because of tobacco withdrawal and that the improvement in performance occurred
because smoking relieves tobacco withdrawal (Schachter 1979; Silverstein 1982). This
latter interpretation assumes that overnight deprivation induces withdrawal, although
this assumption has not been tested directly, withdrawal effects can occur after only 12
hours of deprivation (Hughes, Higgins, Hatsukami 1990).

Ideally, studying smokers before initiation would allow comparison of this baseline
with before and after a smoking episode. As this is impractical, one solution has been
to add a control group of nonsmokers (Hughes, Higgins, Hatsukami 1990). For
example, smokers performed better after smoking and the same as nonsmokers in
several studies of errors on a vigilance task (Taylor and Blezard 1979; Hughes, Keenan,
Yellin 1989; Lyon et al. 1975; Heimstra et al. 1980; Tong et al. 1977; Tarriere and
Hartmann 1983; Keenan, Hatsukami, Anton 1989) and a tracking task (Lyonet al. 1975)
(Figure 1, upper panel). The effect was attributed to relief of withdrawal.

One study provided evidence for enhancement of performance from smoking inde-
pendent of reversing withdrawal. Wesnes and Warburton (1978) reported a pattern
consistent with enhancement when errors on vigilance tasks were studied (Figure 1,
lower panel).

Other indirect evidence can be used to test the withdrawal relief versus enhancement
models. Two studies reported enhancement of tracking or motor skills when smokers
were not deprived (Parrott and Winder 1989; Hindmarch, Kerr, Sherwood 1990;
Larson, Finnegan, Haag 1950; Pomerleau and Pomerleau 1986). Several studies have
examined the effect of cigarette smoking or nicotine administration on the performance
of nonsmokers (Dunne, MacDonald, Hartley 1986; Hindmarch, Kerr, Sherwood 1990,
Wesnes, Warburton, Matz 1983; Wesnes and Revell 1984; West and Jarvis 1986,
Wesnes and Warburton 1984). In two studies, the improvement in nonsmokers was
similar to that of deprived smokers (Wesnes, Warburton, Matz 1983; Wesnes and
Revell 1984). One study reported performance to be similar between deprived smokers
and nonsmokers (Warburton 1990). Finally, nicotine appears to improve the perfor-
mance of animals not previously exposed to nicotine (Clarke 1987; Emley and Hutchin-
son 1984).

In summary, the results of studies to assess tf smoking increases performance through
withdrawal relief or by direct enhancement appear contradictory. One possible ex-
pianation of this discrepancy is that smoking may increase performance through both
withdrawal relief and direct enhancement. The specific mechanism that is operative
may vary not only among smokers but also within smokers across situations.

Variability in Withdrawal

Whereas the necessary and sufficient condition to establish dependence is repeated
exposure to the drug, other factors may exacerbate nicotine withdrawal symptoms.
Although several investigators have commented on the variability of postcessation
symptoms, it is unclear that this variability is greater than with other drug withdrawal
syndromes (Hughes, Higgins, Hatsukami 1990; US DHHS 1988). The results of
retrospective and postcessation studies on self-reported withdrawal symptoms (e.g.,
hunger, restlessness, or inability to concentrate) among smokers who have a greater

526
PERCENT ERRORS

STIMULUS SENSITIVITY P(1)

70

60

50

40

30

0.9

o
@

2°
N

2°
fo.)

  
     
 
 

 

  
  

 

 

 

 

DEPRIVED
~ SMOKERS
NONSMOKERS
SMOKERS
i 1 1 1 | j
1 2 3 4 5 6
HOURS
SMOKERS
NONSMOKERS
DEPRIVED SMOKERS
1 l 1 L 1 i i J
10 20 30 40 50 60 70 80
MINUTES

FIGURE 1.—Upper panel: Performance on a meter (i.e., visual) vigilance task

SOURCE: Heimstra et al. L980,

Lower panel: Performance on the continuous clock task, a visual
vigilance task

NOTE: Increased stimulus sensitivity refers to fewer errors.

SOURCE: Wesnes and Warburton (1978).

527
nicotine intake are inconclusive (Goldstein, Ward, Niaura 1988, Hughes, Higgins,
Hatsukami 1990: Shiffman 1979: US DHHS 1988; Williams 1979). Withdrawal
effects, including weight gain, have not been found to differ consistently by gender or
age (Hughes, Higgins, Hatsukami 1990).

Several studies have suggested that expectancy influences the effects of abstinence;
that is, some individuals may amplify, deny, or misattribute their withdrawal symptoms
(Barefoot and Girodo 1972; Gottlieb et al. 1987; Hughes and Krahn 1985, Hughes et
al. 1989). According to the misattribution model, at times the individual can “mistake”
withdrawal symptoms for other possible events. For example, in one study a labeling
mistake was made when individuals were told that a placebo they were taking was
alleged to have side effects similar to the effects of cigarette withdrawal (Barefoot and
Girodo 1972).

Three direct tests of expectancy have been published (Gottlieb et al. 1987; Hughes
and Krahn 1985; Hughes et al. 1989). In one study, subjects in a double-blind trial of
nicotine polacrilex gum were asked if they thought they had received nicotine or
placebo gum. Those who believed they had received placebo gum had more abstinence
discomfort than those who could not differentiate what they had received; this latter
group had more discomfort than those who thought they had received the nicotine
polacrilex gum (Hughes and Krahn 1985). Because this study used post hoc ratings, it
is unclear that the belief in which gum had been received modified the level of
abstinence effects, or that the level of abstinence effects modified the belief of which
gum had been received.

Two experimental trials have manipulated instructions and thereby directly tested if
expectancy influences abstinence effects. The first study randomly assigned smokers
to a 2x2 design of contrasting instructions; subjects were told that they received either
nicotine polacrilex gum or placebo gum, and actually received either nicotine polacrilex
gum or placebo gum (Gottlieb et al. 1987). Most of the measures of abstinence effects
were unchanged by instructions or by actual drugs. The physical symptoms and
stimulation scores on the Shiffman-Jarvik Withdrawal Scale were less only on some
days in the group told they were receiving nicotine than in the group told they were
receiving placebo. A second study used a similar design and found that abstinence
symptoms were fewer among those who received nicotine polacrilex gum than among
those who received placebo gum, but found no effect of instructions (Hughes et al.
1989). In summary, the seemingly valid proposition that abstinence effects are in-
fluenced by expectancy has not been completely supported by empirical tests.

Abstinence effects have been hypothesized to be greater in more dependent smokers.
However, the scales for dependence used to test this hypothesis vary according to
whether they are quantifying physical dependence (withdrawal), behavioral depen-
dence (desire for tobacco or tendency to relapse), or dependence on tobacco or on the
nicotine in tobacco (Hughes 1984). The Fagerstrém Tolerance Scale (TQ) is the most
widely used dependence scale (Fagerstré6m 1978). TQ consists mostly of items that
refer to behavioral dependence on tobacco. The total TQ score predicted total
abstinence discomfort in one study (Fagerstrém 1980) and weight gain in another study
(Tennesen et al. 1988). However, two detailed studies failed to indicate that TQ

528
predicted weight gain (Emont and Cummings 1987) or self-reported withdrawal
symptoms.

The Reasons for Smoking Scale has two scales relevant to the dependence con-
struct—the addiction scale and the negative affect scale (Ikard, Green, Hom 1969).
Neither of these has been shown to predict weight gain (Bossé, Garvey. Costa 1980),
self-reported withdrawal (Hughes and Hatsukami 1986), or relief by nicotine polacrilex
gum (Hughes and Hatsukami 1986).

Russell's Smoking Motivation Questionnaire has a subscale for dependence (Russell,
Peto, Patel 1974). In one study, the scale predicted total abstinence discomfort and
irritability but did not predict restlessness, depression. hunger, or inability to con-
centrate (West and Russell 1985).

