Guy H. Montgomery et al.HYPNOTICANALGESIA
A META-ANALYSIS OF
HYPNOTICALLY INDUCED ANALGESIA:
How Effective is Hypnosis?
GUY H. MONTGOMERY, KATHERINE N. DUHAMEL,
AND WILLIAM H. REDD
Mount Sinai School of Medicine, New York, New York
Abstract: Over the past 2 decades, hypnoanalgesia has been widely
studied, however, no systematic attempts have been made to deter
mine the average size of hypnoanalgesic effects or establish the gener
alizability of these effects from the laboratory to the clinic. This study
examines the effectiveness of hypnosis in pain management, compares
studies that evaluated hypnotic pain reduction in healthy volunteers
vs. those using patient samples, compares hypnoanalgesic effects and
participants’ hypnotic suggestibility, and determines the effectiveness
of hypnoticsuggestionforpain relief relativeto other nonhypnotic psy
chological interventions. Meta-analysis of 18 studies revealed a moder-
ate to large hypnoanalgesic effect, supporting the efficacy of hypnotic
techniques for pain management. The results also indicated that hyp-
notic suggestion was equally effective in reducing both clinical and
experimental pain. The overall results suggest broader application of
hypnoanalgesic techniques with pain patients.
Hypnosis has been used as a psychological technique for treatment of
a broad range of disorders and illnesses. For example, it has been used in
treatingphobias, depression, anorexia nervosa, dissociative identity dis-
order, psychotic disorders, posttraumatic stress disorders, obesity,
smoking, and somatization disorders (see Rhue, Lynn, & Kirsch, 1993).
Even given the breadth of treatment contexts, hypnosis is perhaps best
known as a pain management technique. Indeed, this popularity may be
due to longstanding reports of pain relief with hypnosis during aversive
medical procedures (e.g., Esdaile, 1957). Hypnosis has been demon
strated to relieve pain in patients with headache (Spinhoven, Linssen,
Van Dyck, & Zitman, 1992; ter Kuile et al., 1994; Van Dyck, Zitman, Lins
sen, & Spinhoven, 1991), burn injury (Patterson, Everett, Burns, &
Marvin, 1992; Patterson & Ptacek, 1997; Wakeman& Kaplan, 1978), heart
Manuscript submitted September 5, 1999; final revision received September 7, 1999.
Address correspondence to Guy H. Montgomery, Ph.D., Assistant Professor, Cancer
Prevention and Control Program, Derald H. Ruttenberg Cancer Center, Mount Sinai
School of Medicine, Box 1130, One Gustave L. Levy Place, New York, NY 10029-6574 or
The International Journal of Clinical and Experimental Hypnosis, Vol. 48, No. 2, April 2000 134-149
© 2000 The International Journal of Clinical and Experimental Hypnosis
disease (Weinstein & Au, 1991), arthritis (Haanen et al., 1991; Horton &
Mitzdorf, 1994), cancer (Katz, Kellerman, & Ellenberg, 1987; D. Spiegel &
Bloom, 1983; Syrjala, Cummings, & Donaldson, 1992; Wall & Womack,
1989), dental problems(Stam, McGrath, & Brooke, 1984), eczema (Hajek,
Radil, & Jakoubek, 1991), and chronic back problems (Melzack & Perry,
1975; Spinhoven & Linssen, 1989). Although there are a number of stud
ies that examine the effects of hypnotic suggestion for clinical pain man
agement (e.g., Barabasz & Barabasz, 1989), and there have been several
review articles attempting to summarize these studies (e.g., Wadden &
Anderton, 1982; and National Institutes of Health technology assess
ment panel on integration of behavioral and relaxation approaches into
the treatment of chronic pain and insomnia, 1996), to our knowledge
there have been no attempts to rigorously quantify the analgesic impact
of hypnotic suggestion across studies.
That is, it is unknown what proportion of people demonstrate pain
relieffollowing hypnotic interventions. This lack of quantification is sur
prising, given the managed health care industry’s emphasis on treat-
ment efficacy and the general public’s rising interest in, and use of,
unconventional therapies for the treatment of physical problems (Eisen-
berg et al., 1993). The purpose of this study is: (a) to determine the per-
centage of people who benefit from hypnoanalgesic suggestions, (b) to
explore whether results based on empirical work in the laboratory gen-
eralize to medical settings and whether the effects of hypnosis are lim-
ited to a particular segment of the population (e.g., highly suggestible
individuals), and (c) to explore whether the effects of hypnoanalgesic
suggestions are less effective, equally effective, or more effective than
other psychological interventions (e.g., relaxation training) in providing
Meta-analysis (Hunter & Schmidt, 1990; Smith, Glass, & Miller, 1980)
provides an established methodology for evaluating the effectiveness of
hypnosis for pain relief across study samples in the published literature.
In summary, meta-analysis is the organization and integration of previ
ously published data through calculation of effect sizes. An effect size is
a standardized estimate of the magnitude of a study effect that permits
directcomparisons of effects acrossstudies. In addition, statistical analy
ses of effect sizes can provide a better understanding of cumulative
research findings for an area, or subject, of research interest than tradi
tional narrative review articles (Hunter & Schmidt, 1990). Although pre
vious meta-analyses have supported the use of hypnosis as an effective
adjunct to both cognitive-behavioral (Kirsch, Montgomery, & Sapirstein,
1995) and psychodynamic psychotherapies (Smith et al., 1980) for a vari
ety of psychological disorders, the specific effects of hypnotic interven
tions for pain relief have not been addressed with this statistical
HYPNOTIC ANALGESIA 135
The primary aim of this study is to estimate the proportion of people
who benefit from hypnotic suggestions for pain relief. To achieve this
goal, we will use meta-analytic statistical techniques to estimate effect
sizes for studies that compared hypnoanalgesic suggestion to no-
treatment or standard-treatment conditions. Secondary aims of this
study are to explore potential differences in the magnitude of pain relief
due to participant sample characteristics (i.e., clinical vs. experimental
pain, level of suggestibility) and to compare the treatment efficacy of
hypnotic interventions with that of nonhypnotic psychological inter
ventions in providing pain relief.
