Decision aids for cancer survivors’ engagement with survivorship care services after primary treatment: a systematic review

Purpose To elucidate existing decision aids (DAs) in supporting cancer survivors’ decisions to engage in cancer survivorship care services after primary treatment. Secondary objectives are to assess the DA acceptability, impact of DAs, and implementation barriers. Methods Databases (PubMed, Embase, PsycINFO, CINAHL) were searched to collect publications from inception through September 2021. Studies describing the development or evaluation of DAs used for survivorship care services after primary cancer treatment were included. Article selection and critical appraisal were conducted independently by two authors. Results We included 16 studies that described 13 DAs and addressed multiple survivorship care domains: prevention of recurrence/new cancers in Hodgkin lymphoma survivors and breast cancer gene mutation carriers, family building options, health insurance plans, health promotion (substance use behavior, cardiovascular disease risk reduction), advanced care planning, and post-treatment follow-up intensity. The electronic format was used to design most DAs for self-administration. The content presentation covered decisional context, options, and value clarification exercises. DAs were acceptable and associated with higher knowledge but presented inconclusive decisional outcomes. Implementation barriers included lack of design features for connectivity to care, low self-efficacy, and low perceived DA usefulness among healthcare professionals. Other survivor characteristics included age, literacy, preferred timing, and setting. Conclusions A diverse range of DAs exists in survivorship care services engagement with favorable knowledge outcomes. Future work should clarify the impact of DAs on decisional outcomes. Implications for Cancer Survivors DA characterization and suggestions for prospective developers could enhance support for cancer survivors encountering complex decisions throughout the survivorship continuum. Supplementary Information The online version contains supplementary material available at 10.1007/s11764-022-01230-y.


Introduction
Cancer survivors continually encounter a myriad of health decisions in diagnosis, survivorship, and the end of life [1].According to the National Cancer Institute, the primary treatment is the first cancer treatment that is typically a combination of surgery, chemotherapy, and radiation [2].As survivors transit into the post-primary treatment phase, the focus shifts from providing treatment to maintaining health and maximizing the quality of life.Thus, survivors face decisions in engaging a diverse range of survivorship care services, addressing surveillance, physical symptoms, psychosocial issues, and preventive health [3][4][5][6][7].The participation of survivors in such decisions is increasingly being advocated as the perceived involvement in decision-making about follow-up care is associated with better quality of life up to 10 years post-diagnosis [8].Contrary to treatmentrelated decisions, decision-making over service engagement involves a broader range of health disciplines and care settings, survivors' self-efficacy to implement the chosen option, and self-management to sustain behavioral changes [9,10].Since informational needs vary by survivorship phase, suboptimal available information, associated health risks, and long-term adverse treatment pose challenges to decision-making [1,[11][12][13].Therefore, interventions supporting decision-making should be tailored to the survivorship phases.
Clinicians could assist survivors in making health decisions post-treatment through shared decision-making (SDM), a process by which physicians inform survivors regarding the potential healthcare interventions [14][15][16].Decision aids (DAs) are supporting tools that provide evidence-based available care options and their outcomes while incorporating value clarification components to guide users about the care options based on their preferences [17].In the post-treatment phase, the potential benefit of DAs for decision-making is two-fold.First, systematic reviews have shown that DAs for cancer-related decisions increase knowledge, reduce decisional conflicts, and enhance satisfaction [18][19][20].DAs could address the informational needs unique to the survivorship phase post-primary treatment.In addition, DAs could better align decisions and personalize survivorship care to maximize value to each survivor [21].DAs are empowering tools to improve survivors' self-management [22,23] during survivorship with typically reduced clinical touchpoints from the active treatment phase.
Despite the promising utility of DAs in oncology, existing systematic reviews have disproportionately focused on decisions during the active treatment phase [18-20, 24, 25].The available DAs and evidence that support their application in the post-treatment phase are still unclear.To address this knowledge gap, this systematic review elucidates the existing DAs developed to support cancer survivors' decisions to engage in cancer survivorship care services after primary treatment until the end of life.Secondary objectives are to assess the acceptability and impact of the DAs and to outline the implementation challenges.The results would reveal the survivorship care areas where DA development efforts could synthesize available evidence to support the DA usage and implementation strategies in promoting routine care integration.

Methods
This systematic review was performed following a protocol drafted by the research team (unpublished) and was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses checklist (Supplementary File 1) [26].

