822 Journal of Pain and Symptom Management Vol. 45 No.

5 May 2013

Original Article

Music Therapy Reduces Pain in Palliative

Care Patients: A Randomized Controlled Trial
Kathy Jo Gutgsell, RN, MT-BC, Mark Schluchter, PhD,
Seunghee Margevicius, MA, MSN, Peter A. DeGolia, MD, Beth McLaughlin, MD,
Mariel Harris, MD, JD, Janice Mecklenburg, CNP, CHPN, and
Clareen Wiencek, PhD, CNP, CHPN
University Hospitals Case Medical Center (K.J.G., P.A.D., B.M., M.H., J.M.) and Case Western
Reserve University (M.S., S.M.), Cleveland, Ohio; and Virginia Commonwealth University (C.W.),
Richmond, Virginia, USA

Abstract
Context. Treatment of pain in palliative care patients is challenging. Adjunctive
methods of pain management are desirable. Music therapy offers a nonpharma-
cologic and safe alternative.
Objectives. To determine the efficacy of a single music therapy session to
reduce pain in palliative care patients.
Methods. Two hundred inpatients at University Hospitals Case Medical Center
were enrolled in the study from 2009 to 2011. Patients were randomly assigned to
one of two groups: standard care alone (medical and nursing care that included
scheduled analgesics) or standard care with music therapy. A clinical nurse
specialist administered pre- and post-tests to assess the level of pain using
a numeric rating scale as the primary outcome, and the Face, Legs, Activity, Cry,
Consolability Scale and the Functional Pain Scale as secondary outcomes. The
intervention incorporated music therapist-guided autogenic relaxation and live
music.
Results. A significantly greater decrease in numeric rating scale pain scores was
seen in the music therapy group (difference in means [95% CI]
1.4 [
2.0,

0.8]; P < 0.0001). Mean changes in Face, Legs, Activity, Cry, Consolability scores
did not differ between study groups (mean difference
0.3, [95% CI]
0.8, 0.1;
P > 0.05). Mean change in Functional Pain Scale scores was significantly greater in
the music therapy group (difference in means
0.5 ([95% CI]
0.8, 0.3;
P < 0.0001).
Conclusion. A single music therapy intervention incorporating therapist-guided
autogenic relaxation and live music was effective in lowering pain in palliative care
patients. J Pain Symptom Manage 2013;45:822e831. 2013 U.S. Cancer Pain
Relief Committee. Published by Elsevier Inc. All rights reserved.

Address correspondence to: Kathy Jo Gutgsell, RN, 5065, Cleveland, OH 44106, USA. E-mail:
MT-BC, Music Therapy Department, Seidman [email protected]
Cancer Center at University Hospitals Case Medi- Accepted for publication: May 14, 2012.
cal Center, 11100 Euclid Avenue, Mailstop: wrn

2013 U.S. Cancer Pain Relief Committee. 0885-3924/$ - see front matter
Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpainsymman.2012.05.008
Vol. 45 No. 5 May 2013 Music Therapy Reduces Pain in Palliative Care 823

