Characteristics and Components of Self-Management Interventions for Improving Quality of Life in Cancer Survivors: A Systematic Review

Simple Summary Self-management interventions can improve clinical and psychosocial outcomes for cancer survivors. However, we do not know which intervention characteristics (i.e., how they are delivered) and components (i.e., what they deliver) are beneficial. This can influence the implementation of such interventions into routine cancer care. We aimed to identify existing self-management interventions for adult cancer survivors, describe their characteristics and components, and investigate associations with quality of life. We identified 32 interventions. Studies had varying quality. A total of 22 studies reported significant improvements in quality of life, associated most often with combined individual and group delivery. Self-management interventions showed promise for improving the quality of life in cancer survivors; however, caution is required because the intervention characteristics and self-management components delivered varied considerably. Still, we highlight what may be worth adapting from existing interventions. Overall, these findings provide the foundations to help inform the development and implementation of self-management interventions for cancer survivors. Abstract Self-management can improve clinical and psychosocial outcomes in cancer survivors. Which intervention characteristics and components are beneficial is unclear, hindering implementation into practice. We systematically searched six databases from inception to 17 November 2021 for studies evaluating self-management interventions for adult cancer survivors post-treatment. Independent reviewers screened for eligibility. Data extraction included population and study characteristics, intervention characteristics (TIDieR) and components (PRISMS), (associations with) quality of life (QoL), self-efficacy, and economic outcomes. Study quality was appraised, and narrative synthesis was conducted. We identified 53 papers reporting 32 interventions. Studies had varying quality. They were most often randomised controlled trials (n = 20), targeted at survivors of breast (n = 10), prostate (n = 7), or mixed cancers (n = 11). Intervention characteristics (e.g., provider, location) varied considerably. On average, five (range 1–10) self-management components were delivered, mostly “Information about condition and its management” (n = 26). Twenty-two studies reported significant QoL improvements (6 also reported significant self-efficacy improvements); these were associated most consistently with combined individual and group delivery. Economic evaluations were limited and inconclusive. Self-management interventions showed promise for improving QoL, but study quality was variable, with substantial heterogeneity in intervention characteristics and components. By identifying what to adapt from existing interventions, these findings can inform development and implementation of self-management interventions in cancer.


Introduction
Due to advances in treatment, the number of people living with, and beyond a cancer diagnosis, is growing in developed countries [1,2].Despite improving prognoses, longterm consequences of the diagnosis and its treatment mean cancer survivors often face physical or psychosocial/emotional problems, such as cancer-related fatigue, anxiety and depression, physical impairments, or social challenges that can be detrimental to quality of life (QoL) [3][4][5].Individuals need to learn to manage these challenges, which may persist for years following treatment [5,6].
In cancer, self-management has been defined as "awareness and active participation by the person in their recovery, recuperation, and rehabilitation to minimise the consequences of treatment, promote survival, health and well-being" [7].Engagement in self-management is important for adjustment to a "new normal", managing issues with healthcare, psychological well-being, and re-establishing routine and social roles [8].For an individual to effectively self-manage, they are likely to require support from others to ensure that they are appropriately equipped with the necessary knowledge and skills [9].There is a maturing evidence base on self-management interventions in cancer survivors; it is suggested they can improve numerous clinical, psychosocial and economic outcomes in cancer patients, such as QoL, physical and psychological well-being [10][11][12][13], and reduce healthcare utilisation [14,15].Underpinning social cognition theories indicate that this is achieved by empowering self-efficacy, through training, education, and skill development [16].
The Practical Reviews in Self-Management Support (PRISMS) taxonomy [17,18] comprehensively classifies possible self-management support components (e.g., training/rehearsal for psychological strategies, monitoring of condition with feedback).However, we lack knowledge about the components that comprise existing interventions [10].In addition, there is substantial heterogeneity in cancer survivors' design preferences for the characteristics of a self-management intervention (e.g., design, content, mode of delivery) [19].Identifying similarities in findings across interventions that have been tested would facilitate clearer judgements of which characteristics (e.g., number and length of sessions) and components (e.g., training, equipment) might best contribute to effectiveness.Such information could be valuable to researchers looking to adapt existing interventions (e.g., for different populations of cancer survivors or different contexts); adaptation can be an efficient approach to intervention development and is becoming increasingly common [20,21].
Ultimately, the goal is for self-management interventions to be adopted into routine cancer care, and there is a recent call to action for advances in this area [22].Rimmer et al. [23] highlight five key areas as essential prerequisites for translation from research into practice, namely improving adaptability, establishing acceptability and feasibility, ensuring description of characteristics and components, conducting process evaluations, and assessing cost-effectiveness.Interventions also need to be replicable and scalable.
Our systematic review sought to establish the characteristics and components of self-management interventions that aim to improve QoL in adult cancer survivors.To achieve this, we examined: (1) descriptions of intervention characteristics and components; (2) QoL outcomes; and (3) the association of characteristics and components with QoL improvements.Our secondary aims were to assess implementation issues, self-efficacy and economics, and the quality of available evidence.
was not confirmed, the paper was excluded.Screening of the search update was conducted by BR and IB.

