8. Psycho-oncological Interventions – Psycho-oncology

Systematic literature review

The evaluation of the effectiveness of psychotherapeutic interventions in cancer patients and survivors is based on 16 systematic reviews as well as 21 RCTs for individual psychotherapy, and 18 RCTs for group psychotherapy. One of these 39 RCTs was already included in one of the systematic reviews, therefore, the results of this study are not reported separately in the results.

The 16 systematic reviews (13 meta-analyses, three systematic reviews) included between six and 245 primary studies in the analysis. Regarding the tumor entities studied, seven reviews focused exclusively on breast cancer and one each on lung and prostate cancer. The remaining reviews (n=7) examined mixed cancer populations. The primary studies included were from the US, Asia, Australia, UK, and the EU, among others. Several reviews focused on only one endpoint: three meta-analyses reported exclusively on the efficacy of psychotherapeutic interventions for reducing depression [583], [591], [592], one for increasing quality of life [586], one for reducing anxiety [585], and one for reducing fatigue [631].

The sample size of the 21 RCTs that examined individual psychotherapy varied from 30 to 1012 patients (N=3,624), of whom 14.5-100% were female. The studies were mostly from the United States (n=3), Canada (n=2), Iran (n=2), China (n=2), and Germany (n=2). The remaining studies were from the Netherlands, Italy, Denmark, Turkey, Norway, Greece, New Zealand, Qatar, South Korea, and Brazil (n=1 each). Regarding the tumor entities investigated, eight studies focused exclusively on breast cancer, two on gynecologic cancers, and one study each on leukemia, esophageal cancer, colorectal cancer, head and neck cancer, and patients undergoing hematopoietic stem cell transplantation. The remaining studies (n=6) examined patients with various cancer diagnoses without focusing on a specific cancer type. Most individual psychotherapies were multimodal interventions with different components. Elements of psychosocial counseling (n=10), skills and management training (n=6), cognitive (-behavioral) elements (n=3), and problem-solving training (n=2) were frequently implemented. The number of intervention sessions presented ranged from one session to 35 sessions. Most frequently, depressiveness (n=10), quality of life (n=9), anxiety (n=8), and psychological distress (n=5) were collected as primary outcomes. Furthermore, pain (n=3), sexual function (n=2), nausea (n=1), and fatigue (n=1) represented primary endpoints in the RCTs.

The sample size of the 18 RCTs that investigated group psychotherapy varied from 25 to 392 patients (n=2,299), of whom 47-100% were female. The studies were mostly from the USA (n=4) and Iran (n=4), as well as China (n=3), Canada (n=2), China (n=2), and Germany (n=2). The remaining studies were from the Netherlands, Spain, and Brazil (n=1 each). Regarding the tumor entities investigated, 11 studies focused exclusively on breast cancer and one study on gynecologic cancer and breast cancer. Furthermore, one study each focused exclusively on lung cancer and prostate cancer, and one study on prostate cancer and gastrointestinal tumors. The remaining studies (n=3) examined patients* with different cancer diagnoses. In the group psychotherapies, cognitive-behavioral (n=4), social-supportive (n=3), and supportive-expressive and psychodynamic therapy elements (n=3) were most frequently used. The primary endpoints most frequently assessed in the RCTs represented depressiveness (n=7), anxiety (n=7), and quality of life (n=7). In addition, sexual function/satisfaction (n=2), psychological distress (n=1), and fatigue (n=1) were collected as primary endpoints.

Results

Of the 16 included reviews, only two differentiated between individual and group psychotherapy in the presentation of results [582], [593]. Therefore, the results of the systematic reviews are also reported below for psychotherapeutic interventions as a whole, i.e., not differentiated between individual and group therapy.

The results of the additionally searched RCTs are presented separately for individual psychotherapy and group psychotherapy.

Since no studies were available which directly compare psychotherapeutic interventions in individual versus group settings, Recommendation 8.1 was formulated in the guideline group on the basis of expert consensus. The choice between individual and group psychotherapy should be based on the patient's preference. However, since group offerings are not as common in the reality of care, the choice will essentially depend on which offerings are available locally.

Quality of life, psychological distress, depressiveness, anxiety, and nausea (Recommendation 8.14)

The short-term effectiveness of psychotherapeutic interventions on global quality of life has been examined in seven meta-analyses [584], [586], [588], [589], [590], [593], [582], all of which demonstrated a significant short-term effect in favor of psychotherapeutic interventions (0.09 ≤ d ≤ 0.65). In contrast, one systematic review [697] describes that only one of the six RCTs included in the review found a significant effect in favor of psychotherapeutic interventions. In addition, our research found ten RCTs on individual psychotherapy [597], [600], [604], [605], [607], [610], [595], [603], [609], [606] and eight RCTs on group psychotherapy [614], [618] [619], [616], [623], [622], [624], [617] which investigated short-term efficacy with regard to global quality of life. Of these, five RCTs on individual psychotherapy and one RCT on group psychotherapy were able to show a significant effect with regard to global quality of life.

The medium-term efficacy of psychotherapeutic interventions with respect to global quality of life was examined in three meta-analyses [588], [589], [582], two of which demonstrated a significant medium-term effect in favor of psychotherapeutic interventions (d = 0.15 and d = 0.38). In addition, our research found eight RCTs on individual psychotherapy [597], [604], [605], [610], [595], [603], [609], [606] and six RCTs on group psychotherapy [618], [619], [616], [622], [624], [617] which examined medium-term efficacy with respect to global quality of life. Three RCTs on individual psychotherapy and one RCT on group psychotherapy showed a significant effect on global quality of life.

The long-term efficacy of psychotherapeutic interventions regarding global quality of life was examined in two meta-analyses [589], [582] and could only be demonstrated in Faller et al. (d=0.37). In addition, our research found two RCTs on individual psychotherapy [597], [600] and three RCTs on group psychotherapy [618], [626], [616] that examined long-term efficacy with respect to global quality of life. None of the RCTs on individual psychotherapy and one RCT on group psychotherapy showed a significant effect on global quality of life. One RCT [618] on group psychotherapy described mixed findings (significant effects for 6 and 12 months, but not for 18 months).

The short-term efficacy of psychotherapeutic interventions regarding psychological distress was examined in four meta-analyses [584], [588], [590], [582], all of which demonstrated a significant short-term effect in favor of psychotherapeutic interventions (-0.23 ≤ d ≤ -0.48). Furthermore, three systematic reviews describe that two of two [593], one of two [696], and eight of 13 [697] RCTs included in the respective reviews found a significant, short-term effect in favor of psychotherapeutic interventions. In addition, our research found six RCTs on individual psychotherapy [596], [597], [610], [595], [608], [606] and five RCTs on group psychotherapy [611], [619], [612], [624], [617] which investigated the short-term efficacy regarding psychological distress. Two RCTs on individual psychotherapy and two RCTs on group psychotherapy showed a significant effect on psychological distress.

The medium-term effectiveness of psychotherapeutic interventions with regard to psychological distress was examined in two meta-analyses [588], [582], both of which demonstrated a significant medium-term effect in favor of psychotherapeutic interventions (d=-0.18 and d=-0.21). In addition, our research found six RCTs on individual psychotherapy [596], [597], [610], [595], [608], [606] and five RCTs on group psychotherapy [611], [619], [612], [624] which investigated the medium-term efficacy regarding psychological distress. Two RCTs on individual psychotherapy and two RCTs on group psychotherapy showed a significant effect on psychological distress. One RCT [619] on group psychotherapy showed mixed findings (significant effect after three months, but not after six months).

The long-term effectiveness of psychotherapeutic interventions on psychological distress was examined in one meta-analysis [582], which demonstrated a significant long-term effect in favor of psychotherapeutic interventions (d=-0.15). In addition, one RCT on individual psychotherapy [597] was available from our search that examined long-term efficacy with respect to psychological distress. No significant effect in favor of psychotherapy could be demonstrated.

The short-term efficacy of psychotherapeutic interventions regarding depressiveness has been examined in eight meta-analyses [583], [584], [588], [589], [591], [592], [593], [582], six of which demonstrated a significant short-term effect in favor of psychotherapeutic interventions (-0.25 ≤ d ≤ -1.11). Furthermore, one systematic review [696] describes that in four of six RCTs included in the review, a significant, short-term effect in favor of psychotherapeutic interventions was found. In addition, our search found 12 RCTs on individual psychotherapy [596], [597], [598], [599], [600], [628], [601], [604], [610], [594], [595], [609] and nine RCTs on group psychotherapy [611], [615], [627], [619], [620], [625], [612], [621], [622] which investigated short-term efficacy with regard to depression. Of these, seven RCTs on individual psychotherapy and five RCTs on group psychotherapy were able to show a significant effect with regard to depressiveness.

The medium-term efficacy of psychotherapeutic interventions with respect to depressiveness was examined in three meta-analyses [588], [589], [582], one of which demonstrated a significant medium-term effect in favor of psychotherapeutic interventions (d=-0.13 (individual), d=-0.21 (group)). In addition, our search found nine RCTs on individual psychotherapy [596], [597], [599], [628], [604], [610], [594], [595], [609] and seven RCTs on group psychotherapy [611], [615], [627], [619], [612], [621], [622] which examined medium-term efficacy with respect to depression. Of these, four RCTs on individual psychotherapy and three RCTs on group psychotherapy were able to show a significant effect with regard to depressiveness. One RCT [619] on group psychotherapy showed mixed findings (significant effect at three months but not at six months).

