Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Long-term effectiveness and cost-effectiveness of an 18-week supervised exercise program in patients treated with autologous stem cell transplantation: results from the EXIST study

Long-term effectiveness and cost-effectiveness of an 18-week supervised exercise program in... Purpose To evaluate the long-term effectiveness and cost-effectiveness of a supervised 18-week high-intensity exercise program compared with usual care in patients treated with autologous stem cell transplantation. Methods One hundred nine patients were randomly assigned to the exercise intervention (n = 54) or the usual care control group (n = 55). Data on cardiorespiratory fitness (VO peak), handgrip strength, general fatigue, and health-related quality of life (quality-adjusted life years [QALYs]) were collected at baseline (T0), after completion of the exercise intervention or at a similar time point in the control group (T1) and 12 months later (T2). Cost questionnaires were used to assess societal costs. Long-term effectiveness (at T2) was evaluated using linear mixed model analyses. For the economic evaluation, missing data were imputed using multiple imputation, and data were analyzed using linear mixed models. Results At T2, no statistically significant differences were found between the intervention and control group for VO peak (0.12; 95%CI − 1.89; 2.14 ml/min/kg), handgrip strength (− 1.08; 95%CI− 2.47; 2.31), and general fatigue (− 0.69; 95%CI − 2.52; 1.14). During 12-months follow-up, no significant between-group differences in QALYs and societal costs were found (QALYs − 0.07; 95%CI − 0.17; 0.04; costs 529; 95%CI − 3205;4452). Intervention costs were €1340 per patient. For all outcomes, the probability of the intervention being cost-effective was low at reasonable values of willingness-to-pay. Conclusion We found no evidence for the exercise intervention being effective on physical fitness and fatigue, nor cost-effective from a societal perspective. Trial registration The study was prospectively registered on 27 May 2010 at the Netherlands Trial Register (NTR2341). Implications for Cancer Survivors The current exercise intervention should not be recommended to patients recently treated with autologous stem cell transplantation. . . . . Keywords Long-term effectiveness Cost-effectiveness Exercise intervention Multiple myeloma Lymphoma * Laurien M. Buffart Amsterdam UMC, University of Amsterdam, Department of l.buffart@vumc.nl Haematology, Amsterdam, The Netherlands Lymphoma and Myeloma Center Amsterdam (LYMMCARE) and Vrije Universiteit Amsterdam, Department of Health Sciences, Cancer Center Amsterdam, Amsterdam, The Netherlands Faculty of Science, Amsterdam Public Health Research Institute, National Institute for Public Health and the Environment, Amsterdam, The Netherlands Bilthoven, The Netherlands Amsterdam UMC, University of Amsterdam, Department of 7 Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Rehabilitation, Amsterdam Movement Sciences, Public and Occupational Health, Amsterdam Public Health Research Amsterdam, The Netherlands Institute, Amsterdam, The Netherlands 3 8 Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Epidemiology and Biostatistics, Amsterdam Public Health Research Medical Oncology, Cancer Center Amsterdam, Institute, De Boelelaan 1089a, 1018HVAmsterdam, The Netherlands Amsterdam, The Netherlands J Cancer Surviv (2019) 13:558–569 559 Introduction for cost-effectiveness analyses, it is important that the time frame of analyses extends beyond the intervention period to Cancer and cancer treatment may have a negative impact on ensure that the most important costs and consequences physical fitness, fatigue, and quality of life, which may be flowing from the intervention are covered [7]. This is because counteracted by exercise interventions. Several systematic re- patients with relatively large health effects directly after the views reported favorable effects of exercise interventions on intervention typically have relatively low levels of healthcare cardiorespiratory fitness [4, 9, 28], muscle strength [9, 28], consumption and productivity losses during the consecutive fatigue [4, 9], and health-related quality of life (HRQoL) [3, period and vice versa. To cover this research gap, the present 30] in patients with cancer during and following cancer treat- study aimed to evaluate the long-term effectiveness and cost- ment. In various countries, this has resulted in the develop- effectiveness of the exercise intervention versus usual care for ment of exercise guidelines for patients with cancer [2]. physical fitness, fatigue, and HRQoL in patients treated with Beneficial effects on these outcomes were also found in pa- autologous stem cell transplantation. tients who received a stem cell transplantation [20, 22]. A recent study has shown that the effects of exercise inter- ventions following cancer treatment for solid tumors on phys- Methods ical fitness and HRQoL can be sustained at 1 year, while the intervention effects on fatigue disappeared [12]. To the best of The EXIST study was a multicenter randomized controlled our knowledge, effects of exercise interventions in patients trial that evaluated the effectiveness of an 18-week supervised who received a stem cell transplantation on these outcomes high-intensity combined resistance and interval exercise pro- at long-term (≥ 1 year) are unknown. gram compared with usual care on physical fitness and fatigue Information on the cost-effectiveness of exercise interven- as primary outcomes as well as its cost-effectiveness in pa- tions is important for decision-makers, as this provides insight tients with a hematologic malignancy treated with autologous into the additional cost per unit of effect gained, and may thus stem cell transplantation. The study procedures were approved provide guidance to decision-makers when deciding how to by the Medical Ethics Committee of the Academic Medical allocate scarce resources in healthcare [7, 27]. Nonetheless, Center (METC AMC 10/106) and by the boards of the Antoni there is only limited evidence available on the cost- van Leeuwenhoek Hospital (Amsterdam), St. Antonius effectiveness of exercise interventions in patients with cancer. Hospital (Nieuwegein), Haga Teaching Hospital (Den A previous systematic review on this topic [18] included only Haag), University Medical Center (Utrecht), Isala (Zwolle), three studies. Since then, a small number of additional studies Erasmus MC/Daniel den Hoed (Rotterdam), VU University have evaluated the cost-effectiveness of exercise interventions Medical Center (Amsterdam), and Leiden University Medical during or after cancer treatment [1, 10, 12, 17, 19, 35], but Center (Leiden). The study protocol as well as the baseline they were heterogeneous in the type of cancer, type of inter- and short-term results have been published previously [21, 23, vention and control condition, study results, and/or they were 24]. not based on patient-level data. Given the importance of this Patients were recruited between March 2011 and February topic and the scarcity of literature, more research on the cost- 2014. Patients were eligible for the study if they were (1) effectiveness of exercise interventions in patients with cancer treated with autologous stem cell transplantation for multiple is warranted. myeloma or lymphoma 6–14 weeks earlier or treated with Recently, we have published the results of the EXercise autologous stem cell transplantation and finished successive Intervention after Stem cell Transplantation (EXIST) random- consolidation chemotherapy or radiotherapy 2–6 week earlier; ized controlled trial [23]. Within this study, an 18-week super- (2) recovered from treatment (Hb > 10.5 g/dL, platelets > vised high-intensity combined resistance and interval exercise 80 × 10 /L); and (3) able to undergo exercise testing and to program was compared with usual care in patients with a participate in an exercise intervention. All patients provided hematologic malignancy treated with autologous stem cell written informed consent. transplantation. At the short-term (directly after the interven- After the baseline assessments and confirmation of eligibil- tion), the exercise intervention did not have beneficial effects ity, patients were randomized to the intervention or usual care on physical fitness, fatigue, and HRQoL [23]. Some may ar- group by an independent data manager using a validated soft- gue that the exercise intervention will therefore not be effec- ware program (TENALEA Clinical Trial Data Management tive, nor cost-effective at the long-term, and that publication of System; Netherlands Cancer Institute, Amsterdam, the these results is therefore not necessary. However, the EXIST Netherlands). Randomization was concealed, stratified for exercise intervention also included counseling sessions transplant center and diagnosis using block randomization aiming to promote compliance and maintenance of exercise with block sizes varying randomly between 2, 4, and 6. The after the program [21, 23, 24] and it is currently unclear study personnel that assessed long-term outcomes and per- whether the intervention has delayed effects. Additionally, formed the analyses were not blinded for treatment allocation. 560 J Cancer Surviv (2019) 13:558–569 Exercise intervention and usual care between T1 and T2. The physical therapists kept a training log during the intervention program. The 18-week exercise intervention was supervised by instructed physical therapists and took place at local physical therapy practices. The program consisted of 30 exercise and 6 Effect measures counseling sessions. Each exercise session lasted approxi- mately 60 min and included 6 resistance exercises targeting The primary effect measures were physical fitness, including the large muscle groups (vertical row, leg press, bench/chest cardiorespiratory fitness and handgrip strength, and general press, and pull over/flies and two additional exercises for the fatigue. abdominal muscles and the upper legs) and 2 bouts of 8 min Cardiorespiratory fitness was expressed as the highest con- cycling interval training (Table 1). The indirect one repetition tinuous 15 s interval values for oxygen uptake (VO in ml/ 2peak maximum (1-RM) test [15] and the steep ramp test [5]were kg/min; MasterScreen CPX, CareFusion, Hoechberg, performed every 4 weeks in order to tailor and adjust the Germany) measured during a cardiopulmonary exercise test exercise intensity prescription. Counseling sessions lasted 5 performed on a cycle ergometer (Lode Excalibur, Groningen, to 15 min each and took place in week 1, 4, 10, 12, 18, and 22 the Netherlands). [21]. The counseling sessions were provided by the physical Handgrip strength of the dominant hand was assessed therapist who supervised the exercise intervention and aimed using a grip strength dynamometer (Hydraulic Hand to improve compliance to the exercise intervention and to Dynamometer, North Coast Medical Inc., Morgan Hill, encourage patients to pursue a physically active lifestyle dur- USA), and the highest score out of three attempts in kilogram ing and following the program [21]. From week 12 onwards, was used in analyses. patients were encouraged to meet physical activity levels as General fatigue was determined using the subscale of recommended by guidelines [11]. Specific program elements Multidimensional Fatigue Inventory (MFI) questionnaire included the provision of general and motivational informa- [29]. The patients could score between 4 and 20, with higher tion, both verbally and via folders, about the desired frequen- scores indicating more fatigue. cy, duration, and intensity of physical activity [21]. For the economic evaluation, HRQoL was assessed using Usual care varied according to patients’ and physicians’ the EQ-5D-3L [8]. The EQ-5D-3L was administered at T0, preferences. Control group patients were not restricted in their T1, T2, and in an additional cost diary half-way between T1 physical activities or in their use of healthcare services. and T2. The EQ-5D-3L consists of five questions evaluating the following health dimensions: mobility, self-care, usual ac- Timing of assessments tivities, pain/discomfort, and anxiety/depression. The patients’ EQ-5D-3L health states were transformed into utility scores The effect measures were assessed at baseline (T0), after com- using the Dutch tariff [14]. QALYs were calculated by multi- pletion of the intervention or at a similar time point in the plying the patients’ utility scores by the time they spent in a usual care control group (T1; ~ 22 weeks after T0), and 1 year certain health state using linear interpolation between mea- after T1 (T2). Patients visited one of two test centers to par- surement points. More QALYs indicate a better HRQoL, with ticipate in exercise testing and filled out questionnaires at approximately 1.3 QALYs as the maximum number (i.e., home. Cost data were assessed using cost questionnaires: pa- 1.3 years in optimal health). tients filled out five monthly cost questionnaires between T0 Effects on all primary outcomes and HRQoL occurring and T1, and