Health Psychology © 2009 American Psychological Association

2009, Vol. 28, No. 5, 579 –587 0278-6133/09/$12.00 DOI: 10.1037/a0016124

Effects of Supportive-Expressive Group Therapy on Pain in Women

With Metastatic Breast Cancer

Lisa D. Butler, Cheryl Koopman, Eric Neri, Janine Giese-Davis, Oxana Palesh, Krista A. Thorne-Yocam,
Sue Dimiceli, Xin-Hua Chen, Patricia Fobair, Helena C. Kraemer, and David Spiegel
Stanford University School of Medicine

Objective: To examine whether a group intervention including hypnosis can reduce cancer pain and trait
hypnotizability would moderate these effects. Design: This randomized clinical trial examined the effects
of group therapy with hypnosis (supportive-expressive group therapy) plus education compared to an
education-only control condition on pain over 12 months among 124 women with metastatic breast
cancer. Main Outcome Measures: Pain and suffering, frequency of pain, and degree of constant pain
were assessed at baseline and 4-month intervals. Those in the treatment group also reported on their
experiences using the hypnosis exercises. Results: Intention-to-treat analyses indicated that the inter-
vention resulted in significantly less increase in the intensity of pain and suffering over time, compared
to the education-only group, but had no significant effects on the frequency of pain episodes or amount
of constant pain, and there was no interaction of the intervention with hypnotizability. Within the
intervention group, highly hypnotizable participants, compared to those less hypnotizable, reported
greater benefits from hypnosis, employed self-hypnosis more often outside of group, and used it to
manage other symptoms in addition to pain. Conclusion: These results augment the growing literature
supporting the use of hypnosis as an adjunctive treatment for medical patients experiencing pain.

Keywords: cancer pain, randomized trial, hypnosis, group therapy, hypnotizability

Pain is a common symptom of cancer that can be mostly
attributed to tumor involvement and treatment side effects (Foley,
1975). Studies have found, for example, that about one third of
lung and colon cancer respondents had experienced “persistent or
frequent” cancer-related pain in the previous 2 weeks (Portenoy et
al., 1992) and over two thirds of an advanced cancer patient
sample reported pain, with most reporting pain as 5 or higher on a
10-point scale (Yun et al., 2003). Pain can also be episodic; it can
flare up even when baseline pain is relatively well controlled.
Almost two thirds of an international sample of 1,095 cancer
patients reported breakthrough pain (Caraceni et al., 2004). The
level of pain has been associated with more advanced disease,
which makes pain “one of the most feared consequences of can-
cer” (Foley, 1996, p. 164). It is understood by patients as a sign of
deterioration in their condition (D. Spiegel & Bloom, 1983b), or as
a signal of impending death (Ferrell & Dean, 1995) or a recurrence
in their cancer (Coward & Wilkie, 2000).
The occurrence of pain in cancer may contribute to further
somatic consequences. Pain is one of the strongest predictors of
fatigue in breast cancer patients (Haghighat, Akbari, Holakouei,
Rahimi, & Montazeri, 2003), and the experience of fatigue and
pain together have been found to be the most distressing symptoms
to patients (e.g., in lung cancer; Cooley, Short, & Moriarty, 2003).
Sleep problems are also associated with greater pain (Koopman et
al., 2002), even after treatment ends (Dow, Ferrell, Leigh, Ly, &
Gulasekaram, 1996).
Pain also has psychosocial effects. Cancer patients with high
pain (vs. low pain) show much higher rates of depression, sug-

Lisa D. Butler, Cheryl Koopman, Eric Neri, Janine Giese-Davis, Oxana
Palesh, Krista A. Thorne-Yocam, Sue Dimiceli, Xin-Hua Chen, Patricia
Fobair, Helena C. Kraemer, and David Spiegel, Department of Psychiatry
& Behavioral Sciences, Stanford University School of Medicine.
Lisa D. Butler is now at the School of Social Work, University at
Buffalo, The State University of New York. Janine Giese-Davis is now at
the University of Calgary Department of Oncology, Division of Psycho-
social Oncology, Tom Baker Cancer Centre, Division of Psychosocial
Resources. Oxana Palesh is now at the Department of Psychiatry and
Radiation Oncology, University of Rochester Medical Center, James P.
Wilmot Cancer Center.
This study was made possible by National Institute for Mental Health
Grant 5R01MH047226 with additional funding from the National Cancer
Institute, the American Cancer Society, the John D. and Catherine T.
MacArthur Foundation, and the Fetzer Institute. Previous analyses of these
data were presented at the annual meeting of the Society for Clinical and
Experimental Hypnosis, San Antonio, TX in November 2001. We ac-
knowledge the contributions of Elaine Miller, Jane Benson, Karin Calde,
Leslie Kinder, Elizabeth T. Klinger, Lynne LoPresto, Julie Seplaki, Thai
Nguyen, Trina Kurek, Rita Halbach, Christopher Biggs, Catherine Classen,
Susan Diamond Moore, Susan Weisberg, Frank Stockdale, Amanda
Kovattana, Catherine Byers, Bob Carlson, Bernadette Paolini, referring
community physicians, the Stanford Oncology Day Care Center, and the
patients and their spouses and families who participated in the study. We
also thank Jackie Tilston for her much appreciated contributions to the
introduction and Jay Azarow for his feedback on a portion of this paper.
ClinicalTrials.gov Identifier: NCT00226928; http://www.clinicaltrials
.gov (registered 09/12/05).
Correspondence concerning this article should be addressed to Lisa D.
Butler, who is now at the Buffalo Center for Social Research, University at
Buffalo, School of Social Work, 229 Parker Hall, Buffalo, NY 14214-
8004. E-mail: ldbutler@buffalo.edu