The long-term effectiveness of psychotherapeutic interventions with respect to depressiveness was examined in three meta-analyses [588], [589], [582], one of which demonstrated a significant long-term effect in favor of psychotherapeutic interventions (d=-0.28). In addition, our research found two RCTs on individual psychotherapy [600], [594] and two RCTs on group psychotherapy [626], [621] that examined long-term efficacy with respect to depression. One RCT on individual psychotherapy and one RCT on group psychotherapy showed a significant effect on depression.

The short-term effectiveness of psychotherapeutic interventions regarding anxiety was examined in seven meta-analyses [584], [585], [587], [588], [589], [593], [582], six of which demonstrated a significant short-term effect in favor of psychotherapeutic interventions (-0.21 ≤ d ≤ -1.1). Furthermore, one systematic review [696] describes that in two of the three RCTs included in the review, a significant short-term effect in favor of psychotherapeutic interventions was found. In addition, our search found nine RCTs on individual psychotherapy [596], [597], [598], [600], [604], [610], [594], [595], [602] and nine RCTs on group psychotherapy [611], [613], [615], [627], [619], [625], [621], [622], [617] that examined short-term efficacy with respect to anxiety. Three RCTs on individual psychotherapy and six RCTs on group psychotherapy showed a significant effect on anxiety. One RCT [598] on individual psychotherapy showed mixed findings (significant effect on HADS-A, but no significant effect on SAS).

The medium-term effectiveness of psychotherapeutic interventions on anxiety was examined in four meta-analyses [587], [588], [589], [582], two of which demonstrated a significant medium-term effect in favor of psychotherapeutic interventions (d=-0.19 and d=-0.33 (individual) and d=-0.22 (group), respectively). In addition, our search found six RCTs on individual psychotherapy [596], [597], [604], [610], [594], [595] and seven RCTs on group psychotherapy [611], [615], [627], [619], [621], [622], [617] which investigated the medium-term efficacy regarding anxiety. One RCT on individual psychotherapy and four RCTs on group psychotherapy showed a significant effect on anxiety. In one RCT on group psychotherapy [622] patients in the psychotherapy group experienced more anxiety than patients in the control group three months after the intervention.

The long-term efficacy of psychotherapeutic interventions with regard to anxiety was examined in two meta-analyses [589], [582], one of which demonstrated a significant long-term effect in favor of psychotherapeutic interventions (d=-0.46). In addition, our search found three RCTs on individual psychotherapy [597], [600], [594] and one RCT on group psychotherapy [621] which examined long-term efficacy with respect to anxiety. Only two RCTs on individual psychotherapy showed a significant long-term effect on anxiety.

The short-term effectiveness of psychotherapeutic interventions regarding nausea was investigated in one meta-analysis [588]. A significant short-term effect in favor of psychotherapeutic intervention was demonstrated (d=-0.39). In addition, our search found one RCT on individual psychotherapy [701] which investigated the short-term efficacy regarding nausea. A significant short-term effect in favor of psychotherapy was demonstrated.

Regarding the medium-term efficacy of psychotherapeutic interventions on nausea, only one RCT [701] on individual psychotherapy was available according to our research. A significant medium-term effect in favor of psychotherapy could be demonstrated.

According to our search, no studies on the long-term effectiveness of psychotherapeutic interventions regarding nausea were available.

Fatigue (Recommendation 8.15)

The short-term efficacy of psychotherapeutic interventions regarding fatigue has been examined in two meta-analyses [631], [588], one of which demonstrated a significant short-term effect in favor of psychotherapeutic interventions (d=-0.42⁶). Furthermore, two systematic reviews describe that in one of five [697] and two of two [698] RCTs included in the review, a significant short-term effect in favor of psychotherapeutic interventions was found. In addition, our search found three RCTs on individual psychotherapy [597], [604], [607] and two RCTs on group psychotherapy [702], [624] that examined short-term efficacy with respect to fatigue. Two RCTs on individual psychotherapy and one RCT on group psychotherapy showed a significant effect on fatigue.

The medium-term efficacy of psychotherapeutic interventions regarding fatigue was only investigated in one meta-analysis [588]. No significant medium-term effect in favor of psychotherapeutic intervention could be demonstrated. In addition, our search found two RCTs on individual psychotherapy [597], [604] and one RCT on group psychotherapy [624] which investigated the medium-term efficacy regarding fatigue. Only one RCT on individual psychotherapy showed a significant medium-term effect in favor of psychotherapy.

According to our search, only one RCT [597] on individual psychotherapy was available for the long-term efficacy of psychotherapeutic interventions with regard to fatigue. No significant effect in favor of individual psychotherapy could be demonstrated.

⁶ Effect only significant for patients after end of treatment, not significant for patients receiving treatment.

Pain (Recommendation 8.16)

The short-term effectiveness of psychotherapeutic interventions regarding pain was investigated in two meta-analyses [588], [589]. No significant short-term effect in favor of psychotherapeutic intervention was demonstrated in either of the two. However, one systematic review [697] describes that a significant short-term effect in favor of psychotherapy was demonstrated in four of the five RCTs included in the review. In addition, our search found four RCTs on individual psychotherapy [596], [604], [607], [603], which investigated short-term efficacy with regard to pain. Of these, only one RCT was able to show a significant effect with regard to pain.

The medium-term efficacy of psychotherapeutic interventions regarding pain was only investigated in one meta-analysis [589]. No significant medium-term effect in favor of psychotherapeutic intervention could be demonstrated. In addition, our search found three RCTs on individual psychotherapy [596], [604], [603] which investigated the medium-term effectiveness regarding pain. Only one RCT demonstrated a significant medium-term effect in favor of psychotherapy.

According to our research, no studies on the long-term efficacy of psychotherapeutic interventions regarding pain were available.

Sexual dysfunction (Statement 8.17)

No systematic reviews were available on the efficacy of psychotherapeutic interventions regarding sexual function, according to our search.

The short-term effectiveness of psychotherapeutic interventions regarding sexual function was examined in one RCT on individual psychotherapy [700] and two RCTs on group psychotherapy [616], [623]. Only the RCT on individual psychotherapy was able to demonstrate a significant short-term effect in favor of psychotherapy.

The medium-term effectiveness of psychotherapeutic interventions regarding sexual function was investigated in one RCT on individual psychotherapy [699] and one RCT on group psychotherapy [616]. Both RCTs were able to demonstrate a significant medium-term effect regarding sexual function.

The long-term effectiveness of psychotherapeutic interventions regarding sexual function was only examined in one RCT on group psychotherapy [616]. No significant long-term effect could be demonstrated.

The evidence base on the effectiveness of psychotherapeutic interventions to reduce sexual dysfunction in cancer patients is heterogeneous and not sufficiently robust. Sexual function was the primary outcome in all RCTs, and two of the four RCTs had a sufficiently large sample size after a priori power analysis. One RCT provided information on side effects and interactions, but these did not occur in either the control or the psychotherapy groups. The methodological quality of all RCTs is considered moderate to low. The guideline group therefore reserves the right to refrain from recommending for or against the use of psychotherapeutic interventions in cancer patients with these complaints. Further studies of high methodological quality are needed to formulate a recommendation.

Assessment of study quality

In total, 13/16 (81%) SRs had deficiencies in more than one critical assessment domain (critical low quality), 2/16 (13%) had deficiencies in one critical assessment domain (low quality), and one (6%) SR had no critical deficiencies and only one non-critical deficiency (high quality). It should be noted that 5/16 SRs had a downgraded quality assessment due to the lack of a list of excluded primary studies and their reasons for exclusion (critical domain 7, AMSTAR-2).

17/21 RCTs (81%) on individual psychotherapy used an adequate method to generate the randomization sequence (selection bias) and 9/21 RCTs (43%) ensured secrecy and unpredictability of group allocation (selection bias); in 11/21 RCTs (52%), the amount of missing data was sufficiently small (<20%) or missing data were appropriately imputed (attrition bias); and in 6/21 RCTs (29%), selective reporting of endpoints could be excluded by an existing study protocol (reporting bias). Due to the nature of the interventions used and the collection of endpoints by self-assessment, blinding of study personnel/participants (performance bias) and thus blinding of endpoint collection (detection bias) were not possible in any of the RCTs.

In total, 13/21 RCTs performed a priori sample size estimation. 6 RCTs achieved their a priori planned sample size.

13/18 RCTs (72%) on group psychotherapy used an adequate method to generate the randomization sequence (selection bias); 4/18 RCTs (22%) ensured secrecy and unpredictability of group allocation (selection bias); in 11/18 RCTs (61%) the amount of missing data was sufficiently small (<20%) or missing data were imputed accordingly (attrition bias). In 1/18 RCTs (6%), selective reporting of end points could be excluded by an existing study protocol (reporting bias). Due to the nature of the interventions used and the collection of endpoints by self-assessment, blinding of study personnel/participants (performance bias) and thus, with the exception of one RCT, blinding of endpoint collection (detection bias) was not possible in any of the RCTs.

In total, 8/18 RCTs performed a priori sample size estimation. 6 RCTs achieved their a priori planned sample size.

Systematic Literature Review
The evaluation of the effectiveness of couple and family interventions in cancer patients and survivors is based on seven systematic reviews and five RCTs. One of the five RCTs was already included in one of the systematic reviews, therefore, the results of this study are not reported separately in the results.