four 3-monthly questionnaires in the year after 1-year follow-up were discounted at a rate of 1.5% [31]. Table 1 Structure of the exercise program Week Frequency of exercise session Type of training Frequency, intensity, and duration of the exercises/sessions 1–8 Twice a week Resistance exercises 2 sets of 10 repetitions at 65–80% of 1-RM* Cycling interval training 2 × 8 min, alternating 30s at 65% and 60s at 30% MSEC 9–12 Twice a week Resistance exercises See week 1–8 Cycling interval training 2 × 8 min, alternating 30s at 65% and 30s at 30% MSEC 12–18 Once a week Resistance exercises 2 sets of 20 repetitions at 35–40% of the 1-RM* Cycling Interval training See week 9–12 1, 4, 10, 12, and 18 Physical activity counseling 5–15 min per session 1-RM one repetition maximum, MSEC maximal short exercise capacity, i.e., the highest workload achieved during the steep ramp test *For the two additional exercises the protocol included the performance of 2 sets of 0.7 times the maximal number of repetitions J Cancer Surviv (2019) 13:558–569 561 Measurement and valuation of resource use replace a sick worker (i.e. friction period), which was estimat- ed to be 23 weeks at time the intervention was provided [31]. Costs were measured from a societal perspective and included After truncating the number of sickness days at the friction intervention costs, healthcare costs, costs of informal care, period, sickness days were valued using age- and gender- sports costs, unpaid productivity costs, and absenteeism costs. specific price weights [31]. Clinical trial–related costs (e.g., costs related to the physical All costs were expressed in 2014 Euros using consumer measurements) were not included. price indices. Costs occurring after 1-year follow-up were Intervention costs were estimated using a micro-costing discounted at a rate of 4% [31]. approach, meaning that detailed information was collected on the use of intervention-related resources as well as their Statistical analyses unit prices. Information on all other kinds of resource use was collected using cost questionnaires administered on a 1- Long-term effectiveness and cost-effectiveness analyses were monthly basis between T0 and T1 and a 3-monthly basis be- performed in accordance with the intention-to-treat principle. tween T1 and T2. To cover the complete duration of follow- Descriptive statistics were used to describe baseline character- up, cost questionnaires administered between T0 and T1 had a istics of patients from the intervention and control group, and recall period of 1 month, while cost questionnaires adminis- of patients with complete data and those with incomplete data. tered between T1 and T2 had a recall period of 3 months. The Linear mixed model analyses were conducted to evaluate cost questionnaire was developed by the research team in the intervention effects on cardiorespiratory fitness, handgrip close collaboration with experts in health economic strength, and fatigue at long-term (T2). The intervention was evaluations. regressed on the outcome value at the short-term (T1) and Intervention costs included all costs related to the imple- long-term (T2) simultaneously, adjusted for baseline levels mentation and execution of the exercise intervention, i.e., of the outcome variable in the model, age, gender, and educa- costs related to the training of the physical therapists, the pro- tion level. This procedure automatically takes into account vision of the exercise and counseling sessions, and the inter- missing values using maximum likelihood estimation. A ran- vention materials. Time investments of the study team and dom intercept for transplant center was added to take into physical therapists were valued using standard costs or, if account the clustering of patients within centers. Regression unavailable, salary data derived from their respective coefficients and 95% confidence intervals (CI) of clinical in- Collective Labor Agreement [31]. Data on the attendance of tervention effects at long-term were presented. exercise and counseling sessions were retrieved from the In the main cost-effectiveness analysis, patients with mul- training logs and the associated correspondence between the tiple myeloma using the extremely expensive drug study team and the physical therapists. Invoices were used to lenalidomide (Revlimid) as maintenance treatment after autol- value material costs. ogous stem cell transplantation (i.e., about €100,000 per year) Healthcare costs included the costs of primary healthcare were treated as being lost to follow-up from the moment they (i.e., general practice, physical therapist), secondary started using lenalidomide. From that point onwards, all of healthcare (i.e., outpatient care, hospitalization, professional their cost and effect measure values were set at missing and home care), and prescribed as well as over-the-counter medi- imputed using multiple imputation. This was done to prevent cation. Primary and secondary healthcare utilization were val- the cost estimates from being biased by the inclusion of these ued using Dutch standard costs [31]. If unavailable, prices of extremely high costs. This was considered to be an appropri- professional organizations were used. Medication use was ate strategy, as we do not expect the exercise intervention to valued using unit prices of the Royal Dutch Society of have an impact on a patient’s need for lenalidomide. Pharmacy. Missing data in the cost-effectiveness analysis were han- Informal care was defined as care provided by family and dled using multiple imputation by chained equations stratified friends, and was valued using a recommended Dutch shadow by treatment group. The imputation model included variables price [31]. that differed between patients with complete and incomplete Sports costs included the patients’ expenses on sporting data (i.e., education level, anti-cancer medication, and goods and sports memberships. smoking), those predicting the “missingness” of data (i.e., Unpaid productivity costs consisted of costs related to lost age, gender, diagnosis, sport history, and time between the hours of domestic tasks, educational activities, and volunteer start of the study and autologous stem cell transplantation), work. Unpaid productivity losses were valued using a recom- baseline effect values and all available follow-up cost and mended Dutch shadow price as well [31]. effect values. Using predictive mean matching, twenty-five Absenteeism costs related to paid work were estimated complete data sets were created in IBM SPSS resulting in a using the friction cost approach (FCA). The FCA assumes that loss of efficiency smaller than 5% (v23.0, Chicago, IL). All costs are limited to the period Dutch companies need to twenty-five datasets were analyzed separately as specified 562 J Cancer Surviv (2019) 13:558–569 below. Pooled estimates were subsequently calculated using patient of 30 (SA2). This was done because the intervention Rubin’srules[36]. was intended to last for 30 training sessions. In the third sen- For calculating ICERs, the mean difference in total costs sitivity analysis, only healthcare costs were analyzed (i.e., the between the intervention and control group was divided by the healthcare/NHS perspective was applied)(SA3). In the fourth mean difference in effects, expressed in terms of cardiorespi- sensitivity analysis (SA4), no correction was made for the use ratory fitness, grip strength, fatigue, and QALYs. Cost and of lenalidomide. effect differences were estimated using linear multilevel anal- yses, with a two-level structure: observations, transplant cen- ter. Within these analyses, cost and effect differences were Results corrected for age, gender, and education level. To deal with the highly skewed nature of cost data, joint uncertainty around Patients costs and effects was estimated using the Bias Corrected Bootstrap method with 5000 replications. To graphically illus- In total, 109 patients were randomized to the intervention (n = trate the uncertainty surrounding the ICERs, bootstrapped in- 54) or control group (n = 55). At baseline, relevant differences cremental cost-effect pairs were plotted on cost-effectiveness were found between both groups for gender and education planes. A summary measure of the joint uncertainty of costs level (Table 2). The percentage of patients with complete ef- and effects was presented using cost-effectiveness acceptabil- fect and cost data at the different measurement points can be ity curves (CEACs). CEACs indicate the probability of an found in Fig. 1. Data on the total number of training sessions intervention being cost-effective in comparison with the con- provided were complete. The main reasons for loss to follow- trol condition for a range of willingness-to-pay values (i.e., the up were relapse of (non-)Hodgkin lymphoma or progression maximum amount of money decision-makers are willing to of multiple myeloma (Fig. 1). Differences in education level, pay per unit of effect gained). Except for the multiple imputa- smoking, and time since autologous stem cell transplantation tion, analyses were performed using Stata v12. Statistical sig- were found between patients with complete and incomplete nificance was set at p <0.05 [33]. follow-up data (data not shown). To test the robustness of the results, four sensitivity analy- In total, 75% of the patients attended ≥ 80% of the training ses were performed. First, analyses were performed using the sessions and 87% of the patients attended the counseling ses- Human Capital Approach (HCA), instead of the FCA (SA1). sions (Table 3)[25]. The HCA regards each hour not worked by the patient as an hour of lost productivity. In contrast to the FCA, the HCA Long-term effectiveness: primary outcomes does therefore not truncate costs to the friction period. Second, analyses were performed in which intervention costs At T2, no statistically significant differences were found be- were calculated using a total number of training sessions per tween the intervention and control group for Table 2 Baseline demographics and clinical characteristics of the patients All (n = 109) Intervention group (n = 54) Control group (n =55) Male ([n (%)) 69 (63) 32 (59) 37 (67) Age (mean (SD)) 52 (11) 52 (11) 53 (12) Married/living together (n (%)) 91 (84) 45 (83) 46 (84) Higher education level (n (%)) 39 (36) 15 (28) 24 (44) Smoker (n (%)) 14 (13) 7 (13) 7 (13) Active (n (%)) 69 (63) 33 (61) 36 (66) Cancer type MM (n (%)) 58 (53) 29 (54) 29 (53) (N)HL (n (%)) 51 (47) 25 (46) 26 (47) Time since ASCT (mean (SD)) 86 (45) 84 (46) 88 (43) Number of co-morbidities (mean (SD)) 2 (2) 2 (2) 2 (2) VO peak (mean (SD)) in ml/kg/min 22 (5) 21 (5) 22 (6) Hand grip strength (mean (SD)) in kg 36 (11) 36 (12) 37 (10) General Fatigue (mean (SD)) (range 0–20) 13 (4) 13 (4) 14 (4) n number, SD standard deviation, MM multiple myeloma, (N)HL (non-)Hodgkin lymphoma, ASCT autologous stem cell transplantation Patients who participating in sports at least once a week before diagnoses/relapse J Cancer Surviv (2019) 13:558–569 563 Fig. 1 Flow diagram of patients Excluded (n=149): Physical (co)morbidies 38%, diseased/progressive disease 11%, second ASCT 11%, other in the EXIST study Paents screened for 39% eligibility (n=468) Non-response (n=112): No interest 30%, already following a rehabilitaon program 19%, no reason given 18%, other 33%. T0 baseline Unknown (n=98) measurement (n=111) Excluded (n=1): Paent is not able to cycle Non-response (n=1) Randomizaon (n=109) Intervenon (n=54) Usual care (n=55) Lost to follow up Lost to follow up T1 Reasons for loss to follow up: Reasons for loss to follow up: aer baseline (n=4) aer baseline (n=8) - Progression MM (n=5) -Progression MM (n=3) - Relapse (N)HL (n=4) - Relapse (N)HL (n=4) Lost to follow up Lost to follow up - Bad health situaon (n=3) - Not agree with allocaon (n=2) T2 aer baseline (n=19) aer baseline (n=16) -Other (n=7) -Other (n=7) Complete data Complete data Effect data; n (%) Effect data; n (%) T0 45 (83.3%) T0 42 (76.4%) T1 47 (87.0%) T1 45 (81.8%) T2 28 (51.9%) T2 31 (56.4%) Cost data; n(%) Cost data; n(%) CQ 1 41 (75.9%) CQ 1 33 (60.0%) CQ 2 39 (72.2%) CQ 2 35 (63.6%) CQ 3 39 (72.2%) CQ 3 42 (76.4%) CQ 4 37 (68.5%) CQ 4 36 (65.5%) CQ 5 39 (72.2%) CQ 5 33 (60.0%) CQ 6 34 (63.0%) CQ 6 28 (50.9%) CQ 7 37 (68.5%) CQ 7 29 (52.7%) CQ 8 29 (53.7%) CQ 8 26 (47.3%) CQ 9 28 (51.9%) CQ 9 28 (50.9%) Mulple imputaons Mulple imputaons Imputed dataset Imputed dataset n=54; 100,0% n=55; 100,0% Analysis cardiorespiratory fitness (0.12 ml/kg/min; 95%CI − 1.89; the exercise intervention was on average more costly and 2.14, p = 0.90), handgrip strength (−0.08 kg; 95%CI − more effective than usual care (Table 3;Fig. 2(1a)). The 2.47; 2.31, p = 0.95), and general fatigue (−0.69 points; CEAC in Fig. 2(2a) indicates that the exercise intervention’s 95%CI − 2.52; 1.14, p =0.46). probability of being cost-effective compared with usual care was 0.39 at a willingness-to-pay of €0 per ml/kg/min, increas- Costs ing to a maximum of 0.73 at a ceiling ratio of €20,000 per ml/ kg/min. On average, intervention costs were €1340 per patient The ICER for handgrip strength was − 427, indicating that (Table 4). During follow-up, secondary healthcare costs were the intervention was dominated by usual care (i.e., more costly statistically significantly higher in the intervention group than and less effective)(Table 3; Fig. 2(1b)). The CEAC in Fig. in the control group (Table 5). Unpaid lost productivity costs 2(2b) indicates that the exercise intervention’s probability of and informal care costs were statistically significantly lower in being cost-effective compared with usual care was 0.39 at a the intervention group than in the control group (Table 5). willingness-to-pay of €0 per kg. At increasing levels of will- Total societal costs were higher in the intervention group than ingness-to-pay, this probability decreased. in the control group, but this difference was not statistically The ICER for general fatigue was − 279. This suggests that significant (€529; 95%CI − 3205 to 4452). a 1-point reduction in general fatigue in the intervention group was associated with a societal cost of €273 compared with the Societal perspective: cost-effectiveness control group (Table 3; Fig. 2(1c)). Please note that a reduc- tion in general fatigue indicates an improved health effect. The The ICER for cardiorespiratory fitness was 972, suggesting CEAC in Fig. 2(2c) indicates that the exercise intervention’s that an increase of 1 ml/kg/min in VO peak in the intervention probability of being cost-effective compared with usual care group was associated with a societal cost of €972 compared was 0.39 at a willingness-to-pay of €0 per point improvement, with the control group. The majority of cost-effect pairs were increasing to a maximum of 0.96 at a willingness-to-pay of located in the northeast quadrant (i.e., 44%), indicating that €31,000 per point improvement. 