Table 3 RCT Experiment Outcomes. MOCA = Montreal Cognitive Assessment, PDQ-39 = Parkinsons Disease Questionnaire, TUG = Time Up and Go Test. Results are presented as mean (standard deviation) unless stated otherwise. Outcomes are adimensional where no units are stated

Pompeu et al. (2012) [35]

• UPDRS-II Score (ADL)

• BBS

• MOCA Test (Cognition)

• RM-ANOVA

• 8.9 (2.9) / 10.1 (3.8)

• 51.9 (4.6) / 52.9 (4.1)

• 21.7 (4.6) / 20.6 (4.5)

• 7.6 (2.9) / 8.1 (3.5)

• 53.1 (3.4) / 54.4 (2.2)

• 23.1 (4.6) / 22.2 (4.5)

• 2 months FU

• 8.1 (3.2) / 8.3 (3.6)

• 53.1 (3.1) / 54.1 (2.0)

• 23.3 (3.4) / 21.8 (4.5)

Post-hoc Tukey tests comparing before and after training comparing control and intervention not statistically significant.

No significant group differences between intervention and control groups

Exergames as effective as traditional balance therapy

Allen et al. (2017) [36]

• Nine hole peg test (s)

• Horizontal Tapping test (taps/60s)

• Horizontal Tapping test (error score)

• ANCOVA

• 28.8 (5.7) / 29.9 (7.3)

• 124.1 (34.9) / 119.0 (29.4)

• 0.047 (0.064) / 0.048 (0.042)

• 29.0 (7.8) / 30.4 (7.5)

• 130.1 (30.4) / 114.6 (26.3)

• 0.070 (0.059) / 0.041 (0.037)

Two-sided t-test, intervention minus control values not statistically significant for nine hole peg test (p = 0.84), statistically significant for horizontal tapping test (p = 0.006) and error score (p = 0.02)

No significant group differences between intervention and control groups in primary or secondary outcome. Improvements in some areas, decline on others

Exergames should consider task specificity

Liao et al. (2015) [37]

• Obstacle Crossing Performance speed (cm/s)

• TUG (s)

• PDQ-39

• ANOVA

• 80.4 (16.1) / 77.5 (21.8) / 75.2 (11.4)

• 11.9 (2.7) / 12.1 (2.1) / 12.6 (4.1)

• 78.2 (23.3) / 82.2 (27.3) / 84.5 (26.0)

• 78.5 (17.0) / 85.8 (18.0) / 87.0 (16.5)

• 12.6 (3.6) / 11.0 (1.8) / 9.7 (2.1)

• 79.0 (24.3) / 70.8 (27.1) / 68.8 (20.0)

• 1 month FU

• 78.2 (17.3) / 84.7 (21.4) / 91.1 (20.0)

• 12.9 (3.8) / 10.7 (1.5) / 9.7 (2.3)

• 80.2 (24.5) / 70.0 (26.5) / 65.8 (18.3)

Statistically significant differences between intervention and control groups for obstacle crossing at post (p < 0.001) and follow-up (p < 0.001)

Statistically significant differences between intervention and control groups for TUG at post (p < 0.001) and follow-up (p < 0.001)

Statistically significant differences between intervention and control groups for PDQ-39 at post (p = 0.004) and follow-up (p = 0.001) but no differences between treadmill and Wii groups.

Wii group shows greater improvement than the other intervention group and control group

Significant improvement of patients with Wii training

Shih et al. (2016) [38]

• BBS

• TUG (s)

• LOS Reaction time (s)

• 50.9 (5.32) / 50.4(4.79)

• 9.5 (2.45) / 10.05 (4.66)

• 0.96 (0.33) / 0.88 (0.24)

• 53.2 (2.86) / 53 (1.89)

• 8.71 (1.8) / 9.18 (3.42)

• 0.74 (0.24) / 0.79 (0.18)

t-test statistically significant for the control and intervention groups in BBS and TUG, and only in the intervention group for LOS.

Intervention group shows better results than control group, both having positive effects.