The seven systematic reviews (five meta-analyses, two systematic reviews) included between ten and 198 primary studies in the analysis. Six of the reviews examined mixed cancer populations and one paper specifically included breast cancer patients. The primary studies included were from the United States, Australia, Canada, Iran, and the EU, among other countries. One meta-analysis reported exclusively on the effectiveness of psychotherapeutic couple and family interventions for increasing quality of life [706]. All other reviews reported multiple relevant endpoints.

The sample size of the five RCTs varied from 17 to 237 patients (N=572), 0-100% were female. The studies were from Denmark (n=2), the USA, Australia, and the UK (n=1 each). In terms of tumor entities examined, three studies focused exclusively on prostate cancer, one study focused on breast cancer, and one study examined patients with multiple cancer diagnoses. Four studies examined couple-based treatment and counseling approaches, and one study examined a family coping intervention. The number of intervention sessions presented ranged from four to six sessions over a maximum period of five months. The primary endpoints collected by the RCTs were mental distress (n=3), quality of life (n=2), and depressiveness (n=1).

Results

Quality of life, mental distress, depressiveness, and anxiety (Recommendation 8.19)

The short-term effectiveness of psychotherapeutic couple and family interventions on global quality of life has been examined in four meta-analyses [703], [704], [706], [582], one of which demonstrated a significant short-term effect (d=0.25). In addition, Griffin and colleagues [707] described in detail in a systematic review that a significant effect in favor of couple and family interventions was not detectable in any of the six RCTs included in the review. In addition, the results of two RCTs were available from our search [709], [712], but they did not find a significant short-term effect in favor of psychotherapeutic couple and family interventions.

The medium-term effectiveness of psychotherapeutic couple and family interventions with regard to global quality of life was examined in three meta-analyses [703], [704], [582] and in two RCTs [709], [712]. Neither the meta-analyses nor the RCTs demonstrated significant medium-term effects in favor of psychotherapeutic couple and family interventions.

The long-term effectiveness of psychotherapeutic couple and family interventions with regard to global quality of life was also examined in three meta-analyses [703], [704], [582]. However, no significant long-term effect in favor of psychotherapeutic couple and family interventions could be demonstrated.

The short-term effectiveness of psychotherapeutic couple and family interventions on mental distress was examined in two meta-analyses [704], [582], one of which demonstrated a significant short-term effect in favor of psychotherapeutic couple and family interventions (d=-0.16). In addition, results of three RCTs [710], [711], [712] were available, but none of them could detect a significant short-term effect in favor of psychotherapeutic couple and family interventions.

Regarding the medium-term effectiveness of psychotherapeutic couple and family interventions with regard to mental distress, the results of two meta-analyses [704], [582] and three RCTs [710], [711], [712] were available. Neither the meta-analyses nor the RCTs demonstrated significant medium-term effects in favor of psychotherapeutic couple and family interventions.

The long-term effectiveness of psychotherapeutic couple and family interventions on psychological distress was examined in two meta-analyses [704], [582], none of which, however, demonstrated a significant effect in favor of psychotherapeutic couple and family interventions. No complementary RCTs were available.

The short-term efficacy of psychotherapeutic couple and family interventions with respect to depressivity was examined in four meta-analyses [703], [704], [705], [582], two of which demonstrated a significant short-term effect in favor of psychotherapeutic couple and family interventions (d=-0.26 and d=-2.15). In addition, one systematic review [707] descriptively describes that in two of nine included RCTs there was a significant effect in favor of psychotherapeutic couple and family interventions. In addition, the results of three RCTs [709], [710], [711] were available from our search, but none of them demonstrated a significant short-term effect in favor of psychotherapeutic couple and family interventions.

The medium-term efficacy of psychotherapeutic couple and family interventions on depressive symptoms was examined in three meta-analyses [703], [704], [582], two of which demonstrated a significant medium-term effect in favor of psychotherapeutic couple and family interventions (d=-0.29 and d=-0.37). In addition, the results of three RCTs were available [709], [710], [711], but they could not determine a significant medium-term effect in favor of psychotherapeutic couple and family interventions.

The long-term efficacy of psychotherapeutic couple and family interventions with regard to depressiveness was examined in two meta-analyses [704], [582] but they were unable to demonstrate a significant long-term effect in favor of psychotherapeutic couple and family interventions.

The short-term effectiveness of psychotherapeutic couple and family interventions regarding anxiety was examined in four meta-analyses [703], [704], [705], [582] two of which demonstrated a significant short-term effect in favor of psychotherapeutic couple and family interventions (d=-0.30 and d=-1.16). In addition, the result of one RCT [709] was available, but it did not find a significant short-term effect in favor of psychotherapeutic couple and family interventions.

The medium-term effectiveness of psychotherapeutic couple and family interventions with regard to anxiety was examined in a meta-analysis [704]. This demonstrated a significant medium-term effect in favor of psychotherapeutic couple and family interventions (d=-0.43). In addition, the results of one RCT [709] were available, but did not detect a significant, medium-term effect in favor of psychotherapeutic couple and family interventions.

The long-term effectiveness of psychotherapeutic couple and family interventions with regard to anxiety was examined in two meta-analyses [704], [582], one of which demonstrated a significant effect in favor of psychotherapeutic couple and family interventions (d=-0.26). No complementary RCTs were available.

Fatigue, pain, and sexual dysfunction (Recommendation 8.20)

Both the short- and medium-term efficacy of psychotherapeutic couple and family interventions on fatigue were examined only in the meta-analysis by Hu et al. [703]. However, no significant short- or medium-term effects in favor of psychotherapeutic couple and family interventions could be demonstrated. No complementary RCTs were available according to our search.

Both short- and medium-term efficacy of psychotherapeutic couple and family interventions on pain were also examined only in the meta-analysis by Hu et al. [703]. However, no significant short- or medium-term effects in favor of psychotherapeutic couple and family interventions could be demonstrated. No complementary RCTs were available according to our search.

Regarding sexual dysfunction, the systematic review by Brandao et al. [708] describes in detail that in eight of ten included RCTs, significant effects on at least one relevant outcome (psychological functioning level, physical functioning level, marital quality, sexual satisfaction) in favor of couple and family interventions were demonstrated. Four of these eight studies showed significant effects on sexual function/satisfaction in favor of the couple and family interventions.

No other systematic reviews or RCTs were available on this endpoint.

Assessment of study quality

All seven SRs (100%) had deficiencies in more than one critical assessment domain (critical low quality).

In total, 2/5 RCTs (40%) used an adequate method to generate the randomization sequence (selection bias), 3/5 RCTs (60%) ensured secrecy and unpredictability of group allocation (selection bias), in 2/5 RCTs (40%), the amount of missing data was sufficiently small (<20%) or missing data were appropriately imputed (attrition bias), and in 2/5 RCTs (40%), selective reporting of end points could be excluded by an existing study protocol (reporting bias). Due to the nature of the interventions used and the collection of endpoints by self-assessment, blinding of study personnel/participants (performance bias) and thus blinding of endpoint collection (detection bias) were not possible in any of the RCTs.

In total, 4/5 RCTs performed a priori case sample size estimation of which one RCT achieved their a priori planned sample size.

Patients with a non-curable, progressive cancer are confronted with a variety of physical, mental, social and spiritual distress factors during the course of the disease [732]. Good palliative care therefore requires addressing the needs of patients in all four dimensions. It places the quality of life of patients with non-curable cancer at the center of its work (Extended S3 Guideline Palliative Care for Patients with Non-curable Cancer. Psycho-oncological interventions are part of a multi-professional and interdisciplinary palliative care approach. In this regard, we refer to the "Extended S3 Guideline on Palliative Care for Patients with Non-curable Cancer" (AWMF Registry Number:128-001OL, https://www.awmf.org/leitlinien/detail/ll/128-001OL.html).

Specific psycho-oncological interventions for patients in the palliative phase are usually supportive and focus to varying degrees on spiritual, psychological, physical or social aspects of cancer [733], [734]. Relaxation techniques, mindfulness-based interventions and meditation, therapeutic hypnotherapy elements, cognitive-behavioral therapy interventions, supportive-expressive interventions, and systemic family therapy approaches, as well as creative arts therapies may be used [732], [735]. In recent decades, specific interventions for working with palliative patients have also been developed and are briefly introduced below.

Meaning-Centered Psychotherapy is a form of existential psychotherapy developed by William S. Breitbart and colleagues that aims to find meaning and significance in the face of death [736], [737]. Originally conceptualized as a group therapy, Meaning-Centered Group Psychotherapy (MCGP) provides a framework in which (oncology) palliative care patients work together to address issues of meaning, inner peace, and goal orientation in life in the context of their illness [735]. Designed for eight sessions, seven main topics are worked on in a combination of psychoeducation, group discussions and exercises: 1. concepts of meaning and sources of meaning; 2. cancer and meaning; 3. creation of meaning from life history; 4. creation of meaning from the present and future; 5. creation of meaning from attitudes (encountering the limits of life); 6. creation of meaning from creative activity and taking responsibility; 7. experience-based creation of meaning (nature, art, humor) [738], [739]. In order to meet the needs of the palliative situation for flexibility of psychotherapeutic treatments, Individual Meaning-Centered Psychotherapy (IMCP), which works on the content of MCGP in seven individual sessions [740], [726], was also developed.

Managing Cancer and Living Meaningfully (CALM) therapy, developed by Gary Rodin and colleagues, is a manualized, semi-structured, short-term psychotherapeutic intervention in an individual setting [741], [742], [743], [723], [744]. In terms of content, CALM focuses on four areas: 1. symptom management and communication with treatment providers; 2. self-change and relationship with loved ones; 3. spirituality, meaning and significance; 4. preparing for the future, maintaining hope and coming to terms with dying [744].

