Introduction

Nearly 18 million individuals are diagnosed with cancer every year [1]. Cancer is, therefore, a major cause of morbimortality and will continue to impose for long the highest clinical and socioeconomic disease-related burden worldwide for a long time [2]. Patients with cancer face physical impairments during and after treatment, often associated with increased levels of pain and fatigue [3, 4]. In addition, the complex and uncertain course of the disease [5] also leads to psychosocial challenges [6], i.e., anxiety and depression [7]. Cancer treatments usually focus on disease recurrence [8] and ongoing physical symptoms [7]. Yet, people with cancer now demand a more person-centered and comprehensive approach [9] that can address mental health problems [8].

Non-pharmacological therapies are of interest for the clinical management of long-term diseases, as considered to be safe, low-cost, and with minor side effects [10]. Among them, music-based interventions have shown to be useful in chronic conditions to improve the physical and emotional well-being in individuals with fibromyalgia [11] or affective disorders [12] and seem to help to modulate cancer-related symptoms [13,14,15,16]. Music-based interventions can be categorized as ‘music medicine’, i.e., passive listening of recorded music offered by healthcare staff, or ‘music therapy’ that encompasses the clinical use of music in all its forms, as provided by a credentialed therapist [16, 17]. Although both terms are often interchanged [18], a clear distinction is that music therapy involves individualized assessment, intervention, and evaluation, and a patient-therapist relationship that develops through the music [19]. Music-based interventions are characterized for using music in a passive or interactive modality (engaging a patient to create live music) and can be applied alone or within a multimodal program [16, 20,21,22,23,24]. Music is a highly structured language that engages complex cognitive, affective, sensory, and motor control processed in the human brain [25, 26]. Listening to music can reduce the activity of the autonomic nervous system, and improve the synchrony of the neural firing, which promotes brain plasticity [27]. Music can also appeal to strong emotional and social responses [28]. This provides a neural basis for the biological impact of music [29] and its influence on the physical and mental health[30]. Several systematic reviews have recently investigated the effectiveness of music-based interventions in cancer care [31, 32]. An overview of these systematic reviews can provide a high-level synthesis of evidence [33, 34]. It can also address the transparency of information and the methodological biases of previous research [34, 35], which may help to understand the clinical relevance of current evidence. The aim of this overview was to gather and assess the available evidence from systematic reviews with meta-analysis on the effectiveness of music-based interventions on physical and psychosocial outcomes in adults diagnosed with cancer.

Methods

The overview protocol was prospectively registered at the Open Science Framework (https://doi.org/10.17605/OSF.IO/Y67BU). This overview has followed the preferred reporting items for overviews of reviews (PRIO) statement and the PRISMA for abstracts [36, 37]

Deviations from intended protocol

There were no major deviations from the registered protocol.

Search strategy

One researcher (ATM) carried out an electronic search from inception to November 2022 in the following databases: CINHAL, Embase, PEDro, PubMed, Scopus, the Cochrane Library, and Web of Science. Medical Subject Headings (MeSH) terms associated with the intervention (music) and the medical condition (e.g., cancer, neoplasm) were combined. A comprehensive search strategy was first constructed for PubMed and then adapted for other databases. The lists of references of previous overviews were manually checked. The detailed search strategies are listed as Supplementary file A.

Eligibility criteria

The eligibility criteria were established following the PICOs framework (Population, Intervention, Comparison, Outcome, Study):

  • P: Adults diagnosed with cancer without restrictions in body location/system or the cancer stage.

  • I: music-based interventions, used alone or as adjuvant to usual or standard care

  • C: no restrictions regarding the control intervention.

  • O: physical (e.g., pain, fatigue), and psychosocial measures (e.g., anxiety, depression, mood, distress, and quality of life).

  • S: systematic reviews with meta-analysis [38].

Systematic reviews were not included when: a) the publication was written in a language other than Spanish or English; b) there were not meta-analyses for the condition of interest; c) music-based interventions were meta-analyzed together with other experimental treatments; and d) meta-analyses included non-adult participants, population without cancer, or non-randomized controlled trials. Possible outcomes of interest that were not analyzed in at least two systematic reviews were not considered. Congress proceedings, thesis dissertations, and network meta-analyses were also excluded.