Key Words
Music therapy, pain, palliative care, randomized controlled trial

Introduction a randomized trial, Tan et al.6 measured

pain, anxiety, and muscle tension levels of
Pain management in palliative care is very
burn patients undergoing dressing changes
challenging. Although patients desire to have
and found that patients who practiced music-
their pain managed, they also hope for lucidity
based imagery, a form of music-assisted relaxa-
and good quality of life as well as a sense of
tion with patient-specific mental imagery, had
control over their lives. Medications that lower
a significant decrease in symptoms. Loewy
pain may lower patients sense of control and
and Dileo7 add that the music therapist incor-
have unwanted side effects such as sedation,
porates techniques of muscle relaxation and
nausea, and constipation. In addition, patients
instructions for integrating breathing with im-
and families may fear addiction to opioids.
ages of comfort to potentiate the effects of mu-
Pain medications primarily target the sensory
sic in end-of-life care. In a 2011 Cochrane
(intensity) dimension of pain.1 Music therapy,
review of music interventions with cancer pa-
defined as the clinical and evidence-based use
tients, four music therapy trials were examined
of music interventions to accomplish individu-
whose interventions included music combined
alized goals within a therapeutic relationship
with imagery.8e11
by a credentialed professional who has com-
There are few quantitative music therapy
pleted an approved music therapy program,2
studies on pain in hospice and palliative care.
offers a low-risk, low-cost, nonpharmacologic
A 2010 Cochrane review of music therapy at
adjunct to standard care.3 The goals of music
the end of life included five trials. Only two
therapy in pain management are to assist the
small studies with a combined sample size of
patient in regaining self-control and becoming
45 examined the effect of music therapy on
actively involved in the management of his/
pain in hospice patients. Their pooled esti-
her pain. The music therapist engages patients
mate indicated no strong evidence of effect
in different types of music interventions (e.g.,
of music therapy (standardized mean differ-
singing, listening to music, and song writing)
ence
0.33; 95% CI
0.92, 0.26; P 0.27).
to enhance relaxation, provide opportunities
The reviewers determined that more studies
for self-expression, facilitate communication
are needed to further evaluate the effects of
with loved ones, and to bring beauty to suffer-
music therapy on pain at the end of life.12
ing. This helps to relieve the anxiety, fear, and
A 2011 Cochrane review examined the ef-
other components of suffering.4 According to
fects of music interventions on the psycholog-
the American Music Therapy Association
ical and physical outcomes of cancer patients.
(AMTA), A diverse array of underlying theo-
The review did not differentiate between mu-
ries forms the foundation for music therapy in-
sic therapy studies using a trained music thera-
terventions. Examples include frameworks
pist and music medicine studies using
from behavioral, psychodynamic, psychologi-
prerecorded music offered by a medical pro-
cal, and neurobiological theories. For the
fessional. Five trials with a combined sample
topic of pain and pain management, emerging
size of 391 measured the effect of music inter-
findings from neuroscience with applied music
ventions on pain and found a moderate pain-
therapy interventions are trending toward
reducing effect in both music therapy and
a fuller understanding of why certain music
music medicine studies (standardized mean
therapy interventions influence outcomes
difference
0.59; 95% CI
0.92,
0.27;
more favorably than others.5 Examples of mu-
P 0.0003). Evidence of the trials included
sic therapy interventions that incorporate be-
in this review suggests that music interventions
havioral frameworks include: the AMTA fact
may be offered as a complementary treatment
sheet on pain management, which describes
to people with cancer, but because most trials
a music therapy protocol for pain man-
were at high risk of bias, that is, one or more
agement developed by Hanser based on a
of the following criteria were not met, the
cognitive behavioral model of therapy.5 In
824 Gutgsell et al. Vol. 45 No. 5 May 2013

results need to be interpreted with caution. attended Palliative Care rounds. The investiga-
The criteria assessed for risk of bias were ran- tor received daily referrals for patients with ad-
dom sequence generation, allocation conceal- vanced, potentially life-limiting illness who
ment, blinding of participants and personnel, were in pain from the Palliative Care Team
blinding of outcome assessment for objective and from Nursing Services. The UHCMC did
and subjective outcomes, incomplete outcome not have a dedicated Palliative Care Unit
data, selective reporting, and other biases. The when the study was being conducted. The Pal-
main reason for receiving a rating of high risk liative Care Team provides consultative services
of bias was the lack of blinding. Blinding is of- for patients throughout UHCMC in intensive
ten impossible in music therapy and music care, general medical, surgical, rehabilitation,
medicine studies that use subjective outcomes and oncology units.
such as pain. This is especially true for music After the initial referral, the investigator
therapy studies that use active music making. conducted a chart review and interviewed the
When participants cannot be blinded to the in- participant and his or her nurse to determine
tervention, there is an opportunity for bias if the following inclusion criteria were met:
when they are asked to report on these subjec- 1) a diagnosis of advanced, potentially life-
tive outcomes. Therefore, it appears impossi- limiting illness, 2) 18 years or older, 3) pain
ble for these types of studies to receive a low of three or greater as measured on a zero to
or even moderate risk of bias even if all other 10 numeric rating scale (NRS), 4) able to un-
risk factors (e.g., randomization, allocation derstand English, and 5) alert and oriented
concealment, and so on) have been adequately to person and place and able to rate pain on
addressed.13 the numeric scale. Patients were not excluded
Analysis of the 2011 Cochrane review reveals if they were on scheduled pain medications,
that music therapy interventions used in re- although interventions were scheduled around
search varied in frequency (single to multiple the administration of breakthrough pain
in number), length (20e120 minutes), live medications, with the intervention occurring
vs. recorded music, patient- vs. therapist- immediately before the next dose of medica-
selected music, and the intervention itself (in- tion. The UHCMC Institutional Review Board
teractive music making with the participants, approved the study. The investigator obtained
music-guided imagery, music-guided relaxa- written informed consent from all participants.
tion, and music-video making). Palliative care
music therapy needs more rigorous research Outcome Measures
so that interventions are evidence based.14 To Primary Outcome: NRS. The NRS is validated
better understand the impact of specific music for use in adults and children aged nine years
therapy interventions, studies are needed that or older in all patient care settings who are
isolate the effects of one intervention.3,15 The able to use numbers to rate the intensity of
authors of the Cochrane review note as well their pain. It is recommended in the literature
that most studies are compromised by small to measure short-term changes in pain and it is
sample size and lack of statistical power.12 used throughout UHCMC.16 Patients rate
The objective of the present study was to de- their pain from zero to 10, with zero reflecting
termine the efficacy of a single music therapy no pain and 10 reflecting the worst possible
session to reduce pain in palliative care patients. pain.17