Based on the empirical dignity model of Harvey M. Chochinov and colleagues, Dignity Therapy aims to alleviate patients' suffering at the end of life and improve their emotional and spiritual well-being [745], [746], [732]. In this process, patients are guided to reflect on memorable life events in an interview and 1-2 therapy sessions in an individual setting. The tape-recorded interview is then transcribed, read back to the patients, and edited if necessary. Patients are then given a document to leave with family members [732], [735]. The intervention is particularly aimed at patients in the terminal phase of their disease [732].

Systematic literature review

In this chapter, all psychotherapeutic interventions were considered that were explicitly directed at patients with non-curable cancer and addressed the physical, psychological, social and spiritual stresses of the palliative situation. The interventions were delivered exclusively in face-to-face contacts.

The evaluation of the effectiveness of psychotherapeutic interventions in the palliative phase is based on eight systematic reviews and 13 RCTs. One RCT [739] is already included in systematic reviews and is not reported separately in the results.

The eight systematic reviews (seven meta-analyses, one systematic review) included between eight and 68 primary studies in the analysis. All reviews included mixed cancer populations, other palliative populations were also included in three cases [714], [589], [715]. In these cases, the results reported below refer to subgroup analyses. The primary studies included were from the US, China, Canada, Australia, the UK, and the EU, among others.

The sample size of the 13 RCTs varied from 28 to 340 patients (N=1,926), 38.1-87.5% were female. The studies were from the USA (n=4), the Netherlands (n=2), Canada (n=2), the UK (n=2), Brazil, Nigeria, and Denmark (n=1 each). The majority of included studies (n=12) included patients with various advanced or terminal cancer diagnoses. One study was conducted with leukemia patients. Common interventions were meaning-based psychotherapy (IMCP/MCGP, CALM; n=3), elements of cognitive behavioral therapy (CBT; n=3), and psychoeducational elements (n=2). Two studies targeted couples. The number of intervention sessions ranged from a single session to 14 weekly sessions. Depressiveness (n=4) and quality of life (n=3) were most frequently collected as primary endpoints. Furthermore, fatigue, mental distress, and anxiety (n=1 each) were collected as primary endpoints.

Results

Quality of life, mental distress, depression, anxiety, and fatigue (Recommendation 8.21)

The short-term efficacy of specific psycho-oncological interventions in the palliative phase with respect to global quality of life was examined in six meta-analyses [713], [589], [715], [714], [716], [718], of which five demonstrated a significant short-term effect in favor of specific psycho-oncological interventions in the palliative phase (0.17 ≤ d ≤ 0.47). In addition, the results of seven RCTs were available [720], [723], [725], [726], [724], [730], [731], of which three found a significant short-term effect in favor of specific psycho-oncological interventions in the palliative phase.

The medium-term efficacy of specific psycho-oncological interventions in the palliative phase with respect to global quality of life was investigated in two meta-analyses [589], [718], one of which demonstrated a significant, medium-term effect in favor of specific psycho-oncological interventions in the palliative phase (d=0.82). In addition, the results of six RCTs [722], [723], [725], [726], [724], [731] were available, two of which recorded a significant medium-term effect in favor of specific psycho-oncological interventions in the palliative phase. In addition, one RCT [722] provided mixed results. No significant effect was shown at two, four, and eight weeks, but significant effect in favor of interventions in the palliative phase at 6 months was demonstrated.

The long-term effectiveness of specific psycho-oncological interventions in the palliative phase with regard to global quality of life was investigated in one meta-analysis [589] and in one RCT [731]. Within these studies, no significant long-term effects in favor of specific psycho-oncological interventions in the palliative phase could be demonstrated.

The short-term efficacy of specific psycho-oncological interventions in the palliative phase with regard to mental distress was investigated in four meta-analyses [713], [714], [716], [718], all of which demonstrated a significant, short-term effect in favor of specific psycho-oncological interventions in the palliative phase (-0.32 ≤ d ≤ -1.02⁷). In addition, results were available from five RCTs [725], [728], [724], [729], [730], two of which recorded a significant short-term effect in favor of specific psycho-oncological interventions in the palliative phase.

The medium-term effectiveness of specific psycho-oncological interventions in the palliative phase with regard to mental distress was investigated in a meta-analysis [718]. However, no significant medium-term effect in favor of specific psycho-oncological interventions in the palliative phase could be demonstrated. In addition, results of four RCTs [725], [727], [728], [724] were available, one of which demonstrated a significant, medium-term effect in favor of specific psycho-oncological interventions in the palliative phase.

The short-term efficacy of specific psycho-oncological interventions in the palliative phase with regard to depressiveness was investigated in five meta-analyses [713], [589], [715], [716], [718], four of which demonstrated significant, short-term effects in favor of specific psycho-oncological interventions in the palliative phase (-0.25 ≤ d ≤ -1.31) In addition, the results of six RCTs [721], [723], [726], [724], [729], [731] were available, two of which demonstrated significant short-term effects in favor of specific psycho-oncological interventions in the palliative phase.

Significant medium-term effects in favor of specific psycho-oncological interventions in the palliative phase were investigated in two meta-analyses [589], [718], none of which, however, demonstrated significant effects of specific psycho-oncological interventions in the palliative phase. In addition, results were available from six RCTs [721], [722], [723], [726], [724], [731], one of which demonstrated a significant effect in favor of specific psycho-oncological interventions in the palliative phase.

Long-term effects of specific psycho-oncological interventions in the palliative phase were investigated in one systematic review [589] and one RCT [731]. However, none of the studies found a significant long-term effect in favor of specific psycho-oncological interventions in the palliative phase.

The short-term efficacy of specific psycho-oncological interventions in the palliative phase with regard to anxiety was examined in five meta-analyses [713], [589], [715], [716], [718], four of which demonstrated significant short-term effects in favor of specific psycho-oncological interventions in the palliative phase (-0.21 ≤ d ≤ -1.07).  In addition, results were available from five RCTs [723], [726], [728], [729], [731], two of which demonstrated significant short-term effects in favor of specific psycho-oncological interventions in the palliative phase.

The medium-term efficacy of specific psycho-oncological interventions in the palliative phase with regard to anxiety was examined in two meta-analyses [589], [718]. Neither could demonstrate significant effects of specific psycho-oncological interventions in the palliative phase. In addition, results from five RCTs [722], [723], [726], [728], [731] were available, two of which demonstrated significant effects of specific psycho-oncological interventions in the palliative phase.

Long-term effects of psychotherapeutic interventions in the palliative phase have been investigated in only one systematic review [589] and one RCT [731], both of which failed to identify significant effects.

The short-term efficacy of specific psycho-oncological interventions in the palliative phase with regard to fatigue was examined in two meta-analyses [714], [717], one of which demonstrated significant effects of specific psycho-oncological interventions in the palliative phase (d=-0.30). In addition, results from three RCTs [720], [727], [729] were available, two of which demonstrated significant short-term effects in favor of specific psycho-oncological interventions in the palliative phase.

The medium-term effectiveness of specific psycho-oncological interventions in the palliative phase with regard to fatigue was investigated in a meta-analysis [717]. This showed a significant effect in favor of specific psycho-oncological interventions in the palliative phase (d=-0.66).

The long-term efficacy of specific psycho-oncological interventions in the palliative phase with regard to fatigue was only investigated in one meta-analysis [717]. This could not find a significant effect of specific psycho-oncological interventions in the palliative phase. No supporting results from RCTs were available in our search.

⁷ In Li et al. (2020) [713] non-standardized MD=-7.3 in Patient Dignity Inventory [747].

Pain (Statement 8.22)

The short- and medium-term efficacy of specific psycho-oncological interventions in the palliative phase with regard to pain was examined in one meta-analysis [589], and one RCT [724].

The meta-analysis showed neither short nor medium-term effects of specific psycho-oncological interventions in the palliative phase, the RCT could only detect significant medium-term effects of specific psycho-oncological interventions in the palliative phase.

The study data on the effectiveness of specific psycho-oncological interventions in the palliative phase for the reduction of pain in people with non-curable cancer has to be considered as not sufficiently robust.

The meta-analysis by Bauereis et al. [589] includes only five RCTs that consider pain reduction as an outcome. The results have high unexplained heterogeneity, which reduces confidence in the effect estimator. The results on medium-term efficacy are based on only one RCT. Furthermore, the meta-analysis includes populations of cancer patients who are not in the palliative phase, thus decreasing the match to the key question of this chapter. The RCT by Rodin et al. [724] has only a small sample, and the planned sample size could not be achieved. Because of the limitations mentioned above, the guideline group refrains from recommending for or against the use of specific psycho-oncological interventions in the palliative phase to reduce pain in people with non-curable cancer. Additional methodologically high-quality studies are needed to formulate a recommendation.

Supplementary evidence

In addition to the above results, one systematic review [719] describes significant short-term effects on at least one primary endpoint (including quality of life, distress, pain) in 6/9 adequately powered RCTs of meaning-based interventions. In 5/9 RCTs with sufficient stat. Power on dignity-based interventions also showed significant short-term effects on at least one primary endpoint.

Assessment of study quality

The study quality of the SRs was assessed using the AMSTAR-2 tool, and the risk of bias of the RCTs was assessed using the Cochrane RoB tool.