579
580 BUTLER ET AL.
gesting that pain may have a causal role in such depression (e.g.,
D. Spiegel, Sands, & Koopman, 1994). One study reported that
50% of the variance in pain was accounted for by mood distur-
bance, patient’s belief that the pain indicated a worsening of the
illness, and the use of analgesic medication (D. Spiegel & Bloom,
1983b). “A cycle is established whereby the pain is a reminder of
the presence of the illness and of the helplessness of the patient to
master it, and this in turn only reinforces the pain experienced”
(D. Spiegel & Bloom, 1983b, p. 341). This cycle can also lead to
reduced activity levels (Ahles, Blanchard, & Ruckdeschel, 1983),
anxiety (Dalton & Feuerstein, 1989), and fear and anger (Reynolds
& Olsen, 1996). In addition, pain and distress can increase signif-
icantly prior to death in metastatic breast cancer (MBC) patients,
although pain alone does not account for the increase in distress
(Butler et al., 2003).
Poor adaptation to pain has led to the development of psycho-
social therapy protocols to teach coping skills and reduce the
emotional impact of pain. Given the evidence pertaining to the
health benefits of social support (e.g., Levy et al., 1990; Turner-
Cobb, Sephton, Koopman, Blake-Mortimer, & Spiegel, 2000;
Weis, 2003), support groups are now widely employed for chronic
illness sufferers. A number of randomized trials have found pos-
itive benefits for a range of group interventions for cancer patients
(e.g., Antoni et al., 2006; Kissane et al., 2003; Lane & Viney,
2005; Levine, Eckhardt, & Targ, 2005), although some studies
have found little or limited psychological effects (e.g., Edelman,
Bell, & Kidman, 1999; Edmonds, Lockwood, & Cunningham,
1999). Supportive-expressive group therapy (SET)—the interven-
tion examined in the present study—also has demonstrated psy-
chosocial effects, including reductions in mood disturbance, de-
pression, trauma symptoms, emotional control, and maladaptive
coping (Classen et al., 2001; Giese-Davis et al., 2002; Goodwin et
al., 2001; Kissane et al., 2007; D. Spiegel, Bloom, & Yalom, 1981;
reviewed in Boutin, 2007), among others. SET also teaches self-
hypnosis for pain management and has been effective in reducing
pain among cancer patients (Goodwin et al., 2001; D. Spiegel &
Bloom, 1983a). A comparison of SET with and without hypnosis
being taught (D. Spiegel & Bloom, 1983a), so that the effect of that
element could be tested, found that the treatment group with
hypnosis reported less increase in pain sensation experience at
follow-up when compared to the therapy group without hypnosis.
One of the oldest uses of hypnosis is for medical analgesia
(Esdaile, 1846), and current research shows that it is a highly
effective adjunct to treatment for pain and related anxiety (Lang et
al., 2000; Patterson & Jensen, 2003; D. Spiegel & Bloom, 1983a)
across a range of medical patients. Within the context of the
conceptualization of pain as a combination of sensory and affec-
tive aspects, hypnosis for pain aims to help the cancer patient filter
out the emotional or “hurt” (D. Spiegel & Bloom, 1983a, p. 335)
component of the experience by reinterpreting the pain as a dif-
ferent sensation (H. Spiegel & Spiegel, 2004). This effect may also
be seen as a function of focused attention relegating pain signals to
the periphery. There is also growing evidence that hypnosis alters
the perception of pain, reducing the brain response to nociceptive
stimuli, not merely the reaction to it (De Pascalis, 1999; Kosslyn,
Thompson, Costantini-Ferrando, Alpert, & Spiegel, 2000; Rain-
ville, Hofbauer, Bushnell, Duncan, & Price, 2002; D. Spiegel,
Bierre, & Rootenberg, 1989).
A number of reviews looking at nonpharmacologic treatments
concluded that there is much support for using hypnosis for
cancer-related pain management (Mundy, DuHamel, & Montgom-
ery, 2003; Sellick & Zaza, 1998; D. Spiegel & Classen, 2000; H.
Spiegel & Spiegel, 2004; Trijsburg, van Knippenber, & Rijpma,
1992). Moreover, one recent study (Montgomery et al., 2007)
found that a brief presurgery hypnosis session reduced pain inten-
sity and unpleasantness, nausea, fatigue, discomfort, and emo-
tional upset in breast cancer patients in the postsurgical period,
highlighting a range of possible target symptoms for hypnosis in
the psychosocial care of cancer patients.
Previous research also indicated that there is typically a corre-
lation between degree of hypnotizability and extent of pain relief
(DeBenedittis, Panerai, & Villamira, 1989; Hilgard & Hilgard,
1994) suggesting that hypnotic interventions for pain may be most
effective for those higher in hypnotic responsiveness, though scant
research has examined this question in metastatic cancer patients.
Consequently, we sought to examine whether highly hypnotizable
women in the treatment condition would report significantly lower
increases in pain over time than would less hypnotizable women in
the same condition or women randomized to the control group. In
addition, we expected that highly hypnotizable women would find
hypnosis to be more beneficial than their less hypnotizable coun-
terparts in the same group.
The aim of this study was to examine the effect of supportive-
expressive group therapy employing hypnosis on pain experience
(degree of pain and suffering, pain frequency, constant pain) in
MBC patients, and to explore the interaction between hypnotiz-
ability and pain experience in women trained to use hypnosis. We
hypothesized that: (a) women randomized to receive treatment
would report a smaller increase in pain over the first 12 months in
the study than women randomized to the control condition; (b)
those in the treatment group who were higher in hypnotizability
would report the smallest increase in pain over the 12 months,
when compared to those lower in hypnotizability in that group or
those not trained in hypnosis (i.e., the control group); and (c) the
high hypnotizable women in the treatment group would report
more benefit from the hypnosis exercises than their less hypnotiz-
able counterparts.
Method
Participants
One hundred and twenty-five women with confirmed metastatic
or locally recurrent breast cancer were randomized into the parent
study between 1991 and 1996. All women but 3 had MBC; the
remaining 3 women had breast cancer recurrences after breast
conserving therapy as their only site of recurrent/metastatic disease
but were judged by our expert oncological consultant as having
equivalent prognoses. For ease of presentation, we refer to all
participants in the present report as having MBC. Women were
recruited through the Oncology Day Care Center at Stanford
University Medical Center, letters to community oncologists, bro-
chures distributed in the community, and notices in local newspa-
pers and breast cancer newsletters (for details on recruitment, see
Classen et al., 2001). A total of 155 women initially entered the
study; 30 of these women were not randomized: 12 due to disease
progression that precluded participation; 7 were found to be inel-
 GROUP THERAPY REDUCES CANCER PAIN 581

