A systematic review with evidence mapping of supportive care interventions for melanoma patients and caregivers

Abstract Aim We conducted a systematic review and evidence gap mapping to explore the existing supportive care interventions and their impact on well‐being outcomes for melanoma patients and caregivers. Methods We searched MEDLINE, Embase, Web of Science Index Medicus, CINAHL, Lilacs, CENTRAL (Cochrane Library) and PsycINFO in December 2022, including interventional studies assessing the effectiveness of any supportive care intervention among melanoma patients and/or their caregivers. Findings Twenty studies were included in this review. These studies consisted of randomised controlled trials (n = 11, 55%), pre‐post studies (n = 7, 35%) and quasi‐experimental trials (n = 2, 10%). All studies originated from high‐income countries and focused primarily on melanoma patients, with no studies identified that focused solely on caregivers. Educational interventions were the most common (n = 7, 35%), followed by psychoeducational interventions (n = 6, 30%) and psychotherapeutic interventions (n = 4, 20%). Nearly all included studies (n = 18, 90%) reported a positive effect of the intervention on the primary outcome of interest; however, most studies (n = 17, 85%) were judged to be at moderate or high risk of bias. Due to heterogeneity of study designs, intervention characteristics and outcome measures, meta‐analysis was not conducted. Implications Supportive care interventions have positive impacts on melanoma patient well‐being outcomes, while being acceptable and feasible to conduct. More research is needed regarding supportive care interventions for melanoma caregivers. Future research should focus on eliminating sources of bias through rigorous methodology, with the development of standardised outcome measures for psychosocial outcomes to facilitate future meta‐analyses.


| INTRODUCTION
Melanoma, the deadliest form of skin cancer, represents 90% of skin cancer mortality and was responsible for an estimated 57,000 deaths globally in 2020. 1,2 The global incidence of melanoma has steadily increased, 1 and is predicted to increase a further 18% in the 2020-2025 period. 3 However, despite the increasing global incidence of melanoma, the global mortality rate has seen progressive improvements throughout the last decade owing to advancements in targeted and immune therapies. 2,4 In 2017, the overall 5-year survival rate for people diagnosed with melanoma was 93%. 5 As a result, there is a growing emphasis on the well-being of melanoma patients during diagnosis, treatment and into survivorship, as many of the therapies leading to long-term survival are associated with persistent, and often severe, adverse effects that can impact daily functioning and health-related quality of life (HRQOL). 6 Furthermore, several patient-reported outcomes that are common among melanoma patients (≥25%) have been associated with poorer patient wellbeing, such as fear of melanoma recurrence/progression, anxiety, depression and sleeping difficulties. 7 Poor patient well-being is also exacerbated by approximately half of melanoma patients (between 45% and 56%) reporting decreased HRQOL as a result of having at-least one moderate or high severity unmet supportive care need. 8 Of these unmet needs, psychological needs (76%), informational needs (64%) and physical needs (59%), were the most commonly reported, all of which are associated with decreased HRQOL. 8 This highlights the supportive care needs of melanoma patients, not only in the domain of treatment-related toxicities, but also in regard to their well-being. These needs are also influenced by the stage of the patients' melanoma; American Joint Committee on Cancer Stage 0 being melanoma in situ, ranging to Stage IV being wide spread metastatic melanoma. 9 Historically, supportive care interventions for melanoma patients typically involved the management of treatment-related side effects 10 ; however, this field has recently evolved to more broadly improve patient well-being by addressing unmet needs and improving the different aspects of HRQOL. 11 There are several pillars of an effective supportive care intervention, including: patient-centred care; the inclusion of caregivers; temporality of support; multidimensional and holistic care; multidisciplinary and coordinated care; evidence-based interventions; and adaptability. 11 Supportive care interventions need to offer high adaptability to address patient needs, often utilising interventions focusing on psychological support, education, exercise, nutrition and/or behavioural change. Thus, although inherently complex and multifaceted, supportive care interventions may improve aspects of HRQOL of patients throughout their melanoma journey, with evidence demonstrating this effectiveness in other cancer settings. [12][13][14][15] As the supportive care and well-being of melanoma patients and their caregivers is becoming an increasingly vital part of clinical practice, it is crucial that the available evidence regarding the effectiveness of these interventions is synthesised. This systematic review and evidence gap map (EGM) aim to synthesise the existing supportive care interventions aiming to improve well-being that are available for melanoma patients and their caregivers, while highlighting gaps in the available evidence for future research.