At this point in the development of the hypnosis literature, it is not
clear whether patients enduring clinical pain demonstrate similar levels
of hypnotically suggested pain relief as healthy volunteers experiencing
experimental pain. It has been argued that clinical patients should dem
onstrate greater hypnotic pain reduction than healthy volunteers based
on data showing that hypnosis has greater impact on pain suffering than
on pain sensation (see Holroyd, 1996). Because pain patients are likely to
suffer to a much greater degree than healthy volunteers undergoing
experimental pain procedures, pain patients may enjoy greater benefit
fromhypnotic interventions relativeto experimental volunteers. We will
use meta-analytic techniques to explore the possibility that pain patients
demonstrate greater hypnoanalgesic effects.
A second participant characteristic potentially linked to hypnoanal-
gesic effects is the participant’s level of suggestibility. Previous reports
have indicated that hypnoanalgesic effects should be larger for more
suggestible participants (Barabasz & Barabasz, 1989; Hilgard & Morgan,
1975; Miller & Bowers, 1986; H. Spiegel & Spiegel, 1978). Although this
position is widely accepted (Holroyd, 1996), it is not certain that indi
viduals high in suggestibility will demonstrate greater hypnotically
induced pain reduction than those of moderate suggestibility, or
whether those of moderate suggestibility will outperform participants
of low suggestibility. Given that most people score in the middle range
on tests of hypnotic suggestibility (Hilgard, Weitzenhoffer, Landes, &
Moore, 1961), it is important to estimate the benefits of hypnosis (i.e.,
effect size) for people of moderate hypnotic suggestibility, because they
make up the largest segment of the population.
The American Psychological Association, through the Division of
Clinical Psychology, has recommended that the effectiveness of psycho
logical interventions should be emphasized to the public (Task force on
promotion and dissemination of psychological procedures, 1995).
Indeed, criteria for empirically validated psychological treatments have
been proposed for “well-established treatments,” and “probably effica
cious treatments” (Chambless et al., 1998; Task Force on Promotion and
Dissemination of Psychological Procedures, 1995). If hypnosis is
revealed as an effective method for reducing pain, it is then important to
136 Guy H. Montgomery et al.
estimate the effects of hypnosis relative to other frequently used psycho
logical interventions for pain management. In the present study, meta-
analyses will be used to compare the effectiveness of hypnosis to that of
nonhypnotic psychological pain management strategies on the basis of
their ability to provide pain relief. The finding that hypnotic suggestions
for pain relief are at least as effective as other cognitive-behavioral inter
ventions would empirically validate the use of hypnosis for pain man
agement and support its use in situations where hypnosis is more effi
cient or specifically requested by patients suffering pain.
As the primary focus of this paper is to statistically estimate the effec
tiveness of hypnotic procedures for pain relief, we limited this first sam
ple of studies to those that directly compared hypnotherapy to no-
treatment, wait list, or standard-treatment control groups on measures
of pain. For the comparison of hypnosis to other psychological pain
management strategies, we changed the criteria and included studies
that contained both hypnotic and nonhypnotic psychological treatment
groups but not necessarily control groups as described above.
Studies included in the present sample were identified from previous
reviews of this literature (Chaves & Dworkin, 1997; Ellis & Spanos, 1994;
Genuis, 1995; Holroyd, 1996; Spinhoven, 1988). We also conducted a
computer search of the PsycLIT database from 1974 to 1997, using the
search terms hypnosis and pain, hypnosis and analgesia, hypnotherapy and
pain, and hypnotherapy and analgesia. The computer search algorithm was
set to accept plurals (i.e., hypnotherapies, pains, and word variants (i.e.,
hypnotically, hypnotize, hypnotizability, painful, analgesic). Initial inclusion
criteria were as follows: (a) a hypnotic intervention intended to reduce
pain was administered to at least one group of subjects; (b) the inclusion
of a no-treatment or standard-treatment control group; and (c) sufficient
data (including some form of pain measurement on a continuous scale)
were reported to allow calculation of effect sizes (Hunter & Schmidt,
1990; Smith et al., 1980). For the comparison of hypnosis to cognitive-
behavioral treatments, the inclusion of a cognitive-behavioral treatment
was substituted for the inclusion of a no-treatment or standard-
treatment control group (point b above).
In several papers with clinical pain samples, control groups were
treated according to standard clinical practices (“treatment as usual”)
(Lang, Joyce, Spiegel, Hamilton, & Lee, 1996; Patterson et al., 1992; Syr
jala et al., 1992; Wakeman & Kaplan, 1978; Weinstein & Au, 1991). In
these experimental studies, hypnosis was used as an adjunct to the pre
scription of a pain reliever. For example, in the study by Lang and col
leagues (1996), radiology patients received analgesic medications
throughout the study. However, the amounts and types of medications
were controlled for within the study design (i.e., use of medication was a
HYPNOTIC ANALGESIA 137
dependent variable). Medication use was similarly controlled for in
studies of burn patients (Patterson et al., 1992; Wakeman & Kaplan,
1978). In each case, study procedures were identical for control groups
and experimental groups with the exception that experimental group
procedures included theaddition of hypnotic suggestions for pain relief.
