Translating Evidence from Dutch Exercise Oncology Trials in Patients with Breast Cancer into Clinical Practice Using the RE-AIM Framework

Purpose . We aimed to evaluate the potential for implementing exercise interventions for patients with breast cancer in the Netherlands, based on fndings of the Dutch randomized controlled trials in this population. Methods . We evaluated the implementation of four Dutch exercise trials retrospectively, using the fve dimensions of the RE-AIM framework: Reach (exercise participation rate), Efectiveness for physical ftness, fatigue, quality of life, and physical function, Adoption (e.g., satisfaction of physical therapists guiding the exercise intervention), Implementation (cost-efectiveness and exercise adherence correlates thereof), and Maintenance (maintenance of exercise levels by individual patients and sustainability of exercise delivery at organization level). Tereby, we refect on these results using (international) literature to gain better insight in overall barriers, facilitators, and opportunities for further implementation of exercise interventions. Results . Participation rates of 44–52% not only indicated acceptable Reach in the context of a trial but also indicated room for improvement. Efectiveness of exercise during and after treatment was demonstrated in most trials showing benefts for aerobic ftness, physical fatigue, quality of life and physical function, and high patient satisfaction. Adoption of the exercise interventions by physical therapists was adequate (satisfaction score: 7.5 out of 10). Evaluation of Implementation indicated adequate adherence to supervised exercise, inconsistent fndings on potential correlates of adherence, and promising results on cost-efectiveness. Currently, reimbursement for exercise programs is lacking. Maintenance of intervention efects at the patient level was limited and inconsistent. Maintenance of intervention availability at the organizational level was facilitated by an extensive network of specially trained physical therapists, but better communication and collaboration between diferent healthcare professionals are desired. Conclusions . Improved implementation could particularly be achieved by increasing reach and improved focus on exercise maintenance on both the patient and organizational level.


Introduction
Evidence from randomized controlled trials (RCTs) indicates that exercise benefts aerobic ftness [1], fatigue [2], health-related quality of life (HRQoL) [3], and physical function [3,4], during and after cancer treatment.Tis has led to the development of national and international guidelines [5][6][7][8][9] recommending exercise as an integral part of cancer care in a number of countries and professions, including sports medicine [4], medical oncology [10], and physical therapy [11,12].However, widespread implementation of exercise interventions is still limited.Translating research from RCTs into practice has shown to be difcult because of problems with population representativeness, limited (fnancial) resources, and program availability and sustainability [13].
In 1999, the RE-AIM framework was developed to evaluate the potential for dissemination of research into clinical practice and to facilitate this process [14].Since then, RE-AIM has been used to plan, evaluate, and review health promotion and disease management interventions [14,15].In the RE-AIM framework, the overall impact of an intervention is described in fve dimensions: reach, efectiveness, adoption, implementation, and maintenance (Table 1).
In this study, we aimed to evaluate the potential for implementation of exercise interventions for people who have been treated for breast cancer with curative intent, based on these four RCTs, with the use of the RE-AIM framework.We summarize the fndings from the four Dutch trials only, since implementation of interventions is a dynamic, context-specifc process [33], with (country) specifc and more generalizable components.In addition, we refect on these results using (international) literature (e.g., trials and reviews) to gain better insight in overall barriers and facilitators for the implementation of exercise interventions.Finally, we describe opportunities for further optimization of implementation.

Methods
For each dimension of the RE-AIM model, we summarized the fndings from the four Dutch trials, as published before August 2022, using the operationalizations of reach, efectiveness, adoption, implementation, and maintenance described in Table 1.Most results have been published previously [17,19,20,22,23,25,26,32], except for information on patient satisfaction (PACT and UMBRELLA Fit trial).As the REACT trial also included patients with other types of cancer, we performed subgroup analyses on the subpopulation of patients with breast cancer, except for the cost-efectiveness analyses of which results are presented from the original papers.For better comparisons, we calculated Cohen's d efect sizes (ES) of the intervention efects, where signifcant efects (p ≤ 0.05) were considered as no substantial diference for ES ≤ 0.2, small for ES 0.2-0.5, moderate for ES 0.5-0.8, and large for ES ≥ 0.8 [34].

Reach
3.1.1.Summary of Results of Dutch RCTs.Across the four RCTs, 44-52% of eligible patients were willing to participate in the trials and the exercise intervention.Main reasons for nonparticipation were lack of time, mental burden, travel distance to the hospital, not wanting to be randomized, or wanting to exercise on their own (Table 3).
Comparison of participants and nonparticipants indicated that patients with a higher educational level were more likely to participate in exercise trials both during and after treatment (Table 3).Additionally, behavioral motivational factors were associated with participation during chemotherapy.Patients with more expected benefts of exercise, higher self-efcacy, fewer negative attitudes, more social support, and fewer perceived barriers to exercise were more likely to participate.Conversely, for exercise interventions following completion of anticancer treatment, patients who perceived more barriers were more likely to participate (Table 3).

