Copyright B 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins

Hsiu F. Tsai, MS, RN Ying R. Chen, MS, RN Min H. Chung, PhD, RN Yuan M. Liao, PhD, RN Mei J. Chi, PhD Chia C. Chang, PhD, RN Kuei R. Chou, PhD, RN

Effectiveness of Music Intervention in Ameliorating Cancer Patients’ Anxiety, Depression, Pain, and Fatigue A Meta-analysis

K E Y

W O R D S

Background: This is the first study to use meta-analysis as a scientific technique to

Cancer patient

provide an integrated analysis of the effectiveness of music intervention in cancer

Meta-analysis

patients. Objectives: The purpose of this study was, using the meta-analysis

Music intervention

method, to present a summary of existing research and explore the effectiveness of music intervention in ameliorating anxiety, depression, pain, and fatigue in cancer patients. Methods: The present study collected quantitative study designs sought of music intervention for cancer patients published from 2002 to 2012. These studies were then cross-referenced using Medical Subject Headings for topics on music intervention and cancer patients. Outcome indicators were anxiety, depression, pain, and fatigue. The quality of the studies was evaluated using Cochrane Collaboration Guidelines. The effect size on outcome indicators used the formula devised by Hedges and Olkin (1985). Results: Results showed that music interventions were significantly effective in ameliorating anxiety (g = j0.553), depression (g = j0.510), pain (g = j0.656), and fatigue (g = j0.422) in cancer patients. Subgroup analyses revealed that age and who selected the music were major factors influencing the effect size on anxiety reduction. Conclusions: Music interventions significantly ameliorate anxiety, depression, pain, and fatigue in cancer patients, especially adults. Music interventions were more effective in adults than in children or adolescents and more effective when patients, rather than researchers,

Author Affiliations: Graduate Institute of Nursing, College of Nursing, Taipei Medical University, (Mss Tsai and Chen, Drs Chung, Liao, and Chou); Department of Nursing, Taoyuan Armed Forces General Hospital, (Ms Chen); Department of Nursing, Hsin Sheng College of Medical Care and Management, Tauyuan (Ms Tsai); and School of Gerontology Health Management, College of Nursing, Taipei Medical University, Taiwan (Drs Chi and Chang).

Effectiveness of Music Intervention in Cancer Patients

The authors have no funding or conflicts of interest to disclose. Correspondence: Kuei R. Chou, PhD, RN, Graduate Institute of Nursing, College of Nursing, Taipei Medical University, No. 250, Wu-Hsing St, Taipei 110, Taiwan ([email protected]). Accepted for publication October 15, 2013. DOI: 10.1097/NCC.0000000000000116

Cancer NursingTM, Vol. 37, No. 6, 2014

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

n

E35

chose the music. Implications for Practice: Our findings provide important information for future music-intervention planners to improve the design and processes that will benefit patients in such programs.

I

n the past 30 years, cancer has been the leading cause of death in Taiwan.1 Furthermore, cancer continues to be 1 of the most important public health issues worldwide.2 Cancer is clearly a health issue that cannot be ignored. Physical and psychological symptoms are common in cancer patients, especially pain, fatigue, anxiety, and depression.3Y5 The American Cancer Society indicates that more than 30% of cancer patients experience marked anxiety.6 Common treatments for psychological symptoms include psychological support, relaxation training, meditation, medication, and music treatment. Among these, music therapy has been widely recommended as a complementary therapy for cancer patients in clinical settings.

Music Intervention The American Association of Music Therapy defines music therapy as ‘‘the clinical and evidence-based use of music interventions to accomplish individualized goals within a therapeutic relationship by a credentialed professional who has completed an approved music therapy program.’’ Music is an auditory composition of pitch, speed, rhythm, and volume. Thus, music therapy can be described as treating psychological needs with the sounds and rhythms of music. Music therapy may include diverse activities such as music composition, singing, and listening. The design of music interventions is affected by the music style, choice of musical instruments, therapist’s style, and patient characteristics. Commonly used methods are active-passive therapy and individual-group therapy.7 There are 4 major categories of music therapy: appreciation, recreation, improvisation, and creation.8 Not only can music therapy improve a subject’s quality of life, it is also effective in adjusting to pressure, relieving pain, expressing feelings, enhancing memory, improving communication, facilitating physiological rehabilitation, and achieving a harmonic state of body, mind, and spirit.9 Furthermore, music interventions can be applied to patients with various diseases.

Application of Music Interventions in Cancer Patients Originally used to enhance sleep and reduce preoperative anxiety in cancer patients, music interventions are now used to reduce pain, mental stress, and physical discomfort from chemotherapy. Music interventions help patients express and release emotions through actual interaction with music.10 In recent studies applying music interventions to reduce anxiety, the most commonly used assessment tools are the Hospital Anxiety and Hamilton Anxiety Scale, the Hospital Anxiety and Depression Scale, and State-Trait Anxiety Inventory (STAI). In a study by Li et al11 on breast cancer patients undergoing radiotherapy, a music intervention significantly reduced anxiety (mean difference, j4.57; 95% confidence interval [CI], j6.33 to j2.82; P G .0001). ComE36 n Cancer NursingTM, Vol. 37, No. 6, 2014

pared with standard care alone, music interventions significantly reduced anxiety in cancer patients (mean difference, j11.20; 95% CI, 19.95 to j2.82; P = .009).12 The most commonly used tools in recent studies evaluating the effects of music interventions on depression are the Hospital Anxiety and Depression Scale and the Zung Self-rating Depression Scale. Music interventions significantly ameliorated depression in women with breast cancer receiving radiotherapy (P G .001)13 and ameliorated depression in children with cancer (negative states average G2.5/10).14 The Present Pain Intensity, Pain Numeric Rating Scale, the Faces Scale, and a visual analog scale were the most commonly used tools to evaluate pain in recent music intervention studies. Madden et al15 reported that a music intervention significantly reduced pain in children with brain cancer. Music interventions significantly reduced pain in cancer patients than in those not receiving a music intervention (r = 0.45; 95% CI, 0.23Y0.63; P G .0001).16 The most common assessment tools used in recent music therapy literature for fatigue evaluation are the Profile of Mood States and the Functional Assessment of Chronic Illness TherapyY Fatigue scale. Compared with standard care alone, music interventions significantly reduced fatigue in leukemia patients receiving chemotherapy (P G .001)17 or stem cell transplants (Profile of Mood States: music group, from 6.4 to 4.3; control group, from 5.8 to 5.2; P = .02).18 Ferrer19 also found that besides reducing anxiety and fear, familiar live music also significantly reduced fatigue in cancer patients who underwent chemotherapy (P = .001). Anxiety, depression, fatigue, and pain have been found to be particularly amenable to be the effects of music interventions. There is a growing body of research documenting the benefit of music interventions in reducing emotional responses and physical symptoms in patients with cancer.3,11,17,19 Recently, Bradt et al12 and Zhang et al20 conducted a systemic literature review and meta-analyses on music intervention for cancer patients. The differences between and distinctive characteristics of existing studies were the rationale for our analysis. The gold standards for evaluating effects of interventions are true and quasi-experimental studies, whereas Bradt et al20 included more nonYrandomized controlled trials in the systemic review. Besides, according to Grading of Recommendations Assessment, Development, and Evaluation working group criteria, the quality of evidence of the results of Bradt et al12 is low in 3 categories (anxiety-STAI, pain, and heart rate) and very low in the other 4 (anxiety-non-STAI, systolic blood pressure, diastolic blood pressure, and quality of life). They concluded that most trials are at high risk of bias and the results should be interpreted with caution. In their meta-analysis, Zhang et al20 included studies published after 1966 and explored the overall effect size only. With the recent advance of research methodology and the development of structured guidelines