Another measure somewhat related to dependence includes the severity of abstinence
discomfort in the past, which appears to predict self-reported abstinence (Hughes and
Hatsukami 1986). Other generic scales, such as the MacAndrews Scale for Addiction
(MacAndrew 1979) and Eysenk Personality Questionnaire (Eysenk and Eysenk 1975),
do not predict abstinence discomfort and weight gain (Bossé, Garvey, Costa 1980).
Although one study found that self-reported smoking for stimulation predicted
abstinence effects (Niaura et al. 1989), an earlier study had found no such relationship
(West and Russell 1985).

In summary, the evidence that any dependence scale predicts abstinence effects is
quite limited. Further tests that use scales that more specifically determine physical
versus behavioral dependence and dependence on nicotine versus tobacco may provide
more informative data.

Timecourse of Withdrawal

Several recent studies produced concordant results on the timecourse of nicotine
withdrawal. Most signs and symptoms of nicotine withdrawal are readily detected
within 24 hours (Hughes, Higgins, Hatsukami 1990). Previous studies have suggested
that abstinence effects can occur even sooner, for example, within 2 hours (US DHHS
1988). These studies have measured effects during smoking and 2 to 6 hours post-
smoking: it was noted that 2 to 6 hours after smoking, self-ratings of performance were
worse than during smoking. Several investigators have interpreted the scores during
smoking as representing baseline and the postsmoking scores as representing
withdrawal. However, as discussed earlier. an alternate interpretation is possible: the
scores 2 to 6 hours postsmoking represent baseline scores and the scores during smoking
represent the acute effects of smoking (Hughes et al. 1990).

The results of several prospective studies indicate that the signs and symptoms of
nicotine withdrawal peak in the first | to 2 days following cessation (Cummings et al.
1985: Hughes and Hatsukami 1986; West et al. 1984: Shiffman and Jarvik 1976:
Schneider, Jarvik. Forsythe 1984) and last about | month (Gritz, Carr. Marcus 1990:
Cummings et al. 1985: Gross and Stitzer 1989: Hughes 1990: Hughes et al. 1990;
Lawrence, Amoedi. Murray 1982: West, Hajek. Belcher 1987). For each of 10 weeks.
Gross and Stitzer (1989) recorded symptoms of quitters and found a peak during the
first week and a return to baseline 3 10 4 weeks postcessation. Snyder, Davis, and

S29
Henningtield (1989) tracked performance on several tasks over 10 days. Impairment
in performance peaked at | to 2 days. and performance on most tasks returned to
baseline during the 10 days: however. performance on some tasks was still impatred
after 10 days. A study by Cummings and colleagues (1985) included 33 subjects who
kept a daily record of 8 withdrawal symptoms. At 21 diys. few subjects were reporting
withdrawal symptoms, with the exception of an occasional desire for a cigarette. A
fourth study (Hughes 1990) provided a less-detailed timecourse but included groups of
never smokers. ex-smokers, and continuing smokers. The withdrawal scores of
abstinent smokers at | month were equivalent to their baseline scores and to those of
never smokers and continuing smokers (Hughes 1990). Although the average
withdrawal symptom score returned to baseline at | month, 45 percent of subjects
reported symptoms still above precessation levels at |-month followup (Hughes 1990).
Further followup of these subjects indicated that their withdrawal scores had returned
to baseline or below baseline by 6 months postcessation. Craving. hunger. and weight
gain are exceptions to the |-month duration: they may continue at least through the first
6-months after cessation (Gritz, Carr, Marcus 1990; Hughes 1990; Hughes et al. 1990;
West, Hajek. Belcher 1987).

With cessation of other drugs. a prolonged withdrawal syndrome has been postulated
(Martin and Jasinski 1969). There is no evidence of a prolonged nicotine withdrawal
syndrome. In fact, scores on withdrawal scales appear to decrease below precessation
levels at followup (Figure 2): that is, positive mood changes occur after long-term
abstinence from smoking (Chapter 11. see section on long-term psychological and
behavioral consequences and correlates of smoking cessation) (Gritz, Carr. Marcus
1990; Gross and Stitzer 1989: Hughes 1990; Hughes et al. 1990).

Withdrawal as a Cause of Relapse

Seven recent studies have examined nicotine withdrawal as a predictor of relapse.
that is. whether smokers with severe withdrawal are more likely to relapse. Five studies
found that some withdrawal symptoms predicted relapse at some points in time (Gritz.
Carr, Marcus 1990; West, Hajek, Belcher 1990: Hughes 1990: Killen et al. 1990: Swan
et al. 1988). The two studies that did not indicate such a relationship examined the
ability of withdrawal to predict abstinence at very early followup (Hughes and Hat-
sukami 1986) or very late followup (Hughes et al. 1990). In the five positive studies,
mood changes, such as depression and anxiety. were the more common predictors.
However. both across and within the studies, there was no consistent or clear grouping
of symptoms predicting withdrawal at specific points in time. One common finding
was that the number of symptoms appeared to be a predictor (Gritz, Carr, Marcus 1990:
Hughes 1990). For subgroups of smokers. such as more dependent smokers.
withdrawal may be an especially important factor in relapse, but this relationship has
not been demonstrated.

Postcessation weight gain has often been hypothesized to be a major cause of relapse.
especially among women (Hall. Ginsberg. Jones 1986). Contrary to several a priori
hypotheses, three prospective studies have found that more weight gain predicted less
relapse (Duffy and Hall 1990: Hall. Ginsberg, Jones 1986; Hughes et al. 1990). There

530
14 —~

12 —

    

Stresses sccccccrceses f/foene

 

 

8 4
6 wY —~p
Mm |
S
9 “7
_
Z 2-
g
TTT 1 1 THY
BL t 2 3 4 #5 6 7 #8 9g 46 24

WEEKS

FIGURE 2,—Self-reported withdrawal discomfort among abstinent smokers
SOURCE: Gross and Stitzer 1989; Hughes (1990).

was no gender difference in this prediction in any of the three studies. This finding is
further supported by a study in which women who reported eating more in the first 4
days of cessation were more likely to be abstinent at 6-month followup (Guilford 1966).
One explanation for the weight gain—relapse finding is that food deprivation increases
the reinforcing effects of drugs (Carroll and Meisch 1984). Cessation of smoking may
decrease metabolic rate (Perkins, Epstein, Pastor 1990); if this is true, to avoid weight
gain, smokers may deprive themselves of food and thereby increase the reinforcing
effects of cigarettes smoked during periods of relapse.

In summary, this recent evidence shows that smokers with more severe withdrawal
Symptoms are more likely to relapse. However, these results should not be
misinterpreted. First, prediction is not equivalent to causality; withdrawal symptoms
may predict relapse, not because they cause relapse, but because they are associated
with some other variable, such as degree of dependence. Second, those symptoms that
predict the occurrence of relapse and the timin g of relapse—very early (<2 days), early
(2-10 days), or later (10-30 days)—vary across studies. Third, although studies have
shown that withdrawal is an early predictor of relapse, these studies have not shown
that withdrawal predicts eventual outcome (i.e., long-term abstinence).

Summary

Strong evidence indicates that smokers who stop smoking experience a nicotine
withdrawal syndrome that includes the short-term consequences of anxiety, irritability,

531
frustration. anger, difficulty concentrating, and restlessness. These symptoms general-
ly occur within 24 hours and subside after about | month. Smokers also report strong
cravings or urges to smoke when they are not smoking: this symptom will persist among
some former smokers. Hunger and weight gain may also persist longer than | month.
Abstinence does not appear to affect short-term caffeine intake. However, it does
increase caffeine metabolism, which may mimic or potentiate symptoms of nicotine
withdrawal. There are conflicting data on the short-term effects of smoking abstinence
on alcohol intake. However, the data suggest that smokers attempting permanent
smoking abstinence experience decreased alcohol intake.

Research on the effects of smoking abstinence on performance indicates that
abstinence impairs performance on attention tasks. This impairment may persist for at
least 7 to 10 days and is relieved by nicotine replacement. Other more complex types
of tasks as well as memory and learning have not been clearly shown to be impaired by
abstinence. The relation of improvement in attention tasks with nicotine may be due
either to withdrawal relief or to performance enhancement; findings are consistent with
both models. However, evidence more strongly suggests withdrawal relief from
receiving nicotine.

