Rehav. Res. &Therapy,
1974. Vol. I2 pp. 107-I 15. Pergamon Press. Printed in Great Britain
ALCOHOLIC
DRINKING: SUPPRESSION
TIME-OUT PROCEDURE*
GEORGE BIGELOW, IRA LIEBSON
BY A BRIEF
and ROLAND GRIFFITHS
Department of Psychiatry, Baltimore City Hospitals and Department of Psychiatry and Behavioral Sciences.
The Johns Hopkins University School of Medicine, Baltimore, Maryland, U.S.A.
(Received 21 November 1973)
Summary-Volunteer
chronic alcoholics were given daily access to substantial quantities of alcohol within a residential research setting. Drinking was suppressed to an average of approximately
one-half of Baseline levels when 10 or 15 min of physical and social isolation was required as an
immediate consequence to receiving each one-ounce drink of 95-proof ethanol. This time-out procedure suppressed the drinking of nine out of ten subjects. Drinking returned to high levels when
brief contingent time-out was discontinued. The relevance of such reversible controlling relationships to the treatment of alcoholism is discussed.
Experimental study of alcoholics’ drinking has focused primarily upon the characteristics
and consequences of alcohol ingestion (Mello 1968; Mello and Mendelson, 1970; Nathan
and O’Brien, 1971, Schaefer et al., 1971; Gottheil et al., 1972). However, research with alcoholics within experimental drinking situations can also provide data concerning the controllability of alcoholics’ drinking (Bigelow et al., 1972; Gottheil, et al., 1973) and especially concerning specific manipulable environmental conditions which exert control over
such drinking and which might be used to reduce it (Cohen et al., 1971b; Cohen et al.,
1971~; Mills et al., 1971; Bigelow and Liebson, 1972).
The reduction of drinking is a primary goal of any alcoholism treatment program. Yet
treatment efforts suffer from a lack of empirical information concerning factors which influence alcoholics’ drinking. Current practice is based upon tradition, speculation and
personal anecdotes. The basic experimental analysis of and modification of alcoholic
drinking within the laboratory appears to offer the opportunity to gain the needed empirical understanding of environmental factors which support alcoholic drinking and of those
which can be manipulated to alter this behavior. Knowledge of such controlling variables
should provide an empirical base for further advances in the prevention, management and
remediation of alcoholism.
The present report describes a simple behavioral procedure which has suppressed the
drinking of chronic alcoholics given access to substantial quantities of alcohol within a
residential hospital research ward. The procedure utilized has been that of contingent
time-out from positive reinforcement (Leitenberg, 1965), which consists of scheduling as
an immediate consequence to taking an alcoholic drink an interval during which other
reinforcers are unavailable. In the present case time-out has consisted of a brief period of
physical and social isolation. Such a contingent isolation procedure is widely used as a
technique for suppressing undesirable behaviors in other populations and settings (Bur* This research was supported by USPHS grant no. AA-00179 to Baltimore City Hospitals. Reprints may be
obtained from the authors, Department of Psychiatry, Baltimore City Hospitals, Baltimore, Maryland 21224,
U.S.A.
107
108
G. BIGELOW,I. LIEBSONand R. GRIFFITHS
chard and Tyler, 1965; Zeilberger er al., 1968; Tharp and Wetzel, 1969; Bostow and Baily,
1969; O’Leary and O’Leary, 1972; Burchard and Barrera, 1972).
METHOD
Subjects
Ten male volunteer subjects referred from the Emergency Room of Baltimore City Hospitals participated. All were chronic alcoholics with histories of prior hospitalization for
alcoholism.
Subjects were generally admitted to the six-bed Behavior Research Ward in a state of
intoxication or withdrawal from alcohol. They were medically and psychiatrically
screened prior to research participation. Criteria which precluded participation included
central or peripheral neurological abnormalities, physical or laboratory evidence of liver
disease, or histories suggesting seizures, pancreatitis, peptic disease, GI bleeding, or severe
functional psychiatric disorder.
Procedure
Subjects were detoxified under medical supervision and their informed consent obtained
prior to research participation. Informed consent involved subjects’ agreement to participate in ‘behavioral studies on alcoholism’. in which alcohol would be made available for
their consumption and in which the conditions of availability might be altered from time
to time. Details of the experimental procedures were described when they were implemented. Subjects were given no instructions or explanations of what they were ‘supposed
to do, or of what outcomes might be expected.
Following consent to participate each subject was given access to a daily alcohol ration
ranging, between subjects, from 12 to 24 drinks. Drinks consisted of 30 cm3 (1 oz) of 95proof ethanol mixed in 60 cm3 (2 oz) of orange juice, and were provided by the ward staff
upon request between 7:00 a.m. and 1l:OO p.m.