Following these standardized methods, 41 effect sizes were initially
calculated from 18 papers
. Effect sizes were based on the pain reports of
933 participants, nearly all of who were randomly assigned to control or
hypnotic intervention conditions. Exceptions to strict random assign
ment were as follows: participants were assigned to treatment condi
tions sequentially following referral while counterbalancing for order of
group assignment between study therapists (Edelson & Fitzpatrick,
1989); participants were alternately placed in one of three treatment
groups (Elton, Boggie-Cavallo, & Stanley, 1988); and group membership
was determined by odd or even medical chart numbers (Weinstein &
Au, 1991). In order to protect against the possibility that studies with
greater numbers of dependent variables (pain measures) would have
undue influence on the final estimate of hypnotically induced pain
reduction (Hunter & Schmidt, 1990), effect sizes were averaged for treat-
ments within studies. For example, if a single study contained two hyp-
notic treatments and four pain measures, then the four pain measures
were averaged for each treatment condition, and the study was repre-
sented by these two effect sizes in subsequent analyses. In the present
sample of studies, only three studies (Baker & Kirsch, 1993; Girodo &
Wood, 1979; Spanos & Katsanis, 1989) contained morethan a single treat-
ment, and therefore the risks to statistical independence due to multiple
treatments seemed small relative to the benefit of a larger sample of
effect sizes for evaluation of participant characteristics. Three additional
studies are represented by two (Spanos, Perlini, & Roberston, 1989;
Wakeman & Kaplan, 1978), or even three (ter Kuile et al., 1994), effect
sizes. However, in these cases, the multiple effect sizes are due to the
findings of two separate experiments having been reported within a sin
gle paper (Spanos et al., 1989)
and results having been independently
described for separate and diverse sample groups (e.g., community
members, patients, students) within papers (ter Kuile et al., 1994;
138 Guy H. Montgomery et al.
It should be noted that, when possible, we based the calculation of effect sizes on the
standard deviation within the control groups. We took this approach because of the poten
tial for inflated experimental group standard deviations due to interactions between the
hypnotic interventions and participant levels of hypnotic suggestibility. Specifically, if
level of suggestibility interacts with hypnoanalgesic effects (a study question), then one
would anticipate greater variance within experimental groups. Study data cautiously ar
gues that such an interaction exists.
2. In the calculation of the effect size for Experiment 1 in the paper by Spanos, Perlini,
and Robertson (1989), results were collapsed across orders of administration for the com
parison of the hypnotherapy and control groups as the authors reported no significant ef
fects due to order (p. 287).
Wakeman & Kaplan, 1978). Therefore, the primary analysis of the effec
tiveness of hypnosis for pain reduction was performed on 27 effect sizes
derived from18studies. Secondary analysis of differencesbetween clini
cal and student study samples were based on a comparison of 10 and 17
effect sizes, respectively; nine additional effect sizes were recalculated to
compareeffect sizes according to reported levels of hypnotic suggestibil
ity; and 27 additional effect sizes were calculated to compare hypnotic to
nonhypnotic psychological pain management strategies.
Types of pain, sample sizes, and effect sizes are presented in Table 1.
Two outliers from the Wakeman and Kaplan (1978) study were immedi
ately apparent (d = 15.45, d = 17.42). In order to provide a more accurate
estimate of the effects of hypnotherapy for pain reduction and control for
the excessive influence of these extraordinary outliers (Abelson, 1995)
on the overall mean effect size, we winsorized the data with g = 2 (Winer,
1971). All subsequent analyses are based on the winsorized data set
unless otherwise noted. In addition, no differences were found between
studies that included standard-treatment control groups versus true
no-treatment control groups (p > .10).
Results revealed a moderate to large effect size, d = .74, (Cohen, 1992).
To control for possible bias in effect size due to study sample sizes and to
take a conservative approach, we weighted effect sizes by the size of the
samples from which they were derived and calculated the mean
weighted effect size (D) according to procedures described by Hunter
and Schmidt (1990). The calculation of variation of D was also weighted
for study sample sizes. That is, the deviation of individual study effects
sizes from the weighted mean effect size was adjusted for sample size
The results revealed a significant effect of hypnotic interventions in the
treatmentof pain (D = .67, VarD = .26; p < .01), indicating that the average
participant treated with hypnosis demonstrated greater analgesic
response than 75% of participants in standard and no-treatment control
groups. Recalculation of the overall treatment effect with the outliers
removed (d = .62, D = .59, VarD = .25) did not significantly differ from the
winsorized results (F = 0.67, p > .10). Therefore, the winsorized results
were used for all remaining analyses.
As seen in Table 2, participants experiencing both clinical and experi
mental pain significantly benefitted from hypnotic interventions (p <
.01), as indicated by the moderate effect sizes. Interestingly, the two pain
groups did not statistically differ from each other in the degree to which
they experienced pain relief (p > .10), suggesting that hypnotic sugges
tions for analgesia work equally as well in laboratory and medical
HYPNOTIC ANALGESIA 139
Variation of D = Σni (di–D)2 / Σni where n is the study sample size, d is the raw effect
size, and D is the mean weighted (by study sample size) effect size.
In a between-groups analysis of variance, hypnotic analgesic effects
were found to differ according to participants’ levels of suggestibility, F =
5.26, p < .02 (see Table 3). For the purposes of this analysis, study partici
pants not identified according to levels of suggestibility were designated
as being in the mid-range of suggestibility, because most participants in
these studies would fall into this category if tested and the numbers of
highs and lows should be relatively even. Post hoc Tukey tests revealed
that individuals in the high hypnotic suggestibility range have greater
pain relief following hypnotic interventions than those in the low range,
p < .05. No other level of suggestibility between group comparisons were
statistically significant (i.e., participants in the midrange did not differ
from those in the high or low range of hypnotic suggestibility, p > .05).