Variability in tobacco withdrawal symptoms resembles that observed for other drug
withdrawal syndromes. Several studies have suggested that expectancy influences
withdrawal effects. However, this has not been completely supported by empirical
tests. Although abstinence effects have been hypothesized to be greater in more
dependent smokers, the evidence is conflicting. Recent data indicate that smokers with
more severe withdrawal symptoms are more likely to relapse. However, no symptoms
or groups of symptoms consistently predict relapse at any given point in time.

LONG-TERM PSYCHOLOGICAL AND BEHAVIORAL CONSEQUENCES
AND CORRELATES OF SMOKING CESSATION

Introduction

Most long-term studies of selt-quitters or smokers taking part in treatment programs
only include data on smoking behavior or smoking status (Adesso 1979: Gordon and
Cleary 1986; Orleans and Shipley 1982: Shipley. Rosen, Williams 1982); followup
measures of psychological and behavioral consequences are rarely included. Thus.
although former smokers represent a large and growing segment of the U.S. population
(Volume Appendix). the long-term psychological and behavioral consequences of
smoking cessation have not been well studied.

Very few studies of former smokers have employed prospective or longitudinal
designs: rather, most have used retrospective or cross-sectional designs. In the typical
retrospective study, subjects are asked whether after quitting or during their experience
of trying to quit, they were more or less nervous, irritable. depressed. sedentary. or
health conscious than before quitting. While relevant to the experience of a person
abstaining from tobacco, retrospective studies potentially suffer from several limita-
tions, including the absence of information about baseline group similarities or differ-

532
ences and the problem of recall bias. (See Chapter 2 for a discussion of methodologic
problems.) Successful former smokers may minimize or fail to recall their difficulties
or exaggerate their prowess (Heinold et al. 1982); recidivists may exaggerate
withdrawal problems to justify their relapse (Graham and Gibson 1971). Cross-
sectional studies do not permit the establishment of comparability at baseline. Con-
clusions from the data are therefore limited, often identifying the correlates of cessation
rather than the consequences. Both consequences and correlates of cessation will be
discussed in this Section.

Most prospective studies of smoking cessation sequelae have been conducted with
smokers participating in formal treatment programs rather than with smokers quitting
on their own (Hughes, Higgins, Hatsukami 1990). Treatment participants may differ
in several ways from self-quitters. In a recent review of findings concerning short-term
withdrawal effects, Hughes, Higgins, and Hatsukami (1990) noted that self-quitters had
fewer and less severe withdrawal symptoms than treated quitters; they noted. as did
Schachter (1982), that clinic populations may include a higher proportion of hardcore,
highly dependent smokers. On the other hand, treated quitters may learn new coping
skills such as relaxation, self-reward, or exercise and gain additional support for their
initial quitting efforts. Therefore, their short-term postquitting experiences may not be
representative of the 90 percent of former smokers who quit on their own (US DHHS
1988: Fiore et al. 1990). Thus. in drawing conclusions from studies of participants in
treatment programs, it is important to be aware of the possible differences between these
two populations of abstainers.

Mood, Anxiety, Perceived Stress, and Psychological Well-Being

Tobacco use has often been described as a maladaptive response to, or a way to cope
with, life stress and a way to regulate negative affect (Tomkins 1966; Billings and Moos
1981: Ockene et al. 1981; Orleans 1985; Abrams et al. 1987). Smokers often believe
that smoking helps them cope with stress and anxiety (Ikard, Green, Horn 1969). Thus.
in addition to the stress of separation from cigarettes (Tamerin 1972), abstaining from
cigarettes potentially could make the smoker feel less able to cope with stress (Abrams
etal. 1987; Marlatt and Gordon 1985) and thereby constitute a biologically based source
of stress (Grunberg and Baum 1985). If the quitter feels unable to cope with stress
without cigarettes, perceived stress may increase, and self-efficacy may decrease.
resulting in heightened anxiety and an overall negative shift in well-being. Alterna-
tively, Cohen and Lichtenstein (in press) have hypothesized that for smokers who want
to quit smoking, continued smoking may prove more stressful than cessation, and
quitting smoking may result in a more positive self-appraisal and heightened feelings
of self-esteem and personal competence. Similarly, other researchers have proposed
that smoking may cause negative self-evaluations and feelings of guilt and helplessness
among smokers who want to quit, so that quitting would result in an overall long-term
improvement in mood, self-image. and self-esteem (Frerichs et al. 1981: Knudsen et al.
1984; Schwartz and Dubitzky 1968).

Possible long-term changes in anxiety levels after quitting might also reflect quilting-
related changes in physiologic stress reactivity (Abrams et al. 1987). To the extent that
smoking contributes to excess physiologic stress reactivity and more ready arousal to
anxiety (Emmons et al. 1986: Williams, Hudson, Redd 1982: US DHHS 1988),
cessation might lead to stable reductions in general anxiety.

Several models have been proposed to understand the possible long-term conse-
quences of smoking cessation for depression or dysphoria (Frerichs et al. 1981; Hughes
1988; Hughes, Higgins, Hatsukami 1990; Tamerin 1972). Studies of withdrawal effects
have found depressed mood or dysphoria to be a common, transient withdrawal effect,
partly reflecting multiple pharmacologic effects of nicotine abstinence (Backon 1983:
Hughes, Higgins, Hatsukami 1990; US DHHS 1988). Covey, Glassman, and Stetner
(in press) found that smokers with a history of major depression had more severe
symptoms of depression 2 weeks after a behavioral treatment for smoking than those
without such a history. However, some theorists have proposed that for smokers who
want to quit, quitting could result in improved mood, well-being, and self-esteem
(Frerichs et al. 1981).

Research Results

Five cross-sectional studies have compared former smokers with continuing smokers
or relapsers on measures of mood, affect, anxiety, and psychological well-being
(Abrams et al. 1987; Giannetti, Reynolds, Rihn 1985; Orleans et al. 1983; Pederson and
Lefcoe 1976; Pomerleau, Adkins, Pertschuk 1978). Of these five studies, three found
no differences between these groups, and two found differences demonstrating more
healthy outcomes for former smokers. Pederson and Lefcoe (1976) compared 46
former smokers, mostly self-quitters who had not smoked cigarettes for 1 year or longer,
with 46 current smokers volunteering for treatment. These researchers found no
differences on Jackson Personality Inventory scales that included measures of anxiety
and self-esteem. Likewise, Pomerleau, Adkins, and Pertschuk (1978) used the
Symptom Checklist (SCL-56) as a 2-year followup measure of dysphoria among 60
smoking cessation treatment participants and found no differences between quitters and
continued smokers. Mean duration of smoking abstinence was not reported. Giannetti,
Reynolds. and Rihn (1985) compared 47 former smokers who had been abstinent for
at least 6 months with 35 current smokers hospitalized for cardiovascular disease and
found no differences in “habits of nervous tension.”

In the only study to employ multiple self-report, physiologic, and observer measures,
Abrams and colleagues (1987) found no significant differences between 22 former
smokers (mean abstinence approximately 2 years) and 22 relapsers on the State-Trait
Anxiety Inventory, but did find that former smokers reported significantly less anxiety
and had significantly lower heart rates in response to simulated smoking-related
stressors. In a study of worksite health screen participants, Orleans and colleagues
(1983) compared 525 long-term former smokers who had been abstinent for more than
12 months (mean abstinence = approximately 9 years) with 856 current smokers and
found that the long-term former smokers had significantly better age- and sex-adjusted
scores on the Health and Nutrition Examination Survey (HANES) General Well-Being
Index, including its anxiety and depression subscales, and on the Framingham measures
of anger symptoms and anger internalization. However, there were no differences on

534
these measures between current smokers and recent ex-smokers, those who had been
abstinent for less than 12 months.