In total, 5/8 (62.5%) SRs had deficiencies in more than one critical assessment domain (critical low quality), 2/8 (25%) had deficiencies in one critical assessment domain (low quality). One (12.5%) SR had no critical deficiencies and only one non-critical deficiency (high quality). It should be noted that in 4/8 SRs there was a downgrading of the quality assessment due to the lack of a list of excluded primary studies and their reasons for exclusion (critical domain 7, AMSTAR-2).

12/17 RCTs (70%) used an adequate method to generate the randomization sequence (selection bias), and 8/17 RCTs (47%) ensured secrecy and unpredictability of group allocation (selection bias). In 12/17 RCTs (70%), the amount of missing data was sufficiently small (<20%) or missing data were appropriately imputed (attrition bias), and in 8/17 RCTs (47%), selective reporting of endpoints could be excluded by an existing study protocol (reporting bias). Due to the nature of the interventions used and the collection of endpoints by self-assessment, blinding of study personnel/participants (performance bias) and thus blinding of endpoint collection(detection bias) were not possible in any of the RCTs.

In total, 10/13 RCTs performed a priori sample size estimation. 3/10 RCTs achieved their a priori planned sample size.

Four new systematic reviews have been found for art therapy since 2012 [791], [795], [701], [799]. These included were a total of eight randomized trials and ten controlled trials from 2006 to 2019. Three of the primary studies (RCTs) included in the reviews [797], [798], [790] are also reported individually, since they were published in the 2012-2020 search period. The studies were from the United States, Asia, and Europe. With the exception of the studies from Monti et al. (N=184) [797] and Joly et al. (N=320) [796], they are based on smaller numbers of cases (n<60). The applied art therapy methods and also the duration of the interventions are heterogeneous in the various studies. One review [799], for example, compared mindfulness-based art therapy (MBAT) with other mindfulness-based therapy approaches (such as Mindfulness Based Stress Reduction (MBSR) and Mindfulness Based Cognitive Therapy (MBCT) and also included individual studies from the field of palliative care [701].

A review by Wood et al. [800] has already shown that art therapy is helpful for people with cancer in various treatment settings. The art therapy approach can help to cope with treatment-related symptoms and support the process of psychological readjustment to the loss, change, and uncertainty caused by cancer. Art therapy is now often offered in group settings, and less frequently in individual therapy sessions. Supportive techniques are often used, for example, to regulate emotions, to experience relaxation through artistic expression, or, increasingly in mindfulness-based interventions.

Meta-analyses and systematic reviews

With regard to depressiveness, anxiety, and distress (mental illness domain), significant improvements with small to medium effects are found in various reviews ([791], [795], [799]. For example, Zhang [799] found significant improvement (p<.001) in depression scores as well as anxiety scores (p= .005) in 7 RCTs (n=888). Two reviews [791], [795] demonstrated significant improvements in quality of life (p<.01, d= 0.03-1.05). The review by Kim et al. [701], however, showed mixed results regarding quality of life, whereas five RCTs showed significant improvements in quality of life. Pain was examined by Kim et al. [701] in their review and came to different conclusions regarding an improvement in pain in the individual studies. A significant improvement of fatigue was described by Jiang et al. [795] in their review in four studies (p=.01). Furthermore, coping strategies were investigated in the review by Azmawati et al. [791], but no significant improvements could be shown.

Primary studies

An RCT by Joly et al. [796] examined an intervention group with art therapy in addition to standard care compared to a control group (treatment as usual, TAU) during radiation therapy. There were significant improvements in the areas of social well-being and cognitive fatigue. Monti et al. [797] and Jang et al. [794] were also able to demonstrate a significant improvement in quality of life after 36 months (p=.01). Öster et al. [798] found mixed results in the various domains investigating quality of life.

The investigations of the domain „Mental Illness“ showed heterogeneous results in the studies examined [790], [792], [797]. The study by Radl et al. [790] showed no significant improvements on emotional distress (p=.26) and mental well-being (p=.12), but did find significant improvement on spiritual well-being (p=.02). In contrast, the study by Jalambadi et al. [793] showed significant improvement on the different levels of distress (p<.05). The study by Jang et al. [794] measured significant improvement on depressiveness (p<.001) as well as anxiety (p<.001).

For improvement in coping, a recent RCT by Czamanski et al. [792] found significant improvements in emotion perception (p=.027) and emotion acceptance (p=.039), which was similarly demonstrated by Öster et al. [798] in the significant improvement in coping strategy of “social relationships” (p=.04). A significant improvement in fatigue is described by Joly et al. (p=.007) [796]. However, the result here refers only to patients with mild fatigue and not to patients with “severe fatigue” within the one-month study period. Three RCTs are also included in the reviews by Azmawati et al. [791], Jiang et al. [795] and Zhang et al. [799].

Risk of Bias Assessment

The risk of bias assessment of the 7 RCTs for art therapy was conducted by three independent raters (two non-specialist research assistants from medicine and psychology, and one specialist from music therapy; supervised by the Guideline working group CAT (KüTh)) using the RoB 2 tool [801]. In summary, a low risk of bias was found in three primary studies and a high risk of bias in four primary studies. The assessment took place with respect to five categories. The high Risk of Bias was found exclusively in Category 3 “bias due to missing outcome data”, i.e. due to missing data in the post test. The underlying drop-out of more than 20% of the patients is mostly explained clinically by the stressful oncological treatment with complex side effects and changes in life circumstances (relocations, etc.). The missing data was handled with appropriate methods in all four cases (intention-to-treat analysis, etc.). With the exception of these four cases, the methodological quality was homogeneous and high, with a low risk of bias.

Conclusion

Art therapy interventions show significant effects on the improvement of depressiveness, anxiety and distress; likewise from four studies to date for improvement of fatigue. The results of other outcome parameters examined, such as quality of life, coping strategies, pain reduction, and spiritual well-being, are heterogeneous. In addition, the duration of intervention and the type of art therapy interventions varied widely. The study data is characterized by strongly converging evidence of art therapy findings on the four outcomes mentioned above. With the exception of a drop-out rate of over 20% in four primary studies, the risk of bias assessment shows a high quality in the new primary studies (2012-2020).

Music therapy (MT) is „(…) the targeted use of music within the framework of the therapeutic relationship to restore, maintain and promote mental, physical and spiritual health“ [835]. Various music therapy concepts and methods (Neurological MT by M. Thaut; Guided Imagery and Music by Helen Bonny a.o.) are grouped under this heading and can basically be assigned to the following two approaches: Functional MT - for the treatment of primarily physiological complaints and Psychotherapeutic MT – with predominantly psychological treatment goals. The latter makes use of active methods (playing an instrument, singing, songwriting, etc.) and receptive methods (listening to individualized music, musical relaxation interventions, sound journeys, etc.), in which different mechanisms of action, both of the music as well as the therapeutic relationship, can be used. MT sessions can take place in individual and/or group settings, depending on the treatment goal and approach.

Music medicine (MM) refers to the use of music as a therapeutic tool by individuals without therapeutic training and outside of a therapeutic relationship (e.g., medical staff handing out an MP3 player/playlist to patients). Studies on music medicine were not included in the systematic literature search but are reported in some of the reviews that were included.

In the period from 2012 to 06/2020, a total of seven systematic reviews with meta-analyses on the use of music (therapy) and its influence on psychological and physiological parameters in cancer patients were published [803], [804], [802], [805], [806], [807], [808]. In addition, 25 new primary studies on oncological patients (excluding palliative patients) with a LoE 1 are available.

Meta-analyses and systematic reviews

The effect of music-based interventions on reducing anxiety is described as significant in six reviews. Bradt et al. [802] showed significant improvements in anxiety on the IGs compared to CG in 13 RCTs (N=1,028, significant improvements IGs vs. TAU; SMD = -0.71, 95% CI: -0.98; -0.43, p<.001) with a moderate to large effect. Zhang et al. [803] found a significant improvement on anxiety across different measurement tools (7 RCTs, N=607, SAS: M = −2.84, 95% CI:−19.51;−6.17, p<.001; 2 RCTs, N=515, HAMA: M=−1.85, 95% CI: -3.43;-0.27, p<.05; 8 RCTs, N=681, STAI: M=-12.30, 95% CI: -18.93;-5.68, p<.001) with small to moderate effect sizes. Li et al. [808] showed significant improvement on anxiety in the IG compared to CG on the basis of six RCTs (6 RCTs, N=447, SMD = −1.51, 95% CI: −2.27;0.75; p<.05), Gao et al. [805] based on three RCTs (N=121, SMD=0.68, 95% CI: 1.35;0.02; p=.04) and Wang et al. [806], in four RCTs (Anxiety: 2 RCTs, N=200, HAMA: MD= –7.04, 95% CI: –9.31;–4.78, p<.00001; 5 RCTs, N=268; SAS: MD=–7.40, 95% CI: -10.28;-4.52, p<.001; 2 RCTs, N=110, State Anxiety Inventory: MD=-12.40, 95% CI: -21.86;-2.94, p=.01). In their systematic review, Nightingale et al. [804] found a significant anxiety reduction in the IG compared to the CG in 85% (11/13) of the studies included: in the meta-analysis of 4 of the 13 included studies, no significant difference between the IG and the CG with regard to anxiety was found.

