Nature and Role of Change in Anxiety Sensitivity During NRT-Aided Cognitive-Behavioral Smoking Cessation Treatment

Abstract

This study evaluated the associations between change in anxiety sensitivity (AS; fear of the negative consequences of anxiety and related sensations) and lapse and relapse during a 4-week group NRT-aided cognitive-behavioral Tobacco Intervention Program. Participants were 67 (44 women; M _age = 46.2 years, SD = 10.4) adult daily smokers. Results indicated that participants who maintained high levels of AS from pretreatment to 1 month posttreatment, compared to those who demonstrated a significant reduction in AS levels during this time period, showed a significantly increased risk for lapse and relapse. Further inspection indicated that higher continuous levels of AS physical and psychological concerns, specifically among those participants who maintained elevated levels of AS from pre- to posttreatment, predicted significantly greater risk for relapse. Findings are discussed with respect to better understanding change in AS, grounded in an emergent taxonic-dimensional factor mixture model of the construct, with respect to lapse and relapse during smoking cessation.

Keywords:

• smoking
• anxiety sensitivity
• treatment outcome
• smoking cessation

Acknowledgements

Dr. Bernstein recognizes the funding support from the Israeli Council for Higher Education Yigal Alon Fellowship, the European Union FP-7 Marie Curie Fellowship International Reintegration Grant, Psychology Beyond Borders Mission Award, the Israel Science Foundation, the Rothschild-Caesarea Foundation's Returning Scientists Project at the University of Haifa, and a National Institute on Drug Abuse (NIDA) Clinical LRP award. Ms. Assayag recognizes the support of the University of Haifa Graduate School. Dr. Stewart was supported by a Killiam Research Professorship from the Dalhousie University Faculty of Science at the time this research was conducted. Data collection was supported by an Idea Research Grant from the Canadian Tobacco Control Research Initiative (15683) awarded to Dr. Stewart, Dr. Zvolensky, and Mr. Steeves. Dr. Zvolensky recognizes funding support from NIDA (R01 DA027533-01, R01 MH076629-01).

The authors thank the clients and therapists from the Tobacco Intervention Program (at Capital Health Addiction Prevention and Treatment Services) and Ellen Rhyno and Jennifer Mullane who assisted with data collection.

Notes

^1. FMM may have a number of significant advantages relative to coherent cut kinetic (CCK) taxometric methods as well as relative to other latent mixture modeling techniques (Bauer & Curran, 2004; Lubke & Tueller, 2010; Muthén, 2008). Unlike other more commonly employed data analytic strategies such as CCK taxometrics, K-means clustering, latent class and latent profile analyses, and factor analysis, FMM facilitates the concurrent modeling of various models that incorporate latent class (categorical) structure and within-class continuity. Thus, FMM is not limited to testing (a) whether a latent variable is either a dichotomous categorical (taxonic) or continuous (i.e., CCK taxometrics) variable, (b) the relative fit of various continuous models (i.e., factor analysis) assuming a single latent homogeneous population, or (c) relative fit of various categorical models that lack within-class continuity due to assumption of local independence (i.e., K-means clustering, latent class/profile analysis). Rather, FMM permits arguably more construct valid and flexible latent structural modeling of various possible categorical and continuous structures simultaneously. Furthermore, and unlike CCK taxometrics, FMM offers a model-based approach in which latent structural models are compared and contrasted in terms of multiple, well-established, objective fit indices. In addition, unlike CCK taxometrics, and similar to other model-based mixture techniques, FMM imposes no limit on the number of possible latent classes that may underlie a construct's putative population heterogeneity or latent class structure. Finally, because FMM incorporates mixture modeling and factor analytic techniques, it is rooted in extensive and well-established quantitative theory and methods.

^2. We planned to conduct two logistic regression analyses in which baseline (prequit) daily smoking rate and nicotine dependence were (separately) entered in step one of the logistic regression equation; the posttreatment AS group status variable (reduced-normative AS group vs. maintained-maladaptive AS group) was entered as a categorical group variable in step two of the equation; and either rates of lapse (vs. no lapse or abstinence) or rates of relapse (vs. abstinence) served as the dependent variables. In the event of significant associations between pretreatment smoking rate or nicotine dependence and lapse and relapse outcomes, this data analytic approach was intended to facilitate a test of the incremental effect of between-group differences (reduced-normative AS group vs. maintained-maladaptive AS group) on lapse and relapse outcomes, above and beyond the variance accounted for by baseline levels of smoking or nicotine dependence. However, neither pretreatment (baseline) smoking rate nor nicotine dependence were significantly related to the studied dependent variables (lapse and relapse outcomes) and thus were omitted as covariates (Tabachnick & Fidell, 2005).