Tsai et al

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

(eg, Consolidated Standards of Reporting Trials statement in the mid-1990s), clinical trials published more recently are of higher quality in general. Analyzing earlier trials more than 2 decades ago could increase the heterogeneity and affect the quality of the meta-analyses. Furthermore, neither Bradt et al nor Zhang et al performed subgroup analysis on the relationship between study characteristics and effect sizes. Therefore, the purpose of the present study was to perform a meta-analysis on reliable and valid true or quasi-experimental studies published between 2002 and 2012 on the effect of music intervention in cancer patients. It was hoped that the present study could provide deeper insight into (1) the effect size of music therapy on the cancer patients’ anxiety, depression, pain, and fatigue outcome variables; (2) how different variables (study design and method, therapy, therapist, and participant characteristics) affected the effect size and if any variable has a greater effect by using subgroup analysis; (3) whether a continuous variable was a significant predictor factor by using meta-regression; (4) and the quality and publication bias of those studies and the impact on effect size.

n

Materials and Methods

Search Strategies The review was guided by the protocol proposed by Cochrane Collaboration Guidelines.12 Using Medical Subject Headings

for topics on music therapy and cancer patients, we searched the OVID system, the EBSCO Host system (which contains databases such as the Cochrane Library, Medline, PubMed, CINAHL, ProQuest, SCOPUS, PsycARTICLES, and PsycINFO), the International Index to Music Periodicals, and Google for experimental or quasi-experimental quantitative studies on this topic. The key words and strategies used in literature search were Exp Neoplasms, cancer OR neoplasm OR malignant OR carcinoma OR tumor, and music OR melody. In total, 367 studies were eventually identified (Figure).

Selection Criteria By consulting other meta-analyses21Y24 and its own objectives, inclusion criteria for the present analysis were derived: (1) published between January 2002 and December 2012 in any language; (2) used music therapy as an intervention in cancer patients; (3) used quantitative methods to assess results; (4) used quasi-experimental or experimental designs in which music therapy was an experimental intervention and regular care or activities were used as controls, excluding other interventions; (5) reported statistical information sufficient to describe the results of music therapy, such as means, standard deviations, mean differences, sample sizes, t values, f values, or P values for an effectsize analysis; and (6) evaluated the effects of music intervention on anxiety, depression, pain, or fatigue. Qualitative studies, duplicate publications, and single cases or single-group experimental studies were excluded. In total, 21 studies were eventually identified (Figure).

Figure n Preferred reporting items for systematic reviews and meta-analyses (PRISMA) 2009 flow diagram.

Effectiveness of Music Intervention in Cancer Patients

Cancer NursingTM, Vol. 37, No. 6, 2014

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

n

E37

E38 n Cancer NursingTM, Vol. 37, No. 6, 2014

Tsai et al

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1. ML: distraction RCT intervention Blinding: Y 2. UC Allocation concealment: Y

ML+ support therapy/ RCT support therapy Blinding: Y Allocation concealment: Y

Kwekkeboom (2003)

Li (2004)

RCT Blinding: Y Allocation concealment: yes

ML/UC

RCT Blinding: nil Allocation concealment: Y

Design

Cassileth et al (2003)

Group music intervention Smolen et al (2002) ML/UC

Study

Intervention (Experimental/ Control)

Table 1 & Characteristics of Included Studies (n = 21)

Gastric cancer/stages II-IIIa

Cancer/Y

Hematologic malignancies/Y

Colorectal cancer/Y

Diagnosis/Severity

Intervention Characterization

Music selection: patient preferred Therapy type: passive Each group size (people): individual Each group length (min): Y Treatment times: Y Frequency (times/day): Y Music therapist: health professional (physician) Total n = 69 Music selection: Complete N(T/C): 36/33 therapist selected Mean age: 52 y Therapy type: auto + passive Setting: inpatient Each group size (people): individual Therapy: HDT/ASCT Each group length: 20Y30 min Treatment times: Y Frequency (times/day): varied Music therapist: certified professional Total n = 58 Music selection: patient Complete N(T/C): preferred 24/14/20 Therapy type: passive Mean age: 53.59 y Each group size (people): individual Setting: outpatient Each group length: 5Y15 min Therapy: tissue biopsy or Treatment times: Y vascular port placement Frequency (times/day): Y Music therapist: health professional (nurse) Total n = 60 Music selection: patient preferred Complete N(T/C): 30/30 Therapy type: passive Each group size (people): Mean age: 53 y individual Setting: inpatient Therapy: surgery Each group length: 20Y30 min Treatment times: 10 Frequency (times/day): 2 Music therapist: Y

Total n = 32 Complete N(T/C): 16/16 Mean age: 59.84 y Setting: inpatient Therapy: colonoscopy

Participants

(1,1,2,2,0)/3

Study Quality (Jadad Score)

Anxiety (SAS)

Anxiety (STAI-S) Pain (VAS)

(continues)

(1,1,2,2,0)/1

(1,1,2,2,0)/2

(1,1,2,2,0)/3 Anxiety/ depression/ fatigue (POMS)

Anxiety (SAI) Physiological (BP, HR)

Outcome

Effectiveness of Music Intervention in Cancer Patients

Cancer NursingTM, Vol. 37, No. 6, 2014

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

n

E39

ML/UC

Clark et al (2006)

RCT Blinding: Y

Ferrer (2007)

ML + S/UC

RCT Blinding: yes Allocation concealment: yes

Hanser et al (2006) ML + AMI/UC + support therapy

RCT Blinding: Y Allocation concealment: Y

1. ML RCT /1. MM Blinding: Y 2. Not interested Allocation in music therapy concealment: Y

Design

Burns et al (2005)

Study

Intervention (Experimental/ Control)

Nil

Breast cancer/ stage IV

Cancer/Y

cancer/Y

Diagnosis/Severity

Table 1 & Characteristics of Included Studies (n = 21), Continued Intervention Characterization