Data Extraction
Data extraction was conducted and cross-checked by several members of the review team (BR, MB, LS, FB, TS, LD, MiB), using structured forms that were first piloted on two papers and revised as needed.

Study Characteristics
Study characteristics extracted included: author and year, intervention name; general characteristics (country, study design, total participants, eligible population); group characteristics for intervention and comparator arms (number analysed, age, sex, cancer site, ethnicity, stage of cancer, time since diagnosis/treatment); and whether a comparator was included, and details of comparator.
Where relevant additional papers (i.e., health economic evaluations) were identified and informed data extraction, study characteristics were extracted from the main evaluation study.Corresponding authors of included papers were contacted to request intervention protocols and relevant missing information.If a reply was not received within two weeks, data extraction decisions were informed by available published material.Published protocols and intervention development papers are acknowledged in Table S3.

Intervention Description
Intervention characteristics were assessed using the 12-item Template for Intervention Description and Replication (TIDieR) checklist [26] TIDieR aims to increase transparency in intervention description to improve replicability, encompassing: why, what materials and procedures, who provided, how, where, when and how much, and tailoring.
Intervention components were mapped to Lorig and Holman's self-management tasks of medical, role, and emotional management, [27] and the PRISMS taxonomy [17].PRISMS is a 14-component framework (e.g., information about available resources) that can be used to support self-management.
To understand acceptability and feasibility, we assessed implementation issues, including take-up rate, non-participation reasons, intervention adherence (e.g., number of sessions attended, withdrawal rates), withdrawal reasons, intervention modifications, and fidelity to the protocol.

Risk of Bias Quality Appraisal
Included RCTs were appraised using a 6-item modified version of the Critical Appraisal Skills Programme (CASP) RCT checklist [28].We considered the 6-item section A "Are the results of the study valid?"; section B "What are the findings?" is already reported in "Outcomes", while section C "Will the results help locally?" was not appropriate for our research question.Response options were "yes", "can't tell", or "no".Non-randomised studies were appraised using the 9-item Joanna Briggs Institute (JBI) critical appraisal checklist for quasi-experimental studies [29].Response options were "yes", "no", "unclear", or "not applicable".For both checklists, more "yes" responses indicated lower risk of bias (RoB) (CASP: ≥5 = low, 3-4 = medium, ≤2 = high; JBI: ≥7 = low, 4-6 = medium, ≤3 = high).

Outcomes
While additional outcomes may have been reported, we were primarily interested in QoL, self-efficacy, and economic factors (e.g., resource use and intervention cost).For these outcomes of interest, where relevant, we extracted data including outcome name; measurement timepoint(s) and instrument(s) used; baseline and follow-up scores for intervention and control groups; significant differences reported, with mean differences and confidence intervals.