Exergaming at least as effective as traditional balance therapy

Ribas et al. (2017) [39]

• BBS

• Fatigue severity scale

• 6MWT(m)

• RM-ANOVA

• 48.4 (2.63) / 50.4 (2.79)

• 3.55 (1.68) / 3.80 (1.66)

• 384 (86.43) / 352 (91.99)

• 48.2 (2.89) / 52.3 (2.26)

• 3.02 (1.22) / 1.83 (0.57)

• 437 (89.69) / 408 (97.27)

• 2 months FU

• 46.9 (2.72) / 47.7 (4.80)

• 3.23 (1.31) / 3.05 (1.11)

• 392 (80.24) / 376 (98.68)

Post-hoc Bonferroni tests before vs. after for intervention p = 0.033, for control p < 0.001, before vs. Follow-up p = 0.022 for intervention, p = 0.037 for control.

Increase in balance for the intervention group, not sustained at follow-up. Significant improvement in fatigue for both groups, not sustained at follow-up

Exergames seem to empower motivation and achieve significant results in balance. Future studies should study the effect of exergames on fall risk.

Zimmermann et al. (2014) [40]

• Neurophysiological tests for alertness, working memory, executive function (Cognition)

• MANOVA

• 272 / 291 (Median)

• −0.16 / -0-05 (Median)

• 2.3 / 2–17 (Median)

• 266 / 275 (Median)

• − 0.14 / -0.16 (Median)

• 2.44 / 2.37 (Median)

t-test statistically significant for alertness (p = 0.024) but not significant for working memory (p = 0.431) or executive function (p = 0.462).

Better performance in attention in intervention than in control group

Non-cognitive-specific exergame therapy may deliver the same degree of cognitive benefit than cognition-specific computerized training

Song l et al. (2017) [41]

• Stepping performance CSRT

• TUG (s)

• MOCA Test (Cognition)

• ANCOVA

• 847 (221) / 824(176)

• 9.51 (2.27) / 9.57 (2.38)

• 26.5 (2.70) / 26.4 (2.77)

• 794 (88) / 798 (169)

• 9.02 (1.70) / 9.72 (2.14)

• 26.7 (2.3) / 27.3 (2.8)

Difference between groups not statistically significant for CSRT (p = 0.59), and MOCA (p = 0.69). Difference statistically significant for TUG (p = 0.02).

No effect in step performance, mobility, muscle power, cognition, reaction time, FOG, or falls

Task-specific training may be required. Sessions may need to be longer. Further research with severe PD patients required. Ensuring safety in a home scenario may hamper effective results.

Ferraz et al. (2018) [42]

• 6MWT (m)

• 10MWT (s)

• PDQ-39

• ANOVA

• 354.9 (98.9) / 405.2 (97.3) / 365.4 (81.1)

• 1.3 (0.3) / 1.3 (0.3) / 1.2 (0.3)

• 47 (25.1) / 38.1 (19.8) / 44.7 (26.7)

• 1 month FU

• 391.7 (107.5) / 440.2 (90.2) / 401.2 (77.9)

• 1.4 (0.4) / 1.4 (0.3) / 1.4 (0.3)

• 41.7 (21.7) / 32.9 (19.1) / 33.9 (25.2)

t-tests pre- vs. post intervention statistically significant for control (p = 0.08), bicycle (p = 0.001) and exergaming (p = 0.005) groups for 6MWT. For 10MWT, not significant (p = 0.068, p = 0.101) for control and bicycle groups but significant (p = 0.011) in exergaming group. For PDQ-39 not significant (p = 0.069, p = 0.185) for control and bicycle groups but significant (p = 0.004) in exergame group.

All groups show significant improvement in 6MWT, only exergame group shows significant improvement in 10MWT. PDQ improvement not significant.

Exergame training has similar outcomes to functional training and bicycle exercise, all therapies present improvements on walking capacity, ability to stand up, sit and functionality.

Tollar et al. (2018) [43]

• UPDRS-II

• BBS

• 6MWT (m)

• ANOVA

• 19.0(4.67)/18.2(3.85)/18.9(3.11)

• 26.3(5.21)/23.6(3.60)/22.7(4.24)

• 270.2(90.66)/204.6(34.94)/222.4(40.85)

• 18.9(2.19)/13.7(2.45)/15.7(2.59)

• 24.9(5.91)/32.4(4.61)/26.9(4.17)

• 253.9(81.61)/334.2(68.90)/364(51.53)

Tukey’s post hoc contrast.

Significant (p < 0.05) improvement in all outcomes except UPDRS, with no difference between exergame and cycling groups. Improvement in UPDRS is present (−4.5 points for exergame and − 3.2 for cycle) but not significant.

Exergame therapy has similar outcomes to bicycle training