.g., physical activity, dietary habits, tobacco use, dental hygiene, cancer screening, vaccination, treatment adherence), the theories most widely applied and tested are those that center on the individual.^1–3 The basic question is: What personal characteristics increase the likelihood of engaging in health promotion activities (e.g., wearing sunscreen) and/or decrease the likelihood of engaging in risky health behaviors (e.g., binge drinking)? There is some consensus in terms of the personal characteristics identified in the most popular and empirically supported theories. For example, the Health Belief Model,⁴ Theories of Reasoned Action and Planned Behavior⁵ and Transtheoretical Model⁶ all indicate that intention, self-efficacy, and attitudes towards change are relevant to understanding and predicting health behavior. As another example, the Social Cognitive Theory⁷ and Social-Ecological Model⁸ focus on the interplay between individuals and their environment and highlight social norms, social/peer facilitation, and observational learning as important variables to consider. In summary, there appear to be some characteristics that many theorists agree are germane to an individual’s health behavior performance and change.

According to the Teachable Moment (TM) heuristic,^9–11 a major health or life event has the potential to serve as a powerful catalyst for health behavior change,^12,13 but only if certain conditions are met. An individual must experience a strong affective response, a change in perception of personal risk, and either a challenge to her or his self-concept or social role to capitalize on the cueing event.^9–11 Thus, the TM contains many of the same general ideas and overlapping personal characteristics found in some health behavior theories. What is relatively unique is the TM’s direct linkage to an external, cueing event and its inclusion of affective, cognitive, and social factors. The strength of the TM for the selection of predictors in observational studies and targets for intervention, especially in comparison to alternative theories of health behavior, maybe its parsimony, though it remains necessary to demonstrate its empirical support.

Theoretically, a cancer diagnosis has the potential to function as a TM. A cancer diagnosis is an objective event to which a person has a subjective response, and the individual response is what may ultimately facilitate or hamper health behavior change. Cancer has the capacity to impact the three main components of the TM (specifically, affective response, risk perception, and self-concept and social role), though there is no predetermined order in which these changes may unfold. First, cancer survivors are likely to perceive their diagnosis as a risk to their physical health and wellbeing, especially in light of the common symptoms of cancer as well as its treatment side- and late-effects (e.g., pain, fatigue, insomnia, anorexia, neuropathy, infertility).^14–16 These physical problems, in concert with the sometimes serious and imminent threat to survival, oftentimes result in cancer having a clear, negative impact on the perception of personal vulnerability. Second, cancer diagnosis and its sequela may result in the experience of distress, which can present as fear, anxiety, depression, or shock, among other things.^17–20 Finally, individuals might experience a change in self-concept as they adopt an identity of “cancer patient” or “cancer survivor” ²¹ or learn to cope with disruptions in their ability to carry out their normal social roles and responsibilities (e.g., parent, employee, community leader).²²

The ability of cancer to influence tobacco use behavior specifically is of paramount importance, as tobacco use after a cancer diagnosis has an undeniable effect on cancer prognosis. To explain, cigarette smoking and other tobacco use play a causal role in risk for second primary cancer, cancer-specific mortality, and all-cause mortality and it is associated with increased risk of cancer recurrence, poorer response to cancer treatment, and treatment complications.²³ Despite this clear negative impact, studies find that approximately 20–30% of cancer survivors classify as current tobacco users, with wide variation in these estimates across disease site, time since diagnosis, and other variables.^24–26 Many studies have examined demographic, clinical, and psychosocial correlates of cancer survivors’ tobacco use (for example, see^25,27,28), but the application of a theoretical framework to the selection of predictor variables is uncommon, which makes it difficult to synthesize results across studies and design interventions that capitalize on those variables that are most impactful. Simply put, if the TM holds promise in explaining and predicting cancer survivors’ tobacco cessation, both research and clinical pursuits stand to benefit.