Eligibility criteria
Inclusion criteria were as follows: (1) targeted patients of any age diagnosed with cancer and/or family caregivers involved in the decision of patients' care; (2) described the development or evaluation of a DA for engaging with survivorship care services for the post-primary treatment phase; and (3) set in any country or setting.No age limits were set to achieve comprehensive search results.Family caregivers that play active roles in decision-making included parents of pediatric, adolescents, and young adults (AYAs), partners, or adult children [27,28].Survivorship care services were defined as health services that addressed any domain of the Quality of Cancer Survivorship Care Framework [29].Survivorship care services are characterized by survivors' treatment (prevention and surveillance of new cancers and management of physical and psychosocial effects) and general healthcare (surveillance and management of chronic medical conditions, health promotion, and disease prevention).Survivorship care services also feature the contextual domains of healthcare delivery (clinical structure, communication/decisionmaking, care coordination, patient/caregiver experience).The systematic categorization of services would reveal gaps or opportunities in clinical care, using consistent terminology in cancer survivorship research.Study protocols were included; however, commentaries, author replies, editorials, and expert opinions were excluded.Studies describing DAs for clinical trial participation were excluded as these decisions did not involve the delivery of established routine care services.Non-English studies were excluded from this review.

Search strategy
Four electronic databases (PubMed, Embase, PsycINFO, and CINAHL) were searched systematically, and studies that fulfilled the eligibility criteria from inception to 29 September 2021 were selected.Additionally, we searched Google Scholar and reviewed reference lists of all included studies.The following keywords were used for the search: cancer survivor, decision aid, and post-treatment phase.The complete search strategy can be found in Supplementary File 2. Two authors (Y.K. and H.Z.) independently assessed the eligibility of each study by reviewing the titles, abstracts, and entire texts.All discrepancies were discussed and resolved by the two authors.

Data extraction and result synthesis
One author (H.Z.) extracted data from all included studies using a standardized data collection form, and a second author (Y.K.) performed data accuracy checks.The following data were collected: bibliometrics, study characteristics (objective, design, and setting), study participant characteristics, DA characteristics (title, target user, decision of interest, administration format, development framework/methodology, summary of content, and value clarification exercise), key findings of evaluated measures, implementation challenges (anticipated and encountered), and study limitations.Value clarification methods were described using Witteman et al.'s taxonomy [30].Implementation challenges were mapped to construct the Consolidated Framework for Implementation Research [31].Evaluation measures were categorized into acceptability, knowledge, and decision-related outcomes according to the Ottawa Patient Decision Aids Research Group [32].
Unique DAs were identified and characterized to fulfill the primary objective.To fulfill the secondary objectives, evaluation outcomes were summarized using descriptive statistics, and measures of association and their associated uncertainty were reported.The qualitative themes and implementation challenges are summarized.

Study quality assessment
Quality assessments for the various study designs were performed using the following appraisal tools: the Revised Cochrane risk-of-bias tool for randomized controlled trials (RCTs) [33], the Risk Of Bias in Non-randomized Studies of Interventions for non-randomized studies [34], the Joanna Briggs Institute checklist for cross-sectional studies [35], the Critical Appraisal Skills Programme (CASP) qualitative checklist for qualitative studies [36], and Mixed Methods Appraisal Tool for mixed-method studies [37].Study assessments were performed independently by two authors (Y.K. and H.Z.), and discrepancies were resolved through discussions until a consensus was reached.

Study quality
All five RCTs were characterized by a high risk of bias [48,[51][52][53] except one RCT [43].The two non-randomized trials assessed were at low-risk bias [46] and high-risk bias [49], respectively.Loss-to-follow-up and measurement of subjective outcomes by unblinded investigators led to a high risk of bias.The methodological quality of one cross-sectional study was unclear due to poor reporting of eligibility criteria, study period, and statistical plan [42].One cross-sectional study was of good quality, but the unclear validity and reliability of its outcome measure might have led to a misclassification bias [47].All three qualitative studies addressed 9/10 items on the CASP checklist but did not include reflexivity statements [41,44,45].Of mixedmethods studies, one had high methodological quality [50]; the other received a poor rating because they failed to justify the value of mixed-methods and presented poor integration of the quantitative and qualitative components [40].Detailed assessments are summarized in Supplementary File 3.