Data Synthesis
Eligible papers were included in a narrative synthesis [30].This was structured around population and study characteristics, description of intervention characteristics and components, implementation issues, RoB quality appraisal, and outcomes.Associations with QoL were examined by study design, cancer site, TIDieR (selected characteristics), PRISMS, self-efficacy and economic factors.This assessed which intervention characteristics and components were (most) consistently associated with improvements in QoL, and whether self-efficacy improvements and economic benefits were concurrent with QoL improvements.

Search Results
Database searches identified 7770 hits.Following deduplication, 4053 titles and abstracts were screened from which 180 full texts were assessed for eligibility.Of these, 34 papers were included.Hand searches and expert consultation identified 19 additional papers, mainly providing additional details of the intervention or economic aspects.Altogether, 53 papers reporting 32 studies (32 interventions) were included  (Figure 1).
While additional outcomes may have been reported, we were primarily interested in QoL, self-efficacy, and economic factors (e.g., resource use and intervention cost).For these outcomes of interest, where relevant, we extracted data including outcome name; measurement timepoint(s) and instrument(s) used; baseline and follow-up scores for intervention and control groups; significant differences reported, with mean differences and confidence intervals.

Data Synthesis
Eligible papers were included in a narrative synthesis [30].This was structured around population and study characteristics, description of intervention characteristics and components, implementation issues, RoB quality appraisal, and outcomes.Associations with QoL were examined by study design, cancer site, TIDieR (selected characteristics), PRISMS, self-efficacy and economic factors.This assessed which intervention characteristics and components were (most) consistently associated with improvements in QoL, and whether self-efficacy improvements and economic benefits were concurrent with QoL improvements.

Search Results
Database searches identified 7770 hits.Following deduplication, 4053 titles and abstracts were screened from which 180 full texts were assessed for eligibility.Of these, 34 papers were included.Hand searches and expert consultation identified 19 additional papers, mainly providing additional details of the intervention or economic aspects.Altogether, 53 papers reporting 32 studies (32 interventions) were included  (Figure 1).
Additional outcomes considered in the eligible studies are listed in Table S11.
Six of the seven studies [39,45,52,54,58,66] that reported significant self-efficacy improvements also reported significant QoL improvements.When only these six studies were considered, improvements were associated most consistently (≥5 studies) with individual delivery, inclusion of tailoring, "Information about condition and its management", "Training for psychological strategies", and "Lifestyle advice and support".
Within the four studies [47,56,70,85] that assessed economics and found significant QoL improvements, only Van der Hout et al. [70] indicated possible economic benefits.Six of the seven studies [39,45,52,54,58,66] that reported significant self-efficacy improvements also reported significant QoL improvements.When only these six studies were considered, improvements were associated most consistently (≥5 studies) with individual delivery, inclusion of tailoring, "Information about condition and its management", "Training for psychological strategies", and "Lifestyle advice and support".
Within the four studies [47,56,70,85] that assessed economics and found significant QoL improvements, only Van der Hout et al. [70] indicated possible economic benefits.

Summary of Findings
This systematic review aimed to identify studies reporting self-management interventions in adult cancer survivors, primarily for description of intervention characteristics and components, and their association with QoL.We identified 53 papers reporting 32 studies (and 32 interventions) of varying, albeit largely average, quality.Included studies were most often RCTs (n = 20) or pre-post design (n = 10); targeted at mixed (n = 11), breast (n = 10), or prostate cancer survivors (n = 7); with usual care (n = 17) or waiting list (n = 6) comparators.Intervention characteristics (e.g., mode of delivery) varied considerably; on average, five (range 1-10) self-management components were included in the interventions, most commonly "Information about condition and its management" (n = 26).Twenty-two studies reported significant QoL improvements (six of which also reported significant improvements to self-efficacy).These improvements were associated most consistently with combined individual and group delivery and "Practical support with adherence".It is worth noting that some included studies were proof of concept or pilot studies so they may not have been powered to detect a significant difference in outcomes.Economic evaluations were limited and inconclusive.