Current Study

The current study is a scoping review of the literature that explores the subordinate constructs of the TM as correlates of smoking or other tobacco use among cancer survivors. To the authors’ knowledge, no published literature review has examined the TM in this context and due to the exploratory nature of the study and desire to examine if there are gaps in the literature, a scoping review is appropriate.^29–31 Like other reviews, scoping reviews aim to provide a knowledge synthesis, though they often differ from systematic reviews in terms of the breadth of the research question (i.e., broad versus well-defined, respectively), the heterogeneity of studies included in the review (i.e., more versus less, respectively), and the typical manner in which the major findings are presented (i.e., narrative versus quantitative, respectively).^29–31 Here, the goal is to provide a “big picture” narrative synthesis of the relevant literature, with the awareness that there exist very few studies that have the express purpose of evaluating the capacity of the TM to explain the continuation or cessation of smoking among cancer survivors or other patient samples (for exceptions, see^9,32–34). Consequently, this study has these aims: 1) examine the extent to which the individual TM constructs have been evaluated as correlates of cancer survivors’ smoking, 2) provide guidance about how to measure the TM in cancer survivors, and 3) explore what empirical support the TM has for understanding cancer survivors’ smoking.

Methods

Data Source

This study follows the Arksey and O’Malley framework for scoping reviews.³¹ A search of peer-reviewed, English language publications over the last 20 years was conducted in the CINAHL, PsycINFO, PubMed, and Web of Science databases by a reference librarian. Within each database, separate searches were conducted for each TM construct to aid identification of potentially relevant results, and then all search results were combined in a single bibliographic reference management tool, and duplicate records were removed to yield the final database. First, to capture the TM as a whole, two terms were used: teach* moment and teach* opportunity. Second, to capture affective response, four terms were used: affective response, expressed emotion, mood*, and emotion*. Third, to capture risk perception, two terms were used: risk belief and risk perception*. Finally, to capture social role and self-concept, two terms were used: self-concept and social role. Each of the aforementioned TM-related searches was combined with keywords meant to capture the population (cancer*, neoplasm, carcinoma*, oncology*, survivor*, and patient*) and behavior (smoking, smoker*, tobacco, tobacco use, tobacco smoking, smoking cessation, quit attempt, motivate*) of interest using the Boolean operator “AND.” In sum, the literature search focused on the TM within the context of cancer survivors’ smoking and other tobacco use.

Eligibility Criteria

Eligibility for this review required the record to: 1) be written in English; 2) be published between 01/01/1999 and 12/31/2018; 3) be a peer-reviewed empirical article that described a quantitative or qualitative study; 4) contain a sample of cancer survivors; 5) measure a construct from the TM; 6) measure one or more aspects of tobacco use, such as smoking status or quit attempts, after cancer diagnosis; and 7) report the association between one or more TM to construct and tobacco use behavior.

Search Results and Coding Procedures

The database search resulted in 599 unique records. The titles and abstracts (and as needed the full-length article) of these records were assessed for eligibility by two independent coders (GEP and TB or WB). In the case of discrepancy between coders, the consensus was reached through facilitated discussion by a third coder (JLB). Of the 599 records, 106 were considered further for inclusion in the review. At this stage, the full-length article was reviewed against the aforementioned eligibility criteria and duplicate data from other records. To identify additional relevant literature not yet considered, a manual search of the reference lists of select prior reviews and meta-analyses was conducted,^35,36 in addition to the review of the reference lists from provisionally included articles (for example,^27,37–39) and publication records from select investigators in this research area. These search strategies led to the addition of 65 new records. The full-length article of the now 171 records was reviewed against the eligibility criteria and examined to identify duplicate data, ultimately leading to 47 records in this review.^27,37–82

Each record in this review was examined thoroughly and data were extracted to describe the study design, sample characteristics, and construct/behavior measurement and elucidate the key finding related to the TM and cancer survivors’ tobacco use. A pair of coders were employed to ensure a high-quality review such that one coder served as the primary coder (GEP) for the 47 records in this review while a second (TB or WB) provided quality control by independently coding 20% of those records. For the 10 records that were double-coded, there was 82% agreement across all variables, indicative of moderate-to-high inter-rater reliability.⁸³

Data Extraction

Data were securely stored in a Research Electronic Data Capture (REDCap) database. The variables coded for each record include publication year, study design (e.g., cross-sectional), study location, sample size, demographic (e.g., age, race), and clinical (e.g., cancer type, cancer stage, and treatment stage) characteristics of the sample, TM construct measurement (e.g., single item, standardized scale), and tobacco use (e.g., prevalence of lifetime use, tobacco product type). Additionally, descriptive information about the analytic approach to examining the relationship between TM and tobacco use was extracted (e.g., bivariate correlation). Finally, key findings were summarized in narrative form for each record.