Study selection

Duplicate records were removed using the Mendeley desktop software, v2.72.0. and manually checked. One researcher (ATM) screened the remaining records based on the title and the abstract. The full text of eligible studies and those lacking an abstract were then revised. A consensus was achieved for three studies with a second researcher (AMHR) who independently double-checked the entire selection process.

Data extraction

Data were extracted with a standardized form that included: a) first author plus et al., the year of publication, and the number of clinical trials of interest; b) sample size (total and the experimental group); c) the characteristics of participants (age, sex, type of cancer); d) description of the experimental and control interventions; e) music style used; f) outcome measures; and g) main results from meta-analysis. We aimed to extract the overall effect size from each meta-analysis. When this was not reported, results from sub analyses were included. Two corresponding authors were contacted by e-mail to clarify some information [39, 40]. A reminder was sent, if necessary, one week after the first message. None of those contacted responded.

Methodological quality

Two independent reviewers (ATM and MJCH) evaluated the methodological quality of systematic reviews using the AMSTAR-2 tool [41]. As recommended, individual ratings of the 16 items were not combined to obtain an overall score [42]. Instead, the attention was given to critical weakness domains, namely: item 2, prospective review protocol; item 4, comprehensive search strategy; item 7, justification of the excluded studies; item 9, risk of bias; item 11, appropriateness of statistical analysis; item 13, interpretation of results based on the risk of bias; and item 15, publication bias [42].

Spin in abstracts of systematic reviews

The abstracts of the systematic reviews were assessed in isolation to quantify the occurrence of spin of information. Two independent reviewers (ATM and PGG) utilized a 7-item checklist [43], where each item was assigned a score of ‘yes’ or ‘no’.

Data synthesis

Findings have been narratively described based on the outcomes of interest. To identify the most relevant key terms across systematic reviews, the VOSViewer software, v. 1.6.18 (Leiden University, The Netherlands) was used to conduct a co-occurrence analysis and bibliometric mapping. The degree of overlap of primary studies among included systematic reviews was evaluated with the Graphical Representation of Overlap for OVErviews (GROOVE)[44]. The GROOVE tool provides a simple, graphical and comprehensive representation, including the number of overlapped and non-overlapped primary studies and the overall assessment of the “Corrected Covered Area” (CCA), along with the CCA value for each pair of systematic reviews. For the CCA, the degree of overlap is considered to be slight (0–5%), moderate (6–10%), high (11–15%), and very high (CCA > 15%) [45]. Additionally, the CCA was measured taking into account chronological structural missingness, i.e., when primary studies were published after a systematic review [44].

Results

Search strategies retrieved a total of 926 eligible records. After removing duplicates, 466 records were screened. We eventually included 13 systematic reviews and 34 meta-analyses in the qualitative synthesis (Fig. 1). A list including the reports excluded during the final screening phase (n = 29) is described in the Supplementary file B.

Fig. 1
figure 1

PRISMA flowchart

Full size image

Table 1 shows the characteristics of the included systematic reviews [39, 40, 46,47,48,49,50,51,52,53,54,55,56]. The most common types of cancer were breast and haematological, i.e., lymphoma and leukaemia. Music-based interventions were often combined with specific cancer treatments (e.g., surgery, chemotherapy or radiotherapy) or with standard or usual care, and involved passive listening of live or recorded music or patient’s active engagement (e.g., singing, clapping, and guided music imagery). Different music styles, selected by therapists or patients’ preferences, were used. A 23% of systematic reviews judged the overall certainty in the evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach [49, 55, 56]. Most reviews (77%) assessed the risk of bias of the clinical trials using the Cochrane Risk of Bias tool.

Table 1 Characteristics of the included systematic reviews
Full size table

Methodological quality

Results for the AMSTAR 2 tool are described in Table 2 (inter-rater agreement, 78.8%). The most important methodological concerns were ‘the lack of comprehensive search strategies’, ‘no information of the excluded studies’, and the ‘interpretation of the review findings without accounting for the risk of bias of primary research’. More than 90% of systematic reviews did not inform of why they included a certain type of study design or about their funding sources.