igible after medical record review; and 11 decided that they did not
want to continue in the study.
Data for all participants are included in the present analyses
except for 1 participant in the treatment group at the Stanford site
who refused to have her hypnotizability assessed at baseline.
Intention-to-treat analyses on this sample of 124 were conducted
using slopes computed over the first 12 months of treatment.
Slopes were calculated for 94 participants for whom we had at
least one usable follow-up assessment. A zero slope was assumed
for the remaining 30 participants. Regarding these 30 participants,
20 of them did not complete any postbaseline assessments (13
were too ill to complete questionnaires or had died [4 treatment
and 9 control]; 1 was too busy [control]; 4 withdrew from the study
because they were not assigned to a support group; 1 woman
withdrew because she did not like the support group; and 1 woman
assigned to the treatment condition withdrew for no stated reason).
With respect to the other 10 women (6 treatment and 4 control),
each had completed only one postbaseline assessment, but it was
proximal to death. In these cases the final assessment point was
removed from analyses and, consequently, the slope was assumed
to be zero. This step was taken because we have previously
identified a “spike” in pain (and distress) that occurs proximal to
death independent of receiving group therapy (Butler et al., 2003)
and that can obscure treatment effects examined by slopes (e.g.,
Classen et al., 2001). Seventeen other women (9 treatment and 8
control) who had postbaseline assessment data in addition to an
assessment just prior to death were included in the analyses,
though the assessment point proximal to death was excluded in the
creation of their slopes. Demographic and medical status charac-
teristics are summarized in Tables l and 2.
Inclusion and Exclusion Criteria
All participants gave written informed consent for participation
in a protocol approved by the Stanford University School of
Medicine Institutional Review Board. Women were eligible for the
study if they had documented metastatic or recurrent breast cancer,
were proficient enough in English to be able to complete ques-
tionnaires and participate in a support group, were living in the
Greater San Francisco Bay Area, and had a Karnofsky score of at
least 70 (Karnofsky & Burchenal, 1949). A patient with a Karnof-
sky score of 70 is able to care for herself, but unable to carry on
normal activity or do active work. We did not include women with
positive supraclavicular lymph nodes as the only metastatic lesion
at the time of initial diagnosis; active cancers within the past 10
years other than breast cancer, basal cell or squamous cell carci-
nomas of the skin, in situ cancer of the cervix (severe cervical
intraepithelial neoplasia or squamous intraepithelial lesion II); or
melanoma with a Breslow depth below 0.76 mm. We also ex-
cluded those with other concurrent medical conditions likely to
influence short-term survival.
Procedure
Baseline Assessments and Randomization
Baseline assessments were conducted at our three study sites
(Stanford, San Francisco, and San Jose, California) and included
measures of pain, distress, coping, social support, and physical
activity as well as evaluations of immune and endocrine function.
The present analyses focus on changes in four pain outcomes
(described below). On completion of baseline testing, participants

Table 1
Demographic Characteristics of the Sample Analyzed by Treatment Condition

Education onlya Group therapy plus educationb

Age, M ⫾ SD (range) 53.1 ⫾ 10.8 (30 to 80) 52.7 ⫾ 10.5 (33 to 73)
Education in years, M ⫾ SD (range) 15.9 ⫾ 2.4 (12 to 20) 16.2 ⫾ 2.7 (12 to 26)
Ethnicity, n (%)
Asian 7 (11.5) 1 (1.6)
African American/Black 1 (1.6) 0 (0)
Hispanic/Latina 2 (3.3) 1 (1.6)
American Indian 1 (1.6) 1 (1.6)
White 50 (82.0) 58 (92.1)
Other 0 (0.0) 2 (3.2)
Marital status, n (%)
Married 31 (50.8) 39 (61.9)
Never married 8 (13.1) 3 (4.8)
Separated or divorced 18 (29.5) 17 (27.0)
Widowed 3 (4.9) 4 (6.3)
Other 1 (1.6) 0 (0.0)
Household income, n (%)
⬍$20,000 10 (16.4) 7 (11.1)
$20,000 to 39,999 11 (18.0) 7 (11.1)
$40,000 to 59,999 16 (26.2) 16 (25.4)
$60,000 to 79,999 6 (9.8) 8 (12.7)
$80,000 to 99,999 5 (8.2) 10 (15.9)
⬎$100,000 13 (21.3) 14 (22.2)
Not reported 0 (0.0) 1 (1.6)

Note. Some percentages do not sum to 100% due to rounding. No differences between the conditions were statistically significant.
a
n ⫽ 61. b n ⫽ 63.
582 BUTLER ET AL.

Table 2
Summary of Medical Status Characteristics Analyzed by Treatment Condition

Education onlya Group therapy plus educationb

Age at initial diagnosis in years, M ⫾ SD (range) 48.0 ⫾ 10.3 (24.9 to 72.1) 47.8 ⫾ 10.2 (28.2 to 67.3)
Age at metastatic diagnosis in years, M ⫾ SD (range) 51.6 ⫾ 10.2 (30.5 to 73.5) 51.8 ⫾ 10.6 (30.8 to 73.2)
Disease-free interval in months, M ⫾ SD (range) 44.5 ⫾ 34.7 (0.0 to 146.1) 48.0 ⫾ 36.6 (0.0 to 162.3)
Time from metastatic diagnosis to study entry in months, M ⫾ SD (range) 28.7 ⫾ 48.2 (1.1 to 244.7) 21.9 ⫾ 27.3 (1.0 to 138.5)
Estrogen receptor negative, n (%) 12 (19.7) 13 (20.6)
Treatment for metastatic disease as of study entry, n (%)
Chemotherapy within the previous 2 months 29 (47.5) 26 (41.3)
Radiation within the previous 2 months 30 (49.2) 24 (38.1)
Hormonal therapy was used for metastasis 47 (77.0) 49 (77.8)
Dominant site of metastasis at study entry, n (%)
Chest wall 20 (32.8) 18 (28.6)
Bone 22 (36.1) 27 (42.9)
Viscera 19 (31.1) 18 (28.6)

Note. Some percentages do not sum to 100% due to rounding. No differences between the conditions were statistically significant.
a
n ⫽ 61. b n ⫽ 63.

