Table 4.
Results of completed economic evaluations
| Study ID, main reference | Clinical effectiveness (IG vs. CG) | Total costs in EUR (2020) (IG vs. CG) | Cost-effectiveness based on direction of effectsa | Risk of bias/quality |
|---|---|---|---|---|
| Results from model-based economic evaluations | ||||
| Results from CEAs | ||||
| Dholakia 2021 [41] | Mortality: 397/4415 (9.0%) vs. 441/4415 (10.0%); RDb -1.0% | Mean 59,849 vs. 65,304; MDb -5455 | Cost-effective | ISPOR-Q: Insufficiently credible |
| Results from CBAs | ||||
| Chen 2022 [40] | Morbidity: 24/240 (10.0%) vs. 16/240 (6.7%); RDb -3.3% | Meanb 3292 vs. 3742; MDb -450 | Cost-effective; total cost–benefit: 108,022 EUR (2020) | ISPOR-Q: Sufficiently credible |
| Van Wijk 2020 [54] | Morbidity: not reported | 28,001 vs. 30,242; difference -2,241c | Unclear (effectiveness not reported); return of investment: 1.55 | ISPOR-Q: Insufficient information |
| Results from trial-based economic evaluations | ||||
| Results from CUAs | ||||
| AlShewaier 2016 [37] | QALYsd: Median 0.679 (IQR 0.10) vs. 0.573 (0.05); Difference in mediansb 0.106 | Median 20,790 vs. 19,952; difference in mediansb 838 | Unclear; incremental analysis required: ICER: 7906 EUR (2020) per QALY gained, but no WTP reported |
RoB 2: High CHEC: 13/19 items |
| Fernandes 2017 [42] | QALYsd: Mean 0.66 ± BS SE 0.04 vs. 0.61 ± 0.04e; MD 0.04 (95% CI 0.01 to 0.07) | Mean 17,432 ± BS SE 1265 vs. 17,574 ± 1480; MD -142 (95% CI -3952 to 3668) (331b when including patient expenses) | Cost-effective (unclear when including patient expenses); Probability of CEA at a WTP of 40,000 EUR: 84% (approx. 79% when including patient expenses) |
RoB 2: Some concerns CHEC: 16/19 items |
| Nguyen 2022 | QALYsd: Mean 0.7 ± SD 0.3 vs. 0.6 ± 0.3; MDb 0.1 | Mean 15,071 ± SD 7014 vs. 15,472 ± 6309; MD -401 | Cost-effective |
RoB: Some concerns (QALY), high (costs) CHEC: 9/19 items |
| Tveter 2020 [53] | QALYsd: Not reported per group; difference 0.07c | Not reported per group; difference -508c | Cost-effective |
RoB 2: Some concerns (QALY), high (costs) CHEC: 11/19 items |
| Results from CEAs | ||||
| McGregor 2004 [47] | HrQoL: EQ-5D-3L VAS (0–100%)d: mean 75.80 ± SD 14.86 vs. 72.15 ± 22.20f; MDb 3.65%, EQ-5D-3L utilitiesd: mean 0.72 ± SD 0.13 vs. 0.60 ± SD 0.31f; MDb 0.12 | Mean 4148 vs. 5005; MDb -856 | Cost-effective |
RoB 2: High CHEC: 8/19 items |
| Tew 2017 [52] | HrQoL: EQ-5D-5L utilitiesd: mean 0.837 vs. 0.760; MD 0.077 (95% CI 0.005 to 0.148) | BS mean 14,269 ± SD 3542 vs. 13,688 ± 3542; BS MD 582 (95% CI -1588 to 2848) | Unclear; incremental analysis required, but ICER not reported |
RoB 2: High CHEC: 14/19 items |
| Results from CCAs | ||||
| Barberan-Garcia 2019 [38] |
HrQoL: SF-36 PCSd: mean 47 ± SD 7 vs. 44 ± 8; MDb 3 SF-36 MCS scored: mean 51 ± SD 9 vs. 50 ± 9; MDb 1 Morbidity: 19/62 (30.6%) vs. 39/63 (61.9%); RDb -31.3% Mortality: 1/62 (1.6%) vs. 4/63 (6.3%); RDb -4.7% PROMs: YPAS scored: mean 46 ± SD 13 vs. 39 ± 15; MDb 7 HAD total score: mean 6 ± SD 5 vs. 8 ± 6; MDb -2 |