Prospective longitudinal studies of smokers who become former smokers or remain
continuing smokers are needed to establish whether any differences between former
and current smokers existed prior to quitting, especially since baseline or “prequitting”
measures of psychological well-being and self-esteem have been found to predict
success in quitting smoking (Hall et al. 1983; Ockene et al. 1982: Schwartz and
Dubitzky 1968; Straits 1970; West et al. 1977). The few prospective studies (Table 2)
that have been conducted have either documented no significant change in psychologi-
cal factors from baseline among former smokers, or no difference in the magnitude of
change for former and continuing smokers, or have indicated improvements for former
smokers. None of these studies demonstrated long-term negative psychological chan-
ges for former smokers.

Two of the prospective studies found no significant changes ina variety of mood and
psychological measures from a prequitting baseline to long-term followup among
former smokers and no significant differences between quitters and continuing smokers
in the magnitude of such change. Pertschuk and coworkers (1979) asked 24 participants
in a nonaversive cognitive-behavioral treatment to complete pretreatment and 2-month
followup ratings of psychological functioning. These researchers found no significant
changes in stress, affect, symptoms of psychological distress, or utilization of
psychiatric treatment as indicated by need for psychotropic medication or mental health
services. Changes from baseline to followup were not evaluated separately for quitters
and nonquitters, but these groups did not differ on 4-month followup ratings. Emmons
and associates (1986) studied the effects of smoking cessation on cardiovascular
reactivily to stress among quit-smoking clinic participants and found no significant
changes from baseline to a 6-month followup among 16 abstainers or 8 relapsers.
However, this study noted that an average weight gain of 5 pounds among abstainers
may have masked improvements in reactivity scores. Because weight was related to
baseline and followup cardiovascular measures, it is possible that in each of these
studies, treatment assisted quitters in avoiding persistent unwanted side effects.

Two studies of nicotine withdrawal effects that extended measurement beyond 4
weeks of abstinence have yielded no evidence for a withdrawal syndrome beyond 4 to
5 weeks (Hughes, Gust, Pechacek 1987: Gross and Stitzer 1989). These studies,
teviewed in detail by Hughes, Higgins. and Hatsukami (1990), found that adverse
postquitting changes in levels of anxiety, restlessness, impatience, irritability, and
dysphoria peaked during the first 2 weeks after quitting, returned to baseline or
below-baseline levels by 4 weeks, and remained at those levels at 10- to 26-week
followups.

Gross and Stitzer (1989) studied 40 smokers who quitafter a 3-session cessation class
and maintained biochemically validated smoking abstinence for 10 weeks while using
nicotine polacrilex gum or a placebo. Subjects completed weekly ratings of withdrawal
symptoms, including symptoms of psychological distress such as uritability, anxiety,
and impatience. Weekly followup ratings were adjusted for baseline ratings and
baseline smoking rate. For the 20 placebo subjects. mean ratings for irritability, anxiety.
and impatience increased from baseline to the first postquit week. returned to baseline
ors

TABLE 2.—Prospective studies of quitting-related changes in mood, anxiety, stress reactivity, perceived stress, self-image, and
psychological well-being

 

Reference

Sample size

Type of study

Findings

Suengths or limitations

 

Pertschuk etal.
(1979)

Emmons et al. (1986)

Gross and Stitzer
(1989)

24 smoking cessation clinic
participants

24 smoking cessation clinic
participants

40 abstainers using nicotine
polacrilex gum or a placebo
following a 3-session
treatment

Stress. affect, psychological
distress, and utilization of
psychiatric treatment were
assessed at the start of treatment
and 2 mo posttreatment

Cardiovascular reactivity

(SBP, DBP, HR) in response to
cognitive and physical stressors
were assessed | wk prior to
treatment and 6 mo after
treatment

A 15-item withdrawal symptom
measure was completed weekly
for 10 postquit weeks

No significant pre- to posttreatment
change in self-reported anxiety,
depression, anger, irritability,
appetite loss, insomnia,
hopelessness, difficulty
concentrating, apathy, use of
psychotropic medication

No significant pre- to posttreatment
change for abstainers (N=16) in
mean SBP, DBP, or HR, and no
difference in amount of change
between abstainers and recidivists
(N=8)

For placebo subjects, rated
symptoms of psychological distress
(irritability, anxiety, impatience)
increased from baseline to first
postquit week, returned to baseline
by week 4, then declined below
baseline initially, stabilizing after
5 wk; scores for active gum users
declined below baseline initially,
stabilizing after 3 wk at
below-baseline levels

Although posttreatment
scores did not differentiate
abstainers (N=16) and
recidivists (N=8), these
groups were not compared
On pre- to posttreatment
changes

Only abstainers had a
significant weight increase
during the following
period: this may account
for lack of reduction in
cardiovascular reactivity

Self-reported abstinence
biologically confirmed and
baseline scores and
baseline smoking rate used
as covariates, but no
control for repeated
Measurement
TABLE 2.—Continued

 

Reference

Sample size

Type of study

Findings

Strengths or limitations

 

Hughes. Gust.
Pechacek (1987)

Hath etal. (1983)

Orleans et al. (1983)

315 smokers followed for 6 mo At [-2 wk, [| mo, and 6 mo,

after a contact treatment with
physician advice and active
nicotine polacrilex or placebo
gum

35 participants in a cessation
clinic for smokers with
chronic cardiopulmonary
disease

72 ex-smokers (N=7 mo
abstinent) who had quit
during the year following a
worksite health screen

(49 at companies with health
promotion programs, 23 at
control companies)

subjects rated 5 withdrawal
symptoms relevant to mood and
psychological functioning
(anger, anxiety, difficulty
concentrating, impatience,
restlessness)

POMS was administered before
and 6 mo after treatment

HANES well-being, anxiety,
and depression scales and the
Framingham anger symptom
scales were administered at a
baseline health screen and
t-yr followup

Among abstinent subjects, these
ratings peaked at [-2 wk
postquitting, returned to baseline by
I mo, and declined further to
below-baseline at 6 mo

A measure of total mood
disturbance (anger/irritability

+ tension + anxiety + fatigue +
confusion + depression/dejection

~ vigor) at 6 mo was significantly
negatively correlated with smoking
reduction: parallel significant
relations were noted for the scales

anger/irritability and tension/anxiety

Significant baseline to |-yr
improvements in the HANES
well-being and depression scales
were observed for new ex-smokers
ut treatment sites only: no changes
in Framingham anger measures
were observed

Below-baseline 6-mo
ratings among nonquitters
suggest a drift in measures
due to a repeated testing
etfect

Analyses controlled for
pretreatment measures

Analyses controlled for
age. sex, baseline values,
and duration of abstinence:
comparisons with never
smokers, long-term former
smokers, or recidivists at
treatment sites were not
conducted
ges

TABLE 2.—Continued

 

Reference

Sample size

Type of study

Findings

Strengths or Timo rlitrons

 

Prochaska et al. din
press}

Cohen and Lichtenstein
(in press)

63 smokers quitting on their
own tated their self-change
processes semiannually for

2A vr

ISO smokers planning to quit

on their own

Self-reevaluation scale items
assessed changes in self-image
related to smoking

Smoking status and perceived

Stress were assessed at baseline.

1. 3. and 6 mo

T-scores declined progressively for

smokers going from action to
maintemance stages

Smokers who never quit (N=57)
and those who quitand relapsed
(N=8 1) maintained baseline stress
levels over the 6-mo followup:
smokers who quit and remained
abstinent (N=12) showed a
Significant decrease in perceived
stress from baseline to followup

Analyses assessed

stage based patterns of
change: comparisons with
smokers who did nat

progress were Hot reported

Causality is unclear

sfress may have
contributed te the tarlure to
quit smoking and failure to
quit may Ieve raused
perceived siress

 

NOTE: SBP= systolic blood pressure: DBP= diastolic blood pressure: HR = heart rc: POMS=Profile of Mood States: HANES-Health and Nutrition Exaninatnon Survey.
levels by week four, then continued to decline, stabilizing at below-baseline levels by
week six. There were significant interactions between use of the gum and the weeks
during which it was used for each of these symptoms, with nicotine polacrilex gum
significantly suppressing postcessation ratings only during the first 4 to 5 weeks after
quitting. The authors concluded that several of the most disturbing aspects of the
tobacco withdrawal syndrome appear to resolve within 4 to 5 weeks after quitting
(Gross and Stitzer 1989). Although findings suggest positive changes over baseline for
these recent quitters, below-baseline 6- to 10-week scores may reflect the effects of the
initial treatment or a repeated-testing effect.