Music selection: patient preferred Complete N(T/C): 25/25 Therapy type: auto

Total n = 50

Music therapist: certified professional Total n = 70 Music selection: therapist selected Complete N(T/C): 35/35 Therapy type: auto + passive Mean age (y): Y Each group size (people): Setting: inpatient individual Each group length: 45 min Therapy: Y Treatment times: 3 Frequency (times/day): Y Music therapist: certified professional

Music selection: patient preferred Therapy type: passive Each group size (people): individual Each group length (min): Y Treatment times: 2 Frequency (times/day): Y Music therapist: certified professional Total n = 63 Music selection: patient preferred Complete N(T/C): 35/28 Therapy type: passive Each group size (people): Mean age: 57.66 y individual Setting: inpatient Each group length: 45 min Therapy: RT Treatment times: Y Frequency (times/day): Y

Total n = 65 Complete N(T/C): ML: 44; MM: 11/10 Mean age: 53.03 y Setting: outpatient Therapy: CT (80%)

Participants

Study Quality (Jadad Score)

(1,1,2,2,1)/3

(1,1,2,2,0)/3

(continues)

Comfort/relaxation/ (1,1,2,2,0)/2 anxiety/fatigue (VAS) Physiological (BP, HR)

Quality of life (FACT-G) Anxiety/depression (HADS) Physiological (BP, HR) Comfort/relaxation (VAS)

Anxiety/depression (HADS) Fatigue (POMS) Pain/distress (pain numeric rating scale/distress numeric rating scale)

Anxiety (STAI) (1,1,2,2,0)/2 Positive/negative affect (PANAS) Fatigue (FACIT-F)

Outcome

E40 n Cancer NursingTM, Vol. 37, No. 6, 2014

Tsai et al

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

MI + standard care/ RCT standard care Blinding: Y Allocation concealment: Y

Burns et al (2008)

RCT Blinding: Y

RCT Blinding: Y Allocation concealment: Y

ML/no music

Zhao et al (2008)

Bulfone et al (2009) ML/standard care

N-RCT Blinding: nil Allocation concealment: Y

Allocation concealment: Y

Design

Kemper et al (2008) UC + rest + ML/ UC + Rest

Study

Intervention (Experimental/ Control)

Breast cancer/stage I-II

Leukemia and high-grade non-Hodgkin lymphoma/Y

Cancer/Y

Cancer/Y

Diagnosis/Severity

Table 1 & Characteristics of Included Studies (n = 21), Continued Intervention Characterization

Each group size (people): 25 Each group length: 20 min Y Treatment times: 3 Frequency (times/day): Music therapist: health professional (researcher) Total n = 63 Music selection: Complete N(T/C): Y therapist selected Mean age: 9 y Therapy type: passive Setting: outpatient Each group size (people): Y Therapy: Y Each group length: 20 min Treatment times: Y Frequency (times/day): Y Music therapist: Y Total n = 95 Music selection: patient Complete N(T/C): 49/46 preferred Mean age: 54.13 y Therapy type: passive Setting: inpatient Each group size (people): Therapy: RT individual Each group length: 30 min Treatment times: Y Frequency (times/day): Y Music therapist: Y Total n = 49 Music selection: patient Complete N(T/C): 25/24 preferred Mean age: 54 y Therapy type: auto Setting: inpatient Each group size (people): individual Therapy: CT Each group length: 45 min Treatment times: 8 Frequency (times/week): 2 Music therapist: certified professional Total n = 60 Music selection: patient Complete N(T/C): 30/30 preferred Mean age: 50.95 y Therapy type: passive

Mean age: 55 y Setting: outpatient Therapy: CT

Participants

Study Quality (Jadad Score)

Anxiety (STAI)

Positive/negative affect (PANAS) Fatigue (FACIT-F)

Anxiety (STAI)

Anxiety (SAS) (HAMA) Physiological (BP, HR, RR)

(continues)

(1,1,2,2,0)/2

(1,1,2,2,0)/2

(1,1,2,2,0)/1

Comfort/relaxation/ (1,1,2,2,0)/1 anxiety/fatigue (VAS) Physiological (HRV)

Outcome

Effectiveness of Music Intervention in Cancer Patients

Cancer NursingTM, Vol. 37, No. 6, 2014

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

n

E41

CAT/volunteer’s attention

ML/UC

Shabanloei et al (2010)

N-RCT Blinding: no

RCT Blinding: Y Allocation concealment: Y

RCT Blinding: yes Allocation concealment: yes

Nguyen et al (2010) ML/standard care

Madden et al (2010)

RCT Blinding: nil Allocation concealment: nil

Allocation concealment: Y

Design

Huang et al (2010) ML/UC

Study

Intervention (Experimental/ Control)

Blood disorder/ solid tumors/Y

Cancer/Y

Leukemia/Y

Cancer/stages I-IV

Diagnosis/Severity

Table 1 & Characteristics of Included Studies (n = 21), Continued Intervention Characterization

Study Quality (Jadad Score)

Pain (VAS) Anxiety (STAI) Physiological (BP, HR, RR)

(continues)

(1,1,2,2,0)/2

(1,1,2,2,0)/2

(1,1,2,2,0)/5

Sensation/distress (1,1,2,3,0)/3 (VAS) Pain (oral numerical scale)

Outcome

Music selection: patient Quality of life preferred (Peds QL) Therapy type: auto + passive Each group size (people): Y Each group length: 60 min Treatment times: Y Frequency (times/week): 3 Music therapist: certified professional Total n = 100 Music selection: therapist Anxiety (STAI) Complete N(T/C): 50/50 selected Mean age: 33.46 y Therapy type: passive

Total n = 48 Complete N(T/C): Y Mean age (y): Y Setting: outpatient Therapy: CT/medication infusion/blood product transfusions

Each group size (people): individual Each group length: 15 min Treatment times: 1 Frequency (times/day): no Music therapist: Y Total n = 126 Music selection: patient Complete N(T/C): 62/64 preferred Mean age (y): Y Therapy type: passive Setting: inpatient Each group size (people): Therapy: RT/CT individual Each group length: 30 min Treatment times: Y Frequency (times/day): Y Music therapist: researcher Total n = 40 Music selection: patient Complete N(T/C): 20/20 preferred Mean age: 8.8 y Therapy type: passive Setting: inpatient Each group size Therapy: lumbar puncture (people): individual Each group length: 23.1 min Treatment times: Y Frequency (times/day): Y ` Music therapist: U

Setting: inpatient Therapy: CT

Participants

E42 n Cancer NursingTM, Vol. 37, No. 6, 2014

Tsai et al

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

ML/1. RX, 2. UC RCT Blinding: Y

Lin et al (2011)

Allocation concealment: Y

Allocation concealment: Y

RCT Blinding: no

ML/routine nursing care

Li et al (2011)