Within this framework an experiment evaluating the effect of a brief, contingent timeout procedure on the drinking of alcoholics was conducted. Time-out consisted of a brief
period of physical and social isolation. A within-subject experimental design was
employed: each subject was exposed to both control and experimental conditionsusually followed by a reversal.
During the initial Baseline control condition subjects consumed their drinks in the main
ward social area, and were free to participate in other activities while they drank. During
the immediately following experimental Contingent Isolation condition subjects were
required to sit for 10 min in an isolation booth immediately upon receiving each drink
(15 min for subjects G and D). Six subjects were subsequently returned to the Baseline
condition.
The isolation booth was a three-sided booth with a curtain across the fourth side, 30
in square inside. It was located on the ward about 15 ft from the main social area. The
effect of the time-out procedure was to eliminate virtually all concurrent behavioral alternatives as an immediate consequence to receiving an alcoholic drink. Socializing, reading,
talking, recreation, etc. were not permitted during time-out. The only activities permitted
were drinking and smoking.
Tables 1 and 2 present a summary of the sequence of experimental conditions and the
durations of exposure to each for the ten participants. Initial exposures to baseline conditions ranged in duration between 5 and 10 days. Exposures to contingent isolation ranged
109
Alcoholic drinking: suppression by a brief time-out procedure
Table 1. Drinking by alcoholic subjects exposed only to Baseline and Contingent
Isolation conditions. The percentage of available drinks consumed in each condition
is shown. Figures in parentheses show the number of days exposure.
Subject
No. Drinks
available
H
24
c
24
M
24
Sa
18
Sm
15
Ho
12
G
12
D
12
Total
Baseline 1
Contingent
isolation
75.8
(5)
100.0
(7)
87.5
(10)
100.0
(8)
98.0
(6)
96.7
(5)
100.0
(6)
100.0
(5)
37.5
(9)
43.8
(11)
88.3
93.4
45.6
(6)
1.7
(6)
46.7
(3)
40.0
(5)
70.0
(13)
0.0
(5)
Baseline 2
63.3
(4)
100.0
(3)
95.8
(4)
100.0
(6)
100.0
(3)
90.4
in duration between 3 and 13 days. Variations in the duration of exposure to conditions
and in whether a reversal was performed were substantially determined by the duration
for which the particular subject agreed to participate in the research.
Two subjects were selected because the time-out procedure alone produced negligible
suppression of their drinking and were exposed to a further behavioral manipulation. A
temporal spacing requirement was imposed such that at least one hour had to elapse
between receipt of successive drinks. Preliminary data (unpublished) suggested this manipulation could decrease subjects’ disposition to drink. Under this ‘spacing’ condition the
effects of Contingent Isolation were re-assessed according to the sequence of conditions
shown in Table 2.
Subjects participated in the study consecutively rather than concurrently. Thus, during
each subject’s participation other subjects on the six-bed research ward were involved in
a variety of other experiments involving alcohol self-administration.
Table 2. Drinking by alcoholic subjects exposed to temporal spacing requirement. The percentage of available
drinks consumed in each condition is shown. Figures in parentheses show the number of days exposure.
Contingent
isolation
only
Spacing &
contingent
isolation
Spacing
only
Baseline 2
100.0
(6)
Subject
No. drinks
available
S
18
100.0
(7)
97.2
(4)
30.0
(5)
71.1
(5)
Mi
15
100.0
(6)
100.0
(5)
34.7
(6)
66.7
(6)
100.0
98.6
32.2
62.2
Total
Baseline
100.0
110
G. BIGELOW,I. LIEBSONand R. GRIFFITHS
RESULTS
Data for individual subjects are shown in Tables 1 and 2. Table 1 presents data for the
eight subjects exposed only to baseline and contingent isolation conditions. Table 2 presents data for the two subjects who were also exposed to the temporal spacing requirement.
On the average, contingent isolation suppressed drinking to about one-half of control
levels. Although the magnitude of contingent isolation’s suppressive effect upon drinking
has varied somewhat across subjects, the individual data indicate that the average effect
was also the modal effect.
Contingent_
Isolation
‘1
PO0
‘0
$
I6-
0
2
4
1
6
6
IO
I2
I4
I6
I6
DAYS
Fig. 1. Brief contingent isolation suppressed subject H'sdrinking to approximately
control levels.
one-half of
The ten subjects consumed 94.6 per cent of the available drinks during the initial baseline period. During the subsequent contingent isolation period (without temporal spacing
of drinks) they consumed only 52.1 per cent of the available drinks. This difference is highly
significant statistically (p < 0.0001 according to a test for difference of proportions). The
six subjects who were subsequently returned to baseline conditions again consumed 91.9
per cent of the available drinks.