Secondary Outcome: The Face, Legs, Activity, Cry,

Methods Consolability Scale. The Face, Legs, Activity,
Setting and Participants Cry, Consolability (FLACC) Scale is a behavioral
All participants were inpatients at University pain assessment in which pain is rated by observ-
Hospitals Case Medical Center (UHCMC) in ing the patient and assigning a number to ones
Cleveland, Ohio between September 2009 findings. The scale is scored between a range of
and August 2011. The principal investigator zero and 10, with zero representing no pain.
(K. J. G.), hereafter called the investigator, col- The scale has five criteria: face, legs, activity,
laborated with the Palliative Care Team (three cry, and consolability, to which each is assigned
physicians and two nurse practitioners) and a score of zero, one, or two.18 Originally
Vol. 45 No. 5 May 2013 Music Therapy Reduces Pain in Palliative Care 825

validated in children with postoperative pain, the CNS. Randomization assignments were
the FLACC Scale has been recently validated generated using SAS software (SAS Institute,
in assessing pain in critically ill adults who are Inc., Cary, NC) by the study statistician, using
unable to self-report pain. Because Voepel- a permuted block scheme with random block
Lewis et al.19 and others found that FLACC sizes of 20 or 30. Because the protocol speci-
scores were comparable with those of the com- fied the presence of a music therapist to facil-
monly used NRS, the authors selected this itate the music therapy intervention, it was not
pain assessment to provide the behavioral com- possible for the participant to be blinded to his
ponent of the patients pain experience. In ad- or her group assignment. If the participants
dition, the FLACC Scale was shown to have pain was less than three on the NRS, he or
excellent interrater reliability, criterion validity, she was excluded from the study.
and construct validity. Health care professionals
who are trained in its use are qualified to per- Music Therapy Group. The investigator, a pro-
form the assessment. Because the FLACC Scale fessional music therapist, informed the patient
has not been validated in adults who are able to of his or her assignment to the music therapy
self-report pain, the present study used the group and then proceeded with the interven-
FLACC Scale as a secondary outcome. tion. After placing a Do Not Disturb sign
on the door and preparing the patient and
Secondary Outcome: The Functional Pain Scale. the environment (adjusting the lights, offering
Patients are asked if their pain is tolerable or a blanket, turning off cell phones, and so on),
intolerable. From there, they describe whether the therapist briefly played the ocean drum to
or not pain keeps them from engaging in daily give the patient the choice of whether or not
activities. A rating of zero reflects no pain. A to include it in the intervention because
rating of one indicates tolerable pain with no some patients express aversion for it and find
impact on activity. A rating of five reflects intol- that it inhibits their ability to relax. The thera-
erable pain with a resulting inability to verbally pist then facilitated a single 20-minute music
communicate.20 The Functional Pain Scale therapy intervention directed at lowering
(FPS) assesses both the patients subjective per- pain. The intervention, a standard protocol
ception of pain and its impact on his or her for all participants, began with verbal instruc-
level of functioning. Although the FPS was de- tions for autogenic relaxation. The music ther-
veloped to determine pain in older people apist asked the patient to pay attention to
who are cognitively intact, the authors selected breathing for approximately one minute.
it as a secondary outcome for the present study Then the therapist led the patient in autogenic
because of its ability to help professionals un- muscle relaxation by asking the patient to pay
derstand how pain affects daily functioning attention to the scalp muscles and allow them
in all their adult patients. to release, and moving down with similar focus
on specific muscle groups, ending with the
Intervention feet. Next, the patient was invited to imagine
After the investigator obtained informed a safe place of his or her own choosing. The
consent from an eligible participant, the inves- therapist asked the patient to imagine what
tigator summoned a clinical nurse specialist he or she saw, smelled, heard, tasted, and felt
(CNS) research assistant who assessed the pa- on the skin at the safe place. Then the music
tients pain using the three measures: the therapist informed the patient that she would
NRS, the FLACC Scale, and the FPS. The begin to play first the ocean drum, if chosen,
CNS then left the hospital unit. If the partici- and then the harp to support his or her explo-
pants pain score was still three or greater on ration of the safe place. The therapist played
the NRS, the investigator immediately thereaf- the same harp pieces for every patient. The
ter opened a serially numbered, sealed, opa- pieces for the present protocol were chosen
que envelope to obtain the patients assigned based on the therapists clinical experience
group. The investigator opened the sealed en- in which patients had described them as sooth-
velope containing group assignment (music ing, peaceful, and calming. All pieces were
therapy or control) in the presence of the pa- played at a soft volume in a slow tempo and
tient but not the CNS to ensure blinding of are described as follows: 1) an improvisation
826 Gutgsell et al. Vol. 45 No. 5 May 2013