RCT Blinding: Y Allocation concealment: Y

ML/routine care

Allocation concealment: Y

Design

Fengjuan and You (2011)

Study

Intervention (Experimental/ Control)

Cancer/stages I-IV

Breast cancer/Y

Hepatoma/Y

Diagnosis/Severity

Table 1 & Characteristics of Included Studies (n = 21), Continued

Each group size (people): Y

Intervention Characterization Outcome

(1,1,2,2,0)/2

(1,1,2,2,0)/2

Study Quality (Jadad Score)

(continues)

Total n = 98 Music selection: Anxiety (C-STAI) (1,1,2,2,0)/3 Complete N(T/C): therapist selected MT: 34; RX: 30; UC: 34 Therapy type: auto + passive Behavior state (RBSS) Each group size (people): 34 Immediate anxiety Setting: inpatient Each group length: 60 min (EVAS) Therapy: CT Physiological (HR, Frequency (times/day): Y temperature) Music therapist: Y Mean age: 52.93 y Treatment times: 1

Each group length: 10Y20 min Treatment times: Y Frequency (times/day): Y Music therapist: Y Total n = 102 Music selection: patient Anxiety (STAI) Complete N(T/C): 50/52 preferred Physiological Mean age: 56.71 y Therapy type: auto + passive (BP, HR, RR) Setting: inpatient Each group size (people): individual Therapy: TACE Each group length: 65 min Treatment times: Y Frequency (times/day): Y Music therapist: health professional (physician) Total n = 120 Music selection: Pain (SF-MPQ) Complete N(T/C): 60/60 patient preferred Mean age: 45.01 y Therapy type: passive Each group size (people): individual Setting: inpatient Each group length: 30 min Therapy: radical Treatment times: Y mastectomy Frequency (times/day): 2 Music therapist: researcher

Setting: inpatient Therapy:

Participants

Effectiveness of Music Intervention in Cancer Patients

Cancer NursingTM, Vol. 37, No. 6, 2014

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

n

E43

ML+ routine nursing RCT Blinding: nil care/routine Allocation nursing care concealment: Y

Li et al (2012)

Breast cancer/Y

Diagnosis/Severity

Therapy type: passive

Complete N(T/C): 60/60 Mean age (y): Y

Each group size (people): individual Setting: inpatient Each group length: 30 min Therapy: CT Treatment times: Y Frequency (times/day): 2 Total n = 120 Music therapist: researcher Complete N(T/C): 60/60 Music selection: patient Mean age: 45 y preferred Setting: inpatient Therapy type: passive Therapy: radical Each group size (people): mastectomy individual Each group length: 30 min Treatment times: Y Frequency (times/day): 2 Music therapist: researcher

Music selection: patient preferred

Intervention Characterization

Total n = 120

Participants

Study Quality (Jadad Score)

Anxiety (SAI)

(1,1,2,2,0)/2

Depression (ZSDS) (1,1,2,2,0)/5

Outcome

Abbreviations: AMI, active music involvement; BP, blood pressure; CAT, creative arts therapy; C-STAI, Chinese State-Trait Anxiety Inventory; CT, chemotherapy treatment; EVAS, Emotional Visual Analog Scale; FACIT-F, Function Assessment of Chronic Illness TherapyYFatigue Scale; FACT-G, Functional Assessment of Cancer TherapyYGeneral; HAD, Hospital Anxiety and Depression Scale; HAMA, Hamilton Anxiety Scale; HDT/ASCT, high-dose therapy with autologous stem cell transplantation; HR, heart rate; HRV, heart rate variability; MI, music imagery; ML, music listening; MM, music making; MT, music therapy; N-RCT, Non-Randomized Controlled Trial; PANAS, Positive Affect and Negative Affect Schedules; Peds QL, Pediatric Oncology Quality of Life Inventory; POMS, Profile of Mood States; RBSS, Resting Behavioral State Scoring System; RCT, randomized controlled trial; RR, respiration rate; RT, audio therapy treatment; RX, relaxation; S, singing; SAI, Spielberger’s State Anxiety Inventory; SAS, Zung’s Self-rating Anxiety Scale; SF-MPQ, Chinese version of the Short-Form of McGill Pain Questionnaire; STAI: State-Trait Anxiety Inventory; T/C, treatment group/control group; TACE, transcatheter arterial chemoembolization; temperature, body temperature; UC, usual care; VAS, Visual Analog Scale; ZSDS, Zung Self-rating Depression Scale.

ML+ routine nursing Blinding: yes care/routine nursing care Allocation concealment: Y

Design

Zhou et al (2011)

Study

Intervention (Experimental/ Control)

Table 1 & Characteristics of Included Studies (n = 21), Continued

Data Extraction To ensure a reliable analysis and prevent subjective sampling errors, 2 analysts independently handled 2 stages of data abstraction: inclusion of studies and recording the variables of key study characteristics. When basic data were analyzed, we grouped items if they were similar. We deleted certain outcomes (quality of life and physiological) or characteristic variables (gender, frequency of therapy, and theory) if they did not appear in most articles. Consequently, we summarized intervention characteristics (music selection, therapy form, therapy type, group size, therapy time, and nature of the control group), participant characteristics (eg, mean age, diagnosis, and patient setting), and therapist characteristics (treatment provider) for all included characteristics (Table 1).

Analytic Approach 1. In this study, outcome variables were sorted according to their property so several effect sizes on different outcomes were obtained, rather than a single effect size per study. 2. We calculated the average effect size of a single variable as the analytic unit for overall effect size.

Quality Assessment This study adapted Cochrane Collaboration Guidelines’ study quality assessment tools as described in a study by Brodaty et al.25 The quality assessment table included 5 aspects: study design, study subjects, outcome measurement, statistical analysis, and study results (Table 1). Studies with a total score of 6 or more out of 10 were included in our meta-analysis. Agreement between raters was assessed with the . statistic. In case of inconsistencies, the researcher and the collaborative rater further discussed the issue to reach a consensus. In terms of reliability, the . value between the researcher and other expert was found to be 0.90.

Risk of Bias across Studies Publication bias occurs as editors tend to accept and publish studies that report significant results. We used a funnel plot and the Egger regression intercept to examine publication bias. Publication bias is unlikely if the funnel plot appears as a symmetrical inverted funnel shape and the Egger regression intercept, as determined by the P value, was 9.05. In the presence of

a potential publication bias or file-drawer effect, the funnel plot would be symmetrical and the P value of the Egger regression intercept would be G.05. A further search for missing studies was recommended.26,27 A sensitivity analysis was also conducted to examine whether the overall effect size had been unduly affected by a single study.