However, these between-group findings must be viewed with some
140 Guy H. Montgomery et al.
Study Characteristics and Mean Effect Sizes by Publication Date
Study Participants Type of Pain nd
Wakeman et al. (1978) patients burn 24 15.45
Wakeman et al. (1978) patients burn 18 17.42
Girodo & Wood (1979) students cold pressor 20 –0.01
Girodo & Wood (1979) students cold pressor 20 1.50
Girodo & Wood (1979) students cold pressor 20 –0.45
Stam et al. (1980) students cold pressor 20 0.12
Spanos et al. (1984) students cold pressor 45 0.47
Spanos et al. (1985) students cold pressor 42 0.94
Tripp & Marks (1986) students cold pressor 28 0.88
Stam et al. (1987) students ischemic 45 0.12
Elton et al. (1988) students ischemic 137 0.70
Edelson et al. (1989) patients chronic pain 16 0.94
Spanos et al. (1989) students focal pressure 96 0.29
Spanos et al. (1989) students focal pressure 60 0.37
Zeltzer et al. (1989) students cold pressor 37 1.19
Spanos & Katsanis (1989) students focal pressure 20 1.35
Spanos & Katsanis (1989) students focal pressure 20 1.81
Spanos et al. (1990) students focal pressure 30 1.67
Weinstein & Au (1991) patients coronary 32 0.65
Syrjala et al. (1992) patients cancer 22 0.10
Patterson et al. (1992) patients burn 20 0.72
Baker et al. (1993) students cold pressor 20 0.10
Baker et al. (1993) students cold pressor 20 0.41
ter Kuile et al. (1994) community headache 29 –0.43
ter Kuile et al. (1994) patients headache 36 0.34
ter Kuile et al. (1994) students headache 26 0.62
Lang et al. (1996) patients radiological 30 1.01
Note. Ischemic, cold pressor, and focal pressure are experimental pain stimuli.
caution due to the small sample sizes (high, n = 7 ; low, n = 4) and assign-
ment of the majority or participants to the mid-range of suggestibility.
Types of nonhypnotic psychological treatment, sample sizes, and
effect sizes for comparisons of hypnosis to nonhypnotic psychological
interventions are presented in Table 4. Results of the comparison of hyp-
notically suggested analgesia with other nonhypnotic psychological
pain management strategies found no differences in effectiveness
between these treatment strategies, d = .12, D =.11,VarD = .62; p > .10.
The results indicate that the average participant treated with hypnosis
demonstrated equivalent analgesic responses to participants in alternate
psychological treatment groups.
The results of the present study demonstrate that hypnotic sugges
tion is an effective analgesic based on analyses of 27 effect sizes and more
than 900 participants. For 75% of the population, hypnosis provided
substantial pain relief. The magnitude of the hypnoanalgesic effect did
not differ for clinical and healthy volunteer samples, however, hypnoan
algesic effects seem to differ according to levels of hypnotic suggestibil
ity, especially when people highest in suggestibility are compared to
those lowest in suggestibility. The limitations of our sample size pre
clude stronger conclusions on the impact of hypnotic suggestibility on
analgesic effect, yet it should be noted that our mid-range group (made
HYPNOTIC ANALGESIA 141
Population Effect Sizes as a Function of Type of Pain
Type of Pain ndDVariation of D
Clinical Pain 10 0.80 0.74 0.33
Experimental Pain 17 0.70 0.64 0.24
Note. Effect sizes based on types of pain are significantly greater than zero (p < .01) but did
not differ between clinical and experimental pain (p > .10).
Population effect sizes as a function of level of hypnotic suggestibility.
Level of Suggestibility ndDVariation of D
High 7 1.22 1.16 0.21
Medium 24 0.64 0.64 0.24
Low 4 0.10 –0.01 0.19
Note. Effect sizes for high and low levels of suggestibility significantly differ (p < .05).
up of the largest number of effect sizes) demonstrated an effect size that
did not statistically differ from the high hypnotic suggestibility group
effect size. This finding suggests that a majority of the population (i.e.,
excluding people scoring in the low hypnotic suggestibility range)
should benefit to a large extent from hypnotically suggested analgesia.
Indeed, hypnotically suggested analgesia is at least as effective as non
hypnotic psychological interventions for pain management (e.g.,
cognitive-behavioral) and should be considered as a potential pain man
agement strategy when discussing treatment modality options with
142 Guy H. Montgomery et al.
Study Characteristics and Mean Effect Sizes by Publication Date for
Comparison of Hypnotic to Nonhypnotic Psychological Treatments
Study Type of Nonhypnotic Treatment nd
Girodo & Wood (1979) rationale and self-statement 20 .23
Girodo & Wood (1979) self-statement 20 1.45
Girodo & Wood (1979) task motivational instructions 20 1.58
Girodo & Wood (1979) rationale and self-statement 20 –.83
Girodo & Wood (1979) self-statement 20 .44
Girodo & Wood (1979) task motivational instructions 20 .53
Girodo & Wood (1979) rationale and self-statement 20 –1.27
Girodo & Wood (1979) self-statement 20 0.00
Girodo & Wood (1979) task motivational instructions 20 .08
Stam & Spanos (1980) analgesia suggestion 30 .04
Spanos et al. (1984) instruction 60 –.19
Spanos et al. (1985) “do whatever you want” 42 –.37
Spanos et al. (1985) stress innoculation 42 –.16
Tripp and Marks (1986) relaxation and analgesia suggestions 28 –.28
Tripp and Marks (1986) analgesia suggestion alone 28 .53
Katz et al. (1987) play 36 –.02
Edelson & Fitzpatrick cognitive behavioral training 18 .16
Spanos & Katsanis (1989) nonhypnotic (suggested) analgesia 30 .50
Harmon et al. (1990) relaxation and breathing exercises 60 1.98
Spanos et al. (1990) analgesia instruction 66 –.45
Patterson et al. (1992) attention and information 20 .46
Spinhoven et al. (1992) autogenic training 46 –.19
Syrjala et al. (1992) therapist contact 23 .38
Syrjala et al. (1992) cognitive behavioral training 23 .28
ter Kuile et al. (1994) autogenic training 29 –1.56
ter Kuile et al. (1994) autogenic training 29 –.45
ter Kuile et al. (1994) autogenic training 22 .37
Hypnotically induced pain reduction has been criticized on the basis
that pain reports are subjective and patients may produce reports of
reducedpain due to social demand characteristics. Potentially, the above
effect sizes may represent participants’ desire to please experimenters
rather than “true” estimates of pain relief.Nevertheless, this possibility
seems unlikely for two reasons. First, hypnotic analgesia has beendem
onstrated with behavioral measures of pain (e.g., Lang et al., 1996;
Wakeman & Kaplan, 1978; Weinstein & Au, 1991). Both Lang and col
leagues (1996) and Wakeman and Kaplan (1978) used changes in pain
medication use as a dependent variable. These effect sizes were not sig
nificantly less than the remaining effect sizes in clinical pain studies.
