See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/11838034
Temporal analysis of the relationship of smoking behavior and urges to mood states in men versus women Article in Nicotine & Tobacco Research · September 2001 Impact Factor: 3.3 · DOI: 10.1080/14622200110050466 · Source: PubMed
CITATIONS
READS
87
25
3 authors, including: Ralph J Delfino
Larry Jamner
University of California, Irvine
University of California, Irvine
144 PUBLICATIONS 5,508 CITATIONS
66 PUBLICATIONS 3,281 CITATIONS
SEE PROFILE
All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.
SEE PROFILE
Available from: Larry Jamner Retrieved on: 12 May 2016
Nicotine & Tobacco Research (2001) 3, 235– 248
Temporal analysis of the relationship of smoking behavior and urges to mood states in men versus women Ralph J. Delfino, Larry D. Jamner, Carol K. Whalen
Epidemiological investigations of mood and smoking have relied largely on retrospective self-reports, with little research on real-time associations. We examined the relationship of mood states to contemporaneous smoking urges and to subsequent smoking and also assessed the effects of smoking on subsequent mood. For 2 days, 25 female and 35 male smokers aged 18– 42 made three prompted diary entries per hour plus pre- and postsmoking entries ( 6882 entries). Data were analyzed with generalized estimating equations. We found significant positive associations between smoking urge and anger, anxiety, and alertness in women and men; fatigue in men only; sadness more strongly in men than women; and happiness in women only. Decreased alertness and increased anxiety predicted subsequent smoking in men only. Smoking was followed by decreased anger levels in men and women and decreased sadness in men only. In men with lower overall anger episodes, increased anger was associated with subsequent increased smoking. These findings suggest that smoking is related to negative affect and energy level, more clearly in men, and has palliative effects on sadness in men and on anger in men and women. These data demonstrate that ambulatory research can reveal targets for early intervention and smoking cessation.
Introduction A key question in discovering how to prevent tobacco smoking is why some people ( usually adolescents) who experiment with tobacco go on to become regular users, whereas most do not. Pomerleau, Collins, Shiffman, and Pomerleau ( 1993 ) have proposed that some people are more sensitive to nicotine than others; the reinforcing consequences of nicotine on mood and performance are greater for this subgroup, and tolerance is more likely to develop. Pomerleau et al. ( 1993 ) further propose that, among individuals less sensitive to nicotine who become smokers, environmental factors Ralph J. Delfino, MD, PhD, Epidemiology Division, Department of Medicine, School of Medicine, and Department of Psychology and Social Behavior, School of Social Ecology, University of California, Irvine; Larry D. Jamner, PhD, and Carol K. Whalen, PhD, Department of Psychology and Social Behavior, School of Social Ecology, University of California, Irvine, California, USA. Correspondence to: Larry D. Jamner, Department of Psychology and Social Behavior, University of California, Irvine, 3340 Social Ecology II, Irvine, CA 92697, USA. E-mail:
[email protected]
such as social cues play a much larger role in determining smoking persistence. Host factors, possibly genetic, may underlie individual predispositions to persistent tobacco use and nicotine addiction ( Pomerleau, 1995 ). These individual differences probably explain findings that nicotine strongly influences emotional and cognitive functions in specific groups of people, including those with clinical syndromes such as major depressive disorder ( Hall, Mu˜noz, & Reus, 1994) and Attention-Deficit/ Hyperactivity Disorder ( Conners et al., 1996 ). Nicotine’s effects on the function and metabolism of neurotransmitters such as serotonin and dopamine ( Gamberino & Gold, 1999) may partly explain such individual differences in the addictive nature of tobacco use, given that there are known and expected host differences in neurotransmitter biochemistry and neurophysiology. Neurophysiological differences between individuals in responses to the addictive effects of tobacco may be partly gender-based, including the relationship between smoking and levels of anger and hostility.