Data Synthesis and Statistical Analysis In this study, comprehensive meta-analysis software version 2 (Biostat, Englewood, New Jersey) was used to calculate effect sizes and systemic analysis. We used Hedges as the measure of the effect qgffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2

X jX

2

þðNC j1ÞðSDC Þ B C s i z e : g ¼ SD ,; SDpooled ¼ ðNE j1ÞðSDNEEÞþN pooled C j2 where SDpooled is the pooled standard deviation between the experimental and control groups, SDE is the standard deviation of the experimental groups, and NE and NC are the respective sample sizes of the experimental and control groups. The pooled standard deviation, SDpooled, was used to standardize betweengroup difference in average values. Hedges and Olkin28 referred to this unbiased estimator, Hedges g, as Cohen d, but it is not the same as Cohen d. The exact form for the correction factor, 3 J, involves the gamma function: g = J  d, J ¼ 1j 4dfj1 . Homogeneity was assessed with Cochrane Q. The presence of homogeneity (a nonsignificant Q value) suggests that betweenstudy differences result from sampling errors. The absence of homogeneity between effect sizes (a significant Q value) suggests that between-study differences cannot entirely be attributed to sampling error but that other factors might be in effect, and further analysis of potential influences is required. Higgins and Thompson29 proposed a tentative classification of I 2 values with the purpose of helping to interpret its magnitude. Thus, percentages of around 25% (I 2 = 25), 50% (I 2 = 50), and 75% (I 2 = 75) would mean low, medium, and high heterogeneity.

Additional Analyses To understand the influences of characteristics of music therapy on the effect size and categorical variables, we used a subgroup analysis to identify characteristics that led to a more prominent outcome. Variables were examined with the mixed-effects model, which is a concept that covers the fixed-effect model and the random-effect model. After stratification, subgroups could have a sample size of 5 or smaller. In the subgroup analysis, when QB is significant, the variable is believed to have an effect on effect size and may be a moderator variable. In the present study, we

Table 2 & Effect Size of Music Interventions on Improving Symptoms in Cancer Patients Null Hypothesis Tests (2 Tailed) Symptom

No. of Studies

Anxiety Depression Pain Fatigue

17 8 6 5

Hedges g (95% CI) j0.553 j0.510 j0.656 j0.422

(j0.716 (j0.681 (j1.016 (j0.669

to to to to

j0.389) j0.340) j0.295) j0.175)

Homogeneity Test

Z

P

Q value

P

I 2, %

C2

j6.621 j5.861 j3.566 j3.344

.001 .001 .001 .001

29.295 8.621 14.313 5.597

.02 .28 .01 .23

45.383 18.806 65.066 28.528

0.051 0.014 0.124 0.032

P values of 9 .001 were rounded to 2 digits. Abbreviation: CI, confidence interval.

E44 n Cancer NursingTM, Vol. 37, No. 6, 2014

Tsai et al

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Table 3 & Overall Effect Size of Improvement in Anxiety (n = 17) Statistics for Each Study Study Name Smolen et al (2002) Cassileth et al (2003) Kwekkeboom (2003) Li (2004) Burns et al (2005) Clark et al (2006) Hanser et al (2006) Ferrer (2007) Kemper et al (2008) Zhao et al (2008) Burns et al (2008) Bulfone et al (2009) Nguyen et al (2010) Shabanloei et al (2010) Jin and Zhao (2011) Lin et al (2011) Li et al (2012)

Hedges g

Lower Limit

j0.432 j0.667 j0.147 j0.850 j0.770 j0.305 j0.144 j0.760 j0.063 j0.353 j0.667 j1.070 j0.073 j0.404 j1.215 j0.502 j0.747 j0.553

j1.115 j1.196 j0.731 j1.372 j1.462 j0.798 j0.739 j1.326 j0.625 j0.038 j1.244 j1.605 j0.681 j0.797 j1.635 j0.980 j1.115 j0.716

Upper Limit 0.252 j0.159 0.436 j0.328 j0.078 0.189 0.450 j0.195 0.499 0.069 j0.011 j0.535 0.534 j0.011 j0.795 j0.025 j0.379 j0.389

Z

P

Hedges g and 95% Confidence Interval

j1.238 .216 j2.562 .010 j0.494 .621 j3.190 .001 j2.182 .029 j1.209 .227 j0.476 .634 j2.634 .008 j0.219 .826 j2.435 .015 j2.340 .019 j3.919 .000 j0.237 .813 j2.016 .044 j5.672 .000 j2.062 .039 j3.980 0.000 j6.621 0.000

Random-effect model Music

performed a subgroup analysis and meta-regression analysis to determine any potential moderating variables.

n

Results

Selection and Characteristics of Studies In total, 21 studies were included in the outcome assessments. Outcome assessments included 17 studies that assessed anxiety, 8 that assessed depression, 6 that assessed pain, and 5 that assessed fatigue outcomes. The age of participants ranged from 8 to 57 years. The studies had quality scores of 6 to 7. Intervention characteristics were as follows: (1) the group size was mostly individual (16/21); (2) music selection was mostly by patients (17/21); and (3) music therapy type was mostly passive (13/21) (Table 1).

Control

Effect Sizes for Anxiety, Depression, Pain, and Fatigue Outcomes Overall, the results of effect sizes revealed that music therapy significantly reduced anxiety, depression, pain, and fatigue in cancer patients (Table 2). Seventeen studies on anxiety were included in our analysis, and the results showed that music therapy moderately but significantly reduced anxiety, with an overall effect size of j0.553 (95% CI, j0.716 to j0.398; Table 2). The effect size of each study was negative, meaning that the presence of music therapy was associated with reduced anxiety (Hedges g = j0.073 to j1.215; Table 3). There was heterogeneity between samples and effect sizes (Q = 29.295, P = .02, I2 = 45.383%). A symmetrical funnel plot and the P value of the Egger regression intercept (.90) suggested no publication bias.27,29 The sensitivity analysis

Table 4 & Overall Effect Size of Improvement in Depression (n = 8) Statistics for Each Study Study Name Cassileth et al (2003) Burns et al (2005) Clark et al (2006) Hanser et al (2006) Kemper et al (2008) Burns et al (2008) Huang et al (2010) Kai-na et al (2011)

Hedges g

Lower Limit

j0.232 j0.736 j0.259 j0.151 j0.163 j0.677 j0.862 j0.787 j0.510

j0.737 j1.427 j0.755 j0.746 j0.726 j1.244 j1.019 Y1.157 j0.681

Upper Limit 0.274 j0.045 0.237 0.444 0.401 j0.110 j0.306 j0.418 j0.340

Z

P

j0.898 j2.088 j1.024 j0.496 j0.566 j2.340 j3.640 j4.179 j5.861

.369 .037 .306 .620 .572 .019 .000 .000 .000

Hedges g and 95% Confidence Interval

Fixed-effect model Music

Effectiveness of Music Intervention in Cancer Patients

Control

Cancer NursingTM, Vol. 37, No. 6, 2014

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

n

E45

Table 5 & Overall Effect Size of Pain Relief (n = 6) Statistics for Each Study Study Name Kwekkeboom (2003) Clark et al (2006) Huang et al (2010) Nguyen et al (2010) Madden et al (2010) Li et al (2011)