Range of DAs on survivorship services-related decisions
Table 1 summarizes the key characteristics of the 13 unique DAs described in the included studies.A total of eight decisions on survivorship care service engagement were described and mapped to the Quality of Cancer Survivorship Care Framework domains (Table 2).For cancer surveillance, two DAs addressed the prevention of new and/or recurrent breast and ovarian cancer among survivors with breast cancer gene (BRCA) mutations through risk-reduction strategies [41,44].One DA focused on lung cancer screening using computerized tomography for Hodgkin lymphoma survivors [40].To manage late and long-term treatment side effects, a DA intervention on family-building options for women of reproductive age addressed the physical and psychosocial impact on fertility [38].Another generic DA on insurance plan selection addressed financial toxicity [49].For health promotion and disease prevention, DAs addressed substance use behavior among AYAs [43] and cardiovascular disease risk reduction among high-risk survivors defined by age and primary treatment history [50].Lastly, services relating to healthcare delivery include advanced care planning (ACP) discussion, documentation [42,48,[51][52][53], and the followup intensity of post-treatment consultations [46].
Most DAs were specific to the post-primary treatment survivorship phase, except ACP and insurance plan decisions after diagnosis [42,48,49,[51][52][53].The users of the two DAs of ACP were broad, including the cancer population as a subgroup [42,48].The remaining DAs had different specificity based on (1) cancer type such as breast, ovarian, Hodgkin lymphoma, (2) cancer mutation status, (3) cancer staging, (4) primary treatment received, and (5) the AYA age group.

Design features of DAs
Approximately half (7/13) of the DAs specified the guidelines for DA development, with the International Patient Decision Aid Standards and Ottawa Decision Support Framework most cited [38,40,41,44,46,53].The remaining DAs were developed through a methodology involving a literature review of available evidence and qualitative discussions with target users.The majority (8/13) of the DAs employed digital websites or computer programs for the DAs format [39, 41-43, 46, 49, 51, 52].Two DAs employed the paper format [40,44], and two DAs adopted a hybrid format [48,53].Most DAs were for self-administration at users' convenience, and three DAs were administered during consultations with healthcare professionals [46,50,51].

Contents features of DAs
All DAs generally followed a similar structure for content presentation.The context of the decision was first described, explaining the heightened risk of recurrent or second cancer, late or long-term physical and psychosocial effects, and the importance of goal setting near end-of-life.Next, a list of possible options was presented using methods such as storyboards or survivors' anecdotes [38,49], actual imaging results for the case of low-dose CT scan [40], pictorial representation of risks [41,50], and comparison tables [44].Additionally, the DA on insurance plan selection provided personalized cost estimates for each option [49].Nine DAs used value clarification exercises to help survivors understand their preferences before guiding them about the application of the DA results for consultation with care providers [38,40,41,43,44,46].The methods employed in the exercises included rating the relative importance of different attributes in decision-making [42,46,52], listing pros and cons of available options with (n = 2) [41,52]  weighing (n = 2) [40,43], listing concerns (n = 1) [38], and viewing a pre-identified list of values, options' pros and cons (n = 1) [44].The method was unclear in one DA [48].
The DA addressing follow-up intensity further employed an option grid to visually compare the congruency between the user's preference and intuitive choice [46].The DA on substance use among AYA had additional unique content features [43].Before introducing the decision context, users were to complete a general health decision-making module.Furthermore, the options were not presented as organized information.Instead, users were guided to explore the consequences of each chosen option via dynamic pathways in an interactive program.
In one study which compared DA usage to an active comparator using an informational website, no significant differences were found in the satisfaction of the amount (P = 0.054) and quality (P = 0.119) of information between the groups [52].

Impact of DAs (Table 3)
Knowledge Five studies assessed survivors' knowledge using investigator-designed questionnaires [40,[48][49][50]53].DAs were associated with improved knowledge when assessed immediately (median correct responses for lung cancer screening: pre-DA = 68% vs. post-DA = 93%, P < 0.001) [40] and several weeks after [mean (SD) knowledge score of ACP 7 weeks after baseline: intervention = 5.12 (0.97) vs. control = 4.66 (1.11), P = 0.005] [53].Contrarily, a pilot RCT did not report significant improvement in ACP knowledge after usage, possibly limited by the statistical power [48].A study reported that the improvement in knowledge of health insurance plans immediately after DA usage (P < 0.0001) was sustained when reassessed 3-6 months later (P = 0.002) [49].The last study evaluated the knowledge of the intervention and control groups and revealed gaps in knowledge in cardiovascular disease risk and statins side effects [50].
Choice predisposition Five studies evaluated the impact of DAs on survivors' propensity to make a choice [40,43,46,52,53].Compared with controls, two studies reported no significant differences in the intention to complete advanced healthcare directives [52,53].The DAs did not significantly change the intention of survivors to participate in future lung screening programs [40].Moreover, transient improvement in attitude toward substance use after the DA did not translate to a change in substance use behavior [43].Only the study on DA for breast cancer aftercare follow-up intensity reported a higher proportion of survivors in the intervention group choosing a less intensive follow-up modality (intervention = 51% vs. control = 29%, P = 0.04) [46].
Other outcomes DA usage increased consultation time (minutes) [mean (SD): intervention = 42.3 vs. control = 29.8,P < 0.01] but was not associated with hospital cost difference over a 3-month follow-up period [46].Psychological constructs such as hopefulness, hopelessness, anxiety, and depression did not differ from the controls following DA for ACP [42,53].Additionally, physical and mental quality of life scores did not differ from controls when a similar DA for ACP was evaluated in another study [51].The DA on insurance plan selection neither improved financial toxicity nor reduced the proportion of participants who delayed or avoided care because of the insurance cost [49].