ISSN 1462-2203 print/ ISSN 1469-994X online/ 01/ 030235-14 © 2001 Society for Research on Nicotine and Tobacco DOI: 10.1080/ 14622200 11005046 6
236
GENDER DIFFERENCES IN SMOKING AND MOOD
Gender differences in personality–smoking linkages have been proposed, but the empirical literature is inconsistent. The Bogalusa Heart Study of 2092 children found that hostility was positively associated with cigarette use in white boys, but not in white girls or in African-American boys or girls ( Johnson, Hunter, Amos, Elder, & Berenson, 1989). In contrast, no gender differences emerged in a prospective study of 4700 college students followed-up 20 years later; high levels of hostility at baseline were associated with initiation and persistence of smoking in both men and women ( Lipkus, Barefoot, Williams, & Siegler, 1994). It has been hypothesized that pharmacological effects of nicotine may be less reinforcing for women than men and that non-nicotinic (psychosocial ) aspects of tobacco smoking may be stronger reinforcers in women than men ( Perkins, 1996). Killen et al. ( 1997 ) provided evidence for this hypothesis in a prospective study of 1026 neversmoker adolescents. Boys with higher depressive symptoms and girls with higher sociability scores were more likely to report tobacco use at the 3–4 year follow-up. In summary, although the above referenced evidence to date on smoking-mood links is not entirely consistent, several sets of findings support the value of delineating the putative emotion-regulation and anger-palliative effects of nicotine in males vs. females. There is little research on the short-term effects of smoking on anger or other mood states in everyday settings ( Hatsukami, Morgan, & Pickens, 1990; Shiffman, Hufford, Hickcox, Paty, Gnys, & Kassel, 1997; Shiffman, Paty, Gnys, Kassel, & Hickcox, 1996; Shiffman & Prange, 1988). A recent study of 30 smokers and 30 non-smokers compared 24-h ambulatory diary reports of anger during the administration of placebo vs. nicotine patch ( Jamner, Shapiro, & Jarvik, 1999). High hostile individuals, identified by the Cook–Medley Hostility Scale, reported significantly less anger in the nicotine patch (13% of diary entries) than in the placebo condition ( 24% ). In contrast, among low-hostile individuals the opposite pattern was observed, though these differences did not achieve significance. Results were consistent across smoking status and gender. These results suggest that smoking may have an angerpalliative effect that reinforces smoking behavior among people prone to anger and hostility. Shiffman ( 2000 ) further highlights the value of ambulatory data in a comparison of smoking typology questionnaires administered at study entry with diary results in 275 smokers. The diary reports ( approximately 35 subject reports per week of smoking situations) were collected with an electronic palm-top diary to verify that the time of the report was proximal to the smoking event. The diary reports were summarized by smoking behavior across 14 mood domains, seven location options, and 10 activities. The diary data were compared to the baseline questionnaire, which was modeled on the diary assessments. The average correlations for mood, location, and activity were small (0.09, 0.09 and 0.08, respectively), and all individual correlations were low to negative.
Diary records are potentially more accurate than retrospective questionnaires because they are obtained in close temporal proximity to actual events and are thus less subject to recall bias. Recall of smoking lapses is also prone to notable bias, despite the expectation that such salient single events would be memorable ( Shiffman, 2000). Shiffman et al. ( 1997 ) examined diary data on smoking relapse triggers from 127 people who had quit smoking and found that retrospective accounts given an average of 72 days later correlated poorly with the real-time data ( kappas of 0.18 to 0.27 for the four content domains of mood, activity, episode triggers, and abstinence violation effects). Recall of the most important trigger matched the diary data only around a third of the time ( kappa=0.19 ). The diary data showed that temptation to smoke was associated with greater negative affect, restlessness, attention disturbance, and exposure to smoking cues, especially in situations of eating and drinking ( Shiffman et al., 1996 ). With one notable exception ( Shiffman et al., 1996 ), most of these ambulatory studies have used selfcollected data on smoking cues reported only at the time of smoking. Little is known about the degree to which the same moods or activities occur at other times, particularly before or after smoking events. We hypothesize that anger and negative affect more generally trigger smoking in some people and that smoking in turn has mood-palliative effects that could serve to reinforce smoking behavior. In the case of anger, people with higher levels of anger may show stronger links between anger and smoking and may gain greater palliative benefits from smoking. We also hypothesize that women may show weaker links between smoking and moods or arousal levels because nicotine may be less reinforcing for women than men, with men more likely to use smoking to regulate moods and performance. In the present study, we employed real-time ambulatory methodologies with young adults to examine the relationships of specific moods to smoking urges and subsequent smoking behaviors, and to examine the effects of cigarette smoking on moods. We also compared the strength of associations in men vs. women and examined potential differences in anger–smoking relationships between people with high vs. low rates of anger episodes. This study extends the experimental study discussed above by Jamner et al. (1999) by examining the anger palliative effects of smoking rather than nicotine patches. We also evaluated mood–smoking links after controlling for three factors that are often associated with cigarette smoking: social environment, caffeine intake, and alcohol consumption. Methods Design In this repeated-measures study, events and psychological states were assessed at the time or soon after they