Hedges g

Lower Limit

Upper Limit

Z

P

j0.054 j0.273 j0.676 j0.716 j1.142 j1.180 j0.656

j0.637 j0.767 j1.033 j1.344 j2.149 j1.566 j1.016

0.529 0.220 j0.319 j0.089 j0.134 j0.795 j0.295

j0.181 j1.087 j3.710 j2.238 j2.221 j5.999 j3.566

.856 .277 .000 .025 .026 .000 .000

Hedges g and 95% Confidence Interval

Random-effect model Music

indicated that removal of any study from the study pool did not affect the overall result. The 8 studies on depression had an overall effect size of j0.510 (95% CI, j0.681 to j0.340; Table 2), suggesting that music therapy moderately and significantly reduced depression in cancer patients. The effect size for each study had a negative direction with Hedges g (range, j0.151 to j0.787; Table 4). There was homogeneity among studies (Q = 8.621, P = .28, I2 = 18.806%). The symmetrical funnel plot combined with the P value of the Egger regression intercept (.19)27,29 suggest no publication bias. Removing any study from the analysis did not affect the result. Six studies on pain were included in our analysis, with an overall effect size of Hedges g = j0.656 (95% CI, j1.016 to about j0.295; Table 2), indicating that music therapy moderately but significantly reduced pain. The effect size for each study had a negative direction, with Hedges g ranging from j0.054 to j1.180 (Table 5). The heterogeneity of the studies (Q = 14.313, P = .01, I 2 = 65.066%) could have resulted from sampling error, differences in study designs, or other factors. There was no publication bias as suggested by the symmetrical funnel plot and the P value of the Egger regression intercept (.72).27,29 Removing any study from the analysis did not affect the result, as shown by sensitivity analysis. The effect size of each of the 5 studies reporting results for fatigue had a negative direction, with Hedges g ranging from j0.036 to j0.987 (Table 6). The overall effect size was small (Hedges g = j0.422; 95% CI, j0.669 to j0.175; Table 2).

Control

There was homogeneity among samples (Q = 5.597, P = .23, I 2 = 28.528%). The funnel plot was symmetrical, and the P value of the Egger regression intercept was 9.5 (.55),27,29 suggesting no publication bias. Again, the sensitivity analysis indicated that removing any study from the analysis did not affect the result.

Subgroup Analyses of Anxiety, Depression, Pain, and Fatigue Outcomes Subgroup analyses demonstrated a remarkable relationship between age, music preference, and the effect size for anxiety. It was significantly larger in adults than in children or adolescents (Hedges g = j0.606 vs j0.068; P G .001 vs P = .748). Absence of homogeneity was statistically significant (P = .02), suggesting that heterogeneity might have been a result of age. The effect size was greater when music was chosen by patients rather than by researchers (Hedges g = j0.631 vs j0.322). Again, patients’ preference had a greater effect on outcome indicators than did that of the researchers (P = .04). Subgroup analyses were conducted on other outcomes, and none of them was statistically significant (Tables 7 and 8). n

Discussion

The aim of this study was to determine the effectiveness of music therapy in ameliorating anxiety, depression, pain, and

Table 6 & Overall Effect Size of Fatigue Improvement (n = 5) Statistics for Each Study Study Name Cassileth et al (2003) Burns et al (2005) Clark et al (2010) Ferrer (2007) Burns et al (2008)

Hedges g

Lower Limit

Upper Limit

Z

P

j0.036 j0.274 j0.094 j0.496 j0.987 j0.422

j0.874 j0.952 j0.585 j1.050 j1.572 j0.669

0.142 0.405 j0.397 j0.058 j0.402 j0.175

j1.412 j0.791 j0.376 j1.754 j3.308 j3.344

.158 .429 .707 .079 .001 .001

Hedges g and 95% Confidence Interval

Fixed-effect model Music

E46 n Cancer NursingTM, Vol. 37, No. 6, 2014

Control

Tsai et al

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Effectiveness of Music Intervention in Cancer Patients

Cancer NursingTM, Vol. 37, No. 6, 2014

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

n

E47

Abbreviation: CI, confidence interval. a Subgroup effect on outcome variable. b Heterogeneity between subgroup (moderator).

Study and method characteristics Study quality Good (e6 points) High (Q7 points) Study design Quasi-experimental Experimental control Control Music therapy vs pure control Music therapy vs nonpure care Therapy characteristics Form Group Individual Music selector Patient Researcher Treatment time e30 min 930 min Undisclosed Type Passive Nonpassive Therapist characteristics Therapist background Music therapist Nonmusic therapist Participant characteristics Age Children/adolescents (6Y17 y) Adults (18Y64 y) .001 .634 .826 .001 .001 .001

.001 .001 .001 .01 .001 .003 .016 .001 .001

.001 .001

.748 .001

j0.063 (j0.625 to 0.499) j0.580 (j0.743 to 0.416) j0.536 (j0.718 to j0.354) j0.850 (j1.372 to j0.328)

j0.431 (j0.675 to j0.188) j0.590 (j0.791 to j0.389) j0.631 (j0.823 to j0.439) j0.322 (j0.566 to j0.079) j0.521 (j0.719 to j0.323) j0.590 (j0.981 to j0.200) j0.599 (j1.085 to j0.113) j0.478 (j0.663 to j0.293) j0.691 (j0.987 to j0.395)

j0.498 (j0.749 to j0.247) j0.568 (j0.777 to j0.358)

j0.068 (j0.481 to 0.345) j0.606 (j0.769 to j0.442)

1 16 16 1

4 13 13 4 10 5 2 11 6

5 12

2 15

Pa

j0.573 (j0.740 to j0.407) j0.144 (j0.739 to 0.450)

Hedges g (95% CI)

16 1

No. of studies

Anxiety

Table 7 & Subgroup Analyses on Anxiety and Depression

5.636

0.173

1.423

0.155

3.809

0.965

1.306

2.988

1.851

QB

.02

.68

.23

.93

.04

.33

.52

.08

.17

Pb

1 7

5 3

5 3

4 3 1

6 2

1 7

7 1

1 7

6 2

No. of studies

j0.163 (j0.726 to 0.401) j0.548 (j0.727 to j0.368)

j0.381 (j0.631 to j0.131) j0.626 (j0.860 to j0.393)

j0.578 (j0.780 to j0.376) j0.349 (j0.668 to j0.030)

j0.557 (j0.769 to j0.345) j0.358 (j0.675 to j0.042) j0.770 (j1.462 to j0.078)

j0.587 (j0.775 to j0.399) j0.157 (j0.566 to 0.252)

j0.163 (j0.726 to 0.401) j0.548 (j0.727 to j0.368)

j0.468 (j0.662 to j0.273) j0.662 (j1.019 to j0.306)

j0.163 (j0.726 to 0.401) j0.548 (j0.727 to j0.368)

j0.506 (j0.711 to j0.300) j0.527 (j0.833 to j0.221)

Hedges g (95% CI)

Depression

.572 .001

.003 .001

.001 .032

.001 .026 .029

.001 .452

.572 .001

.001 .001

.572 .001

.001 .001

Pa

7.124

6.78

7.343

7.139

5.244

7.124

0.882

7.124

8.741

QB

.31

.34

.29

.21

.51

.31

.35

.31

.19

Pb

E48 n Cancer NursingTM, Vol. 37, No. 6, 2014

Tsai et al

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Abbreviation: CI, confidence interval. a Subgroup effect on outcome variable. b Heterogeneity between subgroup (moderator).