Refections and Opportunities to Improve
Implementation.Te reported participation rate of 44-52% in the Dutch exercise trials is somewhat higher than the pooled estimate of 30% reported in a metaanalysis of 23 exercise trials in patients with breast cancer [35].Te highest participation rate of 52%, reported by the UMBRELLA Fit trial, is likely related to the trials within cohorts (TwiCs) design, in which patients participating in an observational cohort were randomly invited to participate in an exercise intervention, thereby limiting intervention nonparticipation due to unwillingness to be randomized [31,36].In other trials, this proportion was shown to be approximately 10-15% [17,26] (Table 3).Additionally, participation rates were infuenced to some extent by the eligibility criteria employed, as they most often excluded patients with severe comorbidities and those with cognitive disorders or not fuent in Dutch.Tese patients may beneft even more from exercise guidance, as they may need specifc exercise prescriptions, and may be less aware of health benefts and less able to fnd adequate health information, respectively.Patients with insufcient mastery of the language may beneft from additional health communication strategies, such as including visual aids to improve reach [37,38].
Te fnding that travel time to the hospital was a commonly reported barrier to exercise participation is in line with other studies reporting that cancer survivors rated long travel time to exercise facilities as an important barrier to participation, particularly when supervised sessions were scheduled for 2 or 3 times per week [39,40].Travel time can be reduced by ofering exercise interventions in local physical therapy practices or in the community settings close to patients' homes.In the Netherlands, regional networks of physical therapists working with patients with cancer are expanding, facilitating the accessibility to supervised exercise sessions.In addition, a network of ftness instructors with additional oncology education is developing, which might ease the transition from healthcare to community settings.
Experiencing less "barriers to exercise" was associated with higher participation during chemotherapy, while, after treatment, patients with more barriers were more likely to participate.Tese fndings might indicate that at start of treatment, patients might be too occupied with the burden of diagnosis and treatment to overcome existing barriers to exercise, while, after treatment, patients' declined ftness levels and difculties overcoming these might make them more prone to accept exercise guidance.
Two other commonly reported barriers are time and mental burden (e.g., "having too many things on one's mind") [41].In studies with patients under active treatment, the timing of trial inclusion before the start of chemotherapy is challenging because of the short time window between diagnosis and start of treatment and because patients who were diagnosed recently can be overwhelmed [26].Tese barriers might be reduced by improving knowledge on the content and benefts of exercise during and after chemotherapy and by optimizing the timing of discussing exercise with patients.Shaping knowledge is among the most commonly used behavioral change techniques [42].Specifcally, instructions on how to perform the exercise behavior and information on the health consequences thereof were often part of interventions that were efective in improving exercise behavior in breast cancer survivors [43].Hence, increasing knowledge of health benefts may help patients to restructure priorities.Tis may also be the case for patients with lower educational levels, who were less willing to participate in exercise trials both during and after chemotherapy [26,30].However, for the latter patients, the educational techniques applied might need to be adapted, for example, by breaking down information into small concrete steps and/or by including visual aids [37].
In the Netherlands, an e-learning module is available for nurse (practitioners), which, in addition to addressing common efects of exercise in patients with cancer, also pays attention to how to coach and motivate patients towards improving and maintaining adequate exercise levels [44].Te optimal timing of discussing exercise with patients is unknown while results from the Dutch trials indicated that thinking about exercise shortly after diagnosis may be an additional burden for some patients; for other patients, the diagnosis may be a teachable moment [45] and the right time to discuss exercise at the time of diagnosis [46].It is also likely that patients' information needs and receptivity change over the course of their treatment and recovery although this is currently an understudied subject.Te ACSM's Exercise Is Medicine (EIM) initiative proposes assessing, advising, and referring to physical activity in a recurrent pattern to take into account the diferent preferences and changing needs of patients for referral to

Reach
Refers to the number and characteristics of participants when compared to the target audience Operationalization: reach was evaluated by the number and characteristics of participants included in the exercise trials when compared to the target population Efectiveness Refers to the positive and negative consequences of the intervention under optimal conditions or real-world circumstances, respectively Operationalization: efectiveness was evaluated by the impact of an intervention on aerobic ftness, fatigue, quality of life, self-reported physical function, and patient satisfaction