Results

Study Design

Supplementary Table 2 contains methodological details about the 47 studies. To summarize the studies’ design, 6.4% (n = 3) were interventions,^41–43 36.2% (n = 17) were longitudinal observational studies^38,44–59 and the remaining 57.4% (n = 27) were cross-sectional observational studies. The number of cancer survivors in each study ranged from 49⁴² to 10 969,⁶⁰ with a median of 211. The majority of studies were conducted in North America (70.2%, n = 33), with the remainder coming from Asian countries (12.8%, n = 6),^{37,61,62,78,81,82} European countries (10.6%, n = 5),^{55,58,60,79,80} and Australia or New Zealand (6.4%, n = 3).^43,56,57 Studies primarily recruited participants from clinics and hospitals (78.7%, n = 37), but population-based surveys (21.3%, n = 10),^{44–46,48,56,60,62,72,77,79} cancer registries (2.1%, n = 1),⁸² and online methods were used as well (2.1%, n = 1) ⁷⁸; options not mutually exclusive.

Sample Characteristics

The sample characteristics for the 47 reviewed studies are summarized here and Supplementary Table 1 has these details. Across studies, females comprised an average of 43.2% (SD = 18.6) of the sample. Participants’ average age was 56.3 ± 12.2 years, but as some studies reported age in terms of a median^43,52,66,68 or range,^{45,56,60–62,69,77} this average is not inclusive of all 47 studies. The racial and ethnic background of participants was unclear in 12.5% (n = 6) of studies.^{53,55,56,58,63,82} For studies that reported race and ethnicity (85.1%, n = 41), an average of 81.3% (SD = 23.8) of participants were White, non-Hispanic. Most samples were comprised of participants of varying disease sites, and across all studies, head/neck (42.6%, n = 20), lung (44.7%, n = 21), leukemia and lymphoma (29.8%, n = 14), breast (29.8%, n = 14), and colorectal (27.7%, n = 13) cancer were the most common diagnoses. Fifty-two percent (n = 25) of studies reported participants’ cancer stage, and Stage 1 (local disease) was the most common stage represented across studies (46.8%, n = 22).^{38,40,47–53,57–59,63–68,73–77} Similar to disease site and stage, most studies (61.7%, n = 29) included participants at varying phases of cancer survivorship, though 21.3% (n = 10) only included participants diagnosed in the last year,^{37,43,50–52,54,56,57,65,68} 4.3% (n = 2) included only participants who were 1–4 years post-diagnosis^39,82 and 12.8% (n = 6) only included participants who were 5 or more years post-diagnosis.^{45,46,48,62,69,78}

Measurement

Tobacco Use

All 47 studies measured smoking status (i.e., current versus former and/or never smoker) as the primary indicator of tobacco use behavior. Nearly all studies focused on cigarette smoking (95.7%, n = 45), though 4.2% (n = 2) also measured other forms of tobacco use, such as smokeless tobacco, snus, electronic cigarettes, or hookah.^27,70 Across studies, on average, 78.7% of participants smoked at some point in their lifetime and 34.2% of participants were current smokers. In terms of precisely how “current” smoking was operationalized, 7-day point prevalence was the most common definition (23.4%, n = 11)^{41–43,45–47,50,52,54,59,69} followed by 30-day point prevalence (14.9%, n = 7)^{39,49,65,66,71,72,74} and 1-year point prevalence (2.1%, n = 1),⁷⁰ with the remaining studies not including an operational definition of this outcome (59.6%, n = 28). Only 21.3% (n = 10) of studies^{27,42,47,51,53,54,64,72,73,80} explored number of cigarettes per day as an outcome, and across these studies, current smokers reported an average of 18.5 (SD=5.6) cigarettes per day.