The effect of music-based interventions on reducing depressive symptoms is largely described as significant in five reviews. Wang et al. [806] were able to show a significant effect on depressivity based on the results of 14 studies from MT and MM (6 RCTs in MT, MD=-7.39, 95% CI: -8.35;-6.43, p<.001). Similarly, Zhang et al. [803] (7 RCTs, M=-6.23, 95% CI: -8.85; -3.36; p<.001) and Li et al. [808] (6 RCTs, SMD=-1.12, 95% CI: -1.87;-0.38, p<.05) showed significant results with regard to depressivity. In the Cochrane review by Bradt et al. [802], a short-term, moderate reduction in depressiveness through active music therapy interventions (SMD= -0.40, 95% CI: -0.74;-0.06, p<.02) was described. In their meta-analysis, Köhler et al. [807] found a significant improvement in mental well-being, reaching across aspects of anxiety, depressiveness, mental distress and well-being, mood, emotional functionality, relaxation, hope, and resilience in the IG (active MT) versus CGs (TAU, waiting list, and active CG: listening to music, visiting, mindfulness practice) (19 RCTs, d=0.35, 95% CI: 0.19;0.50, p<.01). It was found that a lower number of sessions had greater effects (d=0.47) than more frequent sessions (d=0.18). There were also greater effects with receptive MT (d=0.33) than with active or multimodal methods (d=0.19). In addition, a positive effect on depressiveness as well as mood was found. A significant improvement of physiological distress with small effect (physiol. distress: d=-0.34, 95% CI: -0.55;-0.13, p<.004) was also described. Gao et al. [805] were able to show a significant reduction in depressiveness in their review (4 RCTs, N=227, SMD: 1.08, 95% CI: 1.64;0.53, p<.001). Here, no significant improvement in social functioning was found (3 RCTs, SMD=0.51, 95% CI: 0.17;1.19, p=.14).

For improvement in quality of life, a significant effect was found in five reviews. The Cochrane Review by Bradt et al. [802] showed a significant increase in quality of life, also comparing MT versus MM, with moderate effects in three studies (SMD=0.42, 95% CI: 0.06;0.78, p=.02). Gao et al. [805] were able to find significant improvement on health-related quality of life in five studies (SMD=0.61, 95% CI: 0.41;0.82, p<.001). Köhler et al. [807] described a significant improvement with a small effect in their review on quality of life (d=0.48, 95% CI: 0.11;0.85, p=.019), although no significant effect was found for session frequency. Similarly, Li et al. [808] (10 RCTs, SMD: 0.54; 95% CI: 0.40;0.69; p<0.05) and Zhang et al. [803] (2 RCTs, M=13.32, 95% CI: 11.01;15.62, p<.001) described a significant effect for the improvement on different facets of quality of life.

Four out of five reviews could demonstrate a significant reduction in pain. Thus, Bradt et al. [802] (SMD: -0.91; 95% CI: -1.46;-0.36; p=.001), Li et al. [808] (SMD: -1.12; 95% CI: -1.87;-0.38; p<.05), Zhang et al. [803] (M:-0.54; 95%CI: -0.88; -0.20; p<.005), and Gao et al. [805] could determine a significant effect from MT on various parameters of pain. In addition, the review by Köhler et al. [807] described a positive effect on pain.

Inconsistent results regarding the significance on fatigue are found in three reviews. In the review by Bradt et al. [802], small to moderate improvements in fatigue were demonstrated in six studies with small to moderate effects (SMD: -0.38; 95% CI: -0.72;-0.04; p=.03). Zhang et al. [803] also report significant results on fatigue (2 RCTs, M: 0.63; 95% CI: -2.52; 3.77; p=.70). In the review by Gao et al. [805], no significant differences were found between intervention and control groups for fatigue (2 RCTs, SMD: 0.17, 95% CI: 0.95;0.61, p=.67).

Primary studies¹⁰

Twenty-two primary studies were included in the analysis. Individual RCTs are already included in one or more of the reviews and meta-analyses described above. Since the overlapping studies also include older primary studies prior to the designated search period, only the evidence from RCTs after 2014 will be presented separately below.

Music therapy significantly reduced anxiety in the majority of available studies. In the study by Palmer et al. [822] with inpatient breast cancer patients (N=201), the applications of listening to music pre- and perioperatively were compared live vs. recorded music compared to TAU. Anxiety was significantly reduced in both music IGs compared to TAU (MD: -26.5; d=-0.57; MD: -25.7; d=-0.55). By Romito et al. [826], anxiety was significantly improved in both the IG as well as the CG (d=0.69; p=.000). Wang et al. [806] (HAMA: p<.001; SAS: p<.001; State Anxiety Inventory: p=.01) as well as Yates and Silverman [833] (p=.009) each found significant improvement on anxiety in the IGs. Rossetti et al. [828], found a reduction in anxietythrough MT (10 sessions à 35 minutes listening to music with therapist, before and during treatment) during radiation treatment in gynecologic patients (MD: 8.2; p=.002). Dóro et al. [814], found significant group differences with regard to anxiety (IG vs EK: p<.001), whereas Bradt et al. [812] failed to find a significant effect with regard to anxiety (SMD: 0.21; 95% CI: -0.09;0.51, p=.18).

A significant reduction in depressiveness in the IG compared to the CG was demonstrated in the study by Alcântara-Silva et al. (p=.011) [809], as well as in Chen et al. [836]. Romito et al. [826] investigated the use of one session of integrative group MT (N=31) à two and a half hours concomitant with outpatient chemotherapy in 62 patients with breast cancer. Twenty-minute talks (N=31) were used as a control group. The MT group showed significant improvements for depressiveness (d=0.52; p=.000) and anger (d=0.58; p=.000). With regard to negative affect (SMD: -2.86; p=.003) and positive affect (SMD: 2.15; p=.001), the IG achieved greater improvements than the CG in the studies by Fredenburg and Silverman [815], [816].

With regard to hope and helplessness, two studies showed improvements. According to research by Chen et al. [836], the use of music-imagery in eight sessions à 60 minutes led to a reduction in helplessness and hopelessness, as well as cognitive improvement in 52 patients with breast cancer compared TAU. A study by Verstegen et al. [831] compared MT with therapeutic dialogue versus TAU in 10 patients with blood and bone marrow transplantation. The two sessions lasted between 30 and 60 minutes. The parameter hope was significantly improved here in the IG as well (post-test: p=.035).

The parameter stress improved significantly in the study by Romito et al. (IG: d=0.62; p=.036) [826]. Rossetti et al. [828] showed a reduction of distress through MT (IG: p=.0064). Regarding symptom-related physical stress, Fredenburg and Silverman [815], [816] found no significant difference between the intervention and waitlist groups.

Regarding quality of life, a study by Bieligmeyer et al. [811] showed significantly lower scores after MT intervention (single session à 10 minutes of vibroacoustic-combination of auditory and sensory sound experiences) compared to the CG (quiet time) (d=0.37; p<.001). A study by Alcântara-Silva et al. [809] showed a greater improvement in quality of life when listening to music together with a therapist before and after radiation therapy for gynecological oncology patients (N=164; 10 sessions of 35 minutes each) compared with TAU (p=.005). In a prospective palliative study by Warth et al. [832], 2 sessions of MT (30 min each at least 2 days apart) vs. MBSR were compared. With regard to quality of life, the authors found no significant group differences between intervention and control groups for the total quality of life scale in the pre-post comparison (F[1;81]=2.23, p=.14, MD= 3.26 [95%, CI: -5.99/ 12.5], η2=.03, SMD=.14 NNT=7). Ratcliff et al. (2014) [824] compared two different MT formats (ISO principle-based vs unstructured music (UM); 4 sessions each à 50 minutes) vs. TAU in patients after stem cell transplantation and found significant group differences in quality of life improvement comparing baseline to follow-up (p=.001) in favor of ISO principle-based MT (MT > UM > TAU).

The investigation of pain showed improvements in the IG, but no significant effect in the group comparison [811], [812], [814], [818], [831]. Bates et al. [810] examined the effect of two MT sessions à 30 minutes for a total of seven days on patients with multiple melanomas or lymphomas during their first stem cell therapy. Compared to the CG (TAU), there was no significant group difference regarding pain at any time (p=.73). The use of analgesics was not significantly different between IG and CG (p=.73), but it significantly increased in both groups at follow-up measurement (p=.001). Tuinmann et al. [830] investigated the effect of playing music, singing and listening to music in 66 lymphoma patients over eight sessions à 20 minutes versus TAU. They were able to demonstrate that the need for analgesics and the subjective pain perception were lower in the EG than in the CG (SMD: 10; 95% CI: 1.2;18.9; p=.027). Wang et al. [806] investigated the effect of perioperative music relaxation and postoperative music listening in five sessions à 60 minutes in 60 patients with lung cancer. The control conditions were TAU. The effect shown for the EG was a greater reduction in the need for analgesics. In 200 patients with advanced-stage cancer, Gutgsell et al. [817] found greater reduction in pain after a 20-minute session of music relaxation than in the wait-list CG (MD: -1.39; 95% CI: -1.95;0.83; p<.001).  In two studies on blood and bone marrow patients  [816], [815], Fredenburg and Silverman examined the effect of music therapy (various techniques) in one session à 30-45 minutes. Regarding pain, a greater decrease was demonstrated in the IG (SMD: -2.86; p=.009). Letwin and Silverman [818], examined 15 patients with different oncological diagnoses (oncology/hematology department) before and after two sessions of music listening à 30-45 minutes. Compared to the waiting-list CG, no significant differences in pain (p=.891) and resilience (p=.079) were demonstrated. In a study by Palmer et al. [822], no reduction in the need for anesthetics was seen with MT compared to the CG (LM vs. CG: MD:-3.3; 95% CI: -19.0;12.4; d=-0.09; RM vs. CG: MD: -8.6; 95% CI: -20.8;3.5; d:-0.24). Warth et al. [832] also found no significant group difference in pain perception between IG vs. the CG (F[1;81]=.40, p=.53, MD= -.18 [95%, CI:0.74/ 0.39], η2=.01, SMD=-.15 NNT=8.40). In contrast, Ratcliff et al. [824] described significant group differences in terms of improvement in cancer-related symptoms (such as pain, nausea, etc.) in the comparison between baseline to follow-up (p=0.01) in favor of ISO principle-based MT. 