Although patients may report less pain to please an experimenter, it
seems less likely that patients would actively short their morphine dose
and endure increased suffering to achieve this goal. Second, a report
published in 1995 supports the existence of physiological correlates of
hypnoanalgesic effects (Kiernan, Dane, Phillips, & Price, 1995). Confir
mation of physiological changes associated with hypnoanalgesic effects
would argue against a simple compliance explanation of the present
findings. Due to the Kiernan and colleagues’ results, it seems more likely
that hypnotic suggestion is effective in eliciting pain relief. Therefore,
our understanding of hypnoanalgesic effects may benefit most by
exploring mechanisms consistent with modern theories of hypnosis,
such as sociocognitive (Kirsch, 1991; Spanos, 1991) or neodissociation
theories (Hilgard, 1991), rather than attributing hypnotic effects entirely
to compliance factors.
It should be noted that few current practitioners of hypnosis view it as
a stand-alone therapy. Rather, hypnotic suggestion is more commonly
used adjunctly to psychodynamic, cognitive-behavioral, or even phar
macological therapies. Recent data strongly support the advantages of
adding hypnosis to cognitive-behavioral psychotherapy in general
(Kirsch et al., 1995), and the present results strongly suggest the efficacy
of the addition of hypnosis to nonhypnotic pain management strategies.
In clinical practice, therefore, the addition of hypnotic suggestions for
pain relief to standard protocols appears prudent as the majority of
patients are likely to benefit. There is no evidence that harm would be
done when working with appropriately trained health professionals,
and effective hypnoanalgesic interventions can be rather brief and cost-
effective (e.g., one session; Patterson et al., 1992).
Whether hypnotically suggested analgesia is considered a primary
treatment modality or an adjunctive treatment, the question of its
empirical validation remains. The present results support the view that
hypnotically suggested analgesia meets the criteria for a “well-
established treatment” (superior to pill or psychological placebo or to
another treatment) (Chambless et al., 1998). Randomized studies with
HYPNOTIC ANALGESIA 143
cancer patients (Syrjala et al., 1992) and burn patients (Patterson et al.,
1992) demonstrated that hypnotic suggestion was more effective in
reducing pain than nonhypnotic psychological interventions. In addi
tion, Edelson and Fitzpatrick (1989) demonstrated that hypnotic sugges
tion was as effective as a previously “well-established treatment” for
chronic pain (cognitive-behavioral therapy) (Chambless et al., 1998), but
patients were assigned sequentially rather than randomly to treatment
groupsin that study. One could argue that the relatively small samples in
some of the studies comparing hypnotic suggestion to other treatments
is a limitation of this area of research (published Task Force criteria have
previously recommend 30 participants per treatment group) (Task
Force,1995). However, the total sample number of more than 800 partici
pants across studies on which the present comparisons are based would
appear to mitigate this criticism.
In addition to superiority to another treatment, Chambless et al.
(1998) also recommend that the characteristics of the client samples be
clearly specified, that the experiments should be conducted with treat-
ment manuals, and that the effects be demonstrated by at least two dif-
ferent investigating teams. Although the study samples were well
defined and the studies were conducted by two separate research teams
(i.e., Patterson et al., 1992; Syrjala et al., 1992), they were not based on
treatment manuals per se. Nevertheless, Chambless and Hollon state
that relatively simple treatment interventions are an exception to the
manual requirement. Because these brief hypnosis interventions mainly
consist of an induction and suggestions for pain relief, the treatment
approaches are readily available in the published literature (e.g., Rhue,
Lynn, & Kirsch, 1993), and the study procedures aredescribed within the
research papers, it would seem that hypnotically suggested analgesia
should still be considered a well-established treatment.
In summary, the present study is the first, to our knowledge, to quan
tify the magnitude of hypnoanalgesic effects. Hypnotic suggestion
relieves pain for the majority of people, regardless of the type of pain
they are experiencing. Conclusions regarding the influence of individ
ual differences (i.e., hypnotic suggestibility) on outcome should be
drawn with some caution due to the relative paucity of controlled stud
ies examining the contributions of individual variables to treatment
effects, but the results of the present meta-analysis argue for a broader
application of hypnotic suggestion for better pain management, espe
cially as an adjunct to standard treatments in clinical settings. A review
of the literature supports the view that hypnotically suggested pain
reductioncan be classified as a “well-established treatment,” and official
designation will encourage the application of hypnotic suggestion to a
wider range of patients with pain.
