The proposed study examines the role of stress in smoking relapse, focussing on interactions between stress and individual differences in reactivity to stress. The central hypothesis is that stress promotes relapse primarily among smokers who are emotionally or physiologically hyperresponsive to stress. It is also hypothesized that tobacco withdrawal enhances stress reactivity, making dependent smokers especially vulnerable to stress-induced relapse. A sample of 200 quitters, including both unaided quitters and clinic participants, will be drawn from a diverse HMO population. Baseline measures include smoking history and biochemical and psychometric measures of dependence. Stress reactivity is assessed via laboratory exposure to multiple stressors and monitoring of multiple response modalities. Stressors include (1) noise, which elicits emotional, but not physiological, responses; (2) public speaking, which elicits marked emotional and physiological arousal; (3) mental arithmetic, which elicits very consistent cardiovascular responses; and (4) isometric exercise, which elicits cardiovascular responses without emotional arousal. Stressors 1-3 also have apparent relevance to smoking or relapse. Multiple response modalities are assessed before and after each stressor: emotional response, cardiovascular response, and desire to smoke. Reactivity assessments are repeated 3-5 days after cessation to assess the effects of withdrawal, and 1 month later to study the effects of protracted abstinence (among abstainers) and of renewed smoking (among recidivists). Intensive assessment of stress during maintenance is implemented by a series of structured telephone interviews extending from baseline through 6-month follow-up.
Cl aims of abstinence will be biochemically verified at 1, 3, and 6 months. Multiple regression analyses will focus on the joint and interactive effects of reactivity and stress experienced at various times in maintenance. Structural modeling will be used to derive appropriate measurement models for the reactivity and stress data and to test substantive models of their interactions in relapse.