564 J Cancer Surviv (2019) 13:558–569 Table 3 Difference in pooled mean costs and effects (95% confidence intervals), incremental cost-effectiveness ratios, and the distribution of cost-effect pairs around the quadrants of the cost- effectiveness planes Sample size Outcome ΔC (95% CI) € ΔE (95% CI) points ICER €/point Distribution CE-plane (%) 1 2 3 4 IC NE SE SW NW Main Analyses 54 55 Cardiorespiratory fitness 529 (− 3205 to 4452) 0.54 (− 1.14 to 2.23) 972 43.7 30.9 11.5 13.9 Hand grip strength 529 (− 3205 to 4452) − 1.24 (− 4.10 to 1.59) − 427 10.8 8.9 33.5 46.8 General Fatigue 529 (− 3205 to 4452) − 1.78 (− 3.94 to − 0.37) − 297 53.3 40.1 37.1 4.4 HRQoL - QALY 529 (− 3205 to 4452) − 0.07 (− 0.17 to 0.04) − 8043 4.6 5.2 37.1 53.1 SA1 Human Capital Approach 54 55 Cardiorespiratory fitness 1152 (− 5585 to 7936) 0.54 (− 1.14 to 2.23) 2117 46.2 28.4 9.8 15.5 Hand grip strength 1152 (− 5585 to 7936) − 1.24 (− 4.10 to 1.59) − 931 10.9 8.8 29.5 50.8 General Fatigue 1152 (− 5585 to 7936) − 1.78 (− 3.94 to − 0.37) − 646 57.4 35.9 2.3 29.5 HRQoL - QALY 1152 (− 5585 to 7936) − 0.07 (− 0.17 to 0.04) − 17,515 5.4 4.4 33.8 56.4 SA2 Maximal training sessions 54 55 Cardiorespiratory fitness 843 (− 2871 to 4779) 0.54 (− 1.14 to 2.23) 1549 48.3 26.4 9.9 15.4 Hand grip strength 843 (− 2871 to 4779) − 1.24 (− 4.10 to 1.59) − 681 12.0 7.7 28.6 51.7 General Fatigue 843 (− 2871 to 4779) − 1.78 (− 3.94 to − 0.37) − 473 59.0 34.4 1.9 4.7 HRQoL - QALY 843 (− 2871 to 4779) − 0.07 (− 0.17 to 0.04) − 12,817 5.2 4.6 31.7 58.5 SA3 Medical costs only 54 55 Cardiorespiratory fitness 1096 (− 114 to 2220) 0.54 (− 1.14 to 2.23) 2016 70.7 4.0 0.8 24.6 Hand grip strength 1096 (− 114 to 2220) − 1.24 (− 4.10 to 1.59) − 886 18.5 1.2 3.5 76.8 General Fatigue 1096 (− 114 to 2220) − 1.78 (− 3.94 to − 0.37) − 615 88.8 4.6 1.3 6.5 HRQoL - QALY 1096 (− 114 to 2220) − 0.07 (− 0.17 to 0.04) − 16,676 8.6 1.2 3.5 86.7 SA4 No correction for lenalidomide use 54 55 Cardiorespiratory fitness 15,646 (7688 to 30,845) 0.97 (− 0.74 to 2.69) 16,135 84.3 1.1 0.0 14.6 Hand grip strength 15,646 (7688 to 30,845) − 0.22 (− 2.96 to 2.51) − 69,892 41.4 0.7 0.4 57.4 General Fatigue 15,646 (7688 to 30,845) − 0.94 (− 3.12 to 1.24) − 12,667 80.0 1.0 0.1 18.9 HRQoL - QALY 15,646 (7688 to 30,845) − 0.07 (− 0.17 to 0.04) − 242,963 10.6 0.1 1.0 88.2 I intervention group, C control group, C costs, E effects, ICER incremental cost-effectiveness ratio, CE-plane cost-effectiveness plane, SA sensitivity analyses, QALY quality-adjusted life years Costs are expressed in 2014 Euros 1Refers to the northeast quadrant of the CE plane, suggesting that the EXIST exercise intervention is more effective and more costly than usual practice 2Refers to the southeast quadrant of the CE plane, suggesting that the EXIST exercise intervention is more effective and less costly than usual practice 3Refers to the northwest quadrant of the CE plane, suggesting that the EXIST exercise intervention is less effective and more costly than usual practice 4Refers to the southwest quadrant of the CE plane, suggesting that the EXIST exercise intervention is less effective and less costly than usual practice J Cancer Surviv (2019) 13:558–569 565 Table 4 Costs of the EXIST intervention per patient Staff Units Unit prices Total costs Total costs per patient Kick-off physiotherapy (42 clinics) Information packet 42 packets €1.97/packet €82.59 €1.53 Information leaflet 42 leaflets €0.83/leaflet €34.78 €0.64 Time investment Physical therapist 21 h €48.81/h €1025.02 €18.98 Project assistant 21 h €32.54/h €683.35 €12.65 Traveling expenses 4451.2 km €0.19/km €858.63 €15.90 Physiotherapy Training sessions Physical therapist 1356 h €48.81/h €66,171.60 €1225.40 Counseling sessions Physical therapist 63.25 h €48.81/h €3087.27 €57.17 Consultation sports physician/intake Sports physician 1 h €91.91/h €91.91 €1.70 Physical therapist 1 h €48.81/h €48.81 €0.90 Information booklet 54 booklets €3.90/leaflet €210.60 €3.90 Registration logbook 54 logbooks €1.16/logbook €62.64 €1.13 Total €1339.92 km kilometer Costs are expressed in 2014 euros The ICER for HRQoL was − 8043, indicating that the in- important difference being an almost 30-fold increase in the tervention was dominated by usual care for QALYs (i.e., more total societal cost difference (from €529 to €15,646). costly and less effective)(Table 3;Fig. 2(1d)). The CEAC in Fig. 2(2d) indicates that if decision-makers are not willing to pay anything for a QALY gained, the probability of the inter- Discussion vention being cost-effectiveness was 0.39. At increasing levels of willingness-to-pay, this probability decreased. Evaluations of the long-term effectiveness and cost- effectiveness of an 18-week supervised high-intensity com- Sensitivity analyses bined resistance and interval exercise intervention versus usu- al care in patients after autologous stem cell transplantation In SA1, SA2, and SA3, the total cost difference became larger presented in this article failed to show statistically significant compared with the main analysis, but remained in favor of the benefits on physical fitness and general fatigue at long-term, control group. In SA4, cost and effect difference estimates and low probabilities of cost-effectiveness at reasonable differed from those of the main analysis, with the most values of willingness-to-pay. To illustrate the latter, at the Table 5 Mean costs per patient in the intervention and control group, and mean cost differences between both groups during follow-up Cost category Intervention group n =54; Control group n =55; Crude cost differences Adjusted cost differences mean (SEM) mean (SEM) mean (95% CI) mean (95% CI) Primary healthcare costs 1437 (225) 1955 (282) − 518 (− 1280 to 154) − 512 (− 1215 to 78) Secondary healthcare costs 2338 (280) 1845 (198) 493 (−97 to 1235) 501 (97 to 1160) Medication costs 1002 (135) 1271 (250) − 269 (−908to206) − 237 (−884 to 246) Unpaid productivity costs 560 (106) 1480 (276) − 920 (− 1571 to 396) − 884 (− 1437 to − 421) Informal care costs 432 (122) 1148 (244) − 716 (− 1298 to − 227) − 669 (− 1159 to − 193) Absenteeism costs 16,818 (1277) 15,823 (1294) 995 (− 2510 to 4550) 1160 (− 2114 to 4585) Sport costs 470 (51) 596 (77) − 126 (−313to42) − 111 (−292to57) Intervention costs 1340 (47) 0 (0) 1340 (1249 to 1432) 1344 (1243 to 1423) Total costs 24,397 (1322) 24,119 (1430) − 278 (− 4105 to 3548) 529 (− 3205 to 4452) n number, SEM standard error of the mean, CI confidence interval, SD standard deviation Costs are expressed in 2014 Euros 566 J Cancer Surviv (2019) 13:558–569 (1a) (2a) (1b) (2b) (2c) (1c) (1d) (2d) Fig. 2 Cost-effectiveness planes indicating the uncertainty around the (€) of willingness-to-pay per unit of effect gained (2) for cardiorespiratory incremental cost-effectiveness ratios (1) and cost-effective acceptability fitness (a), handgrip strength (b), general fatigue (c), and health-related curves indicating the probability of cost-effectiveness for different values quality of life (d) lower and upper bound of the informal Dutch willingness-to- respectively), the probability of the intervention being cost- pay threshold (i.e., 10,000 and 80,000 €/QALY gained, effective compared with usual care was low (< 0.29). J Cancer Surviv (2019) 13:558–569 567 The lack of significant between-group differences in physi- considered cost-effective depending on the decision-makers’ cal fitness and fatigue at long-term is in line with previous willingness-to-pay. studies in patients with breast cancer at 18 weeks [32]or Strengths of the study are the focus on the long-term (1 year 6 months follow-up [34]. However, a previous study among after intervention completion) effectiveness and cost- patients undergoing autologous or allogeneic stem cell trans- effectiveness specifically in patients after autologous stem cell plantation including a 3-month follow-up showed benefits on transplantation, the multicenter randomized controlled trial cardiorespiratory fitness, assessed with a submaximal exercise design, the use of a broad set of valid and reliable outcome test and knee extension strength, but not on handgrip and fa- measures, and the use of state-of-the-art statistical methods for tigue [13]. The lack of significant intervention effects at the evaluating the intervention’s cost-effectiveness (i.e., the use of long-term indicates that delayed effects of the intervention did seemingly unrelated regression, bootstrapping, and multiple not occur, and were comparable with the findings at short-term imputation). Limitations of the study include the use of self- [23]. This may be related to the suboptimal timing of interven- report for collecting cost data which may have introduced tion delivery, as exercise post-stem cell transplantation may not recall bias, the potential problems with generalizability to oth- be able to speed up natural recovery [23, 25]. A recent meta- er countries with different healthcare systems and/or payment analysis strengthens this hypothesis by reporting significant structures [6], and the large amount of missing cost data. effects of exercise on fatigue and muscle strength pre-stem cell However, missing effect and cost data were taken into account transplantation but not post-stem cell transplantation [16]. by the linear mixed model analyses and multiple imputation, Other reasons for the lack of significance may be related to respectively. Also, during follow-up, more intervention group suboptimal exercise compliance in the intervention group or patients used the extremely expensive drug lenalidomide than contamination in the control group [23], but no evidence for control group patients. As we did not expect the exercise in- this hypothesis was found in a post hoc analysis [23] and pro- tervention to have an impact on a patient’s need for cess evaluation [25]. Finally, the lack of long-term effects also lenalidomide, we dealt with this issue by treating patients as suggests that the counseling sessions alongside, and after, the being lost to follow-up from the moment they started using supervised sessions were insufficient to change daily physical lenalidomide, that is, as if their measurements were missing. activity behavior. Due to practical reasons, physical therapists The long-term effectiveness estimates slightly differ from implemented the counseling sessions differently: some provid- those of the cost-effectiveness analyses. This is due to differ- ed the counseling during the exercise sessions, while others ences in the applied statistical methods, including (1) the use made separate appointments [25]. Additionally, most were of mixed effect models in the effectiveness analyses, which scheduled alongside the program to promote compliance and are typically not applied in cost-effectiveness analyses, (2) it might be that the three sessions focused on promoting phys- multiple imputation, which is recommended for imputing ical activity in daily life were insufficient [21]. missing cost data [26] versus maximum likelihood estimation, The finding that the supervised exercise intervention was which is often used in effectiveness analyses, and (3) a cor- not cost-effective versus usual care is probably due to the rection for the possible correlation between costs and effects intervention not being effective at the short-term, nor at the in the cost-effectiveness analyses [37]. long-term. The lack of cost-effectiveness is also in line with The average cost