were randomized to treatment or control conditions by the project
director, a research nurse, using adaptive randomization biased
coin-design method to ensure comparability of medical status in
treatment and control conditions (for details on the randomization
and stratification procedure, see Classen et al., 2001). Sixty-four
women were randomized to the intervention arm of the study, and
61 women to the control arm. Because this was a psychotherapy
trial, blinding of condition was not possible. Postbaseline assess-
ments were conducted every 4 months for the first year and every
6 months thereafter.
Intervention Condition
Three treatment groups were conducted, one at each geographic
site, and they met weekly for 90-min sessions. The size of the
groups varied over time due to participants dying and to rolling
recruitment, with the group size ranging from 3 to 15 participants
in any given group. The intended duration of treatment was 1 year,
and participants were encouraged to remain in the group for that
time. Most women continued participating for as long as their
health permitted.
The therapy sessions were co-led by two psychotherapists;
among these were a psychiatrist, psychologists, and social work-
ers. The supportive-expressive therapy model involved the cre-
ation of a supportive environment in which participants were
encouraged to confront their problems, strengthen their relation-
ships, and find enhanced meaning in their lives. The intervention
was semistructured, with therapists trained to enhance the creation
of social bonds, identify and encourage the expression of emotion,
and facilitate discussion of the themes as the material emerged and
in an emotionally expressive rather than didactic format. The
themes included: (a) fears of dying and death, including dealing
with the deaths of group members; (b) reordering life priorities; (c)
improving support from and communication with family and
friends; (d) integrating a changed self and body image; and (e)
improving communication with physicians (D. Spiegel & Classen,
2000). Through sharing their experiences, group members also
became role models for one another, teaching each other coping
strategies that they found to be effective in managing the illness.
Hypnosis training. Each session ended with a hypnosis exer-
cise to help patients with stress management and to deal with pain.
Patients were encouraged to use this exercise at home. Hypnosis is
a state of highly focused attention with a suspension of peripheral
awareness and heightened responsiveness to social cues (H. Spie-
gel & Spiegel, 2004). The hypnosis exercises involved a rapid
entry into the hypnotic state by counting from one to three while
looking up and slowly closing the eyes, and reducing muscle
tension by imagining floating in a “bath, lake, a hot tub, or just
floating in space.” Then participants were instructed to imagine
changing the sensation in the part of their body that hurt by
imagining that it was “warmer, cooler, tingling, or numb.” Various
images were utilized involving snow or ice, a hot pack, and so
forth. Then participants were told, “Concentrate on what this feels
like to you in a private sense, and then bring yourself out of the
state of hypnosis by counting backward from three to one. On
three, get ready; on two, roll up your eyes; and on one, let your
eyes open.” They were instructed to practice this exercise several
times a day, and any time the pain became noticeable or started to
increase.
Control Condition
To ensure full participation and cooperation, we offered a self-
directed education intervention to women randomized to the con-
trol condition, and the same materials were also offered to the
women randomized to treatment. Thus, all participants were given
a list of materials to select from or take home on loan, including a
selection of books, pamphlets, video- and audiotapes covering a
range of topics related to breast cancer (for details, see Classen et
al., 2001) and also a 1-year membership to a consumer health
library in their community.
Measures
Demographic Characteristics and Medical Status
Demographic characteristics and medical status information
were collected at baseline.
 GROUP THERAPY REDUCES CANCER PAIN
Pain Level
The Pain Rating Scale (D. Spiegel & Bloom, 1983a, 1983b) was
used to query self-reported pain levels. This measure was used to
assess pain in an earlier sample of MBC patients in our laboratory
(D. Spiegel & Bloom, 1983a, 1983b) and in previous analyses with
the present sample (Butler et al., 2003). The pain items include
single Likert-type items assessing “pain in this moment” using a
scale ranging from 1 (not noticeable) to 10 (excruciating—worst
ever) and “suffering in this moment” using a scale ranging from 1
(easily bearable) to 10 (agonizing— unbearable). The constancy
and frequency of pain were also assessed. “Constant pain” was
coded (as “1”) if the current episode of pain had lasted for at least
6 months and was experienced as constant or persistent during that
time (otherwise it was coded as “0”). “Pain frequency” was coded
as the number of days (1 to 7) that were affected by pain in a given
week of a typical episode since the last assessment (also with
values of 0 ⫽ no pain episode since the last assessment, and 7 ⫽
7 or more days affected in a typical episode). At baseline, 48% of
the sample reported some degree of current pain (M ⫽ 2.2, SD ⫽
1.7, range ⫽ 1 to 9). Preliminary analysis showed that “pain in this
moment” and “suffering in this moment” were highly correlated
(r ⫽ .71, 95% confidence interval [CI] ⫽ 0.65, 0.76); therefore the
average of these two scales was computed for use in the analyses.
Hypnotizability
The Hypnotic Induction Profile (HIP; H. Spiegel & Spiegel,
2004), a structured interview assessment of hypnotizability, was
conducted with each participant at baseline by a trained inter-
viewer. Hypnotizability refers to the robust finding of valid and
reliable individual differences in response to hypnotic suggestions
(Cardeña, Maldonado, van der Hart, & Spiegel, 2000). Possible
scores range from 0 to 10. For the present analyses, participants
were divided into low (scores of 0 to 5) and high (scores ⬎5)
hypnotizability groups. HIP scores were widely distributed with a
mean of 5.3 (SD ⫽ 2.8, range ⫽ 0 to 9.5). Forty-six women with
HIP scores at or below 5 were classified as low in hypnotizability
(27 treatment, 19 control); 78 women whose HIP scores were
greater than 5 were classified as high in hypnotizability (36 treat-
ment, 42 control).
Group Feedback
A Group Feedback Questionnaire was created for the present
study to assess group members’ reactions to and feelings about
participating in the group, the therapists, and using hypnosis. In
addition to questions regarding therapist characteristics, respon-
dents rated the extent to which they felt the hypnosis exercises had
been beneficial to them, using a single item, 10-point scale using
a scale ranging from 1 (not at all beneficial) to 10 (extremely
beneficial), how often they used the exercises outside of group (8
category checklist, ranging from “never” to “several times per
day”), and what they used hypnosis for (e.g., pain control, con-
trolling nausea and vomiting, stress management, feeling comfort-
able, considering problems, etc.).
Analyses
The size of the sample in the overall study was based on the
results of a power analysis related to the primary outcome (sur-
583
vival). Results indicated that there was better than 90% power to
detect effects of the treatment condition on this primary outcome.
Slopes analyses were employed to test our present hypotheses
(Gibbons, Hedeker, Waternaux, Kraemer, & Greenhouse, 1993).
Each participant with a prerandomization baseline measure had a
slope constructed across all available assessments regressed on