Adoption
Refers to the staf and settings that participate Operationalization: adoption was evaluated as the representativeness of settings and satisfaction of staf involved in the Dutch exercise trials Implementation Refers to the extent to which the program was implemented as intended, i.e., intervention fdelity and resources (e.g., cost and time) Operationalization: implementation was evaluated by (i) the participants' adherence to an exercise program and (ii) resources and intervention costs Maintenance Refers to the long-term efects, both at the level of the individual patient and at the level of the organization in terms of the sustainability of the program delivery over time in the settings without added resources and leadership Operationalization: we describe maintenance at both the patient (individual) and setting level.At the patient level, maintenance has been defned as the long-term efects (≥6 months) of the intervention.At the setting level, we examined the extent to which the exercise programs are institutionalized or part of the routine organizational practices and policies European Journal of Cancer Care European Journal of Cancer Care exercise programs [47].For trial purposes, with small windows of opportunity for including patients at the start of chemotherapy, an improvement in research infrastructure and a proactive approach to patients would be helpful, e.g., an outpatient research clinic, where patients are being asked at diagnosis for a broad consent for being approached for (future) research participation.Broad consent enables researchers to directly approach patients without intercession of a healthcare provider and outside a medical appointment.
Physicians play an important role in referring patients to exercise, and thus, in increasing the reach, as patients are more likely to participate in exercise, it has been recommended by a physician [40,48,49].However, while most oncologists, including those in the Netherlands, report understanding the importance of exercise, only one in three actually refers patients [50][51][52].Reported barriers for this poor referral rate are lack of time, insufcient knowledge, and safety concerns [53].It has been suggested that the development of a roadmap for oncology clinicians with detailed pathways for exercise programming, in which discussing exercise participation becomes part of routine care, would facilitate referral [47].However, empirical evidence on the efectiveness of such a roadmap and feasibility in Dutch clinical practice is lacking.Additionally, physicians' referral may also be improved by increasing patient awareness of exercise benefts, empowering patients to raise the issue of referral themselves during consultation [54], and improving insurance reimbursement and thereby accessibility [52].[23], but it was not statistically signifcant in the other [17].Supervised exercise limited physical fatigue signifcantly, while an unsupervised exercise program did not (Table 3).Both supervised and unsupervised exercise had a signifcant benefcial efect on physical functioning in one trial [23] but not in the other [17] (Table 3).

Summary of Results of Dutch RCTs. Supervised exercise interventions during chemotherapy had a signifcant positive efect on aerobic ftness in one study
Exercise after completion of treatment signifcantly improved aerobic ftness, physical functioning, global QoL, and general fatigue in one trial (Table 3) and physical fatigue in both trials [32] (Table 3).
Average patient satisfaction was 8.5 (on a 1-10 scale) for supervised exercise during chemotherapy, 8.4 for supervised exercise after treatment, and 7.4 for an unsupervised exercise program during chemotherapy.Up to 25% of patients reported that the exercise program during chemotherapy was "too burdensome."After treatment, up to 16% of patients reported the program as "not being tailored enough," up to 10% reported difculties with scheduling exercise sessions, and some patients reported the exercise program as "(too) heavy or exhausting" (12% up to 17% for patients in the High Intensity (HI) exercise group) (Table 3).

Refections and Opportunities for Further
Implementation.Te benefcial efects of exercise on aerobic ftness, fatigue, and HRQoL found in the Dutch trials correspond with fndings from other studies [4].Of note, the stringent eligibility criteria of RCTs (e.g., excluding patients with serious orthopedic and cardiovascular or pulmonary comorbidities) may hamper generalizability of the benefcial efects to all patients with breast cancer treated with curative intent.Although it may be expected that these patients could also beneft from exercise, more extensive tailoring of the exercise protocols is likely necessary to take specifc comorbidities into consideration [55].
Te benefcial efects on aerobic ftness, fatigue, and HRQoL, both during and after treatment, were also reported by several meta-analyses on aggregated and individual patient data (IPD) meta-analyses [1][2][3] that also reported larger benefts for supervised interventions and patients with lower baseline HRQoL [56].High baseline values of HRQoL may explain the lack of efects on HRQoL in the UMBRELLA Fit trial as the HRQoL in this cohort of patients was already comparable to the Dutch general female population at the start of the intervention, leaving little room for improvement [32].
Te average efects of exercise on aerobic ftness observed in the Dutch trials correspond to the mean peakVO 2 improvements of 1.80 and 2.13 ml/kg/min reported in the literature [57].Strikingly, previous studies with IPD analyses reported that exercise interventions during treatment did not yield benefts for aerobic ftness in patients with a low ftness level (peakVO 2 below 15.4 ml/kg/min, which is the threshold for functional independence in women [58]) at baseline [56].Also, efects on aerobic ftness were smaller in older patients [1].Te limited efects in these subgroups may be related to low adherence or inability to complete exercises as intended and highlight the need for exercise interventions that are specifcally tailored to older and unft cancer patients to further improve implementation.
Te Dutch trial results suggest a dose-response efect for exercise intensity on aerobic ftness (low-to-moderate (LMI) versus HI) (Table 3).Internationally, in the past 5 years, an increasing number of studies successfully examined the efects of high intensity interval training (HIIT) in patients with cancer and found positive results on cardiorespiratory ftness and cancer-related fatigue [59,60].However, results of the Dutch trials suggest that HIIT may not be the best choice for all patients, as up to one quarter of the patients indicated that the exercise intervention conducted during cancer treatment was too burdensome directly after their chemotherapy administration, and 17% of those who participated in HI exercise after cancer treatment found it (too) heavy or exhausting (Table 3).Also, depending on the goal of the intervention, higher exercise intensity may not always be necessary.For example, relatively low volumes of resistance exercises, at moderate-to-high intensity, have been found to yield signifcant benefts in terms of fatigue levels and HRQoL in patients with prostate cancer [61].
Despite the fnding that some patients found the intervention to be too strenuous or burdensome, the vast majority of the patients in the four Dutch trials indicated being very satisfed with the exercise interventions in which they participated.Tis has also been the case in other exercise trials in patients with breast cancer [62][63][64].Patients 6 European Journal of Cancer Care Table 3: Results of the Dutch exercise trials summarized using the dimensions of the RE-AIM framework.