Mean 5428 (range 1792 to 30,532) vs. 6416 (1587 to 34,521); BS MD -955 (95% CI -3109 to 1033) | Cost-effective |
RoB 2: Some concerns (morbidity, mortality), high (PROMs, costs) CHEC: not applicable |
| Gao 2015 [43] | Morbidity: 12/71 (16.9%) vs. 59/71 (83.3%); RDb -66.4% | Mean 9728 ± SD 1130 vs. 8955 ± 888; MDb 773 | Unclear; incremental analysis required, but ICER not applicable for CCA |
ROBINS-I: Serious CHEC: not applicable |
| Gränicher 2020 [44] |
PROMs: Lysholm Scored: mean 87.1 ± SD 9.0 vs. 69.1 ± 14.9; MDb 18 Lysholm Score pain itemd: mean 25.0 ± SD 0.0 vs. 16.5 ± 9.1; MDb 8.5 Tegner Activity Scaled: mean 3.8 ± SD 0.8 vs. 2.5 ± 0.9; MDb 1.3 Physical function: Stair climbing test, time in seconds: mean 12.58 ± SD 4.64 vs. 13.59 ± 5.30; MDb -1.01 Knee ROM (degrees)d: mean 100.5 ± SD 18.7 vs. 103.5 13.7; MDb -3 |
Mean 3187 vs. 4052; MDb -865 | Unclear (inconsistent effectiveness); individual decision required |
RoB 2: Some concerns (PROMs, physical function), high (costs) CHEC: not applicable |
| Howard 2019 [30] |
Morbidity: 12/40 (30.0%) vs. 29/75 (38.7%); RDb -8.7% Mortality: 1/40 (2.5%) vs. 1/75 (1.3%); RDb 1.2% |
Mean 58,300 ± SD 42,590 vs. 75,248 ± 77,516; MDb (incorporating prehabilitation costs) -16,870 | Unclear (inconsistent effectiveness); individual decision required |
ROBINS-I: Serious CHEC: not applicable |
| Huang 2012 [45] |
Morbidity: Infection rate: 2/126 (1.6%) vs. 1/117 (0.9%); RDb 0.7% Rate of deep vein thrombosis: 5/126 (4.0%) vs. 3/117 (2.6%); RDb 1.4% PROMs: Pain (VAS): mean 2.4 ± SD 0.7 vs. 2.5 ± 0.6; MDb -0.1 Physical function: Knee ROM (degrees)d: mean 76 ± SD 22 vs. 74 ± 20; MDb 2 Ambulation statusd: 108/126 (85.7%) vs. 95/117 (81.2%); RDb 4.5% |
Mean 6726 ± SD 283 vs. 6841 ± SD 241; MDb (incorporating prehabilitation costs) -95 | Unclear (inconsistent effectiveness); individual decision required |
RoB 2: High CHEC: not applicable |
| Koh 2021 [46, 86] |
Morbidity: 24/58 (41.4%) vs. 11/23 (47.8%); RDb -6.4% Mortalityg: 0/58 (0%) vs. 0/23 (0%); RDb 0% |
Not reported per group; MD -2584 | Cost-effective |
ROBINS-I: Serious CHEC: not applicable |
| Lai 2017 [32] | Morbidity: 5/51 (9.8%) vs. 14/50 (28.0%); RDb -18.2% | Mean 7677 ± SD 1374 vs. 8608 ± 2482; MDb -931 | Cost-effective |
RoB 2: High CHEC: not applicable |
| Lai 2019 [33] | Morbidity: 4/34 (11.8%) vs. 12/34 (35.3%); RDb -23.5% | Median 10,456 (IQR 9683 to 11,339) vs. 11,285 (10,544 to 13,340); Difference in mediansb -830 | Cost-effective |
RoB 2: High CHEC: not applicable |
| Ploussard 2020 [50] | Mortality: 0/194 (0%) vs. 1/156 (0.6%); RDb -0.6% | Mean 2904 vs. 3282; MDb -379 | Cost-effective |
ROBINS-I: Serious CHEC: not applicable |
| Risco 2022 [51] | Morbidity: Comprehensive complications index: mean 15.1 ± SD 17.1 vs. 16.6 ± 16.9; MDb: -1.5 | Mean 7288 vs. 7142; MD 145 | Unclear; incremental analysis required, but ICER not applicable for CCA |
ROBINS-I: Serious CHEC: not applicable |
| Wang 2020 [55] |
Morbidity: 21/70 (30.0%) vs. 18/34 (52.9%); RD -22.9% Mortality: 1/70 (1.4%) vs. 1/34 (2.9%); RDb -1.5% PROMs (in subsample of n = 33 vs. n = 24): FACT-Hep scored: median 152 (range 102 to 179) vs. 148 (66 to 175); differences in mediansb 4 |