time using months as the unit of time. To conduct an intention-
to-treat analysis, a zero slope was assumed for each of the partic-
ipants who did not have at least one postbaseline assessment point.
As mentioned previously, data from the final assessment point
were excluded from the slope for a given participant if that
assessment was proximal to death (i.e., it was collected at the
assessment closest to her death and that assessment fell in the 4-
[or 6-] month follow-up assessment window prior to her death).
Previous research has demonstrated that there is a significant rise
in reported distress and pain prior to death in women with MBC
(Butler et al., 2003), which adds variance at the final assessment
that may obscure earlier treatment effects examined by slopes, as
it did in a previous report (Classen et al., 2001) and in preliminary
analyses with the present data. In slopes analysis the endpoints
have a greater influence on the slope relative to other assessment
points. Consequently, the effect of proximity to death on the slope
was removed in the primary analyses for testing the hypotheses.
Secondary analyses in which these endpoints were included in
computing the slopes are also presented.
These outcome slopes became the dependent variables in mod-
els that were analyzed by analysis of covariance (ANCOVA) in a
2 (treatment vs. control) ⫻ 3 (geographic sites) ⫻ 2 (low vs. high
hypnotizability level) design, testing also for the two-way interac-
tion of treatment condition with hypnotizability. Covariates in this
analysis included the overall model intercept and the respective
pain variable intercept that was computed along with the pain
variable slope. Because change in pain is typically associated with
initial levels, each ANCOVA included this pain variable intercept
as a covariate, rather using than the baseline value itself because
the intercept is a better estimate of the true baseline value. Separate
analyses were conducted on scores for pain and suffering in this
moment, pain frequency, and constant pain as dependent variables
based on the slopes computed for each participant’s assessments
over the first year.
All hypothesized treatment versus control relationships were
tested with two-tailed tests with an alpha level of .05. For all
outcomes, effect sizes were calculated using Cohen’s d (Cohen,
1977), which is based on the standardized mean difference com-
paring the group therapy condition to the education-only control
condition. A positive effect size indicates that the group therapy
condition had better results than the education-only control con-
dition, and a negative effect size indicates the reverse. Analyses of
demographic and medical status characteristics with independent
sample t tests (on continuous data) and chi-square tests (on cate-
gorical data) found no statistically significant differences between
participants assigned to the experimental and control treatment
conditions.
Results
Table 3 presents the descriptive statistics at baseline, slopes of
change associated with each condition, F test value and probabil-
ity, and the treatment effect size for each outcome.
584
Table 3
Baseline Pain Measure Scores, Slopes of Change, and Treatment Effect Sizes
Education only
n Baseline score M ⫾ SD
Current pain and suffering intensity 61 1.9 ⫾ 1.4
Frequency of pain 61 4.2 ⫾ 3.1
Constant pain 61 0.3 ⫾ 0.5
Note. F test values are presented for the tests of treatment effects in the analysis of covariance models.
Pain and Suffering in This Moment
The ANCOVA on changes in intensity of current pain and
suffering showed that the overall model was statistically signifi-
cant, F(6, 117) ⫽ 3.86, p ⫽ .001, accounting for 17% of the
variance. Women who received group therapy plus education showed
a significantly smaller increase in the intensity of current pain and
suffering compared to women who received the educational interven-
tion alone, F(1, 117) ⫽ 4.61, p ⫽ .034 (see Figure 1). Women who
reported greater intensity of current pain/suffering at baseline
tended to experience greater decreases in its intensity, F(1, 117) ⫽
14.44, p ⫽ .001, compared to women who reported less pain/
suffering at baseline, who tended to experience increases in its
intensity. No significant relationships were found for changes in
current pain/suffering intensity across the recruitment sites, F(2,
117) ⫽ 1.77, p ⫽ .175, or associated with hypnotizability, F(1,
117) ⫽ 0.39, p ⫽ .532; and hypnotizability did not show a
significant interaction with treatment, F(1, 117) ⫽ 0.07, p ⫽ .792.
When this analysis was repeated without removing the assess-
ments proximal to death from the slopes, the overall model ac-
counted for less of the variance (R2 ⫽ 0.10). The treatment effect
was no longer significant, F(1, 117) ⫽ 0.95, p ⫽ .331, but the
baseline level of pain/suffering continued to significantly predict
greater decreases in pain, F(1, 117) ⫽ 7.22, p ⫽ .008.
Control Slope
3.5
Treatment Slope
3.0
2.5
2.0
1.5
1.0
0 4 8 12
Time (Months)
Figure 1. Slopes and mean scores for pain and suffering over the first 12
months analyzed separately for education only (control) and group therapy
plus education (treatment) conditions.
BUTLER ET AL.
Group therapy plus education
Slope n Baseline score M ⫾ SD Slope F test value p d
0.54 63 2.0 ⫾ 1.5 ⫺0.002 4.61 .034 0.31
⫺0.14 63 4.2 ⫾ 3.2 0.32 0.12 .734 ⫺0.13
0.02 63 0.3 ⫾ 0.5 0.06 0.03 .863 ⫺0.07
Pain Frequency
In the analysis of changes in the frequency of pain, the overall
model was significant, F(6, 117) ⫽ 4.95, p ⫽ .001, and accounted
for 20% of the variance. Women who reported the most frequent
pain at baseline tended to show a decrease in their frequency of
pain, whereas women who reported the least frequent pain at
baseline tended to show an increase in the frequency of pain over
the course of the study, F(1, 117) ⫽ 26.83, p ⫽ .001. Changes in
pain frequency were not significantly related to treatment condi-
tion, F(1, 117) ⫽ 0.12, p ⫽ .734; site, F(2, 117) ⫽ 1.57, p ⫽ .213;
hypnotizability level; F(1, 117) ⫽ 0.01, p ⫽ .926; or the interac-
tion of condition with hypnotizability; F(1, 117) ⫽ 0.02, p ⫽ .889.
There were no differences in the significant findings when this
analysis was repeated with the death-proximal data points in-
cluded.
Constant Pain
The overall model was significantly related to the changes in
experiencing constant pain, F(6, 117) ⫽ 4.44, p ⫽ .001, account-
ing for 19% of the variance. Women who reported constant pain at
baseline showed the greatest reduction in constant pain, in contrast
to women who did not report constant pain at baseline, who tended
to show an increase in constant pain, F(1, 117) ⫽ 21.54, p ⫽ .001.
No significant effects on constant pain were found for treatment
condition, F(1, 117) ⫽ 0.03, p ⫽ .863; site, F(2, 117) ⫽ 2.55, p ⫽
.082; hypnotizability level, F(1, 117) ⫽ 0.60, p ⫽ .441; or the
interaction of condition with hypnotizability, F(1, 117) ⫽ 0.01,
p ⫽ .921. There were no differences in the significant findings
when this analysis was repeated with the death-proximal data
points included.
Uses and Perceived Benefit of Hypnosis Within the
Treatment Group
Table 4 presents the descriptive statistics and effect sizes for the
uses and perceived benefit of the self-hypnosis exercises analyzed
by low versus highly hypnotizable women within the treatment
group. Highly hypnotizable women reported greater use of hyp-
nosis exercises for controlling nausea and vomiting (Cohen’s d ⫽
.73), to manage mental stress/anxiety (Cohen’s d ⫽ .66), and to
feel comfortable (Cohen’s d ⫽ .51) than did low hypnotizable
women. Highly hypnotizable women also reported using the self-
hypnosis exercises more often (Cohen’s d ⫽ .55) and, as hypoth-
esized, they perceived significantly more benefit from them
(Cohen’s d ⫽ .62; p ⫽ .04) when compared to those lower in
hypnotizability.
 GROUP THERAPY REDUCES CANCER PAIN 585