Adoption
Guidance of the exercise programs (i) Patients in the Dutch RCTs were referred to a trained PT close to their home to guide the exercise intervention.Te PTs were trained on the specifc exercise protocols by the coordinating researcher before the start of the intervention (ii) In the Netherlands, PTs are the exercise professionals whom are dedicated to supervise the exercise programs for patients during chemotherapy Satisfaction of physical therapists l (i) 20 PTs rated the exercise intervention after chemotherapy with a 7.5 out of 10 (ii) Preference for HI over LMI, but with the suggestion to start with LMI in patients with reduced physical ftness (iii) Preference for a more fexible exercise protocol to be able to improve variation of exercises (iv) Counseling techniques were rated with a 7 out of 10, with remarks that the counseling was not always possible in the available time (20%) and PTs expressed a need for education with regard to counseling techniques (10%) l 8 European Journal of Cancer Care

Implementation-cost-efectiveness
Currently, physical therapy is not reimbursed through basic healthcare insurance in the Netherlands.For patients who have been hospitalized prior to their chemotherapy or had radiotherapy sessions in the past 6 months, physical therapy is covered starting from the 21st session.Hence, patients who receive neoadjuvant chemotherapy are not entitled to any reimbursement from basic coverage, and patients who receive adjuvant chemotherapy face out-of-pocket expenses for the frst 20 sessions, unless they have additional coverage packages During chemotherapy [21,25] After treatment [29] Supervised (i) Te OnTrack intervention (PACES trial) was found to be cost-efective for QALYs Mean intervention costs for the supervised intervention were 756.67 euros per participant.Te probability of OnTrack being cost-efective compared with UC was 45% at a willingness-to-pay of 20.000€/QALY and 79% at a willingness-to-pay of €80.000/QALY (ii) Te probability that the PACT intervention is cost-efective for patients with breast cancer is 2% for a willingness-to-pay 20.000/QALY and 6% for a willingness-to-pay €80.000/QALY (i) Te probability that HI was cost-efective compared to LMI exercise was 0.91 at 20.000€/QALY and 0.95 at 52.000€/QALY Unsupervised (i) Te unsupervised OncoMove intervention (PACES trial) was found not to be cost-efective (ii) Mean intervention costs for the unsupervised intervention were 46 euros per participant.Te probability of cost-efectiveness was 25% at a willingness-to-pay of 20.000€/QALY and 55% at a willingness-to-pay of €80.000/

QALY
European Journal of Cancer Care in the Dutch trials suggested that they would appreciate being able to reschedule missed exercise sessions, add more variety to the prescribed exercises, and combine the exercise sessions with yoga [22].Tis was also found in a study of women with ovarian cancer [65].Taking patient preferences into account can increase enjoyment, which, in turn, can have a benefcial efect on exercise maintenance [66].At the same time, to achieve their goals and the desired health benefts, it is important that patients are informed on the exercise frequency, intensity, type, and time (FITT) required.Patient satisfaction, and thereby potentially exercise maintenance, can be further improved by taking sufcient time for exercise familiarisation, optimizing exercise scheduling in relation to chemotherapy administrations [67], and adequate tailoring of exercise intensity to the individual's ftness level.
Te generally lower level of satisfaction reported for the home-based exercise counselling compared to supervised exercise may be related to the limited time devoted by healthcare professionals (e.g., physical therapists or nurse practitioners) (HCPs) to instructing and motivating patients and to individualizing the home-based exercises.Motivational interviewing appeared to be an efective technique to improve exercise behavior of patients with cancer in some studies [68,69], but not all [70].Dedicated time, and better training in exercise counselling, and development of supportive tools may improve the counselling skills of HCPs delivering exercise programs.