of the 18-week exercise intervention of the results of an economic evaluation of a combined physical €1340 per patient on average was somewhat higher than the activity and dietary intervention in patients after autologous average cost of €858 per patient of the aforementioned 12- stem cell transplantation [10]. This study, however, used week high-intensity exercise intervention offered to patients modeling techniques, rather than patient-level data. Patient- with various types of cancer [12]. This amount is still relative- level studies evaluating the cost-effectiveness of exercise in- ly low as compared with the anti-cancer treatment costs. terventions for patients with solid tumors provided mixed re- Therefore, exercise interventions, if effective, may have the sults [1, 10, 12, 17, 19, 35]. Kampshoff et al. [12] found a 12- potential to be included as part of standard care for patients week high-intensity exercise intervention to be cost-effective with cancer. However, in its current form, and offered to pa- compared with a moderate-intensity exercise intervention in a tients recently treated with autologous stem cell transplanta- group of cancer survivors with solid tumors (mostly breast tion, the exercise intervention is not effective at the short-term cancer), whereas May et al. [17] found an 18-week exercise and long-term compared with usual care, nor cost-effective, program to be cost-effective versus usual care for colon can- and therefore, we do not recommend it to offer it as part of cer, but not for breast cancer. Van Waart et al. [34]found a standard care. Future studies should examine the (long-term) low-intensity physical activity program not to be cost- effectiveness and cost-effectiveness of other types of exercise effective compared with usual care among breast cancer pa- interventions and may concentrate on exercise interventions at tients undergoing chemotherapy, but found that a combined different time points in the treatment trajectory of patients resistance and aerobic supervised exercise program may be scheduled to receive stem cell transplantation. 568 J Cancer Surviv (2019) 13:558–569 program: the advantage of the steep ramp test. Arch Phys Med In conclusion, we found no significant beneficial effects of Rehabil. 2007;88:610–6. the high-intensity combined resistance and interval exercise 6. Drummond M, Barbieri M, Cook J, Glick HA, Lis J, Malik F, et al. program on physical fitness and fatigue at long-term when Transferability of economic evaluations across jurisdictions: compared with usual care, nor was the intervention cost- ISPOR good research practices task force report. Value Health. 2009;12:409–18. effective from a societal perspective. 7. Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care Acknowledgments The authors acknowledge the A-CaRe Clinical programmes: Oxford university press; 2015. Research group and thank all patients who participated in the study and 8. EuroQol G. EuroQol–a new facility for the measurement of health- all hematologists, nurses, data managers, and physiotherapists who con- related quality of life. Health policy (Amsterdam, Netherlands). tributed to the study. We are also grateful to Karen van der Weiden, Julie 1990;16:199. Heeren, Griet van der Slagmolen, Michiel Greidanus, the team of the 9. Fuller JT, Hartland MC, Maloney LT, Davison K. Therapeutic ef- Sports Medical Advice Center Rotterdam, and the members of the fects of aerobic and resistance exercises for cancer survivors: a Clinical Trial Office, Department of Hematology Amsterdam UMC, lo- systematic review of meta-analyses of clinical trials. Br J Sports cation AMC for their assistance with patient recruitment, data collection, Med. 2018:bjsports–2017-098285. and/or data analysis. 10. Gao L, Moodie M, Brown V, Avery S. Cost-effectiveness of a lifestyle modification program in long-term survivors of hemopoi- Funding This study was funded by the Dutch Cancer Society (grant etic stem cell transplantation. Clin Transpl. 2017;31:e13049. number: ALPE 2009-4620). 11. Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the Compliance with ethical standards American Heart Association. Circulation. 2007;116:1081–93. 12. Kampshoff C, van Dongen J, van Mechelen W, Schep G, The study procedures were approved by the Medical Ethics Committee of Vreugdenhil A, Twisk J, et al. Long-term effectiveness and cost- the Academic Medical Center (METC AMC 10/106) and by the boards of effectiveness of high versus low-to-moderate intensity resistance the Antoni van Leeuwenhoek Hospital (Amsterdam), St. Antonius and endurance exercise interventions among cancer survivors. J Hospital (Nieuwegein), Haga Teaching Hospital (Den Haag), Cancer Surviv. 2018;12:417–29. University Medical Center (Utrecht), Isala (Zwolle), Erasmus MC/ 13. Knols R, de Bruin ED, Uebelhart D, Aufdemkampe G, Schanz U, Daniel den Hoed (Rotterdam), VU University Medical Center Stenner-Liewen F, et al. Effects of an outpatient physical exercise (Amsterdam), and Leiden University Medical Center (Leiden). program on hematopoietic stem-cell transplantation recipients: a randomized clinical trial. Bone Marrow Transplant. 2011;46: Conflict of interest The authors declare that they have no conflict of 1245–55. interest. 14. Lamers L, Stalmeier P, McDonnell J, Krabbe P. Measuring the quality of life in economic evaluations: the Dutch EQ-5D tariff. Open Access This article is distributed under the terms of the Creative Ned Tijdschr Geneeskd. 2005;149:1574–8. Commons Attribution 4.0 International License (http:// 15. Levinger I, Goodman C, Hare DL, Jerums G, Toia D, Selig S. The creativecommons.org/licenses/by/4.0/), which permits unrestricted use, reliability of the 1RM strength test for untrained middle-aged indi- distribution, and reproduction in any medium, provided you give appro- viduals. J Sci Med Sport. 2009;12:310–6. priate credit to the original author(s) and the source, provide a link to the 16. Liang Y, Zhou M, Wang F, Wu Z. Exercise for physical fitness, Creative Commons license, and indicate if changes were made. fatigue and quality of life of patients undergoing hematopoietic stem cell transplantation: a meta-analysis of randomized controlled trials. Jpn J Clin Oncol. 2018;48:1046–57. 17. May AM, Bosch MJ, Velthuis MJ, Van Der Wall E, Bisschop CNS, Los M, et al. Cost-effectiveness analysis of an 18-week exercise programme for patients with breast and colon cancer undergoing References adjuvant chemotherapy: the randomised PACT study. BMJ Open. 2017;7:e012187. 1. Braam KI, van Dijk-Lokkart E, van Dongen J, van Litsenburg R, 18. Mewes JC, Steuten LM, IJzerman MJ, Van Harten WH. Takken T, Huisman J, et al. Cost-effectiveness of a combined phys- Effectiveness of multidimensional cancer survivor rehabilitation ical exercise and psychosocial training intervention for children and cost-effectiveness of cancer rehabilitation in general: a system- with cancer: results from the quality of life in motion study. Eur J atic review. The oncologist. 2012:theoncologist. 2012-0151, 2012. Cancer Care. 2017;26:e12586. 19. Mewes JC, Steuten LM, Duijts SF, Oldenburg HS, van Beurden M, 2. Buffart L, Galvão DA, Brug J, Chinapaw M, Newton RU. Stuiver MM, et al. Cost-effectiveness of cognitive behavioral ther- Evidence-based physical activity guidelines for cancer survivors: apy and physical exercise for alleviating treatment-induced meno- current guidelines, knowledge gaps and future research directions. pausal symptoms in breast cancer patients. J Cancer Surviv. 2015;9: Cancer Treat Rev. 2014;40:327–40. 126–35. 3. Buffart LM, Kalter J, Sweegers MG, Courneya KS, Newton RU, 20. Oberoi S, Robinson PD, Cataudella D, Culos-Reed SN, Davis H, Aaronson NK, et al. Effects and moderators of exercise on quality Duong N, et al. Physical activity reduces fatigue in patients with of life and physical function in patients with cancer: an individual cancer and hematopoietic stem cell transplant recipients: a system- patient data meta-analysis of 34 RCTs. Cancer Treat Rev. 2017;52: atic review and meta-analysis of randomized trials. Crit Rev Oncol 91–104. Hematol. 2018;122:52–9. 4. Cramp F, Daniel J. Exercise for the management of cancer-related 21. Persoon S, Kersten MJ, ChinAPaw MJ, Buffart LM, Burghout H, fatigue in adults. Cochrane database of systematic reviews. 2008. Schep G, et al. Design of the EXercise Intervention after Stem cell Transplantation (EXIST) study: a randomized controlled trial to 5. De Backer IC, Schep G, Hoogeveen A, Vreugdenhil G, Kester AD, van Breda E. Exercise testing and training in a cancer rehabilitation evaluate the effectiveness and cost-effectiveness of an J Cancer Surviv (2019) 13:558–569 569 individualized high intensity physical exercise program on fitness analysis of randomised controlled trials. Br J Sports Med. 2018;52:505–13. and fatigue in patients with multiple myeloma or (non-) Hodgkin's lymphoma treated with high dose chemotherapy and autologous 31. Tan SS, Bouwmans-Frijters CA, Hakkaart-van Roijen L. stem cell transplantation. BMC Cancer. 2010;10:671. Handleiding voor kostenonderzoek: methoden en referentieprijzen 22. Persoon S, Kersten MJ, van der Weiden K, Buffart LM, Nollet F, voor economische evaluaties in de gezondheidszorg. Tijdschrift Brug J, et al. Effects of exercise in patients treated with stem cell voor gezondheidswetenschappen. 2012;90:367–72. transplantation for a hematologic malignancy: a systematic review 32. Travier N, Velthuis MJ, Bisschop CNS, van den Buijs B, and meta-analysis. Cancer Treat Rev. 2013;39:682–90. Monninkhof EM, Backx F, et al. Effects of an 18-week exercise 23. Persoon S, ChinAPaw MJ, Buffart LM, Liu RD, Wijermans P, programme started early during breast cancer treatment: a Koene HR, et al. Randomized controlled trial on the effects of a randomised controlled trial. BMC Med. 2015;13:121. supervised high intensity exercise program in patients with a hema- 33. Van Der Meer EW, van Dongen JM, Boot CR, van der Gulden JW, tologic malignancy treated with autologous stem cell transplanta- Bosmans JE, Anema JR. Economic evaluation of a multifaceted tion: results from the EXIST study. PLoS One. 2017;12:e0181313. implementation strategy for the prevention of hand eczema among 24. Persoon S, Kersten MJ, Buffart LM, Vander Slagmolen G, Baars healthcare workers in comparison with a control group: the JW, Visser O, et al. Health-related physical fitness in patients with Hands4U study. Acta Derm Venereol. 2016;96:499–504. multiple myeloma or lymphoma recently treated with autologous 34. van Waart H, Stuiver MM, van Harten WH, Geleijn E, Kieffer JM, stem cell transplantation. J Sci Med Sport. 2017;20:116–22. Buffart LM, et al. Effect of low-intensity physical activity and 25. Persoon S, Chinapaw M, Buffart L, Brug J, Kersten M, Nollet F. moderate-to high-intensity physical exercise during adjuvant che- Lessons learnt from a process evaluation of an exercise intervention motherapy on physical fitness, fatigue, and chemotherapy comple- in patients treated with autologous stem cell transplantation. Eur J tion rates: results of the PACES randomized clinical trial. J Clin Cancer Care. 2018;27:e12779. Oncol. 2015;33:1918–27. 26. Ramsey SD, Willke RJ, Glick H, Reed SD, Augustovski F, Jonsson 35. van Waart H, van Dongen JM, van Harten WH, Stuiver MM, B, et al. Cost-effectiveness analysis alongside clinical trials II-an Huijsmans R, Hellendoorn-van Vreeswijk JA, et al. Cost–utility ISPOR good research practices task force report. Value Health. and cost-effectiveness of physical exercise during adjuvant chemo- 2015;18(2):161–72. therapy. Eur J Health Econ. 2018;19:893–904. 27. Roseboom KJ, van Dongen JM, Tompa E, van Tulder MW, 36. White IR, Royston P, Wood AM. Multiple imputation using Bosmans JE. Economic evaluations of health technologies in chained equations: issues and guidance for practice. Stat Med. Dutch healthcare decision-making: a qualitative study of the current 2011;30:377–99. and potential use, barriers, and facilitators. BMC Health Serv Res. 37. Willan AR, Briggs AH, Hoch JS. Regression methods for covariate 2017;17:89. adjustment and subgroup analysis for non-censored cost-effective- 28. Segal R, Zwaal C, Green E, Tomasone J, Loblaw A, Petrella T. ness data. Health Econ. 2004;13(5):461–75. Exercise for people with cancer: a systematic review. Curr Oncol. 2017;24:e290. 29. Smets E, Garssen B, Bonke Bd, De Haes J. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 1995;39:315–325. 30. Sweegers MG, Altenburg TM, Chinapaw MJ, Kalter J, Verdonck- Publisher’snote Springer Nature remains neutral with regard to de Leeuw IM, Courneya KS, et al. Which exercise prescriptions jurisdictional claims in published maps and institutional affiliations. improve quality of life and physical function in patients with cancer during and following treatment? A systematic review and meta- http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Cancer Survivorship: Research and Practice Springer Journals

Long-term effectiveness and cost-effectiveness of an 18-week supervised exercise program in patients treated with autologous stem cell transplantation: results from the EXIST study

Download PDF
12 pages

Loading next page...