In a similar study of the effects of nicotine polacrilex gum on tobacco withdrawal.
Hughes, Gust, and Pechacek (1987) studied 315 smokers for 6 months after a minimal
contact treatment involving brief physician counseling, instruction in nicotine
polacrilex gum use, and prescription of nicotine polacrilex gum or a placebo. At a
pretreatment baseline, and again at |- to 2-week, 1-month, and 6-month followups,
subjects rated six withdrawal symptoms related to mood and psychological functioning
including anger, anxiety, difficulty concentrating, impatience, and restlessness in
addition to four others—craving, hunger, insomnia, and physical symptoms. For 75
subjects abstinent at 6 months, of whom 57 used nicotine polacrilex gum and 18 used
a placebo, ratings for anger, anxiety, difficulty concentrating, restlessness, and im-
patience peaked at the I- to 2-week followup, returned to baseline at 1 month, then
dipped to below-baseline levels at 6 months. Subjects receiving nicotine polacrilex
gum compared with those using placebo reported smaller increases from baseline to |-
to 2-week and |-month ratings for most withdrawal symptoms, but nicotine polacrilex
gum effects were not explored at the 6-month followup because too few subjects
continued using the gum. However, 6-month ratings were lower on many symptoms
even among 240 nonquitters, suggesting a drift in ratings due to a testing effect. In fact,
the only symptom change from baseline, which differentiated quitters and nonquitters
at 6 months, was that quitters had a greater increase in hunger than did nonquitters
(p<0.001).

Hughes, Gust, and Pechacek (1987) concluded that, with the possible exception of
hunger and craving or an urge to smoke, there was no evidence for prolonged
withdrawal reactions lasting 6 months or more. (See Chapter 11 for discussion of
hunger and weight effects.) However, these researchers also noted that results based
on a select group of smokers who enrolled ina study and the absence of control groups
of long-term former smokers and continuing smokers not trying to quit limit the
generalizations that can be made about the symptoms of long-term abstainers.

Two other prospective studies comparing quitters and nonquitters have documented
6-month improvements in mood and well-being among former smokers who had
Participated in cessation treatments. Hall and associates (1983) administered the Profile
of Mood States to 35 smokers with cardiopulmonary disease both before and 6 months
after, 1 of 2 different 6-session quitting treatments. Controlling for baseline scores,
they found that total mood disturbance. including anger/irritability, tension/anxiety,
fatigue, confusion, and depression/dejection, was negatively correlated with smoking
reduction (p<0.02). That is, smokers achieving the greatest smoking reduction showed
the greatest improvements in overall mood. The same held true for the separate factors
of anger/irritability (p<0.05) and tension/anxiety (p<0.05). Treatment differences were
not explored.

Orleans and colleagues (1983) studied a group of 72 smokers who had quit in the
previous year (mean abstinence, 7 months), and compared the changes in mood and
well-being occurring among 49 quitters at 4 worksites where a range of employee health
promotion programs had been offered including smoking cessation, exercise, weight
control, and stress management. with those occurring among 23 quitters at 4 no-
treatment control worksites. The investigators controlled for age, sex, baseline values.
and months since quitting. Significant improvements in HANES well-being, anxiety,
and depression scores were observed only among former smokers at treatment com-
panies, but not among those at control companies (p<0.01). These results suggest that
treatment may have potentiated positive changes among new quitters. However, never
smokers, long-term former smokers, continuing smokers, or recidivists at treatment
companies were not compared.

Two studies have documented long-term, quitting-related improvements in psycho-
social outcomes among self-quitters. Prochaska and associates (in press) assessed the
processes that smokers undergo during different stages of smoking behavior change in
a 2.5-year longitudinal study of self-change among 63 self-quitters. These researchers
found significant decreases from baseline in smoking-related negative self-evaluations
(e.g., “My dependency on cigarettes makes me feel disappointment in myself’) from a
prequitting baseline for 9 subjects who progressed from the contemplation stage to the
action stage and then to maintenance, and for 54 subjects who progressed from action
to maintenance. Formal comparisons with subjects who did not progress in their stage
of change were not reported. (See Chapter 2 for a discussion of stages of change.)

Cohen and Lichtenstein (in press) found significant long-term reductions in perceived
stress in a prospective study of 150 unaided quitters. They administered the Perceived
Stress Scale (Cohen, Kamarck, Mermelstein 1983) prior to quitting and again at 1, 3.
and 6 months after the quit date. This scale measures the degree to which individuals
perceive the stresses in their lives to exceed their abilities to cope (range=0—-16). For
the 12 subjects who quit and remained continuously abstinent, perceived stress
decreased significantly from a prequitting mean of 5.7 to a 6-month followup mean of
2.9. Among 57 continuing smokers, perceived stress levels increased slightly from 6.1
prior to quitting to 6.3 at 6 months. Likewise, for the 81 smokers who quit but relapsed,
perceived stress levels increased slightly from a prequitting mean of 5.8 to a 6-month
mean of 6.1. There were no significant differences between quitters, coniinuing
smokers, and relapsers in prequitting perceived stress levels. The investigators suggest
that among smokers who want to stop smoking. quitting may have a beneficial influence
on perceived stress. self-esteem, and general self-efficacy (a belief that one has the
ability to perform a specific behavior such as smoking cessation) (Bandura 1982), and
failing to quit may have opposite effects. However, these researchers also noted that a
causal explanation cannot be clearly invoked: It is possible both that perceived stress
contributed to the failure to quit smoking (Marlatt 1985a; Shiffman 1982) and that
failure to quit contributed to stress.

More prospective studies are needed to clarify the long-term postwithdrawal
psychological consequences of smoking abstinence suggested by the research reviewed

540
for this Report. Studies designed specifically to assess long-term abstinence effects
will require longer followup, larger samples of unselected quitters, and contro! groups
of smokers who are not trying to quit. When possible and appropriate, self-report and
physiologic and observer ratings of emotional and psychological changes should be
included (Abrams et al. 1987: Hughes, Higgins, Hatsukami 1990) with measures of
health-related quality of life (Kaplan 1988).

Self-Efficacy and Locus of Control

Self-Efficacy

Self-efficacy has been shown to be a strong mediator of smoking behavior change
and to predict short- and long-term quitting outcomes (Condiotte and Lichtenstein 198] ;
Coelho 1984; McIntyre, Lichtenstein, Mermelstein 1983). As defined by Bandura
(1982), self-efficacy refers to one’s perceived ability to perform a specific behavior,
such as resist temptations to smoke under specific circumstances: that is, self-etficacy
is a response to a stressful event rather than a global sense of personal competence. As
such, self-efficacy related to smoking cessation is likely to influence both the decision
to engage in a quit attempt and perseverance in coping after quitting (Coelho 1984;
Marlatt 1985b).

The self-efficacy measures employed in smoking cessation research have concerned
only expectations for smoking behavior control. However, several researchers have
proposed that successful smoking cessation might itself result in feelings of increased
general self-mastery and self-confidence. That is. generalized self-efficacy may be a
consequence of smoking cessation (Cohen and Lichtenstein, in press; Marlatt 1985b.c:
Prochaska et al.. in press). No studies have yet examined prequitting to postquitting
changes in generalized self-efficacy.

However, the relationship between cessation and self-efficacy around smoking
control has been studied. Cross-sectional studies among smokers wanting to quit have
found that successful quitters score significantly higher on measures of self-efficacy
than either those who tried to quit and failed (Abrams et al. 1987: Barrios and Niehaus
1985; Prochaska et al. 1982) or continuing smokers (Katz and Singh 1986). These
differences may reflect that successful quitters generally have higher efficacy scores to
begin with (Fleisher et al., in press: Mothersill, McDowell, Rosser 1988: Ockene et al.
1982: Prochaska et al. 1985) or that one’s expectations that smoking can be resisted
would rise significantly as a function of actual success in doing so.