Challenges of DA implementation
Regarding DA characteristics, the strengths of an electronic format for updating and dissemination [41] have been highlighted with concurrent concerns over the computer access and cancer survivors' preferences for the paper format [46].Additional challenges included limited users' engagement in discussion forums and real-time connectivity to care providers, compounded by inadequate navigational support [52].One study which explored healthcare professionals' perspectives revealed low self-efficacy and a lack of perceived usefulness of the DA as implementation barriers [47].
Recurring determinants of successful DA implementation were characteristic of an individual cancer survivor.No consensus was reached regarding the appropriate timing to introduce DAs across studies, which depended on factors such as survivors' emotions and time after diagnosis [41,46,48,49,52].Other clinical and demographic factors, such as old age, poor health literacy, cognitive impairment, and low competence to forecast future preferences, were anticipated implementation challenges [42,43].Regarding the implementation of DAs for substance use among AYAs, the impact of the social environment-the behavior of AYAs' family and friends-and the DA's compatibility with school curricula were the anticipated implementation barriers [43].A restricted range of health insurance plans applicable to cancer survivors affected the DA for insurance plan selection [49].Two studies reported survivors' preference for DA usage in the clinical care settings [41,46].

Discussion
To our best knowledge, this systematic review is the first to present the characteristics of the DAs that assist cancer survivors in deciding on survivorship care service engagement after completing primary treatment.The available DA post-treatment is underexamined contrary to treatment decisions but highlights ongoing efforts to identify complex decisions in the entire survivorship continuum.While the number of identified DAs was limited for each domain of the quality survivorship care framework, the range was comprehensive and addressed a spectrum of survivorship care needs.Included DAs were acceptable with favorable outcomes in improving knowledge despite inconclusive decisional outcomes.
The available DAs for the post-treatment phase suggested different decisional needs from the active treatment phase, shifting from disease management to wellness promotion [54].This review highlighted healthcare delivery as a unique decisional aspect during survivorship.Klaassen et al. [46] developed a DA for specialist surveillance follow-up intensity, which reverses the trend of overusing specialist services during survivorship despite marginal benefit on mortality [55][56][57].This review also identified two types of decisions considered during active treatment or previvor stage-fertility management and genetic testing [25,58].The DAs targeting the post-treatment phase underscored the challenges associated with extending the use of DAs developed for earlier phases.First, DA transferability was limited by the side effects of cancer treatment.For instance, •No significant difference in perceived shared decision-making was found between groups (P = 0.31) •No significant differences in decision satisfaction and uncertainty were found between groups, immediately (P = 0.27) and at 3 months (P = 0.40) after aftercare consultations •No significant differences in perceived informed choice were found between groups, immediately (P = 0.83) and at   N: intervention (24), control (22) Acceptability (intervention group) •92% found the information amount to be highly acceptable •58% found the information clarity to be highly acceptable Knowledge (full cohort) •Half of the cohort answered 10-year myocardial infarction risk correctly   family-building goals have shifted from a fertility preservation approach to assisted reproductive technology.Yet, unaddressed fertility information needs and decision distress were reported in post-treatment women of reproductive age, which compromised survivors' quality of life and reinforced the value of decision support in this survivor population [59,60].Apart from the disparate options that affect transferability, the information needs may differ in quality.Specific to BRCA genetic testing, high-risk breast cancer survivors were keen on information related to the recurrence of the affected breast.Moreover, previvors were more interested in the impact of risk-reducing surgery [44].The DA design for the post-treatment phase requires careful consideration of the residual treatment effects, available care options, and unique information for this phase.The DAs' specificity in the post-treatment survivorship phase should be counterbalanced with efficiency considerations.Among the included studies, DAs for ACP and insurance plan selection were applicable from the time of cancer diagnosis [42,48,49,[51][52][53].The sustained ACP applicability throughout life and the integration of a personalized cost calculator are relevant despite non-specificity.Moreover, DAs addressing health promotion and disease prevention present opportunities for DA adaptations.One study modeled its DA after the validated Statin Choice DA [61], illustrating an attempt to adapt an evidence-based generic tool for cancer survivors by supplementing cardiovascular disease risk estimates specific to cancer treatment history.This approach can apply to other DAs for health promotion, such as smoking cessation, physical activity, and weight loss programs, evaluated in non-cancer populations [62][63][64].Based on a wide range of survivorship care domains, developers could tap available DAs to improve efficiency and apply DAs from acute to extended survivorship and from the general population to the cancer population.