 
/lp/springer-journals/long-term-effectiveness-and-cost-effectiveness-of-an-18-week-9ETwSM0kib

References (43)

Publisher
Springer Journals
Copyright
Copyright © 2019 by The Author(s)
Subject
Medicine & Public Health; Public Health; Oncology; Health Promotion and Disease Prevention; Health Informatics; Quality of Life Research; Primary Care Medicine
ISSN
1932-2259
eISSN
1932-2267
DOI
10.1007/s11764-019-00775-9
Publisher site
See Article on Publisher Site

Abstract

Purpose To evaluate the long-term effectiveness and cost-effectiveness of a supervised 18-week high-intensity exercise program compared with usual care in patients treated with autologous stem cell transplantation. Methods One hundred nine patients were randomly assigned to the exercise intervention (n = 54) or the usual care control group (n = 55). Data on cardiorespiratory fitness (VO peak), handgrip strength, general fatigue, and health-related quality of life (quality-adjusted life years [QALYs]) were collected at baseline (T0), after completion of the exercise intervention or at a similar time point in the control group (T1) and 12 months later (T2). Cost questionnaires were used to assess societal costs. Long-term effectiveness (at T2) was evaluated using linear mixed model analyses. For the economic evaluation, missing data were imputed using multiple imputation, and data were analyzed using linear mixed models. Results At T2, no statistically significant differences were found between the intervention and control group for VO peak (0.12; 95%CI − 1.89; 2.14 ml/min/kg), handgrip strength (− 1.08; 95%CI− 2.47; 2.31), and general fatigue (− 0.69; 95%CI − 2.52; 1.14). During 12-months follow-up, no significant between-group differences in QALYs and societal costs were found (QALYs − 0.07; 95%CI − 0.17; 0.04; costs 529; 95%CI − 3205;4452). Intervention costs were €1340 per patient. For all outcomes, the probability of the intervention being cost-effective was low at reasonable values of willingness-to-pay. Conclusion We found no evidence for the exercise intervention being effective on physical fitness and fatigue, nor cost-effective from a societal perspective. Trial registration The study was prospectively registered on 27 May 2010 at the Netherlands Trial Register (NTR2341). Implications for Cancer Survivors The current exercise intervention should not be recommended to patients recently treated with autologous stem cell transplantation. . . . . Keywords Long-term effectiveness Cost-effectiveness Exercise intervention Multiple myeloma Lymphoma * Laurien M. Buffart Amsterdam UMC, University of Amsterdam, Department of l.buffart@vumc.nl Haematology, Amsterdam, The Netherlands Lymphoma and Myeloma Center Amsterdam (LYMMCARE) and Vrije Universiteit Amsterdam, Department of Health Sciences, Cancer Center Amsterdam, Amsterdam, The Netherlands Faculty of Science, Amsterdam Public Health Research Institute, National Institute for Public Health and the Environment, Amsterdam, The Netherlands Bilthoven, The Netherlands Amsterdam UMC, University of Amsterdam, Department of 7 Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Rehabilitation, Amsterdam Movement Sciences, Public and Occupational Health, Amsterdam Public Health Research Amsterdam, The Netherlands Institute, Amsterdam, The Netherlands 3 8 Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Epidemiology and Biostatistics, Amsterdam Public Health Research Medical Oncology, Cancer Center Amsterdam, Institute, De Boelelaan 1089a, 1018HVAmsterdam, The Netherlands Amsterdam, The Netherlands J Cancer Surviv (2019) 13:558–569 559 Introduction for cost-effectiveness analyses, it is important that the time frame of analyses extends beyond the intervention period to Cancer and cancer treatment may have a negative impact on ensure that the most important costs and consequences physical fitness, fatigue, and quality of life, which may be flowing from the intervention are covered [7]. This is because counteracted by exercise interventions. Several systematic re- patients with relatively large health effects directly after the views reported favorable effects of exercise interventions on intervention typically have relatively low levels of healthcare cardiorespiratory fitness [4, 9, 28], muscle strength [9, 28], consumption and productivity losses during the consecutive fatigue [4, 9], and health-related quality of life (HRQoL) [3, period and vice versa. To cover this research gap, the present 30] in patients with cancer during and following cancer treat- study aimed to evaluate the long-term effectiveness and cost- ment. In various countries, this has resulted in the develop- effectiveness of the exercise intervention versus usual care for ment of exercise guidelines for patients with cancer [2]. physical fitness, fatigue, and HRQoL in patients treated with Beneficial effects on these outcomes were also found in pa- autologous stem cell transplantation. tients who received a stem cell transplantation [20, 22]. A recent study has shown that the effects of exercise inter- ventions following cancer treatment for solid tumors on phys- Methods ical fitness and HRQoL can be sustained at 1 year, while the intervention effects on fatigue disappeared [12]. To the best of The EXIST study was a multicenter randomized controlled our knowledge, effects of exercise interventions in patients trial that evaluated the effectiveness of an 18-week supervised who received a stem cell transplantation on these outcomes high-intensity combined resistance and interval exercise pro- at long-term (≥ 1 year) are unknown. gram compared with usual care on physical fitness and fatigue Information on the cost-effectiveness of exercise interven- as primary outcomes as well as its cost-effectiveness in pa- tions is important for decision-makers, as this provides insight tients with a hematologic malignancy treated with autologous into the additional cost per unit of effect gained, and may thus stem cell transplantation. The study procedures were approved provide guidance to decision-makers when deciding how to by the Medical Ethics Committee of the Academic Medical allocate scarce resources in healthcare [7, 27]. Nonetheless, Center (METC AMC 10/106) and by the boards of the Antoni there is only limited evidence available on the cost- van Leeuwenhoek Hospital (Amsterdam), St. Antonius effectiveness of exercise interventions in patients with cancer. Hospital (Nieuwegein), Haga Teaching Hospital (Den A previous systematic review on this topic [18] included only Haag), University Medical Center (Utrecht), Isala (Zwolle), three studies. Since then, a small number of additional studies Erasmus MC/Daniel den Hoed (Rotterdam), VU University have evaluated the cost-effectiveness of exercise interventions Medical Center (Amsterdam), and Leiden University Medical during or after cancer treatment [1, 10, 12, 17, 19, 35], but Center (Leiden). The study protocol as well as the baseline they were heterogeneous in the type of cancer, type of inter- and short-term results have been published previously [21, 23, vention and control condition, study results, and/or they were 24]. not based on patient-level data. Given the importance of this Patients were recruited between March 2011 and February topic and the scarcity of literature, more research on the cost- 2014. Patients were eligible for the study if they were (1) effectiveness of exercise interventions in patients with cancer treated with autologous stem cell transplantation for multiple is warranted. myeloma or lymphoma 6–14 weeks earlier or treated with Recently, we have published the results of the EXercise autologous stem cell transplantation and finished successive Intervention after Stem cell Transplantation (EXIST) random- consolidation chemotherapy or radiotherapy 2–6 week earlier; ized controlled trial [23]. Within this study, an 18-week super- (2) recovered from treatment (Hb > 10.5 g/dL, platelets > vised high-intensity combined resistance and interval exercise 80 × 10 /L); and (3) able to undergo exercise testing and to program was compared with usual care in patients with a participate in an exercise intervention. All patients provided hematologic malignancy treated with autologous stem cell written informed consent. transplantation. At the short-term (directly after the interven- After the baseline assessments and confirmation of eligibil- tion), the exercise intervention did not have beneficial effects ity, patients were randomized to the intervention or usual care on physical fitness, fatigue, and HRQoL [23]. Some may ar- group by an independent data manager using a validated soft- gue that the exercise intervention will therefore not be effec- ware program (TENALEA Clinical Trial Data Management tive, nor cost-effective at the long-term, and that publication of System; Netherlands Cancer Institute, Amsterdam, the these results is therefore not necessary. However, the EXIST Netherlands). Randomization was concealed, stratified for exercise intervention also included counseling sessions transplant center and diagnosis using block randomization aiming to promote compliance and maintenance of exercise with block sizes varying randomly between 2, 4, and 6. The after the program [21, 23, 24] and it is currently unclear study personnel that assessed long-term outcomes and per- whether the intervention has delayed effects. Additionally, formed the analyses were not blinded for treatment allocation. 560 J Cancer Surviv (2019) 13:558–569 Exercise intervention and usual care between T1 and T2. The physical therapists kept a training log during the intervention program. The 18-week exercise intervention was supervised by instructed physical therapists and took place at local physical therapy practices. The program consisted of 30 exercise and 6 Effect measures counseling sessions. Each exercise session lasted approxi- mately 60 min and included 6 resistance exercises targeting The primary effect measures were physical fitness, including the large muscle groups (vertical row, leg press, bench/chest cardiorespiratory fitness and handgrip strength, and general press, and pull over/flies and two additional exercises for the fatigue. abdominal muscles and the upper legs) and 2 bouts of 8 min Cardiorespiratory fitness was expressed as the highest con- cycling interval training (Table 1). The indirect one repetition tinuous 15 s interval values for oxygen uptake (VO in ml/ 2peak maximum (1-RM) test [15] and the steep ramp test [5]were kg/min; MasterScreen CPX, CareFusion, Hoechberg, performed every 4 weeks in order to tailor and adjust the Germany) measured during a cardiopulmonary exercise test exercise intensity prescription. Counseling sessions lasted 5 performed on a cycle ergometer (Lode Excalibur, Groningen, to 15 min each and took place in week 1, 4, 10, 12, 18, and 22 the Netherlands). [21]. The counseling sessions were provided by the physical Handgrip strength of the dominant hand was assessed therapist who supervised the exercise intervention and aimed using a grip strength dynamometer (Hydraulic Hand to improve compliance to the exercise intervention and to Dynamometer, North Coast Medical Inc., Morgan Hill, encourage patients to pursue a physically active lifestyle dur- USA), and the highest score out of three attempts in kilogram ing and following the program [21]. From week 12 onwards, was used in analyses. patients were encouraged to meet physical activity levels as General fatigue was determined using the subscale of recommended by guidelines [11]. Specific program elements Multidimensional Fatigue Inventory (MFI) questionnaire included the provision of general and motivational informa- [29]. The patients could score between 4 and 20, with higher tion, both verbally and via folders, about the desired frequen- scores indicating more fatigue. cy, duration, and intensity of physical activity [21]. For the economic evaluation, HRQoL was assessed using Usual care varied according to patients’ and physicians’ the EQ-5D-3L [8]. The EQ-5D-3L was administered at T0, preferences. Control group patients were not restricted in their T1, T2, and in an additional cost diary half-way between T1 physical activities or in their use of healthcare services. and T2. The EQ-5D-3L consists of five questions evaluating the following health dimensions: mobility, self-care, usual ac- Timing of assessments tivities, pain/discomfort, and anxiety/depression. The patients’ EQ-5D-3L health states were transformed into utility scores The effect measures were assessed at baseline (T0), after com- using the Dutch tariff [14]. QALYs were calculated by multi- pletion of the intervention or at a similar time point in the plying the patients’ utility scores by the time they spent in a usual care control group (T1; ~ 22 weeks after T0), and 1 year certain health state using linear interpolation between mea- after T1 (T2). Patients visited one of two test centers to par- surement points. More QALYs indicate a better HRQoL, with ticipate in exercise testing and filled out questionnaires at approximately 1.3 QALYs as the maximum number (i.e., home. Cost data were assessed using cost questionnaires: pa- 