in the mode of G Mixolydian with a duple me- immediately before and after the music therapy
ter, 2) four precomposed pieces in the key of C or control intervention. Each study participant
Major that can be described as light classical was assessed by the same CNS pre- and postin-
and are unfamiliar to most listeners: An- tervention. In all but four cases, post-test data
dante by Waddington in duple meter, Pass- were obtained within 10 minutes of completion
ing By and Reverie by Grandjany in duple of the intervention. On three occasions, the
meter, and Barcarolle by Grandjany in triple CNS obtained post-test data in 15 minutes
meter. At the conclusion of the music, the ther- and on one occasion in 30 minutes because
apist gently invited the participant to leave his of schedule conflicts. For 11 patients, blinding
or her imagined safe place and re-enter the of the research assistant was broken because
hospital room, realizing that the safe place is the patients revealed their group assignment.
a resource to which he or she can return at To attempt to control for bias, the therapist
any time. Then the music therapist left the remained outside the room while the research
room and notified the same CNS to return to assistant administered pre- and post-tests to
the patient to reassess pain using the same the patient.
three measures: the NRS, the FLACC Scale,
and the FPS. After completion of the post- Statistical Analysis
tests, the therapist re-entered the patients Comparisons of baseline characteristics be-
room to verbally process the music therapy in- tween groups were made using t-tests or Wil-
tervention and offer follow-up treatment. She coxon rank sum tests for continuous
gave each study participant a CD of the inter- variables, and c2 tests for categorical or binary
vention for future use and provided a CD variables. The mean changes from pre- to post-
player on request. Interested readers may con- test in each of the three pain scales (NRS,
tact the investigator to request a recording of FLACC Scale, and FPS) were compared be-
the intervention. tween the music therapy and control groups
using an independent sample t-test. Two-way
Control Group. The therapist informed the pa- analysis of variance was used to examine
tient of his or her assignment to the control whether treatment effects differed according
group and explained that he or she would to patient characteristics such as age, gender,
receive the live music therapy intervention af- and baseline pain level. All tests were two-sided
ter reassessment for pain. Next, she facilitated with a significance level of 0.05. Statistical
the same comfort measures as for the music analyses were carried out using SAS version
therapy group: adjusting the lights, providing 9.2. Because there was a single primary out-
a blanket, and turning off the telephones. come, no adjustment was made for multiple
Then the therapist invited the patient to relax, comparisons.
but gave no special instructions for doing so The sample size of 200 (100 per treatment
because the therapist-guided autogenic relaxa- arm) provided 80% power to detect between-
tion was integral to the music therapy interven- group differences in mean post-test numeric
tion. She left the room and placed a Do Not pain scores of 0.40 standard deviations, using
Disturb sign on the door. After 20 minutes, a two-sided test with a significance level of
she notified the same CNS to return to the pa- 0.05. The sample size of 100 per group was
tient to reassess pain using the three measures: chosen partly on the basis of what was a feasible
the NRS, the FLACC Scale, and FPS. After number to study and was justified by determin-
post-test data were collected, the therapist pro- ing that it would provide 80% power to detect
vided the music therapy intervention for each an effect size of 0.40 standard deviations,
control patient. The therapist gave each pa- which is in-between what Cohen21 considers
tient in the control group a CD of the interven- a small and a medium effect size (0.2
tion for future use and provided a CD player and 0.5, respectively). We thus determined
on request. that this effect size was suitably low to justify
the sample size. Primary analyses were carried
Data Collection Procedure out using intention-to-treat analysis, including
The CNS, blinded to treatment allocation, all randomized patients on whom data were
administered the pain assessment measures obtained. Statistical analysis of the final data
Vol. 45 No. 5 May 2013 Music Therapy Reduces Pain in Palliative Care 827