No improvement in fatigue was found by Fredenburg and Silverman (p>.05) [815]. However, Alcântara-Silva et al. [809] observed a significant reduction in fatigue in the IG compared to the CG (p=.01). In the study by Warth et al. [832], the authors demonstrated significant improvements in fatigue in favor of the IG vs. the CG (F[1;81]=4.74, p=.03). 

The study results on the outcome mood are heterogeneous. The study by Bieligmeyer et al. [811] showed significant effects for current mood (d=0.43; p<.001) as well as for mood and well-being (p=0.47; p<.01), whereas no improvements were found in the CG. Especially the subscales of the BMQ used for inner balance (d=0.73; p<.001), vitality (d=0.51; p<.001) and alertness (d=0.37; p<.001) achieved significant improvements for the IG. Further significance in favor of the IGs was shown for the parameter satisfaction (d=0.43; p<.001). Palmer et al. [822] could not observe any improvement in satisfaction as a result of the live music intervention. In a study by Dóro et al. [814], stem cell therapy patients (N=100) were examined. Here, singing and musical Improvisation took place in eight sessions à 30 minutes. The CG received TAU. Mood improved in the IG more than in the CG (p<.001). Bradt et al. [812] offered music therapy employing different therapeutic techniques versus music listening as an active CG to inpatients and outpatients with different cancer diagnoses (N=31). Two sessions à 20-45 minutes did not lead to any significant group differences with regard to mood (MT: p<.0001 MM: p>.0001). Bates et al. [810] also found no significant group differences in the measurement of mood. Ratcliff et al. [824] showed significant improvement in mood after stem cell transplantation over four sessions of MT for both MT groups (p<.0001), with no significant difference detectable between the different MT interventions. Warth et al. [832] also described a significant improvement in the mood of palliative patients in the IG vs CG (F[1;81]=6.41, p=.01, MD= .98 [95%, CI: 0.28/ 1.68], η2=.61, SMD=.61 NNT=5.25).

Improved spiritual well-being in the IG compared to the CG was found by Cook and Silverman [813] (subscales: Hope: MD: 3.856; p=.011; Peace: MD: 2.194; p=.018; Meaning: MD:0.416; p=.677). They described the use of listening to music listening versus talking for 34 hematologic oncology patients (leukemia and other diagnoses). In three sessions à 15-30 minutes, the IG was compared with a waiting list CG.

Lin et al. [820] were able to show an improvement in relaxation with MT compared to TAU. Yates and Silverman [833] implemented one session of music listening à 20-30 minutes in 26 inpatients with colon, rectal, or uterine cancer. Compared to the waiting list group, increased relaxation was achieved in the experimental conditions (p=.009). Warth et al. [832] were also able to demonstrate a significant improvement in relaxation in favor of the IG vs. CG (F[1;81]=13.7, p<.01, MD=1.19 [CI: 0.50;1.88], η2=.75, SMD=.75, NNT=2.80) in a pre-post comparison after two 30-minute sessions of MT in palliative patients. Bradt et al. [812] did not find significant group differences with regard to anxiety (MT: p=.0002; MM: p<.0001).

For postoperative recovery time, Palmer et al. [822] showed significant results for live music compared to the use of recorded music (p=.018), but not compared to the TAU group. For 86 cancer patients (with various diagnoses) in outpatient treatment, Burns et al. [837] found improved responsiveness and post-traumatic growth (benefit finding/increased resilience) in a single session (45-50 min) with guided visualization with music compared to decreasing scores in the CG.

When examining physiological outcomes, the study by Bates et al. [810] showed no significant effect in regard to nausea. In the follow-up survey, the CG performed significantly better (p=.035). Moradian et al. [821], used active listening with the Nevasic audio program (pre-recorded audio program) during chemotherapy to reduce treatment-related nausea and vomiting in breast cancer patients (N=99). Listening to music in one group and TAU in another group were used as CGs. No group-specific differences were found with regard to nausea and vomiting. In the IG, the use of antiemetics significantly decreased compared to both CGs (p=.003). Warth et al. [832] found in their study in the palliative setting a significantly greater increase in the high-frequency component of heart rate variability in favor of the IG vs. CG (F [1;81]=8.13, p=.01, MD= .44 [95%, CI: 0.12/.77], η2=.09, SMD=.59 NNT=3.50). When blood volume and –amplitude were examined, only a trend toward vasodilation in favor of the IG vs. CG (F(1;81)=3.33, p=.07, MD= .11 [95%, CI:-.03/ .24], η2=.04, SMD=.35, NNT=42) was demonstrated.

¹⁰ For three of the 25 primary studies at evidence level 1, the full text is only available in Chinese, so they could not be described in more detail and no risk of bias assessment was performed. Their results are included in the review by Wang et al. (2018) [806]. 

Risk of Bias Assessment

The Risk of Bias assessment of the 22 RCTs in MT was performed analogously to art therapy and, in summary, found a low Risk of Bias in 15 studies, "some concerns" in four studies, and a high Risk of Bias in three studies. The assessment took place with regard to the five categories of the RoB 2: The high Risk of Bias occurred exclusively in Category 3, "bias due to missing outcome data", due to missing data in the post-test. From a clinical point of view, a high dropout rate (of more than 20% for the classification as high risk) can mostly be explained by the stressful oncological treatment with complex side effects or changed life circumstances. Missing data were handled with adequate methods (intention-to-treat analysis, etc.). Missing or insufficient information in the publications led to the assessment of some concerns for one study in the Category 4, “bias in the measurement of the outcome“, and in three studies in Category 5, “bias in the selection of the reported results“. Except for these cases, the homogeneity of methodological quality was high with a low risk of bias in 15 studies.

Conclusion

The overall findings of music therapy in oncology are positive. There is increasing methodological improvement in conducting clinical trials. Within the creative arts therapies, MT represents the majority of studies on evidence level 1. Across all treatment stages, stable, positive effects on the enhancement of health-related quality of life, the reduction of anxiety and depression, and the reduction of pain can be found for different diagnostic groups and treatment methods. The study findings for fatigue reduction, mood improvement and relaxation, are mixed. Here, further studies are needed to investigate possible positive effects from MT. Other outcomes with hope, satisfaction, positive affect, and spiritual well-being improved with MT; feelings of helplessness and anger could be reduced. With regard to the physiological parameters, the study data is heterogeneous so far. Overall, further studies with high methodological quality are needed in order to determine the process-related influences and mechanisms and contextual factors underlying the improvements.

Five new reviews ([839], [838], [840], [842], [845]) are available for dance and movement therapy for the period 2012 to 2020. The Cochrane review by Bradt et al. (N=207) [838] includes the primary studies by Dibbell-Hope [846], Sandel et al. [847], and Ho and Fong [848], a conference abstract of the later published study by Ho et al. [843]. The other reviews ([839], [840], [842] are summarized by Koch and Bräuninger in a review [845] which included 13 studies, of which six were RCTs with a total of 282 participants. Of these, two studies are at evidence level 1 ([841], [843], [844]). Other studies have evidence levels 2-4 and are not discussed here.

Meta-analyses and systematic reviews

Results on anxiety reduction are mixed. On the one hand, Boing et al. [839] reported a reduction in fear of recurrence (p=.02). On the other hand, dance and movement therapy interventions did not significantly reduce anxiety (p=.18) in two of the RCTs [846] [848] included in the Cochrane review by Bradt et al. [838].

Mixed results were found in the reviews with regard to depressiveness. On the one hand, in their systematic review, Boing et al. [839] found significant results demonstrating a decrease in depressiveness (p=.01) in one RCT [849]. On the other hand, two RCTs [846], [848] from the review by Bradt et al. [838] found no significant reduction in depressiveness (p=.89).

Regarding the outcome measure stress, only one RCT [844] showed a significant reduction in stress through dance and exercise therapy with a medium effect (d=0.33, p=.039). This study was included in the systematic reviews by Koch and Bräuninger [845], and Boing et al. [839]. Two RCTs [846], [848] from the review by Bradt et al. [838] were unable to demonstrate any significant improvement (p=.24).

For quality of life, Boing et al. [839] summarized the outcomes in regard to mental health (p=.01), environment (p=.029), body image (p<.01), and overall quality of life (p<.01). They also report an improvement in the physical aspects of quality of life (p<.001). The systematic review by Koch and Bräuninger [845] found an increase in quality of life in one RCT [847], while two other RCTs did not demonstrate a significant improvement in quality of life [843] [841]. It should be noted that two RCTs [846][847]  from the Cochrane Review by Bradt et al. [838] also found no improvements with regard to body image (p=.58), whereby body image was assessed by different measurement tools in the studies.

With regard to patients' experience of pain, one RCT [843], which was included in the systematic review by Koch and Bräuninger [845] as well as Boing et al. [839], showed a significant reduction in pain intensity (d=.36; p=.035) and pain interference (d=.35; p=.048).