1.3 years in optimal health). tients filled out five monthly cost questionnaires between T0 Effects on all primary outcomes and HRQoL occurring and T1, and four 3-monthly questionnaires in the year after 1-year follow-up were discounted at a rate of 1.5% [31]. Table 1 Structure of the exercise program Week Frequency of exercise session Type of training Frequency, intensity, and duration of the exercises/sessions 1–8 Twice a week Resistance exercises 2 sets of 10 repetitions at 65–80% of 1-RM* Cycling interval training 2 × 8 min, alternating 30s at 65% and 60s at 30% MSEC 9–12 Twice a week Resistance exercises See week 1–8 Cycling interval training 2 × 8 min, alternating 30s at 65% and 30s at 30% MSEC 12–18 Once a week Resistance exercises 2 sets of 20 repetitions at 35–40% of the 1-RM* Cycling Interval training See week 9–12 1, 4, 10, 12, and 18 Physical activity counseling 5–15 min per session 1-RM one repetition maximum, MSEC maximal short exercise capacity, i.e., the highest workload achieved during the steep ramp test *For the two additional exercises the protocol included the performance of 2 sets of 0.7 times the maximal number of repetitions J Cancer Surviv (2019) 13:558–569 561 Measurement and valuation of resource use replace a sick worker (i.e. friction period), which was estimat- ed to be 23 weeks at time the intervention was provided [31]. Costs were measured from a societal perspective and included After truncating the number of sickness days at the friction intervention costs, healthcare costs, costs of informal care, period, sickness days were valued using age- and gender- sports costs, unpaid productivity costs, and absenteeism costs. specific price weights [31]. Clinical trial–related costs (e.g., costs related to the physical All costs were expressed in 2014 Euros using consumer measurements) were not included. price indices. Costs occurring after 1-year follow-up were Intervention costs were estimated using a micro-costing discounted at a rate of 4% [31]. approach, meaning that detailed information was collected on the use of intervention-related resources as well as their Statistical analyses unit prices. Information on all other kinds of resource use was collected using cost questionnaires administered on a 1- Long-term effectiveness and cost-effectiveness analyses were monthly basis between T0 and T1 and a 3-monthly basis be- performed in accordance with the intention-to-treat principle. tween T1 and T2. To cover the complete duration of follow- Descriptive statistics were used to describe baseline character- up, cost questionnaires administered between T0 and T1 had a istics of patients from the intervention and control group, and recall period of 1 month, while cost questionnaires adminis- of patients with complete data and those with incomplete data. tered between T1 and T2 had a recall period of 3 months. The Linear mixed model analyses were conducted to evaluate cost questionnaire was developed by the research team in the intervention effects on cardiorespiratory fitness, handgrip close collaboration with experts in health economic strength, and fatigue at long-term (T2). The intervention was evaluations. regressed on the outcome value at the short-term (T1) and Intervention costs included all costs related to the imple- long-term (T2) simultaneously, adjusted for baseline levels mentation and execution of the exercise intervention, i.e., of the outcome variable in the model, age, gender, and educa- costs related to the training of the physical therapists, the pro- tion level. This procedure automatically takes into account vision of the exercise and counseling sessions, and the inter- missing values using maximum likelihood estimation. A ran- vention materials. Time investments of the study team and dom intercept for transplant center was added to take into physical therapists were valued using standard costs or, if account the clustering of patients within centers. Regression unavailable, salary data derived from their respective coefficients and 95% confidence intervals (CI) of clinical in- Collective Labor Agreement [31]. Data on the attendance of tervention effects at long-term were presented. exercise and counseling sessions were retrieved from the In the main cost-effectiveness analysis, patients with mul- training logs and the associated correspondence between the tiple myeloma using the extremely expensive drug study team and the physical therapists. Invoices were used to lenalidomide (Revlimid) as maintenance treatment after autol- value material costs. ogous stem cell transplantation (i.e., about €100,000 per year) Healthcare costs included the costs of primary healthcare were treated as being lost to follow-up from the moment they (i.e., general practice, physical therapist), secondary started using lenalidomide. From that point onwards, all of healthcare (i.e., outpatient care, hospitalization, professional their cost and effect measure values were set at missing and home care), and prescribed as well as over-the-counter medi- imputed using multiple imputation. This was done to prevent cation. Primary and secondary healthcare utilization were val- the cost estimates from being biased by the inclusion of these ued using Dutch standard costs [31]. If unavailable, prices of extremely high costs. This was considered to be an appropri- professional organizations were used. Medication use was ate strategy, as we do not expect the exercise intervention to valued using unit prices of the Royal Dutch Society of have an impact on a patient’s need for lenalidomide. Pharmacy. Missing data in the cost-effectiveness analysis were han- Informal care was defined as care provided by family and dled using multiple imputation by chained equations stratified friends, and was valued using a recommended Dutch shadow by treatment group. The imputation model included variables price [31]. that differed between patients with complete and incomplete Sports costs included the patients’ expenses on sporting data (i.e., education level, anti-cancer medication, and goods and sports memberships. smoking), those predicting the “missingness” of data (i.e., Unpaid productivity costs consisted of costs related to lost age, gender, diagnosis, sport history, and time between the hours of domestic tasks, educational activities, and volunteer start of the study and autologous stem cell transplantation), work. Unpaid productivity losses were valued using a recom- baseline effect values and all available follow-up cost and mended Dutch shadow price as well [31]. effect values. Using predictive mean matching, twenty-five Absenteeism costs related to paid work were estimated complete data sets were created in IBM SPSS resulting in a using the friction cost approach (FCA). The FCA assumes that loss of efficiency smaller than 5% (v23.0, Chicago, IL). All costs are limited to the period Dutch companies need to twenty-five datasets were analyzed separately as specified 562 J Cancer Surviv (2019) 13:558–569 below. Pooled estimates were subsequently calculated using patient of 30 (SA2). This was done because the intervention Rubin’srules[36]. was intended to last for 30 training sessions. In the third sen- For calculating ICERs, the mean difference in total costs sitivity analysis, only healthcare costs were analyzed (i.e., the between the intervention and control group was divided by the healthcare/NHS perspective was applied)(SA3). In the fourth mean difference in effects, expressed in terms of cardiorespi- sensitivity analysis (SA4), no correction was made for the use ratory fitness, grip strength, fatigue, and QALYs. Cost and of lenalidomide. effect differences were estimated using linear multilevel anal- yses, with a two-level structure: observations, transplant cen- ter. Within these analyses, cost and effect differences were Results corrected for age, gender, and education level. To deal with the highly skewed nature of cost data, joint uncertainty around Patients costs and effects was estimated using the Bias Corrected Bootstrap method with 5000 replications. To graphically illus- In total, 109 patients were randomized to the intervention (n = trate the uncertainty surrounding the ICERs, bootstrapped in- 54) or control group (n = 55). At baseline, relevant differences cremental cost-effect pairs were plotted on cost-effectiveness were found between both groups for gender and education planes. A summary measure of the joint uncertainty of costs level (Table 2). The percentage of patients with complete ef- and effects was presented using cost-effectiveness acceptabil- fect and cost data at the different measurement points can be ity curves (CEACs). CEACs indicate the probability of an found in Fig. 1. Data on the total number of training sessions intervention being cost-effective in comparison with the con- provided were complete. The main reasons for loss to follow- trol condition for a range of willingness-to-pay values (i.e., the up were relapse of (non-)Hodgkin lymphoma or progression maximum amount of money decision-makers are willing to of multiple myeloma (Fig. 1). Differences in education level, pay per unit of effect gained). Except for the multiple imputa- smoking, and time since autologous stem cell transplantation tion, analyses were performed using Stata v12. Statistical sig- were found between patients with complete and incomplete nificance was set at p <0.05 [33]. follow-up data (data not shown). To test the robustness of the results, four sensitivity analy- In total, 75% of the patients attended ≥ 80% of the training ses were performed. First, analyses were performed using the sessions and 87% of the patients attended the counseling ses- Human Capital Approach (HCA), instead of the FCA (SA1). sions (Table 3)[25]. The HCA regards each hour not worked by the patient as an hour of lost productivity. In contrast to the FCA, the HCA Long-term effectiveness: primary outcomes does therefore not truncate costs to the friction period. Second, analyses were performed in which intervention costs At T2, no statistically significant differences were found be- were calculated using a total number of training sessions per tween the intervention and control group for Table 2 Baseline demographics and clinical characteristics of the patients All (n = 109) Intervention group (n = 54) Control group (n =55) Male ([n (%)) 69 (63) 32 (59) 37 (67) Age (mean (SD)) 52 (11) 52 (11) 53 (12) Married/living together (n (%)) 91 (84) 45 (83) 46 (84) Higher education level (n (%)) 39 (36) 15 (28) 24 (44) Smoker (n (%)) 14 (13) 7 (13) 7 (13) Active (n (%)) 69 (63) 33 (61) 36 (66) Cancer type MM (n (%)) 58 (53) 29 (54) 29 (53) (N)HL (n (%)) 51 (47) 25 (46) 26 (47) Time since ASCT (mean (SD)) 86 (45) 84 (46) 88 (43) Number of co-morbidities (mean (SD)) 2 (2) 2 (2) 2 (2) VO peak (mean (SD)) in ml/kg/min 22 (5) 21 (5) 22 (6) Hand grip strength (mean (SD)) in kg 36 (11) 36 (12) 37 (10) General Fatigue (mean (SD)) (range 0–20) 13 (4) 13 (4) 14 (4) n number, SD standard deviation, MM multiple myeloma, (N)HL (non-)Hodgkin lymphoma, ASCT autologous stem cell transplantation Patients who participating in sports at least once a week before diagnoses/relapse J Cancer Surviv (2019) 13:558–569 563 Fig. 1 Flow diagram of patients Excluded (n=149): Physical (co)morbidies 38%, diseased/progressive disease 11%, second ASCT 11%, other in the EXIST study Paents screened for 39% eligibility (n=468) Non-response (n=112): No interest 30%, already following a rehabilitaon program 19%, no reason given 18%, other 33%. T0 baseline Unknown (n=98) measurement (n=111) Excluded (n=1): Paent is not able to cycle Non-response (n=1) Randomizaon (n=109) Intervenon (n=54) Usual care (n=55) Lost to follow up Lost to follow up T1 Reasons for loss to follow up: Reasons for loss to follow up: aer baseline (n=4) aer baseline (n=8) - Progression MM (n=5) -Progression MM (n=3) - Relapse (N)HL (n=4) - Relapse (N)HL (n=4) Lost to follow up Lost to follow up - Bad health situaon (n=3) - Not agree with allocaon (n=2) T2 aer baseline (n=19) aer baseline (n=16) -Other (n=7) -Other (n=7) Complete data Complete data Effect data; n (%) Effect data; n (%) T0 45 (83.3%) T0 42 (76.4%) T1 47 (87.0%) T1 45 (81.8%) T2 28 (51.9%) T2 31 (56.4%) Cost data; n(%) Cost data; n(%) CQ 1 41 (75.9%) CQ 1 33 (60.0%) CQ 2 39 (72.2%) CQ 2 35 (63.6%) CQ 3 39 (72.2%) CQ 3 42 (76.4%) CQ 4 37 (68.5%) CQ 4 36 (65.5%) CQ 5 39 (72.2%) CQ 5 33 (60.0%) CQ 6 34 (63.0%) CQ 6 28 (50.9%) CQ 7 37 (68.5%) CQ 7 29 (52.7%) CQ 8 29 (53.7%) CQ 8 26 (47.3%) CQ 9 28 (51.9%) CQ 9 28 (50.9%) Mulple imputaons Mulple imputaons Imputed dataset Imputed dataset n=54; 100,0% n=55; 100,0% Analysis cardiorespiratory fitness (0.12 ml/kg/min; 95%CI − 1.89; the exercise intervention was on average more costly and 2.14, p = 0.90), handgrip strength (−0.08 kg; 95%CI − more effective than usual care (Table 3;Fig. 2(1a)). The 2.47; 2.31, p = 0.95), and general fatigue (−0.69 points; CEAC in Fig. 2(2a) indicates that the exercise intervention’s 95%CI − 2.52; 1.14, p =0.46). probability of being cost-effective compared with usual care was 0.39 at a willingness-to-pay of €0 per ml/kg/min, increas- Costs ing to a maximum of 0.73 at a ceiling ratio of €20,000 per ml/ kg/min. On average, intervention costs were €1340 per patient The ICER for handgrip strength was − 427, indicating that (Table 4). During follow-up, secondary healthcare costs were the intervention was dominated by usual care (i.e., more costly statistically significantly higher in the intervention group than and less effective)(Table 3; Fig. 2(1b)). The CEAC in Fig. in the control group (Table 5). Unpaid lost productivity costs 2(2b) indicates that the exercise intervention’s probability of and informal care costs were statistically significantly lower in being cost-effective compared with usual care was 0.39 at a the intervention group than in the control group (Table 5). willingness-to-pay of €0 per kg. At increasing levels of will- Total societal costs were higher in the intervention group than ingness-to-pay, this probability decreased. in the control group, but this difference was not statistically The ICER for general fatigue was − 279. This suggests that significant (€529; 95%CI − 3205 to 4452). a 1-point reduction in general fatigue in the intervention group was associated with a societal cost of €273 compared with the Societal perspective: cost-effectiveness control group (Table 3; Fig. 2(1c)). Please note that a reduc- tion in general fatigue indicates an improved health effect. The The ICER for cardiorespiratory fitness was 972, suggesting CEAC in Fig. 2(2c) indicates that the exercise intervention’s that an increase of 1 ml/kg/min in VO peak in the intervention probability of being cost-effective compared with usual care group was associated with a societal cost of €972 compared was 0.39 at a willingness-to-pay of €0 per point improvement, with the control group. The majority of cost-effect pairs were increasing to a maximum of 0.96 at a willingness-to-pay of located in the northeast quadrant (i.e., 44%), indicating that €31,000 per point improvement. 