excluding: 1) the 11 patients who divulged be wakened for the post-test. The subjects as-
group assignment to the CNS, 2) the four pa- signed to music therapy and control groups
tients who had post-test assessments for more did not differ according to gender, ethnicity, di-
than 10 minutes after the intervention, and agnosis, mean age, or baseline pain severity
3) the 10 patients who chose not to hear the (Table 1). The pain duration variable had
ocean drum and the one patient who re- a skewed distribution in both groups, which is
quested to skip the talk and get right to the why the authors used a nonparametric Wilcox-
music, did not alter the results. on rank sum test to compare the groups at base-
line. Because the median is a better measure of
location than the mean for these data, we added
the median pain duration to Table 1 for this var-
Results iable. Note that the medians of the two groups
are quite similar, reflecting the nonsignificant
Of the 400 referred patients, 200 signed in-
P-value from the rank sum test.
formed consent and were enrolled in the study
(Fig. 1). Of the 200 subjects screened but not
enrolled, 20 were ineligible and 180 did not Numeric Rating Scale
give consent. Reasons for ineligibility included Both music therapy and control groups
pain score less than three (n 15), not oriented showed significant declines from pre- to
to person and place (n 3), did not speak En- post-test (mean change [95% CI]
1.94
glish (n 1), and researcher error (n 1). [
2.37,
1.52] for music therapy and
0.56
The 180 subjects who did not consent gave var- [
0.92,
0.19] for control). However, a signifi-
ious reasons including I want to be alone now, cantly (P < 0.0001) greater change was seen in
It is a bad day, I do not like the harp, I am the music therapy group (difference in means
not interested, Music cannot help my pain, [95% CI]
1.39 [
1.95,
0.83]).
I brought my own music to listen to, or Music
is not my thing. Of the 100 subjects assigned to Face, Legs, Activity, Cry, Consolability Scale
the music therapy group, all but one completed The FLACC Scale scores declined signifi-
the music therapy session and completed all cantly in both the music therapy and control
measurements. The patient who did not com- groups. However, the mean change in scores
plete the post-test exhibited symptoms of confu- did not differ significantly between the two
sion and agitation during the intervention and groups (difference in means [95% CI]
0.3
was excluded from the study. Of the 100 sub- [
0.8, 0.1], P > 0.05).
jects in the control group, all completed the
pretest. Postintervention scores were obtained Functional Pain Scale
on 99 subjects. One control patient who had There was a significant decline in the func-
been in severe pain fell asleep during the con- tional pain score in the music therapy group,
trol session. His nurse requested that he not but not in the control group. The mean

Control group therapy group

Fig. 1. Flowchart of patients through the study.

828 Gutgsell et al. Vol. 45 No. 5 May 2013

Table 1
Demographic Variables of the Study Participants
Study Group

Variables All Patients Music Therapy (n 100) Control (n 100) P-value

Age (mean  SD) 56.09  15.08 57.45  14.76 54.72  15.34 0.20a
Gender, n (%)
Male 62 (31) 31 (31) 31 (31) >0.999b
Female 138 (69) 69 (69) 69 (69)
Race, n (%)
White 135 (67.5) 66 (66) 69 (69) 0.65b
Nonwhite 65 (32.5) 34 (34) 31 (31)
Diagnosis, n (%)
Cancer 174 (87) 91 (91) 83 (83) 0.09b
Noncancer 26 (13) 9 (9) 17 (17)
Pain severity (mean  SD) 6.44  1.82 6.48  1.68 6.39  1.95 0.73a
Pain duration (wk)
Mean  SD 14.04  36.83 8.49  14.50 19.58  49.54 0.51c
Median 4.00 3.50 4.00
a
P-value from t-test.
b
P-value from c2 test.
c
P-value from Wilcoxon rank sum test.