Prospective longitudinal studies, with followup periods ranging from several weeks
to 2.5 years postquitting, lend support to the hypothesis that increases in self-efficacy
conceming smoking control are related to smoking cessation both for untreated self-
quitters (Prochaska et al.. in press) and for smokers enrolled in treatment programs
(Coelho 1984: Killen, Maccoby, Taylor 1984: Nicki, Remington. MacDonald 1984:
Schwartz and Dubitzky 1968). Coelho (1984) reported that smoking control self-ef-
ficacy scores increased significantly from a mean of 77.1 at the time of enrolling in
treatment to a mean of 127.4 at 3 months posttreatment for 18 subjects who had quit
smoking. (Abstinence was defined as continuous nonsmoking since a quit date. but

S41
mean duration of abstinence was not reported.) Conversely, pretreatment and posttreat-
ment means for 48 nonquitters were unchanged from 78.1 to 75.1, respectively.

Two studies examined the effects of different types of smoking intervention treat-
ments on self-efficacy ratings. Killen, Maccoby, and Taylor (1984) found no differen-
ces in the amount of positive change in self-efficacy among abstainers of 4 weeks or
longer who took part in different treatments that included nicotine polacrilex gum,
nonsmoking skill training, or combined nicotine polacrilex gum and skill training.
Nicki, Remington, and MacDonald (1984) followed 53 subjects for 1 year after
treatment and found significantly greater increases in smoking control self-efficacy
among quitters and nonquitters randomized to a behavioral smoking intervention
treatment designed explicitly to enhance smoking control self-efficacy than among
those randomized to a standard control treatment (p<0.05). The mean duration of
abstinence for quitters was not reported.

Locus of Control

Measures of locus of control reflect the extent to which an individual believes that
he or she has control over personal happenings and circumstances. Measures of a
generalized locus of control reflect either expectations that one has internal ({i.e.,
personal) control over the reinforcements for one’s behavior, indicating an internal
locus of control, rather than believing that these reinforcements are determined by fate,
luck, or other forces beyond control (Rotter 1966), which reflects a more external locus
of control. Measures of health locus of control reflect beliefs that important health
outcomes can be controlled through behavior rather than by being at the mercy of luck,
fate, or powerful others (Wallston, Wallston, DeVellis 1978). Itis possible that former
smokers would shift toward a more positive or more internal control orientation in
reaction to their successful quitting. Anecdotal evidence suggests that when smokers
quit smoking they feel both more competent and more in contro] of their lives and that
they experience pride in their perceived “strength of will” (Knudsen et al. 1984).

Cross-sectional studies have demonstrated that former smokers, both self-quitters and
treated quitters, exhibit significantly more internal control orientations than either those
who tried to quit and failed (Rosenbaum and Argon 1979) or continued to smoke and
did not attempt cessation (Mlott and Mlott 1974; Orleans et al. 1983; Rosenbaum and
Argon 1979). However, prequitting measures of generalized (Ockene et al. 1982) and
health-specific (Horwitz, Hindi-Alexander, Wagner 1985) locus of control also dif-
ferentiate these groups.

Locus of control may be related to the duration of abstinence. Orleans and associates
(1983) found no significant differences between 1,343 current smokers and 856
short-term ex-smokers (abstinent for <3 months) in a baseline measure of perceived
personal control over preventable illness. However, 89 medium-term former smokers
(abstinent 3-12 months) and 525 long-term former smokers (abstinent for >12 months)
scored significantly higher on personal control than current smokers (p<0.01). A
followup conducted | year later showed a significant (p<0.01) increase toward internal
control among 72 smokers who had quit since baseline (mean abstinence, 7 months).

542
Conversely, Orleans and colleagues (1983) found a significant shift toward more
external health locus of control of similar magnitude among 30 individuals who had
been former smokers at baseline, but who had relapsed by the 1-year followup. A
similar pattern was reported by Horwitz, Hindi-Alexander, and Wagner (1985) who
followed 219 participants in a single-session hypnosis treatment over a 1-year period.
These researchers found a significant shift (p<0.001 ) toward a more external orientation
among 79 smokers who had tried to quit but failed, with the mean falling from 27.6
pretreatment to 24.2 at the l-year followup. The investigators suggested that general-
ized expectancies for control over one’s health might be diminished by failure and by
the “abstinence violation effect” (ie.. when individuals take a cigarette or relapse. they
may feel guilty or depressed or believe that they are lacking in will power and may
decide they are not maintaining control over smoking) (Marlatt 1985b). However,
Horwitz, Hindi-Alexander, and Wagner (1985) found no significant pretreatment to
followup shift toward an internal health locus of control among 56 continuously
abstinent quitters who had quit with hypnosis. This lack of change toward an internal
health locus of control may in part reflect that treatment using hypnosis does not
engender strong personal, internal attributions for success.

Two studies suggest that treatment factors can influence shifts in locus of control.
Orleans and associates (1983) divided 72 recent former smokers into 2 groups: 49 at
4 worksite companies where a comprehensive employee health promotion program had
been introduced and 23 at 4 no-treatment control companies. The significant overall
shift toward an internal health locus of control was accounted for wholly by the former
smokers at treatment companies. It is possible that the intervening health promotion
program emphasizing personal control over health, well-being. and preventable illness
potentiated or hastened this shift. Blittner, Goldberg. and Merbaum (1978) randomly
assigned 54 smokers seeking treatment to 1 of 3 conditions: a stimulus control
treatment coupled with bogus feedback of superior self-control abilities. a stimulus
control treatment alone, or a wait list control. A statistically significant pretreatment
to posttreatment increase in internal orientation was observed only for the subjects who
received feedback to enhance their expectations of inner control ability. This group
also achieved the greatest 14-month smoking reductions (p<0.001).

Thus, most of the available data suggest that smoking cessation ts related to an
increase in a more internal locus of control orientation: no data indicate a shift toward
an external locus of control for abstainers. There is some support to suggest that
treatment method may have a differential effect on an increase in internal locus of
control orientation.

Coping and Self-Management Skills

The relation of abstinence from cigarettes to a generalized improvement in the extent
and use of coping and self-management skills has not been studied. To the extent that
stopping smoking results in an individual's acquiring or strengthening generally ap-
plicable stress-coping and temptation-coping skills, long-term benefits of abstinence
might be expected to include the generalized use of such skills. However. no studies
have assessed whether increases in generalized stress-coping skills occur as a conse-

542
quence of cessation. Longitudinal studies have not included prequitting and postquit-
ting measures of generic coping strategies. A brief review of the relation of coping to
smoking cessation and maintenance of abstinence may help to provide direction for this
line of needed research.

Shiffman and Wills (1985) have developed a conceptual framework of coping that
distinguishes stress-coping skills, that is. skills used to cope with general lite stressors,
and temptation-coping skills, or skills relevant for coping with a situation in which there
is a specific temptation for substance use or an urge to smoke. Folkman and Lazarus
(1988) defined stress-coping as constantly changing cognitive and behavioral efforts to
manage specific external and internal demands that are appraised as taxing or exceeding
the resources of the person to maintain an appropriate balance between environmental
demands and resources available to the individual to meet those demands. Temptation
coping can be separated into what smokers do when faced with the immediate tempta-
tion to smoke and anticipatory coping or the strategies smokers use to maintain
commitment to abstinence and prevent temptation (Shiffman and Wills 1985).

To the extent that smoking constitutes a maladaptive response for coping with stress
and negative affects such as anxiety, depression, anger, frustration, loneliness, or
boredom (Abrams et al. 1987: Marlatt 1985b,c; Ockene et al. 1981), the former smoker
must find alternative strategies for coping. The use of healthy all-purpose coping
strategies such as self-reinforcement. assertive behavior, social support, relaxation, and
exercise has proven important to success in maintaining abstinence in some studies
(Ashenberg, Morgan, Fisher 1984: Grunberg and Bowen 1985; Marlatt 1985c: Shif-
fman 1982).