NICOTINE & TOBACCO RESEARCH
occurred. As noted above, a major advantage of this approach is a reduction in the likelihood of recall bias because of the close or immediate proximity of events and psychological states. Also, each participant serves as his or her own control over time. The design enables investigators to determine the temporality of associations and to observe between- and within-individual patterns of change in exposure and response over time, thereby gaining insight into host susceptibility. The signal-tonoise ratio is enhanced because multiple exposures or exposure conditions/ levels are studied in each participant and because data can be analyzed within the participant’s cluster of exposure–response measurements, which controls the variability in exposure–response relationships due to between-subject characteristics. This variability is commonly overcome with large sample sizes in typical epidemiological designs such as cohort or cross-sectional studies. Repeated measures, on the other hand, reduce the variability of the response variable( s) without reducing the magnitude of association, thereby enhancing both power and precision while allowing smaller sample sizes (Weiss & Ware, 1996). Participants Twenty-five women and 35 men aged 18–42 years were recruited using advertisements in local Orange County, CA newspapers. Eligible participants were smokers who reported smoking at least 10 cigarettes/ day and were otherwise in good health. The daily cigarette consumption from self-reports given in a background questionnaire at intake was: 10 cigarettes/ day for five women and four men, 11–15 cigarettes/ day for eight women and seven men, 16–20 cigarettes/ day for eight women and 16 men, 21–25 cigarettes/ day for three women and six men, and >25 cigarettes/ day for one woman and two men. In their background questionnaires, only three women and two men reported being completely restricted from smoking while at home, and only two women and three men reported being completely restricted from smoking while at work. Nevertheless, during the 2 days of ambulatory monitoring the restricted subjects each smoked 21–56 cigarettes. One man reported both restrictions, and he smoked 21 cigarettes during ambulatory monitoring. Most men smoked regular length cigarettes ( 81% ), whereas nearly half of the women smoked 100s ( 46% ). The sample included 33 white non-Hispanics ( 18 men, 15 women ), one Hispanic male, 15 Asians ( nine men, six women ), five African-Americans ( three men, two women ), and six other race/ ethnicity ( four men, two women ). The educational background of participants was that 76% of women and 71% of men had attended high school only and the rest at least some college. Forty participants were married ( 24 men, 16 women), 11 were never married ( eight men, three women), and nine were divorced or separated ( three men, six women). Sixtyeight per cent of women and 70% of men were employed.
237
The median age at smoking initiation was 17 years for both men and women. Evidence that participants were dependent smokers was found in self-reported data collected at baseline. The median number of attempts to quit was two for both men ( range 0–10, excluding one outlier of 50) and women ( range 0–10; Wilcoxon rank sums test, p=0.40). Participants answered additional questions at intake representing smoking dependence with up to five ordinal levels. Answers to these questions were dichotomized into categorical variables for higher vs. lower levels of dependence, and then compared between men and women with x 2 tests. Men did not differ significantly from women for several dependence variables, including inhaling deeply ( 60% vs. 52%, respectively ); smoking the first cigarette of the day more than half the time within 30 min of waking ( 51% vs. 48%, respectively); difficulty giving up the usual first cigarette of the day ( 49% vs. 60%, respectively); smoking half or more of the time when sick with a cold, flu, or so ill that they are in bed most of the day (49% vs. 52%, respectively); smoking all of each cigarette smoked ( 46% vs. 60%, respectively); or always inhaling ( 86% vs. 80%, respectively). Men were somewhat more likely to report holding cigarette smoke in their lungs a moment or two before exhaling ( 43% vs. 24%, respectively), but the difference was not significant ( p10% of the time. The higher percentage cut-point in women was needed because some of the women frequently used words for personal anger codes for ordinal scores >1 that were less frequently used among men and that convey little or no hostility ( e.g., sulky, moody). It is for this reason that most women ( 15 of 25) who expressed ‘anger’ episodes ( >1 on 0–5 scale) reported such episodes on more than 5% of observation occasions. The frequency distribution of anger episodes in women with episodes >10% of the time showed three women with a frequency of 11–20% and six women reporting anger more than 20% of the time ( range 11–34%). Note that nine men and six women in the low anger frequency groups reported no anger episodes over an ordinal score of 1.
were five times as likely compared with periods without any urge ( both genders models were p10% of time (9 subjects, ³ 780 obs.)
1.33 (0.62–2.91) 0.89 (0.63–1.24) 0.79 (0.63–0.99)* 1.07 (0.89–1.28)
Odds ratios (95% CI) on binary datab
Men with anger frequency >5% of time (12 subjects, ³ 1135 obs.)
1.37 (0.96–1.97) 1.19 (1.01–1.40)* 0.79 (0.59–1.05) 0.91 (0.56–1.26)
Odds ratios (95% CI) on ordinal datac
1.15 (0.58–2.29) 1.14 (0.72–1.80) 0.65 (0.34–1.26) 1.26 (0.77–1.92)
1.05 (0.82–1.33) 0.97 (0.84–1.11) 0.73 (0.52–1.01) 1.05 (0.74–1.50)
Women with anger frequency £ 10% of time (16 subjects, ³ 1770 obs.)
1.39 (0.95–2.91) 1.69 (1.02–2.78)* 1.78 (1.26–2.53)** 0.49 (0.15–1.60)
Odds ratios (95% CI) on binary datab
Men with anger frequency £ 5% of time (23 subjects, ³ 2210 obs.)
* p