Table 2 Risk of bias (AMSTAR 2) of the included systematic reviews
Full size table

Spin of information in abstracts

The overall spin-abstract score was 21, with a mean value of 1.6 ± 1.3 points (inter-rater agreement, 79%). The most common forms of spin were ‘concluding a positive effect despite high risk of bias of primary trials’ (n = 7), and ‘selective reporting or overemphasis on the beneficial effect of music-based intervention’ (n = 5). No spin of information was found in three abstracts [46, 47, 52] (Supplementary file C).

Co-ocurrence analysis

Twelve out of the thirteen systematic reviews were included in the co-occurrence analysis (Figs. 2 and 3). One review did not include key terms [56]. The pattern of association between keywords has been reflected in the network and density visualizations The terms most frequently used were related to the research design (meta-analysis, systematic review), the intervention (music interventions, music), and the disease (cancer, neoplasms).

Fig. 2
figure 2

Density visualization

Full size image
Fig. 3
figure 3

Network visualization

Full size image

Overlapping between primary study

A total of 202 primary studies were identified across the included systematic reviews, out of which 119 were distinct studies. The overall overlap for the entire matrix of evidence was moderate (CCA = 5,81%) and this remained moderate (CCA = 6,92%) even after adjusting for chronological structural missingness. The citation matrix and the CCA calculation can be found in Supplementary file D. The Supplementary file E presents the graphical representation of the GROOVE tool. Three primary studies from one of the systematic reviews could not be retrieved due to the insufficient information and a lack of response from the corresponding author [39].

Music-based interventions on cancer-related pain

All systematic reviews measuring pain as an outcome (n = 6) concluded that music-based interventions plus usual or standard care were more effective than control interventions (e.g., usual or standard care, wait-list, bed rest, or wearing headphones with no music) to reduce cancer-related pain [39, 46, 48, 49, 52, 56].

Music-based interventions on cancer-related fatigue

Among the five systematic reviews assessing cancer-related fatigue, four of them indicated that combining music-based interventions with usual or standard care could yield more benefits than control interventions to improve cancer-related fatigue [39, 50, 54, 56]. However, one systematic review found no differences between groups [49].

Music-based interventions on cancer-related anxiety

Inconclusive conclusions were detected upon evaluating the six systematic reviews assessing cancer-related anxiety [40, 47, 49, 51, 55, 56].

Music-based interventions on cancer-related depression

Inconclusive conclusions were detected upon evaluating the seven systematic reviews investigating cancer-related depression [39, 40, 49, 51, 53, 55, 56].

Music-based interventions on cancer-related mood and distress

Two systematic reviews concluded that music-based interventions together with usual or standard care could be more effective than controls in reducing cancer-related distress [49, 56]. However, findings on patients’ mood varied across studies [49, 56].

Music-based interventions on cancer-related quality of life

Two out of the three systematic reviews on quality of life demonstrated that music-based interventions combined with usual or standard care were superior to usual care alone in improving health-related quality of life [55, 56], while one review found no significant differences between groups [49].

Adverse events of music-based interventions

Four systematic reviews provided information regarding potential adverse events. In all of these reviews, no adverse events were observed following music-based interventions [46, 50, 52, 56].

Discussion

This overview summarized the evidence from systematic reviews with meta-analysis on the effects of music-based interventions to modulate cancer-related symptoms in adults. Overall, our results seem to suggest that adding music interventions to usual or standard care could be more beneficial than usual care alone to reduce cancer-related pain, fatigue, and distress. On the other hand, findings were inconclusive for anxiety, depression, mood, and quality of life.

The present results expand those of previous overviews underlying the importance of including music-based within a multimodal treatment to decrease cancer-related pain [33, 57]. However, this is the first overview specifically focused on music-based interventions. Music involves cognitive engagement and distraction [58, 59]. Listening to music can help to the release of endogenous opioids and dopamine [58], which supports music-induced analgesia and may contribute to reduce the severity of fatigue [54]. Cancer-related pain is a complex, evolving, and multifaceted phenomenon [60], comprised of several dimensions (sensory. discriminatory, emotional, cognitive, and behavioral) [61]. Complementary integrative therapies, such as music-based interventions, can effectively manage cancer pain [61]. However, music may exert distinct influences on the different dimensions of pain, thus contributing to divergent findings observed in both physical and psychological measures. The impact of music on cancer-related fatigue has found to be highly relevant when music is combined with other therapies, e.g., exercise, especially when the intervention involves active patient’s engagement [50]