| METHODS
The protocol for this systematic review was registered with the International Prospective Register of Systematic Reviews (registration ID: CRD42022296812). The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines 16 were used to guide the reporting of this systematic review (PRISMA 2020 Checklist provided in Supplementary Material 1). Furthermore, standard EGM methodology was used to synthesise evidence from existing studies assessing the effectiveness of supportive care interventions for melanoma patients and caregivers. 17

| Types of studies
Interventional studies assessing the effectiveness of supportive care interventions in melanoma patients and their caregivers were included. Eligible study designs included individually and cluster-randomised controlled trials (RCTs), quasi-experimental trials, controlled pre-post studies, uncontrolled pre-post studies and intermittent time series. Observational studies, cross sectional surveys, case reports, case studies, opinions, editorials, commentaries, letters, conference abstracts, books, grey literature and narrative or systematic reviews were excluded. Qualitative studies were also excluded.

| Types of participants
Studies conducted among melanoma patients and/or their caregivers of any age or sex were included. Studies that included patients diagnosed with other cancers as well as melanoma were only included if melanoma-specific outcomes were reported.

| Types of interventions
Supportive care interventions included (but were not restricted to) educational (i.e. focusing primarily on the provision of information regarding melanoma diagnosis, treatment and prevention), psychoeducational (i.e. providing information, skills and techniques to facilitate better coping), psychotherapeutic (i.e. providing therapeutic intervention to facilitate better coping), exercise, nutritional and behavioural interventions, or any combination of these. Pharmacological interventions were excluded.

| Types of outcome measures
Primary outcomes of interest included psychological distress (including but not restricted to stress, anxiety and depression), fatigue, fear of new or recurrent melanoma and HRQOL. Secondary outcomes of interest included acceptability, feasibility, coping strategies, melanomarelated knowledge, melanoma-related health behaviours (including but not restricted to sun exposure, sun protection and skin self-examination), satisfaction with care or information and unmet supportive care needs. No studies were excluded based on reported outcomes.

| Search methods for the identification of studies
Searches were conducted in November-December 2022 in the following databases: MEDLINE, Embase, Web of Science Index Medicus, CINAHL, Lilacs, CENTRAL (Cochrane Library) and PsycINFO (search strategy provided in Supplementary Material 2). The reference list of all included studies and relevant systematic reviews were also reviewed to identify studies missed during the electronic search. In the case of missing/unpublished studies or data, corresponding authors were contacted. No language or date restriction for publication was employed.

| Selection of studies
Search results were exported into Covidence 19 before duplicates were removed for screening. Within Covidence, three investigators (RAS, SH, JRT) independently screened titles, abstracts and full texts for potential inclusion. Any discrepancies were resolved by consensus or contacting a fourth investigator (IB). Attempts were made to contact authors of included studies to obtain clarifications or additional data if required.

| Data extraction
Data were extracted for key study characteristics and outcomes in a standardised data collection form. Two investigators (RAS, JRT) independently extracted data and any discrepancies were resolved through discussion until consensus was reached or by consulting a third investigator (IB). Data were extracted on the following categories:

| Data synthesis
Included studies were categorised based on the aforementioned variables and risk of bias assessments. Furthermore, a clinical psychologist (IB) reviewed each included study to categorise the type of intervention, as well as its therapeutic modality. All included studies were then entered into EPPI-Reviewer 20 and its accompanying program to visually synthesise the evidence. Using these programs, an EGM matrix was generated to illustrate the overlap between intervention category and primary outcome, while illustrating study design and whether the intervention had a positive or non-positive impact on the primary study outcome. Where data was available from two or more studies utilising the same category of intervention on the same primary outcome where pooling was considered meaningful (i.e. if the underlying clinical questions were sufficiently similar), a meta-analysis was planned. However, due to heterogeneity of study designs, intervention characteristics and outcome measures, meta-analysis could not be performed.