Contingent isolation exerted a suppressive effect upon the drinking of nine of the ten
alcoholic participants in this study. For seven of the ten subjects this suppressive effect
was demonstrated while drinks were freely available. For two subjects this suppressive
effect became evident only when the additional requirement of one hour temporal spacing
between drinks was superimposed. Subject M, the only subject with whom suppression
was not demonstrated, was not available for exposure to this temporal spacing condition,
The course over days of the time-out effect upon drinking has been variable across subjects. Most subjects have displayed greater variability of drinking during contingent isolation periods than during control periods (due, possibly, to a ceiling effect during control
periods). In general there has been no clear trend in the level of drinking across consecutive days of contingent isolation. Two subjects (H and Ho) showed their greatest magni-
Alcoholic drinking: suppression by a brief time-out procedure
Subj.
111
So
Fig. 2. With subject Sa near-total suppression of drinking occurred during the contingent isolation
phase, although the maximum available alcohol was always consumed during control phases.
tude of suppression on the first day of exposure to contingent isolation. The magnitude
of suppression increased over several consecutive days for four subjects (C, Sm, G and Mi).
Three subjects (Su, D and S) showed no apparent trend over time. Only one subject (C)
showed a large magnitude of suppression and subsequently returned to heavy drinking
while the contingent isolation condition remained in effect. Temporal correlations suggest
that relapse in this subject was related to the fortuitous occurrence of a specific emotional
stress (an impending biopsy). This subject’s drinking had been gradually suppressed to zero,
but drinking resumed on the fifth day of contingent isolation within 10 min of discussing
with the subject the biopsy plan.
The strength of this contingent isolation procedure, and its temporal course, are represented graphically in Figs. l-4 which show daily alcohol intake through the course of experiments with four representative subjects.
Contigent
lsolotion
a-+
Subj
Sm
Fig 3. Subject Sm’s drinking became progressively more suppressed until he terminated
research participation.
his
G. BIGELOW,I. LIEBXINand R. GRIFFITHS
112
*Contingent
Isolation
_
- o-o-o-o-o-o-c
16 -
z
z
I*-
E
0
-6
d
z
_
s-
?
4-
o-o
I
’,
I
I
s:
’
:
’
’
Subj.
S
D
of
,
,
,
,
,
,
,
,
,
,
,
,
,
0
2
4
6
e
IO
12
14
16
18
20
22
24
2e
‘I
28
DAYS
Fig. 4. With subject S the effect of brief contingent
isolation was minimal when drinks
available, but substantial
when 1 hr spacing of drinks was required.
were freely
DISCUSSION
The effectiveness of this simple time-out procedure in suppressing alcoholics’ drinking
is especially striking in contrast to the widespread traditional belief in a ‘loss of control
phenomenon with alcoholics. This study joins a growing body of literature (e.g. Bigelow
et al., 1972; Gottheil et al., 1973) which is resulting in a gradual retraction or reformulation
(e.g. Keller, 1972) of the loss-of-control hypothesis.
More significant than simply demonstrating that alcoholics’ drinking can be restrained,
the present study demonstrates control to be exerted by a specific manipulable variablebrief, contingent isolation. Further, this procedure has effectively suppressed drinking even
in the midst of a drinking episode. Specifically, contingent isolation has been shown to
suppress drinking of alcoholics who have been drinking 36&720 cm3 (12-24 oz) of 95
proof ethanol daily for 48 consecutive days immediatley prior to the introduction of contingent time-out. Clearly, alcoholics’ drinking can remain responsive to and controllable
by environmental consequences even after several consecutive days of substantial drinking.
Previous studies with alcoholics have demonstrated that the non-contingent imposition
of a period of isolation and reinforcer-unavailability
does not itself suppress drinking
(Nathanand O’Brien, 1971; Cohen rt ul., 1971a; Cohen et al., 1971b). The studies by Cohen
et al., represent pioneering demonstrations of the controllability of alcoholics’ drinking.
In this work drinking was suppressed when consumption of more than five drinks resulted
in long-duration (24-48 hr) withdrawal of a wide variety of reinforcers (social, recreational, dietary, therapeutic, and vocational, as well as access to alcohol). The research was
conducted using a drinking baseline procedure which prohibited consecutive days of heavy
drinking either immediately prior to or during the experimental condition. The present
study extends the control of drinking to the more naturalistic baseline of continuous drinking, while at the same time greatly reducing the breadth and duration of reinforcer loss.