Summary of Results of Dutch RCTs.
Most patients in the Dutch RCTs were recruited from both community and university hospitals and were referred to a physical therapist specifcally trained to work with patients with cancer, located close to the patients' homes.Te physical therapists who delivered the intervention after completion of cancer treatment were generally satisfed with the content of the trial intervention (supervised aerobic and resistance exercise two days per week, supplemented by counseling on unsupervised exercise for three other days).Te average satisfaction score was 7.5 (on a 1-10 scale, Table 3).While, overall, the physical therapists were satisfed with the exercise intervention, some reported that they would have preferred to prescribe more variation in the resistance exercises (20%) that the exercise counseling was too timeconsuming (20%) and that physical therapists could beneft from some additional training in this regard (10%) (Table 3).

Refections and Opportunities to Improve Adoption.
Few studies have described experiences of professionals delivering exercise interventions to patients with cancer treated with curative intent in the context of a trial.More variation in exercises has also been suggested by other studies, in order to prevent boredom, better tailor exercises to patients' preferences, needs (e.g., functional training), or capabilities, and to add exercise types other than resistance or aerobic exercises, such as balance exercises [65,71].In addition, physical therapists and personal trainers have reported that guidance of a group of patients can be challenging (e.g., dealing with diferent types of group dynamics; providing sufcient attention to individual patients' abilities and needs) [72,73].Tis indicates the importance of qualifed trainers and preferably small group sizes although the latter needs to be balanced with afordability.
Results of focus groups in various HCPs working in primary or secondary care in the Netherlands (e.g., physicians, nurses, and physical therapists) reported that insufcient evidence about benefts of exercise programs was a barrier for their use [71].Tis can be a reason for not referring patients to exercise programs.Hence, eforts to disseminate the evidence on the efects of exercise on cancer outcomes will likely accelerate the implementation of exercise as part of standard cancer care [74].Additionally, these exercise programs should be adequately tailored to the individual patients' needs, capabilities, and preferences [71] while taking evidence-based exercise frequency, intensity, type, and time (FITT) into account.

Implementation.
Implementation was evaluated based on (a) exercise adherence and (b) resources and intervention costs [75].

Exercise Adherence
(1) Summary of Results of Dutch RCTs.Te median attendance rates of supervised exercise in the Dutch trials varied between 77% and 98% and were 71% for unsupervised exercise (Table 3).Median compliance rates ranged between 81% and 88% for moderate intensity aerobic exercises, between 50% and 87% for HI aerobic exercises, and between 84% and 94% for resistance exercises (Table 3).Te most frequently reported reasons for not attending the sessions during chemotherapy were feeling too ill (53%) and logistical reasons (30%).During chemotherapy, higher disease stage, having a partner, higher educational level, and a lower body mass index (BMI) were signifcantly associated with better attendance [20,22].Results on predictors of compliance to the prescribed exercises during and after treatment suggested a diference between exercise type (resistance versus aerobic), intensity (LMI versus HI), and delivery mode (supervised versus unsupervised) [20].In general, after cancer treatment, psychosocial factors, such as higher self-efcacy and having a more positive attitude towards exercise, were associated with a higher attendance and compliance to HI but not to LMI exercise (Table 3).
(2) Refections and Opportunities to Improve Implementation.Te exercise adherence rates and the diversity in predictors of adherence in Dutch studies are in line with previous fndings from other studies [76][77][78].Tis diversity can be explained by diferences in exercise prescriptions between studies and in the predictors studied.Te fnding that treatment-related adverse efects ("feeling too ill") accounted for over half of the total missed sessions is also in line with other exercise studies in patients with breast cancer European Journal of Cancer Care receiving chemotherapy [79,80].Consideration of side effects as part of exercise program design has been proposed, for example, by using "chemotherapy-periodized" exercise prescriptions that take chemotherapy side efects into account [67].Te fnding that patients with lower exercise selfefcacy and more negative attitudes towards exercise had more difculties with adhering to HI exercise suggests that realistic goal setting and starting at a lower intensity, to gain confdence before progressing to HI exercise, may be useful to improve adherence of these patients.
Previous systematic literature reviews found exercise history to be associated with better exercise adherence [77,78].Tis was not supported by the results from the Dutch trials, which suggests that other factors may be more important.It should be noted that the overall adherence reported in the Dutch trials was relatively high.Tis may not, however, turn out to be the case in clinical practice, due to variation in motivation and less emphasis on required adherence [73,81].Future studies in daily clinical practice that yield real-world data on exercise adherence, collected via electronic medical records, might help elucidate which factors are associated signifcantly with adherence to exercise outside the trial context and to identify subgroups of patients and cancer survivors that might require adjustments to the exercise intervention or psychosocial and behavioral support for improving adherence.