| Quality assessment
The Cochrane risk of bias tool version 2 (RoB 2) 21 was used to assess the risk of bias in the included RCTs, while the risk of bias in non-randomised studies of intervention (ROBINS-I) 22 was used to assess risk of bias in non-randomised studies. Two investigators (RAS, JRT) assessed risk of bias with disagreements resolved by consensus or through consulting a third investigator (IB).

Relevant ancillary publications
Study design Melanoma stage Specificity b

| Intervention effectiveness
The majority of included studies reported a positive impact of their respective interventions on the primary outcome of interest. Interventions that were educational in nature were found to improve the efficacy of both patient and caregiver/relative skin self-examinations and sun protection behaviours, 24,37 increased melanomarelated knowledge, 36 and participant satisfaction with melanoma knowledge at diagnosis and throughout treatment, 25 while being acceptable and feasible to conduct. [32][33][34] The included studies also suggest that psychoeducational interventions reduce distress and improve participant coping, 23,28,29 reduce severity of fear of cancer recurrence, 27 and are acceptable and feasible to conduct. 38 Educational -

Primary outcome result Conclusions
Intervention group reported significantly larger decrease in total mood disturbance (p = 0.04) and significantly more behavioural (p < 0.01) and cognitive (p < 0.01) coping skills at 6-month than control group. These effects were no longer present at 12-month Decreased psychological distress and enhanced coping At 1-year follow-up both patients and FDRs in the intervention group reported significant improvements in thorough full-body skin self-examinations, seeking shade and avoiding the outdoors in strong sunlight (p < 0.01) Improved skin self-examination and sun protection The intervention group had a significant improvement in melanoma-related knowledge (p < 0.01), with a majority (89%) of participants reporting a positive attitude towards the intervention at 6-month The intervention group reported significantly more behavioural and cognitive coping skills postintervention (p < 0.01, p = 0.04) and at 6-month (p < 0.01, p = 0.3). This group also reported significantly less total mood disturbance at 6-month (p < 0.01)

Decreased distress and enhanced coping
Participants who received therapy reported significantly lower total mood disturbance at 3-month compared to baseline (p < 0.01), and when compared to the control group (p < 0.05). The intervention group also reported significant decreases in the general severity index (p = 0.02), positive symptom distress index (p = 0.03) and total score (p = 0.02) of the Brief Symptom Inventory at 3-month compared to baseline The intervention did not have a significant impact The intervention reported a 55% recruitment rate and 90% retention rate. Qualitative feedback from participants indicated the intervention PROMs were easy to complete, and consultations with a nurse specialist were timed appropriately and useful The intervention was acceptable and feasible Adherence to the intervention exceeded 75%, with 93% of participants completing follow-up with supportive care physician, exercise physiologist, and dietician. Exercise adherence was 85%, and 85% of participants also accessed integrative therapies The intervention was acceptable and feasible Ninety-seven percent of participants completed both consultations, with the mean total nursing time to prepare and deliver the intervention being 19 min (SD: 5) and 96 min (SD: 15), respectively. Eighty-six percent of participants stated the intervention met their approval and 83% stated they liked the program

The intervention was acceptable and feasible
Ninety-eight percent of participants agreed the screening questionnaires were easy to complete and understand. The mean time to complete the questionnaire was 3.25 min (SD: 2.30). Participants reported that the intervention was very important and acceptable The intervention was acceptable and feasible The intervention group reported a significant (p < 0.01) increase in melanoma knowledge than the clinical visit control group post-intervention. However, the control group reported a significant decrease in participant anxiety (p < 0.01) and distress (p < 0.01) compared to the intervention group The videotape effectively improved knowledge, but standard care better addressed anxiety/ distress distress, 26,40 improve HRQOL, 40 and are acceptable and feasible to conduct. 31,35 Art therapy was found to enhance patient coping and interpersonal relations when conducted in groups, 41 while supportive counselling was acceptable and feasible. 30 Only two studies reported a non-positive interventional effect. Skovlund et al. 39 reported that the use of patient-reported outcomes to facilitate clinician-led discussions with patients during a clinical appointment did not improve patient confidence in self-management, which may have been due to baseline ceiling effects. Gordon et al. 42 further reported that psychodynamic psychotherapy did not significantly reduce anxiety and depression in melanoma patients, likely due to the poor prognosis of these patients at time of publication.