Study and method characteristics Study quality Good (e6 points) High (Q7 points) Study design Quasi-experimental Experimental control Control Music therapy vs pure control Music therapy vs nonpure care Therapy characteristics Form Group Individual Music selector Patient Researcher Treatment time e30 min 930 min Undisclosed Type Passive Nonpassive Therapist characteristics Therapist background Music therapist Nonmusic therapist Participant characteristics Age Children/adolescents (6Y17 y) Adults (18Y64 y) .001 .038 .001

.026 .002 .001

.002 .157

.002 .026

.001 .277

.002 .016

j0.656 (j1.016 to j0.295) j0.576 (j1.119 to j0.032) j0.728 (j1.064 to j0.391)

j1.142 (j2.149 to j0.134) j0.608 (j0.996 to j0.219) j0.656 (j1.016 to j0.295)

j0.690 (j1.128 to j0.252) j0.592 (j1.412 to 0.228)

j0.608 (j0.996 to j0.219) j1.142 (j2.149 to j0.134)

j0.273 (j0.767 to j0.220) j0.740 (j1.134 to j0.347)

j0.835 (j1.368 to 0.303) j0.579 (j1.049 to j0.109)

0 6 4 2

1 5 6 0 4 2

5 1

1 5

2 4

.009 .001

P

a

j0.650 (j1.135 to j0.166) j0.676 (j1.033 to j0.319)

Hedges g (95% CI)

Pain

5 1

No. of Studies

Table 8 & Subgroup Analyses on Pain and Fatigue

0.499

2.103

0.94

0.042

0.94

0.218

0.007

QB

.48

.15

.33

.84

.33

.64

.93

P

b

0 5

4 1

2 3

2 2 1

5 0

1 4

5 0

0 5

5 0

No. of Studies

j0.486 (j0.734 to j0.238)

j0.484 (j0.761 to j0.207) j0.496 (j1.050 to 0.058)

j0.321 (j0.721 to 0.080) j0.589 (j0.904 to j0.273)

j0.425 (j0.800 to j0.051) j0.463 (j0.839 to j0.087) j0.770 (j1.462 to j0.078)

j0.486 (j0.734 to j0.238)

j0.496 (j1.050 to 0.058) j0.484 (j0.761 to j0.207)

j0.486 (j0.734 to j0.238)

j0.486 (j0.734 to j0.238)

j0.486 (j0.734 to j0.238)

Hedges g (95% CI)

Fatigue

.001

.001 .079

.117 .001

.026 .016 .029

.001

.079 .001

.001

.001

.001

Pa

6.128

5.065

5.367

6.128

QB

.11

.17

.07

.11

Pb

fatigue in cancer patients according to Cohen’s guidelines (d values of 0.2Y0.4 indicating a small effect size, 0.5Y0.7 indicating a medium effect size, and Q0.8 indicating a large effect size). This study adopted the Cochrane Collaboration Guidelines’ study quality assessment tools to evaluate the quality of studies. Only studies published within the last decade with a total score of 6 or more out of 10 were included in our metaanalysis. Therefore, the quality of studies in our analysis was higher than that of Bradt et al.12 Our meta-analysis showed that music interventions were moderately ameliorating cancer patients’ pain, anxiety, depression, and fatigue.

was also a confounder. Music selected by patients had a better effect size compared with that selected by researchers. There are some limitations to the present study. First, because of a restriction in length and data display, statistical analyses in individual studies were often incomplete or inconsistent in format. After stratification, the number of studies in each variable subgroup was even smaller. When that occurs, studies with small sample sizes may or may not be representative of the overall effect of the subgroup.

n

Comparison of Outcome Variables Music therapy reduced cancer patients’ pain, corresponding to the reports by Bradt et al12 and Zhang et al.20 Music interventions can stimulate the release of endorphins into the bloodstream.30 In addition, they can also divert cancer patients’ attention and thereby reduce pain. In this study, the effect sizes of music therapy had a significantly negative medium effect size on anxiety in patients with cancer, which was consistent with previous studies.31,32 A review by Bradt et al12 in 2011 also found a medium or greater effect size of music therapy on anxiety, which corresponded with our results. Therefore, music interventions help relax patients and further reduce their anxiety throughout the treatment course. This meta-analysis showed that music interventions had a small, negative effect size on fatigue in cancer patients, consistent with the review by Bradt et al.12 This suggests that music therapy may have limited use in alleviating fatigue in cancer patients. Our analysis of 8 trials revealed that music interventions had a significantly negative medium effect size on depression in patients with cancer. Similar effects had been shown in breast cancer patients after surgery and in pediatric cancer outpatients.13,14 Recently, Zhang et al20 meta-analyzed 7 moderate-quality studies and demonstrated that music reduced depression. However, there were no effects of music therapy or music medicine on depression in the study by Bradt et al.12 Such inconsistency could be caused by the limited numbers and quality of studies included. Overall, our findings and those of others suggest that cancer patients experiencing pain and depression can significantly benefit from music therapy, in addition to relieving anxiety.

Subgroup Comparisons Participants’ characteristics, therapists’ characteristics, and advancement of study design and intervention protocol over the past decade may account for differences in the effects of music interventions. The results of subgroup analyses indicated that factors such as age, gender, culture, education, and interests should be considered when planning music intervention. Patients should be allowed to decide on the listening time and method. Age had a significant impact on the effectiveness of music intervention in reducing anxiety, with adults benefitting more than children and adolescents. Furthermore, the person who selected the music

Effectiveness of Music Intervention in Cancer Patients

Conclusions

Our study confirmed that music interventions can reduce anxiety, depression, pain, and fatigue in cancer patients. The results of subgroup analysis suggested that music interventions were more effective in adults than in children and adolescents and were more effective when patients, rather than researchers, chose the music. The results of the moderator analysis that empirically supported treatments can be implemented in the clinical setting. Through these research efforts, advocates will be able to use the data to support the use of music interventions in the clinical setting and allow more patients with cancer to benefit from this service. For patients with cancer, medication alone is insufficient to correct patient’s anxiety, depression, pain, and fatigue. The implication of the study is that music interventions can be a good nonmedication therapy to reduce anxiety, depression, pain, and fatigue in clinical settings. Our findings provide important information for future music group planners to improve the design and process to better benefit patients throughout these programs. A better understanding of the factors that influence the effectiveness of music interventions in patients with cancer can serve as a knowledge base for successful research design in the future.