Resources and Intervention Costs
(1) Summary of Results of Dutch RCTs.Two of the Dutch trials assessed the cost-efectiveness of the exercise interventions during treatment.In one trial, the supervised exercise intervention during chemotherapy was found to be cost-efective with a probability of 45% at a willingnessto-pay of 20.000€/quality-adjusted life year (QALY) [25].Te other trial reported that, at a willingness-to-pay of 20.000€/QALY, the probability that the intervention would be cost-efective was very low (2%) [21].Te unsupervised exercise intervention was not cost-efective (25% probability for cost-efectiveness at a willingness-to-pay of 20.000€/ QALY).After completion of chemotherapy, a HI-exercise program was more cost-efective than a LMI exercise program [29], with a probability of 91% at a willingness-to-pay of 20.000€/QALY.
(2) Refections and Opportunities for Further Implementation.Results on the cost-efectiveness of exercise interventions in Dutch trials were mixed, possibly explained by contamination or diferences in follow-up time [21,25].Previous systematic reviews showed that supervised exercise interventions and multimodal interventions were costefective when they yielded signifcant benefcial efects on health outcomes such as energy, fear of recurrence, mood, and pain [82,83].
In the Netherlands, exercise supervision from a physical therapist is currently not reimbursed by basic healthcare insurance.However, exercise sessions for patients who have had surgery prior to their chemotherapy or received radiotherapy treatment in the past 6 months can be reimbursed from the 21st session onwards until 1 to 2 years (depending on the treatment and insurance).Additionally, biweekly one-hour supervised exercise sessions can only be provided in group sessions of 2-10 persons because physical therapists are allowed to invoice for a maximum of 30 minutes per day per person.Fortunately, group sessions are often appreciated by cancer survivors and facilitate peer support [22,62,65] and may consequently improve adherence rates.On the other hand, group sessions are often less fexible with regard to exercise times, which has been reported as barrier to (adhering to) exercise programs [22].Also, group sessions are not suitable for every patient as some patients feel uncomfortable with group exercise and/or may require more intensive coaching than possible in group settings.
More information on the cost-efectiveness of exercise interventions and consideration of other healthcare reimbursement strategies (e.g., bundled-payment models) could be helpful to better inform discussions among health policy-makers and insurers about appropriate reimbursement policy and insurance coverage for exercise interventions during and after chemotherapy.Future costefectiveness evaluations need to take into account that higher chemotherapy completion rates resulting from exercise interventions may result in higher costs of medication and secondary healthcare but also higher survival rates [84] and that work absenteeism may be underestimated when absence, beyond the percentage sick leave that is agreed upon by patients and employers, is not reported as absenteeism days [25].

Maintenance.
In this section, we evaluated maintenance at the patient level¸describing long-term efects of the intervention, and at the organizational level, describing the extent to which the exercise programs have been institutionalized and integrated into routine practice, as well as the policies enabling program sustainability.

Patient Level
(1) Summary of Results of Dutch RCTs.Exercise during chemotherapy did not yield signifcant efects on aerobic ftness, self-reported fatigue, or HRQoL at follow-up (i.e., 6, 8, and 48 months) [19,23].However, patients who participated in an exercise program after completing their oncological treatment successfully maintained their improved levels of cardiorespiratory ftness and HRQoL at one-year postintervention (Table 3).Te positive intervention efects on HRQoL observed at one-year follow-up were signifcantly larger for HI compared to LMI exercise (Table 3).
(2) Refections and Opportunities for Improving Maintenance at the Patient Level.Te limited maintenance of intervention efects on most outcomes might be explained by the uptake of exercise by control group participants after the completion of chemotherapy or the specifc focus on improving outcomes during the intervention period without sufcient incorporation of behavioral change techniques to maintain healthy behaviors in the long term.
Sustained benefts on aerobic ftness were found one year after completion of exercise interventions and after cancer treatment.Nevertheless, peakVO 2 levels were still "poor," as compared to healthy adults [29].Tis might indicate that a 12-week program might be too short for patients to fully return to normative values, and patients may not have received sufcient guidance to continue exercising at home at sufcient intensities after completion of the trial to continue improving their peakVO 2 .Tis is in line with the previously mentioned feedback of physical therapists that they had insufcient time for counseling and expressed a need for additional education (Table 3).Further development of tools to improve the quality of counseling and efcient integration into daily practice might improve maintenance of adequate exercise levels to further improve peakVO 2 .Tis could be achieved by improved incorporation of behavioral change techniques, such as "instruction on how to perform the behavior," "feedback and self-monitoring of behavior," and "goal-setting (behavior)" [43].Education on how to use behavior change techniques may help to overcome some perceived barriers of exercise maintenance that have been reported by patients with breast cancer, including psychological barriers (e.g., lack of motivation, fears, dislike of gym, or not being the "sporty type"), physical barriers (e.g., ageing, side efects of cancer treatment, and other comorbidities, weight gain), and contextual and environmental barriers (related to employment, traditional female care-giving roles, access to facilities, and seasonal weather) [85].