144 Guy H. Montgomery et al.
(References marked with an asterisk indicate studies included in meta-analyses.)
Abelson, R. P. (1995). Statistics as principled argument. Hillsdale, NJ: Erlbaum.
*Baker, S. L., & Kirsch, I. (1993). Hypnotic and placebo analgesia: Order effects and the pla
cebo label. Contemporary Hypnosis, 10, 117-126.
Barabasz, A. F., & Barabasz, M. (1989). Effects of restricted environmental stimulation:
Enhancement of hypnotizability for experimental and chronic pain control. Interna
tional Journal of Clinical and Experimental Hypnosis, 37, 217-231.
Chambless, D. L., Baker, M. J., Baucom, D. H., Beutler, L.E., Calhoun, K. S.,Crits-Christoph,
P., Daiuto, A., DeRubeis, R., Detweiler, J., Haaga, D. A. F., Johnson, S. B., McCurry, S.,
Mueser, K. T., Pope, K. S., Sanderson, W. C., Shoham, V., Stickle, T., Williams, D. A., &
Woody, S. R. (1998). Update on empirically validated therapies II. Clinical Psychologist,
Chambless, D. L. & Hollon, S. D. (1998). Defining empirically supported therapies. Journal
of Consulting and Clinical Psychology, 66, 7-18.
Chaves, J. F., & Dworkin, S. F. (1997). Hypnotic control of pain: Historical perspectives and
futureprospects.International Journal of Clinical and Experimental Hypnosis, 45, 356-376.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.
*Edelson, J., & Fitzpatrick, J. L. (1989). Acomparison of cognitive-behavioral and hypnotic
treatments of chronic pain. Journal of Clinical Psychology, 45, 316-323.
Eisenberg, D. M., Kessler, R. C., Foster, C., Norlock, F. E., Calkins, D. R., & Delbanco, T. L.
(1993). Unconventional medicine in the United States: Prevalence, costs, and patterns
of use (see comments). New England Journal of Medicine, 328, 246-252.
Ellis, J. A., & Spanos, N. P. (1994). Cognitive-behavioral interventions for children’sdistress
during bone marrow aspirations and lumbar punctures: A critical review. Journal of
Pain and Symptom Management, 9, 96-108.
*Elton, D., Boggie-Cavallo, P., & Stanley, G. V. (1988). Group hypnosis and instructions of
personal control in the reduction of ischaemic pain. Australian Journal of Clinical and
Experimental Hypnosis, 16, 31-37.
Esdaile, J. (1957). Hypnosis in medicine and surgery. New York: Julian.
Genuis, M. L. (1995). The use of hypnosis in helping cancer patients control anxiety, pain,
and emesis: A review of recent empirical studies. American Journal of Clinical Hypnosis,
*Girodo, M., & Wood, D. (1979). Talking yourself out of pain: The importance of believing
that you can. Cognitive Therapy and Research, 3, 23-33.
Haanen, H. C. M., Hoenderdos, H. T. W., van Romunde, L. K. J., Hop, W. C. J., Mallee, C.,
Terwiel, J. P., & Hekster, G. B. (1991). Controlled trial of hypnotherapy in the treatment
of refractory fibromyalgia. Journal of Rheumatology, 18, 72-75.
Hajek, P., Radil, T., & Jakoubek, B. (1991). Hypnotic skin analgesy in healthy individuals
and patients with atopic eczema. Homeostasis in Health and Disease, 33, 156-157.
*Harmon, T. M., Hynan, M. T., & Tyre, T. E. (1990). Improvedobstetric outcomes using hyp
notic analgesia and skill mastery combined with childbirth education. Journal of Con
sulting and Clinical Psychology, 58, 525-530.
Hilgard, E. R. (1991). A neodissociation interpretation of hypnosis. In S. J. Lynn & J. W.
Rhue (Eds.), Theories of hypnosis: Current models and perspectives (pp. 83-104). New York:
Hilgard,E. R., & Morgan, A. H. (1975). Heart rate and blood pressurein the study of labora
tory pain in man under normal conditions and as influenced by hypnosis. Acta Neuro
biologiae Experimentalis, 35, 741-759.
Hilgard, E. R., Weitzenhoffer, A. M., Landes, J., & Moore, R. K. (1961). The distribution of
susceptibility to hypnosis in a student population: a study using the Stanford Hypnotic
Susceptibility Scale. Psychological Monographs, 75, 1-22.
HYPNOTIC ANALGESIA 145
Holroyd, J. (1996). Hypnosis treatment of clinical pain: Understanding why hypnosis is
useful. International Journal of Clinical and Experimental Hypnosis, 44, 33-51.
Horton, J. R., & Mitzdorf, U. (1994). Clinical hypnosis in the treatment of rheumatoid
arthritis. Psychologische Beitraege, 36, 205-212.
Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis: Correcting error and bias in
research findings. Newbury Park, CA: Sage Publications.
*Katz, E. R., Kellerman, J., & Ellenberg, L. (1987). Hypnosis in the reduction of acute pain
and distress in children with cancer. Journal of Pediatric Psychology, 12, 379-394.
Kiernan, B. D., Dane, J. R., Phillips, L. H., & Price, D. D. (1995). Hypnotic analgesia reduces
R-III nociceptive reflex: further evidence concerning the multifactorial nature of hyp
notic analgesia. Pain, 60, 39-47.
Kirsch, I. (1991). The social learning theory of hypnosis. In S. J. Lynn & J. W. Rhue (Eds.),
Theories of hypnosis: Current models and perspectives (pp. 439-466). New York: Guilford.
Kirsch, I., Montgomery, G., & Sapirstein, G. (1995). Hypnosis as an adjunct to cognitive-
behavioral psychotherapy: A meta-analysis. Journal of Consulting and Clinical Psychol
ogy, 63, 214-220.