| Risk of bias
The assessment of bias in the included studies are provided in Table 3A for RCTs and Table 3B for non-randomised designs.
Of the 11 included RCTs, three were determined to be at low risk of bias, 27,37,38 while some concerns were raised regarding five [23][24][25]36,40 and the remaining three were classified as high risk. 26,28,29 The most common domain where concerns were raised was the selection of reported results domain (n = 8, 73%), [23][24][25][26]28,29,36,40 as most RCTs did not have a pre-specified analysis plan available prior to the publishing of results. This was only classified as 'some concerns' and not 'high risk' for the effected studies, as it was assumed the included studies had pre-specified analysis plans in place, they were simply not publicly available (i.e. in a published study protocol). Four RCTs raised 'some concerns' or a 'high risk' of bias regarding the Randomisation Process domain, 25,28,29,36 as detailed information on the randomisation process used was not available. Finally, two RCTs were classified as 'some concerns' and 'high risk' due to high loss-to-follow-up rates. 26,40 Of the nine non-randomised studies, two were classified as having a 'moderate risk' of bias, 39,42 while seven were classified as having a 'serious risk'. [30][31][32][33][34][35]41 The majority of 'serious risk' is evident in the Confounding domain (n = 7, 78%), [30][31][32][33][34][35]41,42 as few studies utilised statistical analysis techniques that account for confounding variables. It should be noted, however, that many of the included non-randomised studies were acceptability and feasibility trials, and thus, are more susceptible to inherent biases due to their study design.

| Evidence gap map
The EGM of included study designs, primary intervention categories and primary study outcomes is provided in Figure 2. As evident in the EGM, substantial gaps in the evidence still exist regarding the supportive care interventions in the melanoma space. Educational, psychoeducational and psychotherapeutic interventions have the most evidence for affecting supportive care outcomes, but other types of intervention such as exercise, music and nutrition-based interventions are under-represented. While evidence-based interventions are emerging for addressing distress, improving coping and encouraging melanoma-related preventative behaviours, more research is needed to explore ways to increase melanomarelated knowledge, satisfaction with the provision of this information, adjustment to melanoma diagnosis, patient self-management and decrease fear of cancer recurrence.

| DISCUSSION
To the best of our knowledge, this is the first systematic review to explore the scale and depth of the current, available evidence on supportive care interventions for individuals diagnosed with melanoma and their caregivers, while identifying gaps to guide future research. Findings from this systematic review suggest a positive impact of educational, psychoeducational and psychotherapeutic supportive care interventions on patient distress and coping, melanoma-related knowledge, adjustment to melanoma diagnosis and fear of cancer recurrence, while being acceptable and feasible, in participants aged over 40 years on average and regardless of gender. However, the existence of bias was considerable in both randomised and non-randomised studies. Furthermore, there is a clear scarcity of evidence regarding supportive care interventions aimed at supporting caregivers. Thus, the findings of this systematic review and EGM have several implications. First, a substantial proportion of included studies were acceptability and feasibility pilot studies, [30][31][32][33][34][35]38 which did not seem to progress past the pilot stage. This may be because six of the seven pilot studies were published in the 3 years (2019-2022) prior to the search conducted for this systematic review. 30,[32][33][34][35]38 As the introduction of immune and targeted therapies have resulted in melanoma patients living longer than ever

Primary outcome result Conclusions
Participants who received the in-person, workbook, or electronic intervention reported significantly more skin self-examinations at 4-, 12-, and 24-month post-intervention than the control group (p < 0.01). Participants who received an intervention also identified more recurrent melanomas than the control group (p < 0.05)