Smoking abstinence and/or quit attempt behavior were described in roughly half of the 47 studies (53.2%, n = 25).^{27,37–41,43,44,46,48,51–54,58,60,61,65,66,69,71,73,78,81,82} The variation and ambiguity in outcome measurement was high across studies, with some studies reporting continuous abstinence since cancer diagnosis (16.0%, n = 4),^39,48,54,63 others reporting the occurrence of quit attempts since cancer diagnosis (32.0%, n = 8),^{39,41,45,46,53,72,73,81} and still others not specifying the measurement of these outcomes.^38,44 Consequently, it was not possible to synthesize abstinence-related findings across studies.

Affective Response

Most studies (87.2%, n = 41) measured affective response.^{37–49,51–58,60–62,64,65,67–71,73–82} Of these, 68.3% (n = 28) measured depressive symptoms,^{27,37,39–42,44,47,49,51–58,61,62,64,65,70,71,73,74,76,79,80} 26.8% (n = 11) measured anxiety,^{37,42,44,45,51–53,56,58,64,70} 21.3% (n = 10) measured general distress,^{43,45,46,48,67,69,70,77,78,82} 12.2% (n = 5) measured cancer-related worry,^{54,60,64,75,78} and 9.8% (n = 4) measured emotional well-being.^38,57,74,77

The most frequently used measures of affective response were these standardized measures: the Center for Epidemiological Studies Depression Scale (CES-D) (21.3%, n = 10),^{37,39,44,47,54,65,71,73,76,79} Hospital Anxiety and Depression Scale (HADS) (17.0%, n = 8),^{42,51,53,56,58,61,73,80} Brief Symptom Inventory (8.5%, n = 4),^45,46,57,70 and Impact of Events Scale (IES) (9.5%, n = 4).^41,54,64,68 Less common were other multi-item measures like the Medical Outcomes Survey Short Form scales (6.7%, n = 3)^43,74,77 or single item measures (19.5%, n = 8) such as “Would you call yourself depressed?” ⁵⁵ and “Have you felt constantly sad or hopeless for over 2 weeks in the past one year?”.⁶²

Risk Perceptions

Several studies (21.3%, n = 10) measured risk perception.^{27,37,41,46,50,54,59,66,68,72} Of those studies, the majority 63.6% (n = 7) measured participants’ perceptions of the cancer-specific risks of smoking, such as risk for developing another cancer or poorer cancer treatment outcomes.^{37,46,50,54,59,66,72} Other indices of risk perception included participants’ perception of the cancer-specific benefits of quitting, such as fewer treatment side-effects and lower likelihood of cancer recurrence.^27,66,72 Finally, participants’ perception of smoking-related health risks^37,41,66 and quitting-related health benefits^53,66 that were not directly tied to cancer were also assessed.

The most commonly used standardized measure of risk perception was the Perceived Importance of Health Protection Scale, albeit only used in two studies.^41,46 Many other studies utilized multi-item measures, but with no overlap in the measure used across studies.^{27,37,50,54,59,66,72} Some examples of risk perception items are “If you were to continue smoking, what do you think your chances are of developing another head and neck cancer?” ⁵⁰ and “If I use tobacco, I am more in danger of developing health problems than others my age who were never treated for cancer”.⁶⁶

Self-Concept and Social Role

Only 4.3% (n = 2) of studies explored self-concept^63,71 and none studied social role as a correlate of tobacco use behavior. Of the two pertinent studies, one measured self-perception using the Explanatory Model Interview Catalogue, which examines several aspects of stigma, including a diminished sense of identity due to physical illness,⁶³ and the other measured cancer-related post-traumatic growth using the Post-Traumatic Growth Inventory, a 10-item measure exploring the perception of personal growth experienced after a traumatic event.⁷¹