Median 6138 (IQR 4590 to 8833) vs. 7349 (5328 to 11,026); difference in medians -1210 | Cost-effective |
ROBINS-I: Serious CHEC: not applicable |
| Zhou 2017 [34] | Morbidity: 36/197 (18.3%) vs. 194/742 (26.1%); RDb: -7.8% | Not calculable in EUR (2020) as original currency not reported; original values: mean 7131.8 ± SD 2316.6 vs. 77,266.4 ± 1615.0; MDb -134.60 | Cost-effective |
ROBINS-I: Critical CHEC: not applicable |
| Results from CMAs | ||||
| Beaupre 2004 [39] | Not applicable | Mean 1285 ± SD 1196 vs. 1283 ± 1329; MD 2 | Not cost-effective |
RoB 2: High CHEC: not applicable |
| Englesbe 2017 [29] | Not applicable |
Provider perspective: median 16,900 (IQR 10,162 to 30,365) vs. 23,091 (14,993 to 39,017); difference in mediansb -6191 Payer perspective: median 19,216 (IQR 12,122 to 33,840) vs. 24,519 (17,057 to 37,243); difference in mediansb -5303 |
Cost-effective |
ROBINS-I: Serious CHEC: not applicable |
| Mouch 2019 [31] | Not applicable | Mean 24,435 ± SD 20,024 vs. 26,903 ± 24,935; MDb -2468 | Cost-effective |
ROBINS-I: Moderate CHEC: not applicable |
| Pham 2016 [49] | Not applicable | Reported only for a subset of patients (5/29 vs. 11/21): mean 5081 ± SD 298 vs. 5152 ± 656; MDb -71 | Cost-effective |
RoB 2: High CHEC: not applicable |
BS bootstrapped, CBA cost–benefit analysis, CCA cost–consequence analysis, CEA cost-effectiveness analysis, CG control group, CHEC Consensus on Health Economic Criteria, CMA cost-minimisation analysis, CUA cost–utility analysis, EQ-5D-3L EuroQoL 5 dimensions 3 levels, EQ-5D-5L EuroQoL 5 dimensions 5 levels, FACT-Hep Functional Assessment of Cancer Therapy—Hepatobiliary, HAD Hospital Anxiety and Depression Scale, IG intervention group, ICER incremental cost-effectiveness ratio, ISPOR-Q International Society for Pharmacoeconomics and Outcomes Research Questionnaire, IQR interquartile range, MD mean difference, PROMs patient-reported outcome measures, QALY quality-adjusted life-year, HrQoL health-related quality of life, RD relative difference, RoB 2 revised Cochrane risk-of-bias tool for randomised trials, ROBINS-I tool for assessing risk of bias in non-randomised studies of interventions, ROM range of motion, SD standard deviation, SF-36 Short Form 36, VAS visual analogue scale, YPAS Yale Physical Activity Scale
a ‘Cost-effective’ if better effectiveness and same/lower costs, or same effectiveness and lower costs; ‘unclear’ if better effectiveness and higher costs, or same effectiveness and same costs, or inconsistent/poorer effectiveness and lower costs; ‘not cost-effective’ if same effectiveness and higher costs, or poorer effectiveness and same/higher costs
b Calculated by review authors
c Measure of central tendency (mean, median) not reported
d Higher values indicating higher benefit
e When missing values were imputed by linear trend at point and adjusted for baseline EQ-5D-3 L scores
f Values were obtained through author contact
g Survival was also calculated but not mentioned in the methods as an outcome