Our results accentuate that the AYA population is unique for decision support.Although only one DA for substance use targeted this age group, its unique content reinforced AYAs' distinct characteristics [43].AYA cancer survivors encounter survivorship care decisions while developing full autonomy, and parents may play a role in decision-making [65].The additional health decision-making module in the DA for substance use aimed to improve AYAs' self-efficacy, which strongly correlated with higher degrees of self-regulatory skills and perceived autonomy [66].Furthermore, the innovative presentation via interactive pathways was compatible with AYAs' preferred format of contextualizing information to their way of life, instead of a factual presentation format employed predominantly by DAs targeting the adult population [67].DA development for this age group warrants consideration of additional strategies to promote self-efficacy and ensure compatibility of information with AYAs' preference to improve communication.
Similar to the Cochrane systematic review on cancer screening and treatment decisions [20], studies that evaluated DAs for ACP also demonstrated improved knowledge of available care options [40,49,53].However, non-significant effects on decisional outcomes observed in most studies have possible explanatory factors.First, the quality of most RCTs was compromised with a high risk of bias due to missing outcome data.Second, multiple pilot RCTs examining the feasibility of DA usage lacked statistical power for conclusive inference.Third, the impact of DAs may be short-lived relative to the behavioral change required when engaging with survivorship care services.Despite improved knowledge, DAs did not improve intentions to participate in screening or did not modify risky lifestyle behaviors [40,43].This observation could be attributed to survivors' failure to cognitively associate such health behaviors with potential threats and disproportionate focus on the cancer condition since diagnosis [68].One DA promoted a higher adoption of a less intensive follow-up modality among breast cancer survivors than controls, highlighting the usefulness of DAs for care model selection after primary treatment [46].Apart from the range of outcomes evaluated, levels of self-efficacy and motivation are constructs of interest as behavioral change underlies survivorship care service engagement.
In this review, the anticipated DA implementation challenges were consistent with the system-level factors reported in the SDM literature [69,70].With a wide range of applicable DAs in the survivorship phase, a systematic approach to consolidating DA dissemination through a library catalog could facilitate comprehensive access.Since the content structure of most included DAs was similar, a recurring template could be employed.Additional interactive platforms should be considered to improve connectivity between the user and the survivor community or care providers.Additionally, DA usage should be personalized based on the survivor characteristics highlighted to affect DA uptake.Specifically, healthcare professionals should tailor DA to each survivor's preferred format, readiness for change, literacy, age, and setting.Notably, for AYAs, engagement with family members or friends in the survivor's relationship network is critical for promoting a conducive social environment to achieve behavioral change and decision-making [71,72].
This review has several limitations.First, non-RCTs not ranked the highest on the hierarchy of evidence were included to capture an exhaustive range of DAs [73].Second, the mixed quality of the included studies may affect the reliability of outcomes.Third, the generalizability of the results beyond the American and European populations to other regions with different cultures may be challenging, given that these countries have relatively better SDM awareness [74].Notably, the concept of SDM underlying the utility of DAs may be foreign to survivors and clinicians from settings with a stronger hierarchical culture [75,76].

Conclusions
This systematic review identified a diverse range of DAs to support cancer survivors' decisions over survivorship care service engagement after primary treatment.Despite inconclusive results on decisional outcomes, care settings can implement available DAs for knowledge improvement.Prospective DA developers should examine the need for cancer survivor specificity and explore if existing DAs available for the treatment phase or non-cancer populations could be readily adopted, especially for health promotion.Particularly, this review highlighted AYAs as a special population for tailored DA design and implementation considerations.Future work should clarify the impact of DAs on decisional outcomes or behavior-related constructs.

Identification of studies via databases Identification of studies via other methods Identification Screening Included
and without

Table 2
Survivorship care domains addressed by included decision aids

Table 3
Summary of studies reporting the development or evaluation of decision aids