However, two large worksite studies demonstrated no differences between current
and former smokers in the self-reported use of healthy and unhealthy techniques for
coping with stress (Blair et al. 1980; Orleans et al. 1983). In support of the importance
of coping skills. Katz and Singh (1986) found that 77 former smokers who had abstained
for 6 months or more (mean 6.7 years) had significantly higher scores on the Rosenbaum
Self-Control Schedule (a self-report measure of individual differences in applying
self-control or coping methods) than 52 smokers recruited for a quit-smoking treatment.
Self-cured” and treated former smokers did not differ on this measure. The inves-
tigators concluded that former smokers may have succeeded because they possessed
better self-coping skills initially. The same interpretation could be applied to the study
by Abrams and associates (1987) in which 22 former smokers (mean abstinence 22
months) exhibited better observer-rated skills to resist the temptation to smoke than did
22 recidivists in simulations involving interpersonal smoking triggers. Shiffman (1982)
found that former smokers who reported using cognitive and behavioral strategies to
cope with smoking temptations were less likely to relapse. These tew studies support
the conclusion that use of skills to cope with stress and with temptations or urges to
smoke seem to be more prevalent among former smokers compared with current
smokers.

544
Social Support and Interpersonal Interactions

Research has not addressed how smoking cessation influences the level of general or
quitting-relevant social support available to the quitter or how cessation affects the
quality of the individual's interpersonal interactions. Research on social support
processes has focused on examining baseline or posttreatment measures of social
support as predictors of quitting success (Graham and Gibson 1971: Lichtenstein,
Glasgow, Abrams 1986: Mermelstein et al. 1986; Ockene et al. 1982: US DHHS 1989),
Several studies have demonstrated that successful quitters had significantly fewer
smokers in their social networks at baseline than did continuing smokers (Eisinger
1971; Graham and Gibson 1971; Ockene et al. 1982). Others have demonstrated that
the quitter’s success stimulated quitting by others. especially spouses (Suedfeld and
Best 1977).

A few studies are relevant to the investigation of cessation effects on social support.
A large-scale, cross-sectional and longitudinal worksite study (Orleans et al. 1983)
found no differences among current smokers, former smokers, and never smokers at
baseline in satisfaction with personal relationships and interpersonal communication
or in satisfaction with coworker relationships. However, at 1-year followup. 72
baseline smokers who had quit (mean abstinence, 7 months) showed a significant
decline from baseline in satisfaction with coworker relationships (p<0.01) and scored
significantly lower in satisfaction with personal relationships (p<0.05) than a group of
30 baseline former smokers who had relapsed since baseline. Whether new former
smokers were in no-treatment control companies or in treatment companies where they
benefitted from multiple health promotion programming, designed in part to boost
coworker support, did not affect changes in satisfaction with interpersonal relationships.
These negative changes in interpersonal relationships are difficult to interpret because
former smokers in this study also demonstrated decreases in anxiety and depression and
improvements in coping strategies compared with baseline. One possibility is that new
former smokers may be less tolerant of smokers in their environment. Further study is
needed to replicate and explain this isolated finding.

In contrast, Prochaska and colleagues (in press) monitored a group of 63 self-quitters
who progressed through the stages of smoking behavior change to maintain abstinence
over 2.5 years (mean duration of abstinence was not reported) (Chapter 2). They found
that their use of helping relationships continued to increase with time. Similarly,
Horwitz, Hindi-Alexander, and Wagner (1985) found that 56 successful quitters
reported significantly greater social support from spouses and friends | year after a
single-session hypnosis treatment than they did at baseline. No changes in reported
level of support were noted for 84 continuing smokers, but even 79 recidivists reported
significant increases in spouse support over baseline. Notwithstanding hypnotic sug-
gestions that “other peoples” smoke will not bother you.” successful quitters reported
significantly (p<0.05) more often expressing objections to others smoking around them
(mean=2.38) than either recidivists (mean=0.75) or continuing smokers (mean=0.50)
at the |-year followup. Likewise. more former smokers requested nonsmoking areas
in restaurants (53 percent) and public transport (32 percent) than did recidivists (12
percent and 12 percent, respectively) or continuing smokers (8 percent and 6 percent.
respectively). This practice may have helped to minimize social pressures to smoke
commonly precipitating relapse (Marlatt and Gordon 1985). and helped to assure
support for maintenance. [tis also possible that these practices simply resulted from,
rather than contributed to, smoking abstinence.

The results of these studies, although somewhat conflicting. suggest that former
smokers played an active role in structuring the improved support they reported as a
way of maintaining abstinence. However, given the limited information, no con-
clusions regarding the effect of smoking cessation on social interactions can be made
at present.

Summary

Research findings provide no evidence for any long-term negative psychological
effects beyond hunger and craving. However, the available findings suggest that there
are some postwithdrawal psychological benefits that may increase with duration of
abstinence.

HEALTH PRACTICES OF FORMER SMOKERS

Introduction

Several studies have found that both good health practices and poor health practices
cluster (Belloc and Breslow 1972: Tapp and Goldenthal 1982; Verbrugge 1982:
Marsden, Bray. Herbold 1988). Self-defined former smokers appear more likely than
current smokers to engage in regular exercise and to practice other recommended health
behaviors. In general, smokers who quit and who subsequently or concurrently change
other health behaviors may represent a more distinct health-conscious group. Castro
and coworkers (1989) have suggested that cigarette smokers exhibit less healthy
lifestyles along cognitive. behavioral, and motivational dimensions. As the authors
noted, addictive behaviors seldom occur in isolation but are instead embedded within
complex behavioral chains or lifestyles. Conversely, the data presented in this Section
suggest that when individuals stop smoking. other beneficial health practices also may
emerge. Given the nature of the available data, it is not possible to determine whether
these other beneficial health behaviors reflect the characteristics of a distinct health-
conscious subgroup of smokers, emerge as part of the smokers” efforts to maintain
abstinence (e.g., increased exercise), represent a response to adverse withdrawal
symptoms (e.g.. changes in dietary practices). or are direct effects of quitting.

This Section reviews data on former smokers’ physical activity and dietary practices
and use of other substances such as alcohol and other forms of tobacco, and former
smokers’ profiles with regard to multiple health-enhancing behaviors. Changes in
former smokers’ physical activity and dietary practices, as they relate to postcessation
weight changes, are also reviewed in Chapter 10.

The studies reviewed in Chapter 10 are longitudinal investigations in which former
and continuing smokers are compared. This Section focuses on cross-sectional data

546
from two nationwide surveys, the National Health Interview Survey (NHIS) conducted
by the National Center for Health Statistics (Kovar and Poe 1985: Schoenborn and
Benson 1988) and the Behavioral Risk Factor Surveillance System (BRFSS) coor-
dinated by the Centers for Disease Control and conducted by State health departments
(Remington et al. 1988). Both surveys provide large data sets on health behaviors in
the noninstitutionalized adult population. The limitations of drawing conclusions from
cross-sectional data apply here (Chapter 2).

For its yearly interviews, NHIS uses a multistage probability scheme sampling
technique developed in collaboration with the Bureau of the Census and employs
personnel trained for the decennial census. BRFSS uses a multistage cluster technique
of random digit dialing to select households for its yearly telephone survey. Both
randomly select a respondent from a list of residents identified when a household is
chosen.

A core set of questions each year is used in NHIS, then additional questions are added
in supplements to the core survey in keeping with each year’s chosen focus. In 1985,
the NHIS special topic was health promotion, with variables such as physical activity,
dietary practices, sleep, weight, alcohol use, and smoking that were similar to those
used in the pioneering Alameda County study. The health promotion portion of the
interview was completed by an estimated 90 percent of eligible respondents (Schoen-
born and Benson 1988). In 1987, the special topic was cancer, with questions on diet,
smoking, smokeless tobacco use, alcohol use, vitamin and mineral consumption,
knowledge about cancer risks, cancer screening and preventive care, and family history
of cancer. The cancer-related portion of the interview was completed by approximately
86 percent of eligible respondents (Schoenborn and Boyd 1989). In both NHIS surveys.
a former smoker self-reported as having smoked at least 100 cigarettes and not smoking
at the time of the survey. Mean duration of abstinence was not reported (Schoenborn
and Benson 1988; Schoenborn and Boyd 1989).