Electronic interactive interventions can effectively deliver preventative education
The intervention was found helpful by 72% of participants, with at least 61% of participants utilising meditation each week of follow-up. Participants also reported numerous benefits, challenges, and feedback for the intervention The intervention was acceptable and feasible The intervention did not significantly improve knowledge, skills, or confidence in participant selfmanagement (p = 0.37), coping self-efficacy (p = 0.31), and patient-perceived patient-physician interaction efficacy (p = 0.42) at 12-month post-intervention The intervention did not improve knowledge, skills, or confidence in self-management Both ITT and EOI analysis showed the intervention significantly (p < 0.05) reduced participant anxiety at 2-and 6-month follow-up and increased vitality and general health at 2-month follow-up. EOI analysis demonstrated the intervention significantly reduced distress at 2-month follow-up (p = 0.01)

Reduced distress and improved HRQOL
The intervention significantly (p = 0.02) increased participant levels of secretory immunoglobulin A over a 6-month period. The intervention also significantly reduced total mean avoidance at 6month (p = 0.05) and mental disengagement at 6-(p = 0.02) and 12-month (p = 0.05) Enhanced immunological function, coping skills, and interpersonal relations before, 2,4 a recent shift of focus has occurred regarding the emerging field of supportive care. This may explain why the majority of pilot studies have been conducted in the previous 3 years, with associated RCTs possibly planned or underway in 2023 and beyond. Therefore, if not already planned or underway, these interventions should be further explored in rigorously designed RCTs to investigate their impact on well-being outcomes, with reasons for the failed progression past the pilot study phase investigated if relevant. Second, the identified supportive care interventions were primarily educational, psychoeducational or psychotherapeutic in nature, and concerned with the informational and psychological needs of participants. As supportive care remains an emerging field, there are several therapeutic approaches that have shown promise in other oncological populations that remain unexplored in the melanoma space. These other approaches include exercise, dance, music, massage and nutrition-based interventions in individual and group formats, 61,62 which have shown promise in reducing fatigue, nausea and improving diet in general cancer populations. Thus, interventions based on these approaches may represent viable methods to address prevalent the unmet social and physical needs of people diagnosed with melanoma, 8 and should be investigated.
Third, the majority of psychoeducational and psychotherapeutic interventions utilised behaviour therapy. Two recent systematic reviews have highlighted that third wave cognitive-behavioural therapies, in particular acceptance and commitment therapy (ACT), are effective in treating anxiety, depression, distress, poor quality of life and fear of cancer recurrence. 63,64 Although melanoma patients are reporting similar concerns, 65,66 these interventions are yet to be implemented in this context. To date, ACT has been implemented mostly with women diagnosed with breast cancer, and to a lesser extent in ovarian, colorectal, brain and testicular cancers. 63,64 Melanoma diagnosis and treatment often result in adverse effects, which can affect the ability to participate in meaningful activities. 66 ACT aims to educate and encourage flexible thinking, in order to accept the experiences associated with painful life events, and engage meaningfully and fully with activities in life, 67,68 which appears to be a viable yet unexplored method to address the challenges faced by melanoma patients.
Fourth, the majority of interventions were conducted individually and face-to-face, with few studies exploring the utility of digital technology to deliver individual or group interventions. As a result of the COVID-19 pandemic, telehealth and digital technologies have become increasingly implemented in routine medical care, 69 and represent an opportunity to overcome barriers to accessing supportive care interventions, especially in rural/ remote areas where melanoma is more prevalent. 70,71 Thus, melanoma-specific supportive care interventions delivered through digital technology should be further explored in future research.
Fifth, the majority of interventions were conducted by one facilitator, [24][25][26][27]29,31,32,34,35,[37][38][39][40][41] with few interventions employing multidisciplinary teams. 23,28,30,33,36,42 Evidence from interventions conducted in head and neck cancer patients have demonstrated that the implementation of multidisciplinary teams increases adherence to guidelines and protocols, is cost-effective, reduced time to treatment and further support and increased patient and clinical satisfaction. 72 In future studies, multidisciplinary teams should be employed to implement supportive care interventions within the melanoma space due to these positive outcomes through the co-ordination of support services and increased ability to holistically address unmet needs.