Table 4
Exploratory Comparison of Uses and Perceived Benefit of Hypnosis (Mean Across Three Follow-Up Assessments) Analyzed by Low
Versus Highly Hypnotizable Women in Group Therapy Condition

Low hypnotizable Highly hypnotizable

n M SD n M SD d

Controlling nausea and vomiting 15 1.1 0.3 26 2.1 2.4 0.73

Pain management 17 3.5 3.1 26 3.6 3.0 0.46
Physical stress 17 3.1 2.9 26 3.8 3.1 0.24
Mental stress/anxiety 16 3.8 2.9 26 5.7 2.7 0.66
Considering problems 15 2.5 2.3 26 3.3 2.8 0.32
Discovering solutions to problems 15 2.5 2.3 26 3.3 2.8 0.03
Undergoing treatment 15 1.9 2.1 26 3.0 2.7 0.40
Imagining the cancer going away 16 3.3 3.2 26 3.7 3.3 0.14
Feeling comfortable 16 3.5 2.8 26 5.0 2.7 0.51
Dealing with strong emotions 16 2.7 2.5 26 3.3 2.4 0.26
Total number of times used self-hypnosis exercises 15 25.5 18.9 26 36.2 19.1 0.55
Perceived benefit derived from the self-hypnosis exercises 20 4.5 2.9 27 6.3 3.0 0.62