Key Findings

Affective Response

Current smokers and other tobacco users endorse a variety of negative affective responses, including but not limited to general distress, depressed mood, guilt, anxiety, and fear and worry specifically about future cancer outcomes (e.g., progression, new diagnosis). When compared to former or never smokers, current smokers tended to report heightened levels of general negativity (e.g., ^{39,47,52,56,61,64}), though fear of cancer recurrence and cancer-related intrusive thoughts demonstrated mixed associations with cancer survivors’ smoking status.^{41,58,60,64,78} The relationship between markers of negative affective responses and smoking cessation outcomes was less clear. For instance, a cross-sectional study suggested depressive symptoms are associated with continuing to smoke,⁴⁰ yet in longitudinal studies, depressive symptoms were not uniformly predictive of making a quit attempt or maintaining abstinence.^42,45,46,54 Studies were also mixed with respect to findings of cancer-related fear and worry, with some suggesting it was associated with persistent smoking (or never making a quit attempt), others indicating it was a risk factor for relapse after a quit attempt, and still others suggesting it may function as a protective factor for continuous abstinence among former smokers.^41,53,54 Although less common, some studies described positive affective responses, most of which showed higher levels of emotional well-being in quitters and abstainers relative to smokers.^57,74

Risk Perceptions

Many cancer survivors with and without a tobacco history could identify the health risks associated with smoking in general, with some current and former smokers also able to draw connections between how changes in their tobacco use might impact their long-term health outcomes. Risk perceptions tended to differ by smoking status, such that many studies found never smokers and former smokers tended to ascribe greater risk to smoking than current smokers.^27,54,59,72 Indeed, current smokers were more likely to report low overall risk perceptions, perceive fewer negative consequences from smoking, and perceive few benefits to smoking cessation while also reporting the benefits of continued smoking (e.g., stress reduction, consistency with perceived social norms).^66,72 Across cancer survivors, the perception of greater health risks due to smoking was associated with a higher intention to abstain from smoking in the future,³⁷ although the findings for behavioral outcomes like the occurrence of quit attempts and long-term abstinence were mixed.^{41,46,50,54,59,72}

Self-Concept and Social Role

Only two studies measured self-concept and social role as correlates of cancer survivors’ tobacco use. First, the elements of cancer survivors’ self-concept having to do with self-blame and internalized stigma were associated with positive tobacco-related behavior change (e.g., quitting or reducing tobacco use).⁶³ Second, posttraumatic growth, which was conceptualized in this review as an element of self-concept and is always viewed in positive terms, was not significantly associated with cancer survivors’ tobacco use.⁷¹ However, given the limits inherent in the synthesis of only two studies, the contribution of self-concept and social role to understanding cancer diagnosis as a TM for smoking cessation remains unclear.

Discussion

The TM illustrates how major changes in one’s life can influence health behavior performance and change.^9–11 As a whole, the TM is consistent with the components of many health behavior theories, with its constructs overlapping those from the Health Belief Model,⁴ Theory of Planned Behavior,⁵ Transtheoretical Model,⁶ Social Cognitive Theory,⁷ and Social-Ecological Model⁸ which gives credibility to its basic premise. Despite its promise as a simple, yet potentially adequate explanation for how motivation and preparation for smoking cessation could arise after a cancer diagnosis, the TM has yet to be thoroughly explored in the context of cancer survivorship. This scoping review was designed to help fill this gap in the literature.

The 47 studies in this review were predominantly cross-sectional and conducted in North America. Samples were mainly drawn from clinics and hospitals, and individuals diagnosed with tobacco-related and non-tobacco-related cancers are represented. Time since diagnosis and phase of cancer survivorship varied across studies, but most samples were comprised of cancer survivors diagnosed within the last five years, which may or may not encapsulate the window in which cancer diagnosis would impact smoking behavior. In the studies reviewed, when current tobacco use was operationalized—which was not the norm—it was most often based on 7- or 30-day point prevalence, consistent with the general consensus on the definition of a current smoker.^84,85 The majority of cancer survivors in this review were lifetime tobacco users with approximately one-third described as current tobacco users, as is commonly found in studies with predominate lung or head/neck cancer survivors,⁸⁴ or studies where the focus is on tobacco use after cancer diagnosis^25,27,39 as is the case here.

This review set out to inform the measurement of the TM’s three central constructs (namely, affective response, risk perception, and self-concept/social role) and evaluate their association with cancer survivors’ smoking behavior. Results indicate the most research attention centers on variables that tap into affective response followed by risk perception, with self-concept/social role a distant third. This is interesting as most health behavior theories that focus on the individual give less attention to affective as compared to cognitive and social influences.^1–3 That said, it is possible the literature reviewed herein follows this pattern because cancer survivors’ emotional functioning has been a long-standing topic of scientific inquiry,⁸⁶ taking center stage in cancer survivorship research.