In 1987, BRFSS covered blood pressure, physical activity, weight and dieting. diet,
alcohol use, preventive practices, seatbelt use, stress, pregnancy status, use of oral
contraceptives, and use of smokeless tobacco and cigarettes. The median cooperation
rate (the ratio of completed interviews to the sum of completed interviews and refusals)
among the participating States was 84 percent (Remington et al. 1988). Similar to
NHIS, BRFSS defined a former smoker as an individual who had smoked at least 100
cigarettes in his or her lifetime and was not smoking at the time of the survey. (Mean
abstinence of former smokers cannot be calculated. However. 64 and 54 percent of
men and women, respectively, were abstinent from cigarettes for more than 5 years.)

Although these three surveys are similar, the published data available from them
differ in several respects. Data from the 1985 NHIS, presented in Table 3, are
age-adjusted (Schoenborn and Benson 1988). Data from the 1987 NHIS, presented in
Table 4, are simple proportions with no variables controlled (Schoenborn and Boyd
1989). Data from the 1987 BRFSS were analyzed to assess the relationships between
cigarette smoking and lifestyle and preventive practices (Table 5) and to examine the
same relationships with respect to the duration of cigarette abstinence (Table 6). The
odds ratios, presented in Tables 5 and 6, are controlled for age. ethnicity. and level of
education.

547
TABLE 3.—Summary of data from 1985 NHIS, behaviors of never, former, and
current smokers aged 20 and older

 

 

 

 

. Never Former Current

Behavior smokers smokers smokers
oe Gy %

MEN
Alcohol consumption
Heavier drinker® 79 12.7 18.9
25 drinks” 13.8 21.2 28.7
Weight/diet/exercise
Never eats breakfast 18.9 22.3 33.3
Snacks daily 39.3 40.4 38.5
Less physically active” 13.2 14.6 18.8
Sedentary? 46.6 47.7 57.2
Overweight" 28.1 30.0 21.2
Other
Sleeps <6 hr 21.5 22.5 24.9
WOMEN
Alcohol consumption
Heavier drinker" Ll 3.7 6.1
25 drinks? 22 5.0 8.5
Weight/diet/exercise
Never eats breakfast 17.7 19.8 37.6
Snacks daily 37.6 415 35.3
Less physically active’ 19.9 23.3 24.9
Sedentary4 61.1 58.5 64.3
Overweight® 24.9 23.0 17.9
Other
Sleeps $6 hr 20.4 19.9 24.4

 

NOTE: All percentages are age adjusted; NHIS=National Health Interview Survey.

“Measure developed by the National Institute on Alcohol Abuse and Alcoholism. Categories based on ounces of ethano!
consumed during the past 2 wk: heavier drinker is defined as having an average of 1.0 0z (2 drinks) or more/day.

"Five drinks or more on 10 days or more in the past year.

“Based on perceived level of physical activily relative to others.

“Energy expenditure on leisure activity of 0 to 1.4 kcal ke/day.

“Twenty percent or more above desirable weight based on 1983 Metropolitan Life Insurance Company standards.
according to self-report of weight.

SOURCE: Schoenborn and Benson (1988).

548
TABLE 4.—Summary of data from 1987 NHIS behaviors of never, former, and
current smokers aged 18 and older

 

 

Never smokers Former smokers Current smokers
Behavior % % %
MEN
Alcohol consumption
Drinks beer >S/wk 6.4 12.6" , 17.1°
Drinks 23 beers/episode 36.3 30.5" S32
Drinks wine 25/wk 1.2 38? 1.7
Drinks 23 glasses wine/episode 12.9 11.6°° 202°
Drinks liquor >5/wk L7 4.gab 4.1°
23 drinks/episode 20.4 35 gab 451°
Dietary practices
3 meals/day on weekdays 48.6 50.9 , 32.8"
3 meals/day on weekends 44.3 44.0 35.0"
Avoids snacks weekdays 24.5 30.5°> 26.55
Avoids snacks weekends 21.0 25. gth 365
Has changed diet for health 35.0) 44.34) 26.4°
220% above desirable weight 249 a4.24b 23.85
Preventive care
Digital rectal exam (ever) 59.5 66.8" , 59.4
Blood stool test (ever) 38.6 44.9° 33.9
Proctoscopic exam (ever) 24.0 27.7 24.0
WOMEN
Alcohol consumption
Drinks beer 25/wk 0.9 2.3% 4.0°
Drinks 23 beers/episode 17.1 17.2" 32.7
Drinks wine >S/wk 13 4.3%? 1,9
Drinks 23 glasses wine/episode 7.0 10,9°> 17.8°
Drinks liquor 25/wk 0.7 2.78 27
23 drinks/episode 13.7 14.1> 32.0¢
Dietary practices
3 meals/day on weekdays 50.1 49,5? . 29.5°
3 meals/day on weekends 44.2 41.8" 29.4
Avoids snacks weekdays 26.6 26.9 26.8
Avoids snacks weekends 236 a4 484 23.6
Has changed diet for health 38.7 49.o°? 345°
220% above desirable weight 243 a4.ge 203°
Digital rectal exam (ever) 56.8 67.4°° 60.6"
Blood stool test (ever) 37.9 46.28" 35.7°
Proctoscopic exam (ever) 20.8 27,38? 21.1
Pap smear (within year) 39.2 43,585 40.75
Breast self-exam (within yr) 34.8 40.38% 34.0
Breast exam (monthly) 51.5 §2.2 52.1
Mammogram (ever) 38.5 46.7° 35.1°

 

NOTE: NHIS=National Health Interview Survey.
“Former smokers differ from never (ps0.5).
Former smokers differ from current (ps0.5).
Current differ from never (ps0.5).
This Is (a) but not (b) because of sample size despite sume point estimate.
SOURCE: Schoenborn and Boyd (1989).

549
TABLE 5,—Summary of data from 1987 BRFSS, behaviors of former smokers
and current smokers aged 18 and older

 

Adjusted odds ratios

 

 

 

Former smokers Current smokers Former smokers
. relative to never relative to never relative to current
Behavior smokers smokers smokers
MEN
Alcohol consumption
Any alcohol/mo 1.75" za 0.82"
25 drinks/episode 1.67" 2.64" 0.63"
260 drinks/mo 1.75" 3.03" 0.58"
Drinking and driving 144" 1.99" 0.71"
Weight/diet/exercise
Obese (BMI) 1.05 0.62" 1.68"
Obese (Met. Life)! 1.06 0.64" 1.63"
Trying to lose pounds 1.22" 0.63" 1.92"
More exercise 0.98 0.82" 117°
Eating fewer kcal 0.85° 0.82° 1.04
Physical activity 1.10° 0.69" 1.573
Sedentary oot" 1.44" 0.64"
Preventive care
Cholesterol test 127° 0.94 1.34"
Flu shot past month 1.09 0.87° 1.26"
Other
Use ST La 0.84" 2.09"
Use seatbelt 0.92? 0.58" 1.60"
WOMEN
Alcohol consumption
Any alcohol/mo 2.077 2.34" 0.87"
25 drinks/episode 1.86" 3.35" 0.55"
260 drinks/mo 28x" 5.45" 0.52"
Drinking and driving 187 2,92" 0.65"
Obese (BMI) 0.98 0.63" 1,59"
Obese (Met. Lifes” 0.96 0.65" 1.52"
Trying to tose pounds 1.19" 0.75" 1.60"
More exercise 1.07 0.72" 1.48"
Eating fewer kcal 0.97 0.96 0.99
Physicat activity 1.17 0.81" 1.45*
Sedentary 0.86" 1.24" 0.69"
Preventive care
Cholesterol test 1.15" lie 1.05
Flu shot past month 0.95 0.91° 1.05

550