decline was significantly greater (P < 0.0001) Discussion

in the music therapy group than in the control
The results of this research appear to indicate
group (difference in means [95%CI]
0.52
that a single music therapy intervention low-
[
0.78,
0.25]; Table 2 and Fig. 2).
ered pain in hospitalized palliative care pa-
Further analyses were carried out to exam-
tients. A noteworthy finding is the efficacy of
ine whether baseline characteristics of the pa-
the intervention itself. Evidence-based music
tients were related to the efficacy of the
therapy practice often uses patient-preferred
intervention. These analyses present the
music as part of an individualized treatment
mean change in pain score for both music
plan.22 In contrast, the present research inter-
therapy and control groups, stratified by levels
vention was a standard protocol and varied little
of each of the baseline factors being exam-
from one patient to another. Examples of varia-
ined. Factors examined were age (#55 and
tions included the music therapist giving each
>55 years), gender (male and female), race patient the option of including the ocean
(white and nonwhite), diagnosis (cancer and drum in the intervention. Of 100 patients in
noncancer), pain severity (mild [0e3], moder- the music therapy group, 10 declined its inclu-
ate [4e6], and severe [7e10]), and duration
sion. The rationale for providing this choice
of pain at baseline (#4, 5e12, 13e24, and
was that some patients express aversion for the
>24 weeks). A significant P-value for the test
sound of the ocean drum, finding that it inter-
for interaction indicates that the efficacy of
feres with their ability to relax. In another exam-
the intervention differed across levels of the
ple of variation, one patient requested that the
baseline factor being examined. Interaction therapist skip the relaxation talk and get
tests for the analyses of NRS and FPS scores right to the music. The therapist chose to
were not significant, indicating that effects of honor his request and not add to the distress
music therapy did not vary across levels of
he already experienced from being in pain. In
the baseline factors. In the analysis of FLACC
addition, the therapist individualized each in-
Scale scores, the interaction test for age was
tervention by matching her breathing with the
significant (P 0.03) and results indicate that
patient and adjusting the tempo and cadence
the effect of music therapy was greater in those
of the spoken script to meet the patients needs.
aged #55 years (95% CI
1.57,
0.27) com- Other than the patients described above, the
pared with those aged >55 years (
0.59, verbal instructions, the harp music selections,
0.85). This result should be interpreted with and the length of the intervention were consis-
caution given that multiple tests were done
tent from patient to patient.
and we did not correct for multiple testing.
Vol. 45 No. 5 May 2013 Music Therapy Reduces Pain in Palliative Care 829

Although it is true that music therapists com-

0.52  0.95a (
0.78,
0.25)

0.62  0.96 (
0.81,
0.43)

0.10  0.93 (
0.29, 0.08)

monly assess for patient preferences and then

2.38  1.01 (2.18, 2.58)

1.76  1.01 (1.56, 1.96)

2.30  0.97 (2.11, 2.49)

2.19  1.04 (1.99, 2.40)
Mean  SD (95% CI)
design interventions that include such music,
there are precedents to therapist-selected music
that are documented in the literature. The
Bonny Method of Guided Imagery and Music
FPS

(GIM) was developed by music therapist Helen

Bonny. The GIM is fully integrated into and en-
dorsed by the AMTA. The GIM uses specifically
sequenced classical music programs to stimu-
late inner experience to meet clinical goals.
99
100
99
99

100
99
99
The GIM uses Western classical music because
N

it is the field of expertise of the persons who de-

veloped and tested the programs. This music

1.01  1.56 (
1.32,
0.70)

0.67  1.80 (
1.02,
0.31)

contains elemental, harmonic, rhythmic, and

0.34  1.68 (
0.81, 0.13)
1.81  1.86 (1.44, 2.18)
0.78  1.56 (0.47, 1.09)

1.72  2.09 (1.31, 2.13)

1.04  1.66 (0.71, 1.37)

structural patterns that have stood the test of

Mean  SD (95% CI)

time, effectively engaging persons in explora-

Summary of NRS, FLACC, and FPS Scores by Study Group

tion during altered states of consciousness,

and which consistently evoke imagery responses
FLACC

of therapeutic value.23
Mandel et al.24 found that cardiac rehabilita-
tion patients who listened to prerecorded in-
strumental music interspersed with spoken
suggestions at home for at least three months
NRS numeric rating scale; FLACC Face, Legs, Activity, Cry, Consolability Scale; FPS Functional Pain Scale.