Discussion period. In that study, the induction included instructions to use

The results of this randomized trial indicate that supportive
expressive group therapy that includes hypnosis can significantly
reduce the experience of pain and suffering for women with MBC.
The primary analyses indicate that participants in the treatment
condition plus education compared to the education-only condition
reported significantly less increase in pain and suffering over the
first 12 months of treatment. The magnitude of this treatment
effect was in the small to medium range. Participants did not,
however, differ significantly in their reports of pain frequency or
constancy. These findings of a reduction in aspects of pain, as well
as the lack of an effect on pain frequency and constancy, are in line
with previous research (Goodwin et al., 2001; D. Spiegel &
Bloom, 1983a).
The present study also examined whether trait differences in
hypnotic susceptibility would affect pain experiences and other
outcomes. Contrary to one of our hypotheses, higher hypnotizable
participants in the treatment condition did not report a significantly
smaller increase in pain over time, when compared to those lower
in hypnotizability in the same group or those who did not receive
the intervention. However, among those who received the inter-
vention, highly hypnotizable participants reported significantly
greater benefit from the hypnosis exercises than did those who
were less hypnotizable, suggesting that greater hypnotizability
may facilitate other beneficial effects of the intervention. Explor-
atory analyses also indicated that these highly hypnotizable
women did not differ from the less hypnotizable women in the
frequency with which they used hypnosis for pain control, but they
did use self-hypnosis more often outside of the group setting
overall and applied the exercises to address other symptoms be-
sides pain, including anxiety/stress, discomfort, and nausea/
vomiting. These findings imply that teaching self-hypnosis to
cancer patients may provide them with a tool they can apply to
purposes beyond simply control of pain.
A recent report echoes this finding of multifaceted benefits of
hypnosis for cancer patients. Montgomery et al. (2007) found that
postsurgical patients who had undergone a brief presurgery hyp-
nosis session reported less pain intensity and unpleasantness, nau-
sea, fatigue, discomfort, and emotional upset in the postsurgical
hypnosis for many of these symptoms. In the present study, how-
ever, participants’ use of hypnosis for purposes beyond pain man-
agement was more self-initiated, yet it appeared to result in ben-
efits nonetheless, especially for those more hypnotizable.
Unfortunately, our study design does not allow us to disentangle
the effects of the hypnosis from those associated with being in
group therapy generally. Our findings indicate that these combined
aspects were associated with less increase in pain over time, and
these beneficial pain results were not limited to those higher in
hypnotic responsiveness. The lack of interaction between condi-
tion and hypnotizability is consistent with the suggestion that this
intervention may be useful for those of different levels of hypnotic
susceptibility. Although previous research has found a correlation
between degree of hypnotizability and extent of pain relief (e.g.,
Hilgard & Hilgard, 1994), some researchers emphasized the im-
portance of the patient’s motivation and other personality factors
in the prediction of the hypnotic response (Frischholz, Spiegel, &
Spiegel, 1981). A review by Holroyd (1996) noted several possible
explanations for how hypnosis can benefit low hypnotizables,
including that pain thresholds tend to increase with repeated hyp-
notic analgesia suggestions, hypnotic control of pain can improve
with practice, and low hypnotizables can increase their responsive-
ness to hypnotic suggestions with practice. Consequently, it is
possible that less hypnotizable participants improved in hypnotic
pain control with the training they received.
Such training and improvements may also lead to higher self-
efficacy and feelings of pain-management mastery, which are both
associated with reductions in subjective levels of pain (Bandura,
O’Leary, Taylor, Gauthier, & Gossard, 1987; Litt, 1988; Manning
& Wright, 1983). Moreover, D. Spiegel and Bloom (1983a) sug-
gested that hypnosis, as well as general participation in supportive-
expressive group therapy, can reduce distress and may thereby
have an indirect impact on pain, and treatment effects on distress
have been found in this sample (Classen et al., 2001). In addition,
those in the treatment group may have practiced hypnosis to
increase feelings of relaxation, and thereby lowered their pain
experience through reductions in muscle tension and stress
(De Pascalis, Magurano, & Bellusci, 1999; Gay, Philippot, &
586 BUTLER ET AL.
Luminet, 2002). Finally, participating in a group in which expe-
riences of pain are shared, along with pain management strategies
and encouragement, may have contributed to the present treatment
effect on pain. The relationship between responsiveness to hypno-
sis and pain experience in this population warrants further inves-
tigation. Indeed, studies are needed to clarify the benefits of
training/practice to hypnotic responsiveness and pain control, pos-
sible pathways to pain reduction that hypnosis and/or support
group participation may contribute to (e.g., changes in stress level
or perceived social support), as well as the role of individual
differences in hypnotic responsiveness in determining pain and
other outcomes, so that future clinical interventions that include
hypnosis for cancer pain (and perhaps other objectives) may be
better targeted.
There are a number of important limitations to this study. The
study’s generalizability may be limited to predominantly White,
middle to upper class women with MBC. It is not known whether
SET would be as equally effective in reducing pain and suffering
for people who come from a wider range of ethnic and socioeco-
nomic status backgrounds or who experience other types of cancer
or pain (Edwards, Fillingim, & Keefe, 2001). Moreover, the bur-
den of completing extensive assessments at 4-month intervals and
participating in weekly group sessions may have precluded the
participation of many potential participants who felt that they were
unable to meet these requirements due to the debilitating effects of
their illness.
Because we did not include a group that received the interven-
tion without hypnosis, we are unable to separate the positive
contributions of the SET intervention from those that may be due
to the hypnosis exercises in the sessions (and use of self-hypnosis
outside of the group). However, a previous study found that
members of SET groups that employed hypnosis reported signif-
icantly less pain than those receiving the same intervention but
without hypnosis (D. Spiegel & Bloom, 1983a).
There are also a number of methodological limitations of the
study including the use of self-reported measures of pain, each of
which was composed of a single item. This may, therefore, have
resulted in less reliable assessments of aspects of the pain experi-
ence and could thereby have limited our ability to detect changes
over time. In addition, the study may have been underpowered to
detect interactions, such as the one between hypnotizability and
treatment condition. Applying intention-to-treat analyses to data
from this population that had so many incomplete assessments
(due to illness and death) may also limit the ability to identify
positive effects for those able to participate throughout the inter-
vention.
In conclusion, the results of this study suggest that the experi-
ence of pain and suffering for patients with MBC can be success-
fully reduced with an intervention that includes hypnosis in a
group therapy setting. These results augment the growing literature
supporting the use of group therapy with hypnosis as an effective
adjunctive treatment for medical patients experiencing pain. Our
additional findings that participants higher in hypnotizability who
received the intervention, compared to those lower in hypnotiz-
ability in the same group, reported greater benefits from using
hypnosis, used it more often overall, and in some cases employed
it to address other symptoms, suggest that although hypnosis is not
at present standard practice for treating a wide range of symptoms
that trouble cancer patients, it is worth examining this potential.
Together these findings suggest that there may be a number of
benefits to the use of hypnosis in cancer care including, but not
necessarily limited to, its more traditional application for pain
control.
References
Ahles, T. A., Blanchard, E. B., & Ruckdeschel, J. C. (1983). Multidimen-
sional nature of cancer pain. Pain, 17, 277–288.
Antoni, M. H., Wimberly, S. R., Lechner, S. C., Kazi, A., Sifre, R. Urcuyo,
K. R., et al. (2006). Reduction of cancer-specific thought intrusions and
anxiety symptoms with a stress management intervention among women
undergoing treatment for breast cancer. American Journal of Psychiatry,
163, 1791–1797.
Bandura, A., O’Leary, A., Taylor, C. B., Gauthier, J., & Gossard, D.
(1987). Perceived self-efficacy and pain control: Opioid and nonopioid
mechanisms. Journal of Personality and Social Psychology, 53, 563–
571.
Boutin, D. L. (2007). Effectiveness of cognitive behavioral and supportive-
expressive group therapy for women diagnosed with breast cancer: A
review of the literature. The Journal for Specialists in Group Work, 32,
267–284.
Butler, L. D., Koopman, C., Cordova, M. J., Garlan, R. W., DiMiceli, S.,
& Spiegel, D. (2003). Psychological distress and pain significantly
increase prior to death in metastatic breast cancer patients. Psychoso-
matic Medicine, 65, 416 – 426.