Current clinical practice guidelines recommend the use of music to manage the cancer-related psychological burden during and after treatment [62, 63]. However, the exact mechanisms to understand how music engagement contributes to mental health remain unknown [64]. We found inconclusive results for anxiety, depression, mood, and the quality of life. This is in line with prior findings reported in palliative cancer care [65], but contradicts those for non-adult cancer populations [66, 67]. This might be because children and adolescents with cancer do not have as many comorbidities as adults and often tend to respond better to treatment. The style of music, along with personality, cultural, and contextual factors have an influence on the effects of music [64, 68]. Also, the diversity of tools used to measure anxiety and depression may contribute to the inconsistency of results [68]. In summary, more definite conclusions could be drawn with less heterogeneity in participants’ characteristics, especially age and cancer stage, assessment tools, and music intervention protocols.

Clinical relevance

This overview provides an updated synthesis of evidence about the use of music as an adjuvant therapy for adults in cancer care. Given that music is a potential cost-effective intervention [58], the present findings seem to encourage clinicians to implement its use into daily practice. There are, however, barriers that need to be overcome, mostly related to the lack of practical guidelines for music dose and timing [69]. Researchers have a strong responsibility to provide a complete description of interventions. That is the sole purpose of the TIDieR checklist, designed to advance evidence-based clinical practice [70]. However, none of the included systematic review provided information about how well described music-based interventions were in primary trials, which detracts from their replicability. Other important aspects should be born in mind. First, a clear distinction between music medicine or music therapy can be clinically relevant but it was only made in two of the systematic reviews [55, 56]. Music therapy was superior to music medicine to improve the quality of life and fatigue [56], but worse than music medicine for reducing anxiety [55]. These results suggest that the person who conducts the intervention and the mode of delivery may be clinically relevant. Second, treatment benefits following music interventions may depend on patients’ characteristics, e.g., emotional vulnerability [71]. Third, the lack of adverse effects suggests that music is a safe intervention in this population, although information regarding potential adverse events was only reported in four systematic reviews [46, 50, 52, 56]. Finally, most systematic reviews did not clarify whether ‘standard’ or ‘usual’ care included supportive cancer care, as a paradigm for modern treatment in oncology [72], to manage the physical, psychological, social, and spiritual needs of patients [73], or specific cancer treatments such as chemotherapy. This needs to be clarified in future systematic reviews.

Methodological concerns

We have addressed, for the first time, potential biases, and transparency of information of systematic reviews in this field. The main concerns were related to the search strategy and the interpretation of the results without accounting for potential risk of bias. This may lead to selection bias and to an inaccurate translation of the findings to the clinical setting. It is somehow concerning that 40% of the reviews ‘overemphasized’ the beneficial impact of the music intervention group. Unfortunately, this misleading presentation of results is not new in the context of cancer treatment [74]. The certainty in the evidence using the GRADE framework was only evaluated in three systematic reviews [49, 55, 56]. In addition, as previously stated, the presence of adverse events, which is highly relevant, was poorly reported. Both aspects need to be carefully considered to improve the standard of quality. We incorporated chronological structural missingness to calculate the degree of overlap with the GROOVE tool, which is a novel and interesting approach. The GROOVE may also enable the analysis of overlap for specific outcomes, but this feature was not considered due to the high heterogeneity of measurement tools among the included reviews. Finally, evidence from clinical trials need to be complemented by qualitative studies to get a whole idea of music as individualized therapy.

Limitations

Literature search screening was conducted by a single researcher. Congress proceedings, network meta-analysis and reviews not written in Spanish or English were excluded, thus meaningful information may have been overlooked. The PICOs question considered music-based interventions in general and was not limited to music therapy or music medicine. In addition, the overlap of clinical trials among reviews prevented us to conduct meta-meta-analysis or to evaluate the certainty in the evidence.

Conclusions

Based on our results, we can conclude that:

  • The combination of music-based interventions with standard or usual care could be more effective than standard care alone to reduce cancer-related pain, fatigue, and distress in adults diagnosed with cancer.

  • The additive effect of music-based interventions to standard or usual care remains uncertain for anxiety, depression, mood, and the quality of life.

  • Clinical and methodological concerns have been discussed and should be carefully considered when interpreting our findings in a clinical context.