to relax their body and mind had significantly

99
100
99
99

100
99
99
N

more improvement in systolic blood pressure,

Table 2

anxiety, and stress than those who only at-

tended cardiac rehabilitation. The music ther-

1.39  1.99a (
1.95,
0.83)

1.94  2.14 (
2.37,
1.52)

0.56  1.83 (
0.92,
0.19)

apists carefully selected the music with

6.69  1.72 (6.35, 7.03)
4.74  2.59 (4.23, 5.26)

6.41  1.91 (6.03, 6.79)

5.86  2.42 (5.38, 6.34)
Mean  SD (95% CI)

attention to properties that research suggests

are conducive to relaxation, including slow
tempo, soft dynamics, and long phrases.
In addition, the investigator, a trained music
NRS

therapist, observed in years of clinical practice

that patients in pain are often vulnerable and
their desire to manage pain overrides personal
preferences in music. Many patients reported
lower pain perception after her intervention
of carefully selected music. Therefore, on the
100
99
99

100
99
99

99
N

strength of the literature cited above, on the

clinical experience of the investigator, and to
limit the variable of music selections in order
Difference from

Difference from

Difference in mean change between

music therapy and control groups

to demonstrate scientific rigor, the authors de-

post to pre

post to pre

signed the present study with no assessment

Test

for patient preference in music. The only op-

tions allowed were choice of ocean drum and
Post

Post
Pre

Pre

shortening the autogenic relaxation, but only

P < 0.0001 from t-test.

when their inclusion would have increased pa-

tient distress.
Music therapy
Study Group

A finding of this study is that pain also de-

creased significantly (P < 0.05) in the control
Control

group on two of the three measures (NRS and

FLACC Scale). It appears that the simple act
a
830 Gutgsell et al. Vol. 45 No. 5 May 2013

Fig. 2. Changes in pain scores of the participants.

of inquiring about pain and then instructing were obtained within 10 minutes of the com-
the patient to relax is in some instances enough pletion of the music therapy session. On three
to lower pain significantly, as long as it includes occasions the CNS obtained post-test data in
offering to make adjustments to the environ- 15 minutes and on one occasion in 30 minutes
ment such as turning down the lights, pulling as a result of schedule conflicts; 2) address
the window shades, supplying a blanket, turn- whether patients request fewer breakthrough
ing off cell phones, reassuring the patient that pain medications after music therapy; 3) find
someone will reassess his or her pain in 20 min- out whether successive interventions have a
utes, and putting a Do Not Disturb sign on the cumulative pain-lowering effect; 4) examine
door to ensure privacy. whether a therapist-created recording of an in-
Although all attempts were made to mini- tervention has the same pain-lowering effect if
mize risk of bias, two risks remained, which the patient listens to it after a live session with
are implicit in music therapy research. The the same therapist; and 5) address whether
first is the blinding of participants and person- pain is lowered in control group patients who
nel. Because music therapy requires the pres- later receive music therapy.
ence of the music therapist, both the The strengths of the present study are its
therapist and the patient were not blinded to large sample size, its use of one music therapy
group assignment. The second risk is the intervention, and its attempt to meet scientific
blinding of outcome assessment. When partic- standards of a quality randomized controlled
ipants cannot be blinded to the intervention, trial. Because of these features, it provides
there is definitely an opportunity for bias a valuable addition to the literature. Based
when they are asked to report on subjective on the results, palliative care clinicians may
outcomes such as pain.13 confidently refer trained music therapists to
A limitation of the study is that it may be dif- treat pain in this vulnerable population.
ficult to generalize the results to all palliative
care patients in pain, as 45% of the referred
patients did not consent to participate. For Disclosures and Acknowledgments
consenting patients who choose to be less ac- This research was supported by a grant from
tively involved in a music therapy session, the the Kulas Foundation in Cleveland, Ohio. The
intervention used in this study has clinical sig- authors declare no conflicts of interest.
nificance. Further research is needed to repli- The authors would like to thank the Kulas
cate the study so that its results may be Foundation, all of the patients who partici-
generalized to other music therapists and mu- pated in the study, the Clinical Nurse Special-
sical instruments. ists who assisted in gathering data, the Core
Additional research also is needed to: 1) Library, and the Art and Music Therapy De-
measure the length of time pain is reduced af- partment at University Hospitals Case Medical
ter a music therapy intervention. In the pres- Center for its support and encouragement
ent study, in all but four cases, post-test data throughout the study.
Vol. 45 No. 5 May 2013 Music Therapy Reduces Pain in Palliative Care 831