564 J Cancer Surviv (2019) 13:558–569 Table 3 Difference in pooled mean costs and effects (95% confidence intervals), incremental cost-effectiveness ratios, and the distribution of cost-effect pairs around the quadrants of the cost- effectiveness planes Sample size Outcome ΔC (95% CI) € ΔE (95% CI) points ICER €/point Distribution CE-plane (%) 1 2 3 4 IC NE SE SW NW Main Analyses 54 55 Cardiorespiratory fitness 529 (− 3205 to 4452) 0.54 (− 1.14 to 2.23) 972 43.7 30.9 11.5 13.9 Hand grip strength 529 (− 3205 to 4452) − 1.24 (− 4.10 to 1.59) − 427 10.8 8.9 33.5 46.8 General Fatigue 529 (− 3205 to 4452) − 1.78 (− 3.94 to − 0.37) − 297 53.3 40.1 37.1 4.4 HRQoL - QALY 529 (− 3205 to 4452) − 0.07 (− 0.17 to 0.04) − 8043 4.6 5.2 37.1 53.1 SA1 Human Capital Approach 54 55 Cardiorespiratory fitness 1152 (− 5585 to 7936) 0.54 (− 1.14 to 2.23) 2117 46.2 28.4 9.8 15.5 Hand grip strength 1152 (− 5585 to 7936) − 1.24 (− 4.10 to 1.59) − 931 10.9 8.8 29.5 50.8 General Fatigue 1152 (− 5585 to 7936) − 1.78 (− 3.94 to − 0.37) − 646 57.4 35.9 2.3 29.5 HRQoL - QALY 1152 (− 5585 to 7936) − 0.07 (− 0.17 to 0.04) − 17,515 5.4 4.4 33.8 56.4 SA2 Maximal training sessions 54 55 Cardiorespiratory fitness 843 (− 2871 to 4779) 0.54 (− 1.14 to 2.23) 1549 48.3 26.4 9.9 15.4 Hand grip strength 843 (− 2871 to 4779) − 1.24 (− 4.10 to 1.59) − 681 12.0 7.7 28.6 51.7 General Fatigue 843 (− 2871 to 4779) − 1.78 (− 3.94 to − 0.37) − 473 59.0 34.4 1.9 4.7 HRQoL - QALY 843 (− 2871 to 4779) − 0.07 (− 0.17 to 0.04) − 12,817 5.2 4.6 31.7 58.5 SA3 Medical costs only 54 55 Cardiorespiratory fitness 1096 (− 114 to 2220) 0.54 (− 1.14 to 2.23) 2016 70.7 4.0 0.8 24.6 Hand grip strength 1096 (− 114 to 2220) − 1.24 (− 4.10 to 1.59) − 886 18.5 1.2 3.5 76.8 General Fatigue 1096 (− 114 to 2220) − 1.78 (− 3.94 to − 0.37) − 615 88.8 4.6 1.3 6.5 HRQoL - QALY 1096 (− 114 to 2220) − 0.07 (− 0.17 to 0.04) − 16,676 8.6 1.2 3.5 86.7 SA4 No correction for lenalidomide use 54 55 Cardiorespiratory fitness 15,646 (7688 to 30,845) 0.97 (− 0.74 to 2.69) 16,135 84.3 1.1 0.0 14.6 Hand grip strength 15,646 (7688 to 30,845) − 0.22 (− 2.96 to 2.51) − 69,892 41.4 0.7 0.4 57.4 General Fatigue 15,646 (7688 to 30,845) − 0.94 (− 3.12 to 1.24) − 12,667 80.0 1.0 0.1 18.9 HRQoL - QALY 15,646 (7688 to 30,845) − 0.07 (− 0.17 to 0.04) − 242,963 10.6 0.1 1.0 88.2 I intervention group, C control group, C costs, E effects, ICER incremental cost-effectiveness ratio, CE-plane cost-effectiveness plane, SA sensitivity analyses, QALY quality-adjusted life years Costs are expressed in 2014 Euros 1Refers to the northeast quadrant of the CE plane, suggesting that the EXIST exercise intervention is more effective and more costly than usual practice 2Refers to the southeast quadrant of the CE plane, suggesting that the EXIST exercise intervention is more effective and less costly than usual practice 3Refers to the northwest quadrant of the CE plane, suggesting that the EXIST exercise intervention is less effective and more costly than usual practice 4Refers to the southwest quadrant of the CE plane, suggesting that the EXIST exercise intervention is less effective and less costly than usual practice J Cancer Surviv (2019) 13:558–569 565 Table 4 Costs of the EXIST intervention per patient Staff Units Unit prices Total costs Total costs per patient Kick-off physiotherapy (42 clinics) Information packet 42 packets €1.97/packet €82.59 €1.53 Information leaflet 42 leaflets €0.83/leaflet €34.78 €0.64 Time investment Physical therapist 21 h €48.81/h €1025.02 €18.98 Project assistant 21 h €32.54/h €683.35 €12.65 Traveling expenses 4451.2 km €0.19/km €858.63 €15.90 Physiotherapy Training sessions Physical therapist 1356 h €48.81/h €66,171.60 €1225.40 Counseling sessions Physical therapist 63.25 h €48.81/h €3087.27 €57.17 Consultation sports physician/intake Sports physician 1 h €91.91/h €91.91 €1.70 Physical therapist 1 h €48.81/h €48.81 €0.90 Information booklet 54 booklets €3.90/leaflet €210.60 €3.90 Registration logbook 54 logbooks €1.16/logbook €62.64 €1.13 Total €1339.92 km kilometer Costs are expressed in 2014 euros The ICER for HRQoL was − 8043, indicating that the in- important difference being an almost 30-fold increase in the tervention was dominated by usual care for QALYs (i.e., more total societal cost difference (from €529 to €15,646). costly and less effective)(Table 3;Fig. 2(1d)). The CEAC in Fig. 2(2d) indicates that if decision-makers are not willing to pay anything for a QALY gained, the probability of the inter- Discussion vention being cost-effectiveness was 0.39. At increasing levels of willingness-to-pay, this probability decreased. Evaluations of the long-term effectiveness and cost- effectiveness of an 18-week supervised high-intensity com- Sensitivity analyses bined resistance and interval exercise intervention versus usu- al care in patients after autologous stem cell transplantation In SA1, SA2, and SA3, the total cost difference became larger presented in this article failed to show statistically significant compared with the main analysis, but remained in favor of the benefits on physical fitness and general fatigue at long-term, control group. In SA4, cost and effect difference estimates and low probabilities of cost-effectiveness at reasonable differed from those of the main analysis, with the most values of willingness-to-pay. To illustrate the latter, at the Table 5 Mean costs per patient in the intervention and control group, and mean cost differences between both groups during follow-up Cost category Intervention group n =54; Control group n =55; Crude cost differences Adjusted cost differences mean (SEM) mean (SEM) mean (95% CI) mean (95% CI) Primary healthcare costs 1437 (225) 1955 (282) − 518 (− 1280 to 154) − 512 (− 1215 to 78) Secondary healthcare costs 2338 (280) 1845 (198) 493 (−97 to 1235) 501 (97 to 1160) Medication costs 1002 (135) 1271 (250) − 269 (−908to206) − 237 (−884 to 246) Unpaid productivity costs 560 (106) 1480 (276) − 920 (− 1571 to 396) − 884 (− 1437 to − 421) Informal care costs 432 (122) 1148 (244) − 716 (− 1298 to − 227) − 669 (− 1159 to − 193) Absenteeism costs 16,818 (1277) 15,823 (1294) 995 (− 2510 to 4550) 1160 (− 2114 to 4585) Sport costs 470 (51) 596 (77) − 126 (−313to42) − 111 (−292to57) Intervention costs 1340 (47) 0 (0) 1340 (1249 to 1432) 1344 (1243 to 1423) Total costs 24,397 (1322) 24,119 (1430) − 278 (− 4105 to 3548) 529 (− 3205 to 4452) n number, SEM standard error of the mean, CI confidence interval, SD standard deviation Costs are expressed in 2014 Euros 566 J Cancer Surviv (2019) 13:558–569 (1a) (2a) (1b) (2b) (2c) (1c) (1d) (2d) Fig. 2 Cost-effectiveness planes indicating the uncertainty around the (€) of willingness-to-pay per unit of effect gained (2) for cardiorespiratory incremental cost-effectiveness ratios (1) and cost-effective acceptability fitness (a), handgrip strength (b), general fatigue (c), and health-related curves indicating the probability of cost-effectiveness for different values quality of life (d) lower and upper bound of the informal Dutch willingness-to- respectively), the probability of the intervention being cost- pay threshold (i.e., 10,000 and 80,000 €/QALY gained, effective compared with usual care was low (< 0.29). J Cancer Surviv (2019) 13:558–569 567 The lack of significant between-group differences in physi- considered cost-effective depending on the decision-makers’ cal fitness and fatigue at long-term is in line with previous willingness-to-pay. studies in patients with breast cancer at 18 weeks [32]or Strengths of the study are the focus on the long-term (1 year 6 months follow-up [34]. However, a previous study among after intervention completion) effectiveness and cost- patients undergoing autologous or allogeneic stem cell trans- effectiveness specifically in patients after autologous stem cell plantation including a 3-month follow-up showed benefits on transplantation, the multicenter randomized controlled trial cardiorespiratory fitness, assessed with a submaximal exercise design, the use of a broad set of valid and reliable outcome test and knee extension strength, but not on handgrip and fa- measures, and the use of state-of-the-art statistical methods for tigue [13]. The lack of significant intervention effects at the evaluating the intervention’s cost-effectiveness (i.e., the use of long-term indicates that delayed effects of the intervention did seemingly unrelated regression, bootstrapping, and multiple not occur, and were comparable with the findings at short-term imputation). Limitations of the study include the use of self- [23]. This may be related to the suboptimal timing of interven- report for collecting cost data which may have introduced tion delivery, as exercise post-stem cell transplantation may not recall bias, the potential problems with generalizability to oth- be able to speed up natural recovery [23, 25]. A recent meta- er countries with different healthcare systems and/or payment analysis strengthens this hypothesis by reporting significant structures [6], and the large amount of missing cost data. effects of exercise on fatigue and muscle strength pre-stem cell However, missing effect and cost data were taken into account transplantation but not post-stem cell transplantation [16]. by the linear mixed model analyses and multiple imputation, Other reasons for the lack of significance may be related to respectively. Also, during follow-up, more intervention group suboptimal exercise compliance in the intervention group or patients used the extremely expensive drug lenalidomide than contamination in the control group [23], but no evidence for control group patients. As we did not expect the exercise in- this hypothesis was found in a post hoc analysis [23] and pro- tervention to have an impact on a patient’s need for cess evaluation [25]. Finally, the lack of long-term effects also lenalidomide, we dealt with this issue by treating patients as suggests that the counseling sessions alongside, and after, the being lost to follow-up from the moment they started using supervised sessions were insufficient to change daily physical lenalidomide, that is, as if their measurements were missing. activity behavior. Due to practical reasons, physical therapists The long-term effectiveness estimates slightly differ from implemented the counseling sessions differently: some provid- those of the cost-effectiveness analyses. This is due to differ- ed the counseling during the exercise sessions, while others ences in the applied statistical methods, including (1) the use made separate appointments [25]. Additionally, most were of mixed effect models in the effectiveness analyses, which scheduled alongside the program to promote compliance and are typically not applied in cost-effectiveness analyses, (2) it might be that the three sessions focused on promoting phys- multiple imputation, which is recommended for imputing ical activity in daily life were insufficient [21]. missing cost data [26] versus maximum likelihood estimation, The finding that the supervised exercise intervention was which is often used in effectiveness analyses, and (3) a cor- not cost-effective versus usual care is probably due to the rection for the possible correlation between costs and effects