Critical Appraisal of Evidence
We echo the observation from an earlier review that existing self-management interventions have largely been developed for breast and prostate cancer survivors [10].The interventions are typically either adjustment-(e.g., general self-management skills, such as problem solving or action planning) [8,12,19] or problem-focused (e.g., target

Summary of Findings
This systematic review aimed to identify studies reporting self-management interventions in adult cancer survivors, primarily for description of intervention characteristics and components, and their association with QoL.We identified 53 papers reporting 32 studies (and 32 interventions) of varying, albeit largely average, quality.Included studies were most often RCTs (n = 20) or pre-post design (n = 10); targeted at mixed (n = 11), breast (n = 10), or prostate cancer survivors (n = 7); with usual care (n = 17) or waiting list (n = 6) comparators.Intervention characteristics (e.g., mode of delivery) varied considerably; on average, five (range 1-10) self-management components were included in the interventions, most commonly "Information about condition and its management" (n = 26).Twenty-two studies reported significant QoL improvements (six of which also reported significant improvements to self-efficacy).These improvements were associated most consistently with combined individual and group delivery and "Practical support with adherence".It is worth noting that some included studies were proof of concept or pilot studies so they may not have been powered to detect a significant difference in outcomes.Economic evaluations were limited and inconclusive.

Critical Appraisal of Evidence
We echo the observation from an earlier review that existing self-management interventions have largely been developed for breast and prostate cancer survivors [10].The interventions are typically either adjustment-(e.g., general self-management skills, such as problem solving or action planning) [8,12,19] or problem-focused (e.g., target specific issues, such as managing fatigue) [86].However, cancer is a complex chronic illness, presenting different problems for specific diagnoses (e.g., seizures for brain tumours) [87]; thus, a "one size fits all" approach is likely inappropriate [11].Furthermore, depending on how rapidly a cancer is expected to progress, the optimal timing of intervention delivery might vary for different cancers, though whether self-management interventions are more effectively delivered at certain time points is unclear and requires consideration, with the data in our review too heterogeneous to comment.
Intervention development need not start de novo; existing, effective self-management interventions may be adaptable to another context/setting [20,21], and might include both "core" adjustment-focused elements that are applicable across cancers, before applying problem-focused elements, targeted to individual cancers.Still, researchers must be confident in the appropriate selection of "core" elements to adapt to their intervention.This is hindered by, as shown here, the substantial heterogeneity of intervention characteristics and poor reporting of components, and exacerbated by the lack of intervention protocols available.Existing interventions typically lack clarity in their description, impairing the potential for replicability when considering adaptation or, indeed, large-scale implementation.Further, poor reporting of fidelity assessments, and reasons for non-participation and withdrawal, means it remains unclear whether interventions are ineffective due to poor fidelity, or disadvantageous characteristics and components.Ultimately, this makes it difficult to recommend certain characteristics and components for use in future intervention development efforts.
Nonetheless, it is notable that blended individual and group delivery was consistently associated with improved QoL.While an individual element may be important for privacy around sensitive issues, the review of Coffey et al. [19] indicates that cancer survivors favour the inclusion of a group element in self-management interventions.This could be due to the opportunity to gain support from similar others and the ability to interact and share experiences in a safe space.
Consistent with Cheng et al. [88], this review supports the potential effectiveness of self-management interventions for improving QoL among survivors.However, significant improvements were typically observed for limited, varying symptoms and functioning aspects of QoL.Further, there was a dearth of significant improvements in comparison to control groups, perhaps influenced by the heterogeneity in what was determined as "usual care".Moreover, while we classified comparator groups as "usual care plus" if a passive form of self-management (e.g., leaflet) was provided, this was not always clear; it is, therefore, possible that "usual care" may have contained elements of self-management skill development that diminished the intervention effect.