References 1. The Republic of China Year Book. 2012. http://www.ey.gov.tw/en/cp .aspx?n=CBDA2319F19402AE. Accessed October 30, 2012. 2. American Cancer Society. Cancer Facts & Figures. Atlanta, GA: American Cancer Society; 2010. 3. Kwekkeboom KL. Music versus distraction for procedural pain and anxiety in patients with cancer. Oncol Nurs Forum. 2003;30(3):433Y440. 4. Deng G, Cassileth BR. Integrative oncology: complementary therapies for pain, anxiety, and mood disturbance. CA Cancer J Clin. 2005;55(2): 109Y116. 5. Shabanloei R, Golchin M, Esfahani A, Dolatkhah R, Rasoulian M. Effects of music therapy on pain and anxiety in patients undergoing bone marrow biopsy and aspiration. AORN J. 2010;91(6):746Y751. 6. Advance for Nurses. The Sound of Therapy. 2013. http://nursing.advanceweb .com/Features/Articles/Sound-Therapy.aspx. Accessed October 14, 2013. 7. Nedelut N, Tudor EV. Music therapy in fighting certain behavior disorders. REVART. 2012;1:5Y16. 8. Bruscia K. The fundamentals of music therapy practice. In: Kenneth B, ed. Case Studies in Music Therapy, 1st ed. Gilsum, NH: Barcelona Publishers; 1996:3Y17. 9. American Music Therapy Association. What is music therapy. 2008. http:// www.musictherapy.org/about/musictherapy/. Accessed February 14, 2011. 10. Richardson MM, Babiak-Vazquez AE, Frenkel MA. Music therapy in a comprehensive cancer center. J Soc Integr Oncol. 2008;6(2):76Y81.

Cancer NursingTM, Vol. 37, No. 6, 2014

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

n

E49

11. Li XM, Yan H, Zhou KN, Dang SN, Wang DL, Zhang YP. Effects of music therapy on pain among female breast cancer patients after radical mastectomy: results from a randomized controlled trial. Breast Cancer Res Treat. 2011;128(2):411Y419. 12. Bradt J, Dileo C, Grocke D, Magill L. Music interventions for improving psychological and physical outcomes in cancer patients. Cochrane Database Syst Rev. 2011;10(8):CD006911. 13. Zhou KN, Li XM, Yan H, Dang SN, Wang DL. Effects of music therapy on depression and duration of hospital stay of breast cancer patients after radical mastectomy. Chin Med J (Engl). 2011;124(15):2321Y2327. 14. Kemper KJ, Hamilton CA, McLean TW, Lovato J. Impact of music on pediatric oncology outpatients. Pediatr Res. 2008;64(1):105Y109. 15. Madden JR, Mowry P, Gao D, Cullen PM, Foreman NK. Creative arts therapy improves quality of life for pediatric brain tumor patients receiving outpatient chemotherapy. J Pediatr Oncol Nurs. 2010;27(3): 133Y145. 16. Dileo C. Effects of music and music therapy on medical patients: a metaanalysis of the research and implications for the future. J Soc Integr Oncol. 2006;4(2):67Y70. 17. Burns DS, Azzouz F, Sledge R, et al. Music imagery for adults with acute leukemia in protective environments: a feasibility study. Support Care Cancer. 2008;16(5):507Y513. 18. Cassileth BR, Vickers AJ, Magill LA. Music therapy for mood disturbance during hospitalization for autologous stem cell transplantation. Cancer. 2003;98(12):2723Y2729. 19. Ferrer AJ. The effect of live music on decreasing anxiety in patients undergoing chemotherapy treatment. J Music Ther. 2007;44(3):242Y255. 20. Zhang JM, Wang P, Yao JX, Zhao L, Davis MP, Walsh D, Yue GH. Music interventions for psychological and physical outcomes in cancer: a systematic review and meta-analysis. Support Care Cancer. 2012;20(12): 3043Y3053.

E50 n Cancer NursingTM, Vol. 37, No. 6, 2014

21. Cuijpers P, Smit F, van Straten A. Psychological treatments of subthreshold depression: a meta-analytic review. Acta Psychiatr Scand. 2007;115(6): 434Y441. 22. Cuijpers P, van Straten A, Andersson G, van Oppen P. Psychotherapy for depression in adults: a meta-analysis of comparative outcome studies. J Consult Clin Psychol. 2008;76(6):909Y922. 23. Cuijpers P, van Straten A, Warmerdam L, Smits N. Characteristics of effective psychological treatments of depression: a metaregression analysis. Psychother Res. 2008;18(2):225Y236. 24. Szentagotai A, David D. The efficacy of cognitive-behavioral therapy in bipolar disorder: a quantitative meta-analysis. J Clin Psychiatry. 2010;71(1):66Y72. 25. Brodaty H, Green A, Koschera A. Meta-analysis of psychosocial interventions for caregivers of people with dementia. J Am Geriatr Soc. 2003; 51(5):657Y664. 26. Light RJ, Pillemer DB. Summing Up: The Science of Reviewing Research. Cambridge, MA: Harvard University Press; 1984. 27. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629Y634. 28. Hedges LV, Olkin I. Statistical Methods for Meta-analysis. Orlando, FL: Academic Press; 1985. 29. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557Y560. 30. Dunbar RI, Kaskatis K, MacDonald I, Barra V. Performance of music elevates pain threshold and positive affect: implication for evolutionary function of music. Evol Psychol. 2012;10(4):688Y702. 31. Zhao PT, Liang J, Shao QJ, Liang F, Yuan HQ, You FS. Interventional effects of musical therapy to physiological and psychological conditions in process of radiotherapy for patients with cancer. Nurs Res. 2008;15(14):1097Y1099. 32. Lin MF, Hsieh YJ, Hsu YY, Fetzer S, Hsu MC. A randomised controlled trial of the effect of music therapy and verbal relaxation on chemotherapyinduced anxiety. J Clin Nurs. 2011;20(7Y8):988Y999.

Tsai et al

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Effectiveness of music intervention in ameliorating cancer patients' anxiety, depression, pain, and fatigue: a meta-analysis.

This is the first study to use meta-analysis as a scientific technique to provide an integrated analysis of the effectiveness of music intervention in...
503KB Sizes 0 Downloads 3 Views