Popularity aside, some conclusions and recommendations are warranted regarding the measurement of TM constructs. First, in the studies reviewed, the affective response was principally defined in terms of depressive and anxiety symptoms, and while these are common psychological responses to the cancer experience,^17,18 negative effect is broader than these symptom pools, positive affective responses should not be ignored⁸⁷ and the TM indicates both negative and positive affective responses can spur motivation for health behavior change.^9–11 Consequently, future studies should measure affective response broadly, encompassing negative and positive functioning alike, as that would allow tests of the extent to which dramatic or abrupt changes in cancer survivors’ experiences of distress and/or wellbeing lead to changes in smoking behavior. If response burden is a concern, then one could narrow the focus to measures of depressive symptoms, cancer-specific distress (e.g., fear of recurrence), and emotional wellbeing, as that could help elucidate some of the mixed results observed here (e.g., ^{40-42,45,46,53,54,57,58,74}). Second, measures of smoking-related risk perception varied from single items²⁷ to multi-item standardized measures (e.g., ^41,46,50), with further variability in item content (e.g., cancer-specific benefits of continued smoking^66,72 versus general health risks of smoking⁴¹). Precisely how one measures risk perceptions has direct implications for how one interprets the results of any analysis and how one designs smoking cessation interventions. Best practices for risk perception research demands asking about specific health harms (e.g., “cervical cancer” not “illness”), providing specific timeframes (e.g., “before the end of radiation in six weeks…”), specifying a meaningful comparison group (e.g., “compared to cancer survivors who don’t smoke, how likely are you to…?”), and using well-defined response categories (e.g., “not likely” to “extremely likely”).^88,89 Until studies that examine risk perception in the context of smoking after cancer diagnosis routinely incorporate these more reliable and valid measurement practices, the role of risk perception in cancer survivors’ smoking cessation will remain unclear. Third, in the cancer survivorship literature, there is some consensus about what is meant by “social role” ³¹ yet the same cannot be said for “self-concept” as it is sometimes conflated with self-perception, self-esteem, and other constructs.⁹⁰ Moreover, in some cases, the social role is considered a component of self-concept,⁹¹ which only further complicates tests of their independent and combined influence on health behavior performance. It is paramount that cancer survivorship researchers first agree upon what is meant by social role and self-concept and then create or adapt psychometrically sound measures of these constructs for use in smoking cessation studies. Issues of measurement aside, this scoping review provides new information about the utility of the TM for understanding cancer survivors’ smoking behavior, as discussed below.

Affective Response

Cancer diagnosis and its sequelae have long been recognized as chronic stressors. As with any stressor, an effective response to cancer is expected, and the TM posits that this response could be positive or negative.^9–11 Indeed, prior literature supports the notion that cancer diagnosis can impact psychological functioning in both positive and negative ways, with different manifestations over time and across people.^17,92,93 Key findings from this review are consistent with the literature describing the full range of affective responses after a cancer diagnosis, though most studies reviewed describe negative (e.g., depression, fear) as opposed to positive (e.g., happiness, gratitude) outcomes. As is, this review clearly suggests negative affective responses are strongly associated with smoking status, such that they are more concentrated among current than former and never smokers. This pattern of findings converges with those in the larger literature, which show strong links between negative affect and smoking status and demonstrates tobacco use’s function as a coping mechanism.^94–96 Notably, negative affective responses were not reliably predictive of or associated with cancer survivors’ making a quit attempt or maintaining abstinence, which is a departure from the literature that documents the role of negative affect in lapse and relapse. This difference may be attributable to the methodology and sampling of the studies in this review, as most did not set out to explore causal links between negative affective response and smoking cessation. Consistent with other literature,⁹⁶ though, cancer survivors who were abstinent from smoking exhibited more positive affective responses, as indicated by higher scores on measures of emotional wellbeing. Any strong conclusions or causal inferences about the relationship between cancer survivors’ smoking cessation and positive affective response, though, should be tentative due to the small number of relevant studies. In sum, for cancer survivors, there appears to be an association between negative and positive affective responses on the one hand and smoking behavior on the other.