Online self-management interventions are increasingly popular, in part due to their reach [89,90].However, this should be approached with caution, as online interventions were not consistently associated with improved QoL here.Instead, we conclude that a blended approach (e.g., including face-to-face/group delivery and some form of digital delivery), where possible, may be valuable.Ultimately, patient, public and stakeholder involvement during intervention development is required to consider the design preferences of the target population [91].
There is growing evidence to support the effectiveness of tailoring interventions to the individual [92][93][94].However, while tailoring showed promise, it was not consistently associated with improved QoL.Still, there are several tailorable variables, which can independently moderate effectiveness [95]; the heterogeneity that was tailored by included studies (e.g., personalised goals, number of sessions) might help explain the observed inconsistency.Consequently, it may be beneficial if future interventions incorporate and compare the effectiveness of different elements of tailoring.
Perhaps surprisingly, studies that delivered more/most PRISMS components were not consistently associated with improved QoL.Nonetheless, those that assessed the least components consistently reported neutral or negative impacts on QoL, indicating the value of considering multiple PRISMS components in an intervention.While we did not examine behavioural change techniques (BCTs) included in the interventions, we might speculate that those interventions that delivered more PRISMS components were likely to have included more/multiple BCTs; it has previously been shown that interventions that include a combination of BCTs are more often effective [96].It was promising to observe consistent concurrence of significant self-efficacy and QoL improvements across studies that reported both.This is congruent with the theoretical notion that self-management interventions improve clinical and psychosocial outcomes by empowering self-efficacy [16]; this emphasises the value of considering such skill development.
The interventions included in this review targeted a variety of different areas (e.g., symptoms, psychological well-being, lifestyle behaviours).We chose not to focus on interventions with a specific target as we wanted to provide a comprehensive overview and synthesis of the available evidence.It does, however, raise the question of whether interventions with a common target (e.g., symptom management) were more likely to impact positively on QoL.We considered this post hoc and were unable to reach any clear conclusions.
The paucity of robust evidence on self-management interventions impacts the ability of policymakers and stakeholders to make effective decisions [97].Specifically, evidence on the impact of QoL and resource utilisation informing cost-effectiveness models and budgetary impact is critical, yet health economic evaluations, particularly cost-utility analyses, were rare.Where available, studies largely evaluated health service resource use; however, it is not enough to suggest self-management can reduce healthcare utilisation [15] if this is outweighed by intervention delivery costs, for example.Since implementing healthcare changes may require training, time and material resources, [18] economic factors-and particularly cost-effectiveness analyses-require further consideration.

Implications
We provide a comprehensive overview of the available evidence, informing four of the key influences on intervention implementation proposed by Rimmer et al. [23].Mapping the characteristics and components of TIDieR and PRISMS, respectively, indicates which elements may be adaptable across cancers and offers a systematic description of interventions and their content.Examining associations with QoL provides a starting point for understanding which characteristics and components may be most beneficial.The findings suggest incorporating a combination of individual and group delivery and ensuring the availability of practical support with intervention adherence may be worthwhile.Overall, however, the effectiveness of specific characteristics and components is inconclusive, largely due to the heterogeneity of interventions, measurements and procedures and, probably also, what interventions were trying to influence/change.We also identify directions for future research to complement the recent call to action for advancement in evidence on the effectiveness of self-management in cancer survivors [22].
To improve the replicability and scalability of self-management interventions in cancer, characteristics should be reported more consistently, in accordance with the TIDieR checklist [26].Still, we would suggest that consideration is given to whether TIDieR, as it stands, is appropriate for capturing "dose" for increasingly popular, online interventions.We would further recommend that authors report fidelity and reasons for drop-out more transparently.For enhanced clarity, and to encourage a common language, on what support is being delivered, future research should also explicitly map their intervention components to the PRISMS taxonomy [17].