Regarding fatigue, two RCTs [846], [848] in the Cochrane review by Bradt et al. [838] did not show significant improvements (p=.44). The reviews by Boing et al. [839], Goodill et al. [842], and Koch and Bräuninger [845] also report the results of the aforementioned RCTs.

Other positive findings are reported by Boing et al. [839] for social functioning (p=.04), mental health (p=.04), emotional role (p=.07), well-being, and mood (p<.05). Koch and Bräuninger [845] report the results of an RCT [849], which demonstrated a significant improvement regarding life satisfaction (d=2.69; p<.001).

In addition, the influence of dance therapy interventions on various physiological variables was investigated. In the review by Boing et al. [839], one RCT each reported an improvement in functional capacity (p=.03), physical activity (p=.01), strength (p<.05), grip strength and arm volume (p=.028), as well as physical function (p=.001).

Primary studies

The studies described (RCTs) are already included in the reviews by Boing et al. [839] and Koch et al. [845], but are presented again in detail here.

Fear of Recurrence was significantly reduced (p=.02) in the intervention group from Crane-Okada et al. [841], whereas the intervention by Ho et al. [843] showed no effect on participant’s anxiety. Regarding perceived stress, Ho et al. demonstrated a significant reduction in their study (d=0.33; p=.039) [843]. In a secondary analysis of the data on the question of moderator effects of the interventions on stress parameters, Ho et al. [844] found no significant effect of the intervention on cortisol levels over the course of the day after three weeks (B=-0.55; 95% CI: -1.2;0.08).

There was also no significant effect on the quality of life of the intervention groups after a Mindful Movement Program or after six concomitant sessions of DMT [841], [843]. No significant improvements were found on the severity (d=0.01; p=.98) and interference (d=0.14; p=.43) on fatigue, sleep disturbance (d=0.20; p=.24) and depressiveness (d=0.53; p=.11) [843]. Significant improvements were seen in the aforementioned studies in terms of severity (d=0.36; p=.035) and impairment due to pain (d=0.35; p=.048) [843] as well as mindfulness of the IG (post-test: p=.026) in the study by Crane-Okada et al. [841]. 

Risk of Bias Assessment

The risk of bias assessment of the 2 RCTs for dance and movement therapy was conducted analogously to the process for art and music therapy by three independent raters and showed a consistently low risk of bias (low RoB) across four assessment categories, whereby in Dimension 2, possible bias can be suspected due to the lack of blinding as well as an unequal implementation of the intervention in the study by Crane-Okada et al. [841] (see GL-Report).

Conclusion

The current studies and reviews on dance and movement therapy show an improved, but heterogenous picture with regard to most of the target criteria investigated. Only in a few studies demonstrated a significant reduction of pain and the effects of pain and perceived stress. In addition, in the context of psycho-oncological rehabilitation, there was an increase in life satisfaction, a decrease in the fear of recurrence, and an improvement in mindfulness after dance and movement therapy. The findings on quality of life, depression, fatigue, and upper body mobility were mixed. No effects could be demonstrated for improvement in sleep and body image, although in the latter case, difficulties in the measurability of the construct and a high heterogeneity of the outcome measures should be taken into account.

One should take into account that it is necessary to look not only at the effects but also at the mechanisms of action in order to gain a deeper understanding of the results (for initial findings see: De Witte et al. [850]).

From 2012 to 2020, six high-quality, overlapping reviews of the creative arts therapies (CAT) were published, which include 42 studies, 39 of which were RCTs from art, music, and dance therapies (N=3,726) [851], [852], [853], [854], [855], [856].

Converging empirical evidence  

Findings on anxiety reduction through creative arts therapies (CAT) are largely positive. Hertrampf et al. [853] reported a significant reduction in anxiety trough CAT in five RCTs (N=478; post-test: 0.12[Tang, Yong et al. 2019] included nine RCTs (N=747) and found a significant reduction in anxiety through CAT in four RCTs (N=337, SMD=-0.48; 95% CI: -0.75;-0.21; p=.0005). A significant reduction in anxiety through CAT was also demonstrated in three RCTs from Archer et al. [851] and in four RCTs from Boehm et al. [852]. Rabadán et al. [854] included 11 studies (N=680) with CAT in patients with breast cancer and found no effect of DMT on anxiety, but a significant reduction in anxiety with MT.

Significant results were also recorded for the reduction of depressiveness. A significant reduction in depressiveness with CAT resulted in five RCTs (N=304) (post-test: 0.14[Hertrampf, Ruth- S. et al. 2017]. Tang et al. [856] (SMD=-6.59; 95% CI: -7.70; -5.47; p<.00001) and Archer et al. [851] also reported positive results in terms of a reduction in depressiveness in two RCTs (N=180) and three RCTs (N=213), respectively. Boehm et al. [852] demonstrated an improvement in depressivity in the IG in five RCTs (N=241), but this was not significant (M: -0.30; 95% CI: -0.60;0.00; p=.05). Only one study in the review by Rabadán et al. [854] failed to show any effects of DMT on depressive symptoms (p=.53) [843]. 

A reduction in stress was shown not only for CAT in four RCTs (N=262) from the review by Hertrampf et al. [853] (post-test: 0.07[Ho, Rainbow T. H. et al. 2016]. 

With regard to an increase in quality of life, Hertrampf et al. (post-test: 0.12[Hertrampf, Ruth- S. et al. 2017] and Archer et al. [851] each found improvements in quality of life trough CAT in four RCTs (N=260) and five RCTs, respectively. However, no effect on sleep or quality of life was found by Rabadán et al. [854] after dance and movement therapy. The results from Boehm et al. [852], showed no significant group differences in four RCTs (M=0.15; 95% CI: -0.09;0.40; p=.22) as well.

Positive results are consistently reported for pain reduction. In the review by Hertrampf et al.  [853], a moderate to large effect could be observed regarding the reduction of pain on the post-test in four RCTs (N=430) from MM (post-test: 0.54[Rabadán, Jennifer et al. 2017/1/1] showed that dance therapy significantly reduced both the severity (p=.035; d=0.36) and interference of pain (p=.048; d=0.35) [843].

Non-converging evidence

The findings of the studies on the reduction of fatigue show heterogenous results. While Tang et al. [856] found a significant improvement in fatigue in one RCT (N=60; SMD: -1.90; 95% CI: -2.93;-0.87, p=.0003), Ho et al. [843] found no effect of DMT on fatigue (severity: p=0.98; d=0.01; interference: p=.43; d=0.14). Finally, two RCTs (N=170) from CAT on fatigue found no to a moderate effect at post-test (immediately after treatment in Dibbel-Hope  [846]; no measurement in Ho et al. [843]), and a small to large effect at follow-up after 3 weeks. Follow-up effects were partly larger than post-test effects (post-test: 0.42[Hertrampf, Ruth- S. et al. 2017]. In another study included in the review by Rabádan et al. [854], fatigue was examined from different aspects. Art therapy significantly reduced fatigue (p=0.007; for patients with mild fatigue), it provided an increase in social well-being (p=.02; for patients with severe general fatigue who had received chemotherapy), it led to an increase in quality of life (p=.034), emotional well-being (p=.017), social well-being (p=.025), and motivation (p=.03; in patients with initially low motivation), and significantly reduced fatigue after six months for patients with severe fatigue [Joly, F. et al. 2016].

The results on the improvement of body image are heterogeneous, which can be largely attributed to the heterogeneity and problematic nature of the outcome measures on body image. The specific review on the outcome body image by Lewis-Smith et al. [855] included three CAT studies up to 2017 (N=110). The authors found no significant improvements in body image trough AT and DMT. The review by Hertrampf et al. [853] included two RCTs (N=79) from CAT on body image improvement, both of which showed no effect or a small effect (0.01[Rabadán, Jennifer et al. 2017] and Archer et al. [851], found improvements in body image. 

With regard to strengthening coping strategies (strategies for dealing with illness), no to small effects were reported (0.1[Hertrampf, Ruth- S. et al. 2017]. A review of art therapy also showed no effects on coping strategies [798]. In contrast, the review by Archer et al. [851] found significant improvements in coping strategies in one RCT.

Furthermore, two RCTs (N=79) showed small to medium effects (0.04[Hertrampf, Ruth- S. et al. 2017] of CAT on mental health.

Conclusion 

The number and quality of systematic reviews on CAT has increased significantly in recent years. Nevertheless, in many reviews, no meta-analysis could be performed due to the heterogeneity of the primary studies. The heterogeneity can be explained through the use of a wide range of methods, interventions, and measurement instruments. Only Boehm et al. [852] and Tang et al. [856] calculated overlapping effect sizes. Boehm et al. [852] was able to show a significant reduction in anxiety (four RCTs; M= -1.10; 95% CI: -1.4;-0.88; p<.01), whereas Tang et al. [856], in the more recent study, in addition to also demonstrating a significant reduction in anxiety (four RCTs; SMD=-0.48; 95% CI: -0.75;-0.21; p=.0005), was also able to show a significant reduction in depression (two RCTs; SMD=-6.59; 95% CI: -7.70;-5.47; p<.00001) and a significant improvement in fatigue (one RCT) (SMD=-1.90; 95% CI: -2.93;-0.87; p=.0003) through CAT. Thus, the consistently demonstrated reduction in anxiety is the most consistent finding for CAT, with long-term effects also emerging. These findings on the three main outcomes of the overlapping studies are reflected in the individual recommendations and strengthen the evidence base, especially with regard to depression and anxiety reduction.