Caraceni, A., Martini, C., Zecca, E., Portenoy, R. K., Ashby. M. A.,
Hawson, G., et al. (2004). Breakthrough pain characteristics and syn-
dromes in patients with cancer pain. An international survey. Palliative
Medicine, 19, 177–183.
Cardeña, E., Maldonado, J., van der Hart, O., & Spiegel, D. (2000).
Hypnosis. In E. Foa, T. Keane, & M. Friedman (Eds.), Effective treat-
ments for PTSD (pp. 407– 440). New York: Guilford.
Classen, C., Butler, L. D., Koopman, C., Miller, E., DiMiceli, S., Giese-
Davis, J., et al. (2001). Supportive-expressive group therapy and distress
in patients with metastatic breast cancer: A randomized clinical inter-
vention trial. Archives of General Psychiatry, 58, 494 –501.
Cohen, J. (1977). Statistical power analysis for the behavioral sciences.
Hillsdale, NJ: Erlbaum.
Cooley, M. E., Short, T. H., & Moriarty, H. J. (2003). Symptom preva-
lence, distress, and change over time in adults receiving treatment for
lung cancer. Psycho-Oncology, 12, 694 –708.
Coward, D. D., & Wilkie, D. J. (2000). Metastatic bone pain: Meanings
associated with self-report and self-management decision making. Can-
cer Nursing, 23, 101–108.
Dalton, J. A., & Feuerstein, M. (1989). Fear, alexithymia and cancer pain.
Pain, 38, 159 –170.
DeBenedittis, G., Panerai, A. A., & Villamira, M. A. (1989). Effects of
hypnotic analgesia and hypnotizability on experimental ischemic pain.
International Journal of Clinical & Experimental Hypnosis, 37, 55– 69.
De Pascalis, V., Magurano, M. R., & Bellusci, A. (1999). Pain perception,
somatosensory event-related potentials and skin conductance responses
to painful stimuli in high, mid, and low hypnotizable subjects: Effects of
differential pain reduction strategies. Pain, 83, 499 –508.
De Pascalis, V. (1999). Psychophysiological correlates of hypnosis and
hypnotic susceptibility. International Journal of Clinical & Experimen-
tal Hypnosis, 47, 117–143.
Dow, K. H., Ferrell, B. R., Leigh, S., Ly, J., & Gulasekaram, P. (1996). An
evaluation of the quality of life among long-term survivors of breast
cancer. Breast Cancer Research and Treatment, 39, 261–273.
Edelman, S., Bell, D. R., & Kidman, A. D. (1999). A group cognitive
behaviour therapy programme with metastatic breast cancer patients.
Psycho-Oncology, 8, 295–305.
Edmonds, C. V. I., Lockwood, G. A., & Cunningham, A. J. (1999).
 GROUP THERAPY REDUCES CANCER PAIN
Psychological response to long term group therapy: A randomized trial
with metastatic breast cancer patients. Psycho-Oncology, 8, 74 –91.
Edwards, C. L., Fillingim, R. B., & Keefe, F. (2001). Race, ethnicity and
pain. Pain, 94, 133–134.
Esdaile, J. (1846). Hypnosis in medicine and surgery. New York: Julian.
Ferrell, B. R., & Dean, G. (1995). The meaning of cancer pain. Seminars
in Oncology Nursing, 11, 17–22.
Foley, K. M. (1975). Pain syndromes in patients with cancer. In J. J.
Bonioca, V. Ventafridda, R. B. Fink, L. E. Jones, & J. D. Loeser (Eds.),
Advances in pain research and therapy (Vol. 2, pp. 59 –75). New York:
Raven.
Foley, K. M. (1996). Controlling the pain of cancer. Scientific American,
275, 164 –165.
Frischholz, E. J., Spiegel, H., & Spiegel, D. (1981). Hypnosis and the
unhypnotizable: A reply to Barber. American Journal of Clinical Hyp-
nosis, 24, 55–58.
Gay, M. C., Philippot, P., & Luminet, O. (2002). Differential effectiveness
of psychological interventions for reducing osteoarthritis pain: A com-
parison of Erikson hypnosis and Jacobson relaxation. European Journal
of Pain, 6, 1–16.
Gibbons, R. D., Hedeker, D., Waternaux, C., Kraemer, H. C., & Green-
house, J. B. (1993). Some conceptual and statistical issues in the analysis
of longitudinal psychiatric data. Archives of General Psychiatry, 50,
730 –750.
Giese-Davis, J., Koopman, C., Butler, L. D., Classen, C., Cordova, M.,
Fobair, P., et al. (2002). Change in emotion-regulation strategy for
women with metastatic breast cancer following supportive-expressive
group therapy. Journal of Consulting and Clinical Psychology, 70,
916 –925.
Goodwin, P. J., Leszcz, M., Ennis, M., Koopmans, J., Vincent, L., Guther,
H., et al. (2001). The effect of group psychosocial support on survival in
metastatic breast cancer. New England Journal of Medicine, 345, 1719 –
1726.
Haghighat, S., Akbari, M. E., Holakouei, K., Rahimi, A., & Montazeri, A.
(2003). Factors predicting fatigue in breast cancer patients. Supportive
Care in Cancer, 11, 533–538.
Hilgard, E. R., & Hilgard, J. R. (1975). Hypnosis in the relief of pain. Los
Altos, CA: Kauffman.
Holroyd, J. (1996). Hypnosis treatment in clinical pain: Understanding why
hypnosis is useful. International Journal of Clinical and Experimental
Hypnosis, 44, 33–51.
Karnofsky, D. A., & Burchenal, J. H. (1949). The clinical evaluation of
chemotherapeutic agents in cancer. In C. M. MacLead (Ed.), Evaluation
of chemotherapeutic agents (pp. 191–205). New York: Columbia Uni-
versity Press.
Kissane, D. W., Bloch, S., Smith, G. C., Miach, P., Clarke, D. M., Ikin, J.,
et al. (2003). Cognitive-existential group psychotherapy for women with
primary breast cancer: A randomized controlled trial. Psycho-Oncology,
12, 532–546.
Kissane, D. W., Grabsch, B., Clarke, D. M., Smith, G. C., Love, A. W.,
Bloch, D., et al. (2007). Supportive-expressive group therapy for women
with metastatic breast cancer: Survival and psychosocial outcome from
a randomized controlled trial. Psycho-Oncology, 16, 277–286.
Koopman, C., Nouriani, B., Erickson, V., Anupindi, R., Butler, L. D.,
Bachmann, M. H., et al. (2002). Sleep disturbances in women with
metastatic breast cancer. Breast Journal, 8, 362–370.
Kosslyn, S. M., Thompson, W. L., Costantini-Ferrando, M. F., Alpert,
N. M., & Spiegel, D. (2000). Hypnotic visual illusion alters color
processing in the brain. American Journal of Psychiatry, 157, 1279 –
1284.
Lane, L. G., & Viney, L. L. (2005). The effects of personal construct group
therapy on breast cancer survivors. Journal of Consulting and Clinical
Psychology, 73, 284 –292.
Lang, E. V., Benotsch, E. G., Fick, L. J., Lutgendorf, S., Berbaum, M. L.,
587
Berbaum, K. S., et al. (2000). Adjunctive non-pharmacological analgesia
for invasive medical procedures: A randomised trial. The Lancet, 355,
1486 –1490.
Levine, E. G., Eckhardt, J., & Targ, E. (2005). Change in post-traumatic
stress symptoms following psychosocial treatment for breast cancer.
Psycho-Oncology, 14, 618 – 635.
Levy, S. M., Herberman, R. B., Whiteside, T., Sanzo, K., Lee, J., &
Kirkwood, J. (1990). Perceived social support and tumor estrogen/
progesterone receptor status as predictors of natural killer cell activity in
breast cancer patients. Psychosomatic Medicine, 52, 73– 85.
Litt, M. D. (1988). Self-efficacy and perceived control: Cognitive media-
tors of pain tolerance. Journal of Personality and Social Psychology, 54,
154 –160.
Manning, M. M., & Wright, T. L. (1983). Self-efficacy expectancies,
outcome expectancies, and the persistence of pain control in childbirth.
Journal of Personality and Social Psychology, 45, 421– 431.
Montgomery, G. H., Bovbjerg, D. H., Schnur, J. B., David, D., Goldfarb,
A., Weltz, C. R., et al. (2007). A randomized clinical trial of a brief
hypnosis intervention to control side effects in breast surgery patients.
Journal of the National Cancer Institute, 99, 1280 –1281.
Mundy, E. A., DuHamel, K. N., & Montgomery, G. H. (2003). The
efficacy of behavioral interventions for cancer treatment-related side
effects. Seminars in Clinical Neuropsychiatry, 8, 253–275.
Patterson, D. R., & Jensen, M. P. (2003). Hypnosis and clinical pain.
Psychological Bulletin, 129, 495–521.
Portenoy, R. K., Miransky, J., Thaler, H. T., Hornung, J., Bianchi, C.,
Cibas-Kong, I., et al. (1992). Pain in ambulatory patients with lung or
colon cancer. Cancer, 70, 1616 –1624.
Rainville, P., Hofbauer, R. K., Bushnell, M. C., Duncan, G. H., & Price,
D. D. (2002). Hypnosis modulates activity in brain structures involved in
the regulation of consciousness. Journal of Cognitive Neuroscience, 14,
887–901.
Reynolds, P. T., & Olsen, E. J. (1996). Psychiatric aspects of cancer pain.
Journal of the Florida Medical Association, 83, 705–707.
Sellick, S. M., & Zaza, C. (1998). Critical review of 5 nonpharmacologic strategies
for managing cancer pain. Cancer Prevention and Control, 2, 7–14.
Spiegel, D., Bierre, P., & Rootenberg, J. (1989). Hypnotic alteration of so-
matosensory perception. American Journal of Psychiatry, 146, 749 –754.
Spiegel, D., & Bloom, J. R. (1983a). Group therapy and hypnosis reduce
metastatic breast carcinoma pain. Psychosomatic Medicine, 45, 333–339.
Spiegel, D., & Bloom, J. R. (1983b). Pain in metastatic breast cancer.
Cancer, 52, 341–345.
Spiegel, D., Bloom, J. R., & Yalom, I. D. (1981). Group support for
patients with metastatic cancer: A randomized outcome study. Archives
of General Psychiatry, 38, 527–533.
Spiegel, D., & Classen, C. (2000). Group therapy for cancer patients: A
research-based handbook of psychosocial care. New York: Basic Books.
Spiegel, D., Sands, S., & Koopman, C. (1994). Pain and depression in
patients with cancer. Cancer, 74, 2571–2578.
Spiegel, H., & Spiegel, D. (2004). Trance and treatment: Clinical uses of
hypnosis. Washington, DC: American Psychiatric Press.
Trijsburg, R. W., van Knippenberg, F. C., & Rijpma, S. E. (1992). Effects
of psychological treatment on cancer patients: A critical review. Psy-
chosomatic Medicine, 54, 489 –517.
Turner-Cobb, J. M., Sephton, S. E., Koopman, C., Blake-Mortimer, J., &
Spiegel, D. (2000). Social support and salivary cortisol in women with
metastatic breast cancer. Psychosomatic Medicine, 62, 337–345.
Weis, J. (2003). Support groups for cancer patients. Supportive Care in
Cancer, 11, 763–768.
Yun, Y. H., Heo, D. S., Lee, I. G., Jeong, H. S., Kim, H. J., Kim, S. Y., et
al. (2003). Multicenter study of pain and its management in patients with
advanced cancer in Korea. Journal of Pain and Symptom Management,
25, 430 – 437.