Strengths and Limitations
Our review is the first to map intervention components to PRISMS, examine associations of characteristics and components with QoL, and review economic implications.We provide a novel and comprehensive extension to existing evidence synthesis [13,14], offering greater depth in understanding intervention effectiveness, implementation potential and future directions.
Despite attempts to define self-management in cancer [7], self-management interventions were difficult to identify.The conceptualisation of self-management in existing evidence synthesis has varied: for example, Cuthbert et al. [14] required an education component, whereas we required a more explicit description of "self-management".This perhaps explains the limited overlap of included studies (n = 8) in our respective reviews, emphasising the need for a consensus definition and more clarity in the reporting of future interventions.
Although our review benefitted from thorough searches-including several databases, forward and backward citation searches of included studies and relevant reviews, and expert consultation-we did not search the grey literature or include studies not available in English.Hence, there is a small possibility that a relevant study was missed.A metaanalysis and meta-regression were not feasible due to the aforementioned heterogeneity and, in particular, because 20 different measures of QoL were used.Despite 10 studies using the EORTC-QLQC30, the heterogeneity within this subset was still substantial for study design, population, and time points measured.Similar comments apply in relation to self-efficacy.Therefore, associations with QoL and self-efficacy, and consistencies in these associations within studies that assessed both, were only examined by "vote counting".

Conclusions
Self-management interventions show promise for improving QoL in cancer survivors.However, study quality was variable, with substantial heterogeneity in the characteristics and components used, and insufficient evidence on cost-effectiveness.Nevertheless, our review is comprehensive and, while caution is required, highlights what might be worth adapting from existing interventions (e.g., combining individual and group delivery, practical support with adherence).These findings provide the foundations to inform further development and facilitate steps towards the implementation of self-management interventions for cancer survivors.

Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cancers16010014/s1,Table S1: Search concepts; Table S2: Database searches; Table S3: Protocols and papers to support intervention development; Table S4: Additional population characteristics of cancer survivors; Table S5: Theory and rationale for the intervention; Table S6: Lorig and Holman self-management tasks; Table S7: Implementation issues; Table S8: Critical appraisal skills programme (CASP) risk of bias appraisal; Table S9: Joanna Briggs Institute (JBI) risk of bias appraisal; Table S10: Quality of life instruments and their scoring; Table S11: Self-efficacy and additional outcomes; Table S12: Economics; Supplementary Data S1: Quality of life and self-efficacy outcome data, and intervention component judgements.

Figure 1 .
Figure 1.PRISMA flow diagram of paper selection.

Figure 2 .
Figure 2. Harvest plot of association between TIDieR characteristics and QoL.Figure 2. Harvest plot of association between TIDieR characteristics and QoL.

Figure 2 .
Figure 2. Harvest plot of association between TIDieR characteristics and QoL.Figure 2. Harvest plot of association between TIDieR characteristics and QoL.

Figure 3 .
Figure 3. Harvest plot of association between PRISMS components and QoL.

Figure 3 .
Figure 3. Harvest plot of association between PRISMS components and QoL.

Table 1 .
Study and population characteristics of self-management interventions in cancer survivors.
Intervention included eight weekly 2 h sessions.Participants to complete a variety of assigned tasks (e.g., revising self-narrative, adaption of social roles, adjusting self-expectation) at home between sessions; asked to keep work-related diary to report how they managed it in their daily life.Patients encouraged to observe their thoughts and acts, and the different tasks proposed during and between sessions help them to detect and react to difficult situations.Patients are asked to be actively involved in the process since the aim is to introduce change in their daily routines.Participants introduced to hypnosis in first session and at the end of each session, a 15 min supervised hypnosis exercise is conducted.Participants receive a CD for each exercise to encourage at home practice.

Table 4 .
Quality of life outcomes.