Organizational Level
(1) Summary of Results of Dutch RCTs.For the conduct of the Dutch trials, physical therapists were trained to supervise patients with cancer in exercising during chemotherapy and after treatment, within the initiated Onconet network.After trial completion, the Onconet foundation further educated physical therapists on the content and delivery of exercise programs for patients with cancer, and thereby consolidated and expanded a physical therapist network.Te education also includes mandatory refresher courses where physical therapists are updated on results from recent studies.Currently, the network of physical therapists specialized in guiding patients with cancer is nationwide, with over 700 locations mostly within a 15-minute travel distance from any address.Additionally, MSc-level programs are available to educate physical therapists in oncology.
(2) Refections and Opportunities for Improving Maintenance at the Organizational Level.In the Netherlands, currently, most supervised exercise interventions are ofered by allied healthcare professionals.In primary care, this is primarily via physical therapists working in private clinics.Exercise can also be ofered as part of a multidisciplinary rehabilitation program in secondary (hospitals) or tertiary (rehabilitation clinics) care.Outside of the healthcare system, ftness trainers with oncology specialization are increasingly available.Tese ftness professionals mainly deliver exercise interventions to patients who have completed treatment at least 3 months earlier [71].
Results from a qualitative study in the Netherlands indicate that HCPs working in primary care (e.g., general practitioners, physical therapists) perceive collaboration, communication, and referral between primary and secondary HCPs to be suboptimal [71], whereas HCPs working in secondary care (e.g., physicians, nurses, and paramedics) raised general concerns about inadequate cooperation and networks between healthcare institutes [71].Te HCPs suggested that more use of health information technology, improved access to electronic health records, improved rehabilitation guidelines with recommendations about roles and responsibilities of each HCP, and better networks would improve the implementation of exercise in cancer care [71].
Internationally, the most reported barriers to integrating exercise in oncology settings are at the organizational level [86].Tese barriers are related to the limited capacity and resources of staf, including insufcient time to prescribe and refer patients to exercise programs, and to the organization of care processes (e.g., absence of an established care pathway or structure) [86].To reduce organizational barriers in the Netherlands, the "Taskforce Cancer Survivorship Care" has been established since 2017.In this taskforce, HCPs, policymakers, researchers, and patient organizations join forces aiming to improve attention for and optimization of quality of care over the whole cancer continuum and to improve organizational structures by better coordination between HCPs [87].Te taskforce also pursues an increase of physical therapist participation in multidisciplinary and oncology care networks, enabling further knowledge exchange and improved communication with other HCPs.

Discussion
In this study, we have used the RE-AIM framework to describe the potential for implementation of exercise interventions for patients with breast cancer, based on four RCTs previously conducted in the Netherlands.Results from these RCTs demonstrated that exercise during and after treatment has benefcial efects on aerobic ftness, fatigue, and HRQoL in patients with breast cancer.Additionally, both patients and physical therapists were generally satisfed with the intervention, but there were challenges to exercise maintenance.
Te current network of physical therapists specialized in oncology that was initiated at the start of the trials continues to expand and represents a fruitful interaction between research and clinical practice.Te current evaluation revealed key opportunities to further optimize implementation of exercise programs in the oncology setting.First, there is room to further increase knowledge and awareness among HCPs of the potential benefts of exercise and to improve organizational structures to increase referral to supervised programs.Improving awareness and referral requires more insight into perspectives of organizational European Journal of Cancer Care stakeholders and policymakers and optimal dissemination of patient information between HCPs for which a whole system approach is needed [71,86].
Second, although the interventions in the Dutch exercise trials were tailored to individuals' ftness level and treatment side efects, specifc subgroups of patients, such as the elderly and those who are in poor physical condition, are more prone to nonparticipation and appear not to beneft as much [1,56].Tese patients may beneft from an even more personalized and goal-directed functional exercise training program.Such programs have been shown to be feasible and promising in patients with metastatic breast cancer [88].Similarly, a patient-centred, goal-directed, self-management enhancing functional exercise program that is based on a biopsychosocial model, Coach2move, has shown to be (cost-)efective in improving physical activity and function in community dwelling older adults with mobility problems [89].A personalized program that is tailored to the individual's needs and preferences, including behavioral change techniques such as "goal-setting" and "feedback and self-monitoring of behavior," may also facilitate sustained benefts over time.Successful inclusion of behavioral change techniques as part of exercise supervision may require additional schooling for physical therapists.
Finally, the implementation of exercise programs for all patients with cancer is currently hampered by the lack of reimbursement of physical therapist-guided exercise programs during and after treatment.Te Taskforce Cancer Survivorship aims to improve healthcare during and after cancer treatment, and one of the pillars of the Taskforce is to improve reimbursement of allied healthcare [87].
Some limitations of our study should be noted.Because it has been suggested that implementation strategies must be tailored to its context to improve efectiveness [90], we summarized exercise trials with a homogeneity in settings, circumstances, and conditions, thereby specifcally focusing on patients with breast cancer in the Netherlands.Hence, caution is needed when generalizing our fndings to other countries with diferent healthcare systems and to patients with other cancer types or advanced cancer [91,92].On the other hand, the fndings from Dutch trials seem to echo those of studies conducted in other countries.Additionally, we based our assessments on a retrospective evaluation of the potential impact of exercise intervention trials.Our fndings might have been diferent if data from the clinical practice setting had been collected prospectively (e.g., information on treatment referral and treatment fdelity outside of the context of a trial).Future studies should, therefore, prospectively evaluate the implementation of exercise interventions using the RE-AIM framework, for example, by collecting real-world data to describe characteristics of patients who are referred to exercise interventions and to register the delivered exercise prescription in terms of FITT-factors and the resulting changes in aerobic ftness, physical functioning, HRQoL, and achievement of physical therapy goals.Tis would also facilitate obtaining information about and from patients with comorbidities or those who otherwise would be excluded from trials or referred to less extent to trials [93,94].Such collection of real-world data would be facilitated by an adequate registration system to structurally monitor clinical practice, in order to learn from every patient, and subsequently optimize healthcare [95].Moreover, future research could beneft from hybrid designs, in which elements of clinical efectiveness and implementation research are combined [73,96].Tis might speed up the translation of research's fnding into clinical practice 97.