References 13. Bradt J, Dileo C, Grocke D, Magill L. Music in-

terventions for improving psychological and physi-
1. Kwekkeboom K. Oncology nurses use of non- cal outcomes in cancer patients. Cochrane
drug pain interventions in practice. J Pain Symptom Database Syst Rev 2011;8:CD006911.
Manage 2008;35:83e94.
14. Hilliard RE. The use of music therapy in meet-
2. American Music Therapy Association. Defini- ing the multidimensional needs of hospice patients
tion of music therapy. 2011. Available from and families. J Palliat Care 2001;17:161e166.
http://www.musictherapy.org. Accessed September
16, 2011. 15. Gallagher LM, Lagman R, Walsh D, Davis MP,
LeGrand SB. The clinical effects of music therapy
3. Groen K. Pain assessment and management in in palliative medicine. Support Care Cancer 2006;
end of life care: a survey of assessment and treat- 144:859e866.
ment practices of hospice music therapy and nurs-
ing professionals. J Music Ther 2007;44:90e112. 16. Caraceni A, Cherny N, Fainsinger R. Pain mea-
surement tools and methods in clinical research in
4. Bailey LM. Music therapy in pain management.
palliative care: recommendations of an expert work-
J Pain Symptom Manage 1986;1:25e28.
ing group of the European Association of Palliative
5. American Music Therapy Association. Music Care. J Pain Symptom Manage 2002;23:239e255.
therapy and pain management fact sheet. 2010.
Available from http://www.musictherapy.org. Ac- 17. McCaffery M, Beebe A. Pain: Clinical manual for
cessed February 27, 2012. nursing practice. St. Louis, MO: CV Mosby Co., 1989.

6. Tan X, Yowler CJ, Super DM, Fratianne RB. The 18. Merkel SI, Voepel-Lewis T, Shayevitz JR,
efficacy of music therapy protocols for decreasing Malviya S. The FLACC: a behavioral scale for scoring
pain, anxiety, and muscle tension levels during postoperative pain in young children. Pediatr Nurs
burn dressing changes: a prospective randomized 1997;23:293e297.
crossover trial. J Burn Care Res 2010;31:590e597. 19. Voepel-Lewis T, Zanotti J, Dammeyer JA,
7. Loewy JV. In: Dileo C, Loewy JV, eds. Music ther- Merkel S. Reliability and validity of the faces, legs,
apy at the end of life. Cherry Hill, NJ: Jeffrey Books, activity, cry, consolability behavioral tool in assessing
2005. acute pain in critically ill patients. Am J Critic Care
2010;19:55e61.
8. Allen J. The effectiveness of group music and
imagery on improving the self-concept of breast 20. Gloth FM III, Scheve AA, Stober CV, Chow S,
cancer survivors [unpublished PhD thesis]. Phila- Prosser J. The Functional Pain Scale: reliability, val-
delphia, PA: Temple University, 2010. idity, and responsiveness in an elderly population.
J Am Med Dir Assoc 2001;2:110e114.
9. Burns DS. The effect of the Bonny method of
guided imagery and music on the mood and life 21. Cohen J. Statistical power analysis for the behav-
quality of cancer patients. J Music Ther 2001;38: ioral sciences, 2nd ed. Hillsdale, NJ: Lawrence Erl-
51e65. baum Associates, 1988.
10. Burns DS, Azzouz F, Sledge R, et al. Music imag- 22. Magill L. The use of music therapy to address
ery for adults with acute leukemia in protective en- the suffering in advanced cancer pain. J Palliat
vironments: a feasibility study. Support Care Care 2001;17:167e172.
Cancer 2008;16:507e513. 23. Association for Music and Imagery. Definition of
11. Montserrat G, Domenech M. The effect of mu- the Bonny method of guided imagery and music.
sic and imagery to induce relaxation and reduce 2012. Available from http://www.ami-bonnymethod.
nausea and emesis in cancer patients undergoing org/faq.asp. Accessed May 4, 2012.
chemotherapy treatment [unpublished PhD thesis]. 24. Mandel SE, Hanser SB, Ryan LJ. Effects of a mu-
Stockton, CA: University of the Pacific, 2008. sic-assisted relaxation and imagery compact disc re-
12. Bradt J, Dileo C. Music therapy for end-of-life cording on health-related outcomes in cardiac
care. Cochrane Database Syst Rev 2010;1:CD007169. rehabilitation. Music Ther Perspect 2010;28:11e21.