Risk Perceptions

Many perceive cancer diagnosis as a clear threat to survival and quality of life. Based on the TM, cancer survivors who perceive themselves as vulnerable to poor outcomes like early mortality and functional decline might attempt to regain control of their lives, health, and wellbeing by enacting some measure of health behavior change, including but not limited to smoking cessation. Such efforts will only manifest under certain conditions, though, one of which is holding the viewpoint that smoking is detrimental and abstinence is beneficial. Key findings from this review provide some evidence of an association between risk perceptions and smoking status. More specifically, never and former smokers tended to perceive greater harm from smoking than current smokers. Moreover, current smokers’ impression of smoking was that it held more positives than negatives, with the opposite being true for smoking cessation (i.e., quitting possessed more cons than pros). The larger literature supports these findings such that non-smokers are more likely to associate poor health outcomes with tobacco use than current smokers and current smokers are likely to espouse the benefits of smoking even while acknowledging its inherent risks.^97–99 As is, it is unclear to what extent factors like causal attributions, hindsight bias, and self-protection are at play in shaping the risk perceptions of cancer survivors who smoke. Whatever the case, it is noteworthy that stronger risk perception was positively associated with intention to abstain from smoking, suggesting the threat of smoking-related health problems has the potential to stimulate cancer survivors’ contemplation and preparation for health behavior change, even if the actual quit attempts do not follow.

Self-Concept and Social Role

Cancer survivors are likely to experience a significant change in self-perception and social roles, which may stem from an incompatibility between their new health status and their old social/personal norms. The TM posits that people may derive meaning or purpose from their social roles and strive to comply with social norms to maintain positive self-impression and minimize social judgment. In this way, cancer survivors would evaluate their health behaviors in light of their social role and self-concept, making adjustments in these variables in an effort to maximize overall health. Unexpectedly, this review did not uncover any studies of cancer survivors’ social role and smoking behavior. Two studies did provide data on self-concept, though, with one indicating that its negative aspects (e.g., internalized stigma) might motivate smoking cessation and the other indicating that its positive aspects (post-traumatic growth) may have no relation to smoking behavior. With so little data from which to draw conclusions and the imprecision in how these particular TM constructs were operationalized in this review, no real conclusions can be drawn as to the utility of self-concept and/or social role in understanding cancer survivors’ smoking behavior.

Conclusions

This scoping review’s findings must be considered in light of its limitations. First, the process of a scoping review is a relatively new technique for quickly mapping and synthesizing existing bodies of research in a topic area.^29–31 This type of review is not meant to assess the quality of the literature or meticulously describe all available literature.³⁰ As such, the findings of this review do not offer a definitive statement about the relevance or utility of the TM for explaining or predicting smoking cessation after a cancer diagnosis, especially since the studies reviewed herein did not have that expressed goal. Instead, this review offers clues about how the TM fits into the existing literature on smoking cessation after cancer diagnosis and suggests that at least two of its central constructs (namely, affective response and risk perceptions) warrant further attention. Second, a small number of relevant studies and/or large heterogeneity in measurement across studies in some cases preclude any firm conclusions about associations between TM constructs on the one hand and cancer survivors’ smoking behavior on the other. Third, the selected search terms, while intended to be comprehensive, may have been too restrictive to capture all studies of relevance to the TM. Even with these limitations, this review highlights some room for improvement regarding operationalization and measurement of the TM’s primary constructs, at least as they apply to cancer survivors’ smoking behavior. Given that the TM of cancer diagnosis is widely cited in research and clinical settings,¹⁰⁰ it is imperative to adopt precise definitions and apply rigorous measurement in order to evaluate critically whether or not the TM provides a convincing means of conceptualizing and explaining cancer diagnosis as a cueing event for health behavior change.

Funding

This research was supported by grants from the National Institutes of Health: UL1 TR001998 from the National Center for Advancing Translational Sciences, T32 DA035200 from the National Institute of Drug Abuse, and K07 CA181351 (JLB) from the National Cancer Institute. The content of this review is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Declaration of Interest

None of the authors have any conflicts of interest to declare.