*Lang, E. V., Joyce, J. S., Spiegel, D., Hamilton, D., & Lee, K. K. (1996). Self-hypnotic relaxa
tion during interventional radiological procedures: Effects on pain perception and
intravenous drug use. International Journal of Clinical and Experimental Hypnosis, 44,
Melzack, R., & Perry, C. (1975). Self-regulation of pain: The use of alpha-feedback and hyp
notic training for the control of chronic pain. Experimental Neurology, 46, 452-469.
Miller, M. E., & Bowers,K. S. (1986). Hypnotic analgesia and stress inoculation in the reduc-
tion of pain. Journal of Abnormal Psychology, 95, 6-14.
*Patterson, D. R., Everett, J. J., Burns, G. L., & Marvin, J. A. (1992). Hypnosis for the treat-
ment of burn pain. Journal of Consulting and Clinical Psychology, 60, 713-717.
Patterson, D. R., & Ptacek, J. T. (1997). Baseline pain as a moderator of hypnotic analgesia
for burn injury treatment. Journal of Consulting and Clinical Psychology, 65, 60-67.
Rhue, J. W., Lynn, S. J., & Kirsch, I. (Eds.) (1993). Handbook of Clinical Hypnosis. Washington,
DC: American Psychological Association.
Smith, M. L., Glass, G. V., & Miller, T. I. (1980). The benefits of psychotherapy. Baltimore: Johns
Hopkins University Press.
Spanos, N. P. (1991). A sociocognitive approach to hypnosis. In S. J. Lynn & J. W. Rhue
(Eds.), Theories of hypnosis: Current models and perspectives (pp. 324-361). New York:
*Spanos, N. P., Kennedy, S. K., & Gwynn, M. I. (1984). Moderating effects of contextual vari
ables on the relationship between hypnotic susceptibility and suggested analgesia.
Journal of Abnormal Psychology, 93, 285-294.
*Spanos, N. P., Ollerhead, V. G., & Gwynn, M. I. (1985). The effects of three instructional
treatments on pain magnitude and pain tolerance: Implications for theories of hypnotic
analgesia. Imagination, Cognition and Personality, 5, 321-337.
*Spanos, N. P., & Katsanis, J. (1989). Effects of instructional set on attributions of nonvoli
tion during hypnotic and nonhypnotic analgesia. Journal of Personality and Social Psy
chology, 56, 182-188.
*Spanos, N. P., Perlini, A. H., & Roberston, L. A. (1989). Hypnosis, suggestion, and placebo
in the reduction of experimental pain. Journal of Abnormal Psychology, 98, 285-293.
*Spanos, N. P., Perlini, A. H., Patrick, L., Bell, S., & Gwynn, M. I. (1990). The role of compli
ance in hypnotic and nonhypnotic analgesia. Journal of Research in Personality, 24,
Spiegel, D., & Bloom, J. R. (1983). Group therapy and hypnosis reduce metastatic breast
carcinoma pain. Psychosomatic Medicine, 45, 333-339.
Spiegel, H., & Spiegel, D. (1978). Trance and treatment: Clinical uses of hypnosis. Washington,
DC: American Psychiatric Press.
146 Guy H. Montgomery et al.
Spinhoven, P. (1988). Similarities and dissimilarities in hypnotic and nonhypnotic proce
duresfor headache control: Areview. American Journal of Clinical Hypnosis, 30, 183-194.
Spinhoven, P., & Linssen, A. C. G. (1989). Education and self-hypnosis in the management
of low back pain: a component analysis. British Journal of Clinical Psychology,28, 145-153.
*Spinhoven, P., Linssen, A. C. G., Van Dyck, R., & Zitman, F. G. (1992). Autogenic training
and self-hypnosis in the control of tension headache. General Hospital Psychiatry, 14,
*Stam, H. J., & Spanos, N. P. (1980). Experimental designs, expectancy effect, and hypnotic
analgesia. Journal of Abnormal Psychology, 89, 751-762.
Stam, H. J., McGrath, P. A., & Brooke, R. I. (1984). The effectsof a cognitive-behavioral treat
ment program on temporo-mandibular pain and dysfunction syndrome. Psychosomatic
Medicine, 46, 534-545.
*Stam, H. J., & Spanos, N. P. (1987). Hypnotic analgesia, placebo analgesia, and ischemic
pain: The effects of contextual variables. Journal of Abnormal Psychology, 96, 313-320.
*Syrjala, K. L., Cummings, C., & Donaldson, G. W. (1992). Hypnosis or cognitive behav
ioral training for the reduction of pain and nausea during cancer treatment: A con
trolled clinical trial. Pain, 48, 137-146.
Task force on promotion and dissemination of psychological procedures. (1995). Training
in and dissemination of empirically validated psychological treatments: Reports and
recommendations. Clinical Psychologist, 48, 3-23.
*ter Kuile, M. M., Spinhoven, P., Linssen, A. C. G., Zitman, F. G., Van Dyck, R., & Rooijmans,
H. G. M. (1994). Autogenic training and cognitive self-hypnosis for the treatment of
recurrent headaches in three different subject groups. Pain, 58, 331-340.
*Tripp, E. G., & Marks, D. (1986). Hypnosis, relaxation and analgesia, suggestions for the
reduction of reported pain in high and low-suggestible subjects. Australian Journal of
Clinical and Experimental Hypnosis, 14, 99-113.
Van Dyck, R., Zitman, F. G., Linssen, A. C. G., & Spinhoven, P. (1991). Autogenic training
and future oriented hypnotic imagery in the treatment of tension headache: outcome
and process. International Journal of Clinical and Experimental Hypnosis, 39, 6-23.