The present data do not reveal the mechanism by which the time-out procedure suppresses drinking. The l&15 min time-out duration was selected as being the average
Alcoholic drinking: suppression by a brief time-out procedure
113
duration taken to consume a single drink by this subject population, with the aim being
to eliminate specifically those social and recreational activities which generally occur concurrently with the consumption of alcoholic beverages. If time-out eliminated those reinforcers which normally maintained drinking, drinking would be suppressed. However,
Shipleyetal.( 1972) have demonstrated that contingent time-out from one reinforcer can suppress behavior which is maintained by another reinforcer. Consequently, the time-out procedure may be an effective behavioral suppressant either because it eliminates those
sources of reinforcement which previously maintained drinking, or because the isolation
acts as an aversive event which suppresses drinking directly.
The results with two subjects exposed to the requirement of one-hr spacing of drinks
indicate that such temporal spacing enhances the controllability of alcoholic drinking and
establishes a more sensitive behavioral baseline for evaluating conditions which may influence the disposition to drink. The requirement of temporal spacing might further be
used experimentally to reveal the suppressive effectiveness of other variables whose actions
would be masked if drinking were unrestrained.
Several clinical applications involving time-out procedures with alcoholism have been
reported. Sulzer (1966) and Pickens et al. (1973) have reported case studies in which contingent time-out was utilized. The community-reinforcement
treatment of alcoholism
reported by Hunt and Azrin (1973) and the employee alcoholism program described by
Cohen et al. (1973) are both conceptually based upon the principle of arranging an immediate time-out from reinforcement as a consequence to drinking.
Since one of the defining characteristics of alcoholism is that it has the‘consequence of
ultimate loss of reinforcement (social, economic, occupational, etc.) it may be surprising
that programming reinforcement loss can suppress drinking. The difference appears to be
in the immediacy and reliability of the occurrrence of time-out. Although loss of, or timeout from reinforcement occurs naturally as a consequence to chronic excessive drinking,
its occurrence is generally long delayed and unreliable. The effective application of a contingent time-out procedure is charactierized by immediacy and reliability.
The contingency management approach represented by the present contingent isolation
time-out procedure is both conceptually and operationally different from those behavior
therapy procedures which attempt to produce an aversion to alcohol by pairing drinking
with some trauma. The latter approach is conceptually based upon a respondent conditioning model, and attempts to produce changes in affective processes or autonomic
mechanisms which would subsequently be reflected in behavioral change (Franks, 1966).
The present contingent time-out procedure is conceptually based upon an operant behavior model, which emphasizes that the consequences of a behavior are a primary determinant of its recurrence.
The demonstrated relationship between drinking and contingent isolation is a reversible
one. Drinking is suppressed during the contingency but returns to high levels when the
time-out contingency is discontinued. Recognition of such reversible relationships has
been of great utility to the physical and biological sciences. Reversible relationships (e.g.
the freezing of water) indicate that current circumstances are controlled by the current
values of relevant parameters. Irreversible relationships (e.g. the burning of wood) indicate
that current circumstances are controlled by the past values of relevant parameters. Both
kinds of relationships are important in the control of behavior, just as they are in the control of other physical events. However, there has been a tendency within the mental health
field to overlook behavioral control via reversible relationships, in preference to attribut-
G. BIGELOW,I. LIEB.WNand R. GRIFFITHS
114
ing causation
to less accessible
historical
and intrapsychic
factors.
sible procedures should receive increased attention.
The present demonstration
of the effectiveness of the contingent
dicates that the causes of alcoholics’
vironmental
conditions
rather
drinking
than
solely
are located
within
should serve as an impetus to further exploration
fication
of alcoholic
re-programming
represent
Such therapy
of environmental
the development
tenance
BEELOW
drinking.
individual.
of contingency
chronic
via rever-
procedure
This
in-
within en-
demonstration
management
management
on a relatively
technology
time-out
at least significantly
via contingency
conditions
of a behavioral
the
Intervention
in the modi-
will involve the
basis,
and should
for both achievement
and main-
of sobriety.
G., COHEN M., LIEBSQN I.
REFERENCES
and FAILLACEL. (1972) Abstinence or moderation’? Choice by alcoholics.
Behav. Res. & Therapy 10, 209-214.
BIGELOWG. and LIEBSONI. (1972) Cost factors controlling alcoholic drinking. Psychol. Rec. 22,305-314.