Conclusion
Te RE-AIM framework facilitated a retrospective evaluation of the impact of exercise interventions and their potential for implementation in clinical practice.We found acceptable RE-AIM outcomes in terms of participation rates, intervention efects, satisfaction of patients and physical therapists, and adherences rates within the trial context.Additionally, an established network of physical therapists educated in oncology facilitates the maintenance of exercise interventions outside of clinical trials.We have recommended several steps that could be taken to further improve implementation of exercise programs for cancer patients and survivors, including improved referral (reach), improved tailoring of exercise interventions to individual needs and preferences, improved attention to maintenance of exercise behavior (efectiveness, adoption, implementation, and maintenance), and improved reimbursement (reach and maintenance).

Table 1 :
Defnitions of the RE-AIM framework and operationalization used in the current study.

Table 2 :
Characteristics of the studies.
a Complementary to the supervised interventions, participants were encouraged to be physically active at moderate intensity for at least 30 minutes, 3-5 times per week in all trials.1RM: one-repetition maximum; AE: aerobic exercise; FITT: frequency intensity time type; HI: high intensity; HIIT: high intensity interval training; HR: heart rate; HRR: heart rate reserve; HRrest: resting heart rate; LMI: low-to-moderate intensity; PACES: physical exercise during adjuvant chemotherapy efectiveness study; PACT: physical activity during cancer treatment; MI: moderate intensity; MSEC: maximal short exercise capacity; RE: resistance exercise; REACT: resistance and endurance exercise after chemotherapy; RM: repetition maximum; SD: standard deviation; SRT: steep ramp test; TwiCs: trials within cohorts; UMBRELLA Fit: utrecht cohort for multiple breast cancer intervention studies and long-term evaluation; VT: ventilatory threshold; WLC: wait list control; Wmax: maximal wattage.

Table 3 :
Continued.Existence of a broad network of PTs trained to guide patients with cancer with exercise during and after chemotherapy (ii) All PTs within the network follow mandatory refresher courses and have to pass summative tests related to these courses (iii) Te network covers most of the populated areas in the Netherlands, and an specialized PT is available within a 15 min commute for most people Over 550 PTs working at over 700 locations in the Netherlands have been specifcally educated on exercise treatment during and after treatment, via the Onconet courses executed by the Dutch Institute of Allied Healthcare.Tese PTs have received 67 hours or more of additional training in subjects such as basic oncology, exercise oncology, behavioral support, dealing with cancer-specifc side efects, dealing with comorbidities, using clinimetrics, and clinical reasoning in an oncology context *Indicate efects in favor of the control group.EC: European credits; HI: high intensity; IQR: interquartile range; LMI: low-to-moderate intensity; MHI: moderate-to-high intensity; PT: physical therapist; QALY: quality-adjusted life year; WLC: wait list control.a Unpublished results of the REACT trial (additional analysis on data from patients with breast cancer only); analysis on predictors for participation with multivariable logistic regression; analysis on predictors for adherence performed with univariate logistic regression; b