Sixth, most studies had at least a moderate risk of bias. To address the domains where bias was most prevalent, future RCTs should endeavour to provide sufficient detail regarding the randomisation process and ensure pre-specified analysis plans are publicly available. Non-randomised studies should prioritise the use of analysis methods that account for possible confounding and effect modification, while employing robust follow-up routines to prevent participant loss-to-follow-up. However, removing risk of bias regarding measurement of the outcome in both randomised and non-randomised studies will be difficult, given the widespread adoption of self-reported outcome measures ensuring outcome assessors (i.e. the participants) will be aware of whether they received the intervention or not.
Penultimately, studies were too heterogeneous to conduct meta-analysis. Many interventions did not report a standardised frequency or duration of intervention, with results being unclear regarding what the appropriate frequency or duration of intervention is. Furthermore, each intervention focused on unique ranges of melanoma patients (Stages 0-IV), with early-stage melanoma being the most researched group, with interventions in advanced melanoma patients less prevalent. Each included study also differed not only in outcomes of interest, but in the tools used to measure these outcomes. Future research can mitigate these points of heterogeneity by developing and utilising melanoma-specific standardised outcome measures. Furthermore, future research should investigate whether interventions should be tailored to specific stages of the disease or adapted across different stages. This future research should consider the changing context of melanoma patients due to improvements in available treatments, such as the introduction of immune checkpoint inhibitors and targeted therapy improving long-term survival of Stage IV patients, 73 and the introduction of adjuvant therapy for patients diagnosed with Stages II and III. 74 Through addressing these points of heterogeneity, effective comparison of intervention effectiveness can be conducted through the successful pooling of data for meta-analyses, informing guideline development.
Finally, more research is needed regarding the effectiveness of supportive care interventions for caregivers of melanoma patients, as evidence from other oncological populations has demonstrated the significant burden and supportive care needs of caregivers, 75 with some interventions beginning to emerge. 76,77 Further research is also needed regarding the effectiveness of supportive care interventions for younger melanoma patients, as melanoma is one of the most common cancers among individuals aged 20-39 years in the United States, 78 and the most common cancer among individuals aged 15-24 years in Australia. 79 Therefore, future research should aim to investigate the effectiveness of supportive care interventions in younger melanoma populations.
This literature review and EGM synthesises the effectiveness of supportive care interventions in melanoma patients and caregivers globally, allowing for a comprehensive snapshot of the existing evidence within this field. However, we acknowledge the possibility of bias being introduced at every stage of our review process. To address this, a comprehensive search strategy was utilised with a list of pre-determined electronic databases to capture eligible studies. Furthermore, a minimum of  two investigators assessed eligibility of identified studies, extracted data and evaluated the risks of bias. While we attempted to be as inclusive as possible in our search strategy, all identified studies were published in English and originated from high-income countries. Furthermore, although we attempted to assess reporting bias, this assessment largely relies on information available in published articles and supplementary material, and thus, may be underrepresented. While supportive care interventions in the melanoma context are an emerging field of research, this systematic review and EGM has demonstrated that they are effective in improving patient well-being, and acceptable and feasible. Our confidence in these interventions will increase with higher quality, rigorous research conducted using standardised outcome measures while assessing a broader range of interventions and delivery methods, which may increase choice and accessibility to the appropriate supportive care intervention for both melanoma patients and caregivers.

FUNDING INFORMATION
This work and the researchers time was supported by Melanoma Institute Australia and the Bill and Patricia Ritchie Foundation.

ACKNOWLEDGEMENT
Open access publishing facilitated by The University of Sydney, as part of the Wiley - The University of Sydney agreement via the Council of Australian University Librarians. F I G U R E 2 Evidence gap map of included study designs, intervention categories and outcomes.