BOSTOWD. and BAILYJ. (1969) Modification of severe disruptive and aggressive behavior using brief timeout
and reinforcement procedures. J. appl. Behav. Anal. 2, 31-37.
BURCHARDJ. D. and BARRERAF. (1972) An analysis of timeout and response cost in a programmed environment. J. trppl. Bchar. Anal. 5, 271-282.
BL’K(.HAKI)J. and TYLI.R V. (1965) The modification of delinquent behavior through operant conditioning.
Behuc. Rex & Therapy 2, 241-250.
COHEN
M., BIGELOWG.. HARGETTA., ALLENR. and HALS~EDC. (1973) The use of contingency management
procedures for the treatment of alcoholism in a work setting. Int. J. Addic. Alcohol. Drugs. (In press).
COHEN M., LII’I~SON
I. and FAILLA~FL. (197la) The role of reinforcement contingencies in chronic alcoholism:
An experimental analysis of one case. Behav. Res. & Therapy. 9, 375-379.
COHEN M.. LIEBSON
I.. FAILLACEL. and ALLENR. (197lb) Moderate drinking by chronic alcoholics. J. nerc. ~l~(‘nt.
Dis. 153. 434 444.
COHEN M., LII:~~CIN
I., FAILLAC~L. and SPEERSW. (1971~) Controlled drinking and incentives for abstinence.
Psychol. Rep. 28, 575580.
FRANKSC. (1966) Conditioning and conditioned aversive therapies in the treatment of the alcoholic. In?. J. Addic.
I, 61-98.
GOTTHEIL
E., CORBETTL. 0.. GRASBERGER
J. C. and CORNELI~~N
F. S. (1972) Fixed interval drinking decisions.
I. A research and treatment model. Q. J. Stud. Alcohol 33, 31 l-324.
GOTTHEIL E., CRAWFORD H. and CORNELISON F. (1973) The alcoholic’s ability to resist available alcohol. Dis.
nerv. Syst. 34, 80-84.
HUNT G.
and AZRINN. (1973) A community reinforcement approach to alcoholism. Behav. Res. & Therapy 11,
91-104.
KELLERM. (1972) On the loss-of-control phenomenon in alcoholism. Br. J. Addict. 67, 153166.
LEITENBERG
H. (1965) Is time-out from positive reinforcement an aversive event‘? A review of the experimental
evidence. Psychol. Bull. 64,428~441.
MELLON. K. (1968) Some aspects of the behavioral pharmacology of alcohol. In EFROND. H. (Ed.) Psychopharmacology. A Review of Progress 1957-1967. PHS Publication no. 1836, U.S. Government Printing Office,
Washington, D.C.
MELLON. K. and MENDEL~~NJ. H. (1970) Experimentally induced intoxication in alcoholics: A comparison
between programmed and spontaneous drinking. J. Pharmac. exp. Ther. 173, 101-I 16.
MILLSK. C., SOBELLM. B. and SCHAEFER
H. H. (1971) Training social drinking as an alternative to abstinence
for alcoholics. Behav. Therapy 2, l&27.
NATHANP. E. and O’BRIENJ. S. (1971) An experimental analysis of the behavior of alcoholics and nonalcoholics
during prolonged experimental drinking: A necessary precursor of behavior therapy? Behav. Therapy 2,455
476.
O’LEARYK. and O’LEARYS. (1972) Classroom Management. Pergamon Press. New York.
PICKENSR., BIGELOWG. and GRIFFITHSR. (1973) An experimental approach to treating chronic alcoholism:
A case study and one-year follow-up. Behau. Res. & Therapy 11, 321-325.
SCHAEFERH. H., SOBELLM. B. and MILLS K. C. (1971) Baseline drinking behaviors in alcoholics and social
drinkers: Kinds of drinks and sip magnitude. Behav. Res. & Therapy 9, 23-37.
SHIPLEYC. R., BARONA. and KAUFMANA. (1972) Effects of timeout from one reinforcer on human behavior
maintained by another reinforcer. Psycho/. Rec. 22, 201-210.
Alcoholic drinking: suppression by a brief time-out procedure
SULZER
115
E. S. (1966) Behavior modification in adult psychiatric patients. In ULLMANL. and KRASNERL. (Eds.)
Cascj Sruclirs in Behavior Modification. HoIt, Rinehart of Winston, pp. 196200. New York.
THAKP R. and WETZEL R. (1969) Behavior Modification in the Natural Environment. Academic Press, New York.
ZEILBERGER J., SAMPENS. and SLOANH. (1968) Modification of a child’s problem behaviors in the home with
the mother as therapist. .I. appl. Behau. Anal. 1,47-53.