intervention not being effective at the short-term, nor at the in the cost-effectiveness analyses [37]. long-term. The lack of cost-effectiveness is also in line with The average cost of the 18-week exercise intervention of the results of an economic evaluation of a combined physical €1340 per patient on average was somewhat higher than the activity and dietary intervention in patients after autologous average cost of €858 per patient of the aforementioned 12- stem cell transplantation [10]. This study, however, used week high-intensity exercise intervention offered to patients modeling techniques, rather than patient-level data. Patient- with various types of cancer [12]. This amount is still relative- level studies evaluating the cost-effectiveness of exercise in- ly low as compared with the anti-cancer treatment costs. terventions for patients with solid tumors provided mixed re- Therefore, exercise interventions, if effective, may have the sults [1, 10, 12, 17, 19, 35]. Kampshoff et al. [12] found a 12- potential to be included as part of standard care for patients week high-intensity exercise intervention to be cost-effective with cancer. However, in its current form, and offered to pa- compared with a moderate-intensity exercise intervention in a tients recently treated with autologous stem cell transplanta- group of cancer survivors with solid tumors (mostly breast tion, the exercise intervention is not effective at the short-term cancer), whereas May et al. [17] found an 18-week exercise and long-term compared with usual care, nor cost-effective, program to be cost-effective versus usual care for colon can- and therefore, we do not recommend it to offer it as part of cer, but not for breast cancer. Van Waart et al. [34]found a standard care. Future studies should examine the (long-term) low-intensity physical activity program not to be cost- effectiveness and cost-effectiveness of other types of exercise effective compared with usual care among breast cancer pa- interventions and may concentrate on exercise interventions at tients undergoing chemotherapy, but found that a combined different time points in the treatment trajectory of patients resistance and aerobic supervised exercise program may be scheduled to receive stem cell transplantation. 568 J Cancer Surviv (2019) 13:558–569 program: the advantage of the steep ramp test. Arch Phys Med In conclusion, we found no significant beneficial effects of Rehabil. 2007;88:610–6. the high-intensity combined resistance and interval exercise 6. Drummond M, Barbieri M, Cook J, Glick HA, Lis J, Malik F, et al. program on physical fitness and fatigue at long-term when Transferability of economic evaluations across jurisdictions: compared with usual care, nor was the intervention cost- ISPOR good research practices task force report. Value Health. 2009;12:409–18. effective from a societal perspective. 7. Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care Acknowledgments The authors acknowledge the A-CaRe Clinical programmes: Oxford university press; 2015. Research group and thank all patients who participated in the study and 8. EuroQol G. EuroQol–a new facility for the measurement of health- all hematologists, nurses, data managers, and physiotherapists who con- related quality of life. Health policy (Amsterdam, Netherlands). tributed to the study. We are also grateful to Karen van der Weiden, Julie 1990;16:199. Heeren, Griet van der Slagmolen, Michiel Greidanus, the team of the 9. Fuller JT, Hartland MC, Maloney LT, Davison K. Therapeutic ef- Sports Medical Advice Center Rotterdam, and the members of the fects of aerobic and resistance exercises for cancer survivors: a Clinical Trial Office, Department of Hematology Amsterdam UMC, lo- systematic review of meta-analyses of clinical trials. Br J Sports cation AMC for their assistance with patient recruitment, data collection, Med. 2018:bjsports–2017-098285. and/or data analysis. 10. Gao L, Moodie M, Brown V, Avery S. Cost-effectiveness of a lifestyle modification program in long-term survivors of hemopoi- Funding This study was funded by the Dutch Cancer Society (grant etic stem cell transplantation. Clin Transpl. 2017;31:e13049. number: ALPE 2009-4620). 11. Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the Compliance with ethical standards American Heart Association. Circulation. 2007;116:1081–93. 12. Kampshoff C, van Dongen J, van Mechelen W, Schep G, The study procedures were approved by the Medical Ethics Committee of Vreugdenhil A, Twisk J, et al. Long-term effectiveness and cost- the Academic Medical Center (METC AMC 10/106) and by the boards of effectiveness of high versus low-to-moderate intensity resistance the Antoni van Leeuwenhoek Hospital (Amsterdam), St. Antonius and endurance exercise interventions among cancer survivors. J Hospital (Nieuwegein), Haga Teaching Hospital (Den Haag), Cancer Surviv. 2018;12:417–29. University Medical Center (Utrecht), Isala (Zwolle), Erasmus MC/ 13. Knols R, de Bruin ED, Uebelhart D, Aufdemkampe G, Schanz U, Daniel den Hoed (Rotterdam), VU University Medical Center Stenner-Liewen F, et al. Effects of an outpatient physical exercise (Amsterdam), and Leiden University Medical Center (Leiden). program on hematopoietic stem-cell transplantation recipients: a randomized clinical trial. Bone Marrow Transplant. 2011;46: Conflict of interest The authors declare that they have no conflict of 1245–55. interest. 14. Lamers L, Stalmeier P, McDonnell J, Krabbe P. Measuring the quality of life in economic evaluations: the Dutch EQ-5D tariff. Open Access This article is distributed under the terms of the Creative Ned Tijdschr Geneeskd. 2005;149:1574–8. Commons Attribution 4.0 International License (http:// 15. Levinger I, Goodman C, Hare DL, Jerums G, Toia D, Selig S. The creativecommons.org/licenses/by/4.0/), which permits unrestricted use, reliability of the 1RM strength test for untrained middle-aged indi- distribution, and reproduction in any medium, provided you give appro- viduals. J Sci Med Sport. 2009;12:310–6. priate credit to the original author(s) and the source, provide a link to the 16. Liang Y, Zhou M, Wang F, Wu Z. Exercise for physical fitness, Creative Commons license, and indicate if changes were made. fatigue and quality of life of patients undergoing hematopoietic stem cell transplantation: a meta-analysis of randomized controlled trials. Jpn J Clin Oncol. 2018;48:1046–57. 17. May AM, Bosch MJ, Velthuis MJ, Van Der Wall E, Bisschop CNS, Los M, et al. Cost-effectiveness analysis of an 18-week exercise programme for patients with breast and colon cancer undergoing References adjuvant chemotherapy: the randomised PACT study. BMJ Open. 2017;7:e012187. 1. Braam KI, van Dijk-Lokkart E, van Dongen J, van Litsenburg R, 18. Mewes JC, Steuten LM, IJzerman MJ, Van Harten WH. Takken T, Huisman J, et al. Cost-effectiveness of a combined phys- Effectiveness of multidimensional cancer survivor rehabilitation ical exercise and psychosocial training intervention for children and cost-effectiveness of cancer rehabilitation in general: a system- with cancer: results from the quality of life in motion study. Eur J atic review. The oncologist. 2012:theoncologist. 2012-0151, 2012. Cancer Care. 2017;26:e12586. 19. Mewes JC, Steuten LM, Duijts SF, Oldenburg HS, van Beurden M, 2. Buffart L, Galvão DA, Brug J, Chinapaw M, Newton RU. Stuiver MM, et al. Cost-effectiveness of cognitive behavioral ther- Evidence-based physical activity guidelines for cancer survivors: apy and physical exercise for alleviating treatment-induced meno- current guidelines, knowledge gaps and future research directions. pausal symptoms in breast cancer patients. J Cancer Surviv. 2015;9: Cancer Treat Rev. 2014;40:327–40. 126–35. 3. Buffart LM, Kalter J, Sweegers MG, Courneya KS, Newton RU, 20. Oberoi S, Robinson PD, Cataudella D, Culos-Reed SN, Davis H, Aaronson NK, et al. Effects and moderators of exercise on quality Duong N, et al. Physical activity reduces fatigue in patients with of life and physical function in patients with cancer: an individual cancer and hematopoietic stem cell transplant recipients: a system- patient data meta-analysis of 34 RCTs. Cancer Treat Rev. 2017;52: atic review and meta-analysis of randomized trials. Crit Rev Oncol 91–104. Hematol. 2018;122:52–9. 4. Cramp F, Daniel J. Exercise for the management of cancer-related 21. Persoon S, Kersten MJ, ChinAPaw MJ, Buffart LM, Burghout H, fatigue in adults. Cochrane database of systematic reviews. 2008. Schep G, et al. Design of the EXercise Intervention after Stem cell Transplantation (EXIST) study: a randomized controlled trial to 5. De Backer IC, Schep G, Hoogeveen A, Vreugdenhil G, Kester AD, van Breda E. Exercise testing and training in a cancer rehabilitation evaluate the effectiveness and cost-effectiveness of an J Cancer Surviv (2019) 13:558–569 569 individualized high intensity physical exercise program on fitness analysis of randomised controlled trials. Br J Sports Med. 2018;52:505–13. and fatigue in patients with multiple myeloma or (non-) Hodgkin's lymphoma treated with high dose chemotherapy and autologous 31. Tan SS, Bouwmans-Frijters CA, Hakkaart-van Roijen L. stem cell transplantation. BMC Cancer. 2010;10:671. Handleiding voor kostenonderzoek: methoden en referentieprijzen 22. Persoon S, Kersten MJ, van der Weiden K, Buffart LM, Nollet F, voor economische evaluaties in de gezondheidszorg. Tijdschrift Brug J, et al. Effects of exercise in patients treated with stem cell voor gezondheidswetenschappen. 2012;90:367–72. transplantation for a hematologic malignancy: a systematic review 32. Travier N, Velthuis MJ, Bisschop CNS, van den Buijs B, and meta-analysis. Cancer Treat Rev. 2013;39:682–90. Monninkhof EM, Backx F, et al. Effects of an 18-week exercise 23. Persoon S, ChinAPaw MJ, Buffart LM, Liu RD, Wijermans P, programme started early during breast cancer treatment: a Koene HR, et al. Randomized controlled trial on the effects of a randomised controlled trial. BMC Med. 2015;13:121. supervised high intensity exercise program in patients with a hema- 33. Van Der Meer EW, van Dongen JM, Boot CR, van der Gulden JW, tologic malignancy treated with autologous stem cell transplanta- Bosmans JE, Anema JR. Economic evaluation of a multifaceted tion: results from the EXIST study. PLoS One. 2017;12:e0181313. implementation strategy for the prevention of hand eczema among 24. Persoon S, Kersten MJ, Buffart LM, Vander Slagmolen G, Baars healthcare workers in comparison with a control group: the JW, Visser O, et al. Health-related physical fitness in patients with Hands4U study. Acta Derm Venereol. 2016;96:499–504. multiple myeloma or lymphoma recently treated with autologous 34. van Waart H, Stuiver MM, van Harten WH, Geleijn E, Kieffer JM, stem cell transplantation. J Sci Med Sport. 2017;20:116–22. Buffart LM, et al. Effect of low-intensity physical activity and 25. Persoon S, Chinapaw M, Buffart L, Brug J, Kersten M, Nollet F. moderate-to high-intensity physical exercise during adjuvant che- Lessons learnt from a process evaluation of an exercise intervention motherapy on physical fitness, fatigue, and chemotherapy comple- in patients treated with autologous stem cell transplantation. Eur J tion rates: results of the PACES randomized clinical trial. J Clin Cancer Care. 2018;27:e12779. Oncol. 2015;33:1918–27. 26. Ramsey SD, Willke RJ, Glick H, Reed SD, Augustovski F, Jonsson 35. van Waart H, van Dongen JM, van Harten WH, Stuiver MM, B, et al. Cost-effectiveness analysis alongside clinical trials II-an Huijsmans R, Hellendoorn-van Vreeswijk JA, et al. Cost–utility ISPOR good research practices task force report. Value Health. and cost-effectiveness of physical exercise during adjuvant chemo- 2015;18(2):161–72. therapy. Eur J Health Econ. 2018;19:893–904. 27. Roseboom KJ, van Dongen JM, Tompa E, van Tulder MW, 36. White IR, Royston P, Wood AM. Multiple imputation using Bosmans JE. Economic evaluations of health technologies in chained equations: issues and guidance for practice. Stat Med. Dutch healthcare decision-making: a qualitative study of the current 2011;30:377–99. and potential use, barriers, and facilitators. BMC Health Serv Res. 37. Willan AR, Briggs AH, Hoch JS. Regression methods for covariate 2017;17:89. adjustment and subgroup analysis for non-censored cost-effective- 28. Segal R, Zwaal C, Green E, Tomasone J, Loblaw A, Petrella T. ness data. Health Econ. 2004;13(5):461–75. Exercise for people with cancer: a systematic review. Curr Oncol. 2017;24:e290. 29. Smets E, Garssen B, Bonke Bd, De Haes J. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 1995;39:315–325. 30. Sweegers MG, Altenburg TM, Chinapaw MJ, Kalter J, Verdonck- Publisher’snote Springer Nature remains neutral with regard to de Leeuw IM, Courneya KS, et al. Which exercise prescriptions jurisdictional claims in published maps and institutional affiliations. improve quality of life and physical function in patients with cancer during and following treatment? A systematic review and meta-

Journal

Journal of Cancer Survivorship: Research and Practice – Springer Journals

Published: Jul 8, 2019

There are no references for this article.