Wadden, T. A., & Anderton, C. H. (1982). The clinical use of hypnosis. Psychological Bulletin,
*Wakeman, R. J., & Kaplan, J. Z. (1978). An experimental study of hypnosis in painful
burns. American Journal of Clinical Hypnosis, 21, 3-12.
Wall, V. J., & Womack, W. (1989). Hypnotic versus active cognitive strategies for alleviation
of procedural distress in pediatric oncology patients. American Journal of Clinical Hypno
sis, 31, 181-191.
*Weinstein, E. J., & Au, P. K. (1991). Use of hypnosis before and during angioplasty. Ameri
can Journal of Clinical Hypnosis, 34, 29-37.
Winer, B. J. (1971). Statistical Principles in Experimental Design (2nd ed.). New York: McGraw
*Zeltzer, L. K., Fanurik, D., & LeBaron, S. (1989). The cold pressor pain paradigm in chil
dren: Feasibility of an intervention model (part II). Pain, 37, 305-313.
Eine Meta-Analyse von hypnotischer
Analgesie: Wie effektiv ist Hypnose?
Guy H. Montgomery, Katherine N. DuHamel, und William H. Redd
Zusammenfassung: In den letzten zwei Jahrzehnten wurde hypnotische
Analgesie vielfach untersucht. Es wurden jedoch keine systematischen Ver
suche unternommen, das Ausmaß der hypno-analgetischen Wirkung zu
bestimmen oder die Verallgemeinerbarkeit von Ergebnissen in Relation zu
Patienten-Untergruppen zu setzen. Die vorliegende Studie untersucht
HYPNOTIC ANALGESIA 147
folgendes: die Wirksamkeit von Hypnose bei Schmerzbewältigung; ver
gleicht Untersuchungen zur suggerierter Schmerzminderung bei gesunden
Vpn. mit solchen, die Patientenstichproben verwendeten; vergleicht die
hypno-analgetische Wirkung mit der hypnotischen Suggestibilität der Pro
banden; und bestimmt die Wirksamkeit der suggerierten Schmerzminderung
in Relation zu anderen, nicht-hypnotischen psychologischen Interventionen.
Eine Meta-Analyse von 18 Untersuchungen ermittelte eine mäßige bis starke
hypno-analgetische Wirkung, d. h. die Wirksamkeit von Hypnotherapie bei
Schmerzbewältigung wird bestätigt. Die Ergebnisse deuten weiterhin an, daß
die hypnotische Suggestion bei Minderung von klinischem sowie experi
mentell suggeriertem Schmerz gleichermaßen effektiv war. Allgemein deu
ten die Ergebnisse auf eine breitere Anwendung von hypnotischer Analgesie
University of Tennessee, Knoxville, TN, USA
Une meta-analyse de l’analgésie induite par hypnose:
comment l’hypnose est-elle efficace?
Guy H. Montgomery, Katherine N. DuHamel, et William H. Redd
Résumé: Depuis plus de 20 ans, l’hypnoanalgésie a été largement étudiée.
Cependant aucune tentative systématique n’a été effectuée pour déterminer la
dimension des effets analgésiques ou n’a établi la généralisation des décou-
vertes de laboratoire au cas des patients. Cette étude examine l’efficacité de
l’hypnose sur le traitement de la douleur, compare des études qui ont étudié la
diminution de la douleur en hypnose chez des volontaires en bonne santé ver-
sus des études utilisant des cas personnels de patients, elle compare aussi les
effets hypnoanalgésiques et la suggestibilité hypnotique des participants.
Elle détermine l’efficacité de la suggestion hypnotique pour diminuer la doul
eur face à toute intervention psychologique non hypnotique. la méta analyse
de 18 cas a montré un effet hypnoanalgésique modéré à important prouvant
l’efficacité de l’intervention hypnotique dans le traitement de la douleur. Les
résultats ont montré également que la suggestion hypnotique était également
efficace dans la réduction de la douleur clinique ou expérimentale. La totalité
des résultats suggre une application plus large des techniques hypnoan
algésique avec les patients douloureux.
Psychosomatic Medicine & Clinical
Hypnosis Institute, Lille, France
Un meta-análisis de la analgesia inducida hipnóticamente.
Qué tan eficaz es la hipnosis?
Guy H. Montgomery, Katherine N. DuHamel, y William H. Redd
Resumen: Ha habido un amplio estudio de la hipnoanalgesia en las dos últi
mas décadas. Sin embargo, no se han realizado estudios sistemáticos para
148 Guy H. Montgomery et al.
dterminar el tamaño de los efectos hipnoanalgésicos ni se ha establecido la
generalizabilidad de lo encontrado en el laboratorio a las muestas de pacien
tes. Nuestro estudio examina la efectividad de la hipnosis para controlar el
dolor, compara estudios que evaluaron la reducción hipnótica del dolor en
voluntarios saludables vs. estudios con muestras de pacientes, compara los
efectos hipnoanalgésicos con la sugestibilidad hipnótica de los participantes,
y determina la efectividad de la sugestión hipnótica para aliviar el dolor en
comparación con otras intervenciones psicológicas pero no hipnóticas. Un
meta-análisis de 18 estudios mostró un efecto hipnoanalgésico de moderado a
grande, lo que sustenta la eficacia de las técnicas hipnóticas para el control del
dolor. Los resultados también indican que la sugestión hipnótica fue igual
mente eficaz para reducir tanto el dolor clínico como el experimental. En gen
eral estos resultados sugieren una aplicación más amplia de técnicas hipnoan
algésicas para tratar acientes con dolor.
Uniformed Services University of the Health
Sciences-Bethesda, MD, USA and the
University of Valencia, Spain
HYPNOTIC ANALGESIA 149