Allen 2010.

Methods	Parallel group design. Randomised using a randomisation schedule with randomly permuted block sizes, developed by an investigator not involved in subject recruitment or assessment. Data analysed on an intention‐to‐treat basis. Treated as outpatients and at home for 48‐72 hours over 6 months. Assessed at baseline and post intervention. Assessors were blinded.
Participants	24 participants in the exercise group and 24 in the control group. 3 drop‐outs in the exercise group. Participants' mean age was 66 years (exercise) and 68 years (control); male/female 13/11 (exercise) and 13/11 (control); duration of PD 7 years (exercise) and 9 years (control). Hoehn and Yahr stage not reported. Inclusion criteria: diagnosis of idiopathic Parkinson's disease, able to walk independently (with or without an aid), fallen in the last year or deemed to be at risk of falling, 30‐80 years of age, and on the same PD medication for the past 2 weeks. Exclusion criteria: significant cognitive impairment (Mini Mental State Examination [MMSE] <24), had another neurological/musculoskeletal/cardiopulmonary/metabolic condition that would interfere with safe conduct of the training or testing protocol.
Interventions	Exercise: 40‐ to 60‐minute program of progressive lower limb strengthening and balance exercises (targeted leg muscle strength, balance, and freezing). Once‐monthly exercise classes, with the remaining exercise sessions at home. Control: usual care with advice on fall prevention and falls diary recording any fall. Drug therapy was allowed to vary.
Outcomes	PD falls risk score. Knee extensor strength. Coordinated stability. Sway. Maximum balance range in standing. Alternate step test. Freezing of Gait Questionnaire. Sit‐to‐stand time. Fast walking speed. Comfortable walking speed. Short physical performance battery. Falls Efficacy Scale － International. PDQ‐39. Participants were assessed in their home about 1 hour after taking their usual PD medication, and the order of measurements was standardised.
Notes	Participants in the exercise group who experienced freezing of gait were also instructed in cueing strategies to reduce freezing as part of their exercise program. Exercise group completed a mean of 70% of total prescribed exercise sessions.
Risk of bias
Bias	Authors' judgement	Support for judgement
Eligibility Criteria	Low risk
Randomisation Method	Low risk	Randomly permuted block size.
Concealment of Allocation	Unclear risk	No information provided.
Similarity at Baseline	Low risk
Withdrawals Described	Low risk	6% overall, but all from exercise group.
Intention To Treat Analysis	Low risk	An intention‐to‐treat approach was used for all analyses.
Cointerventions Constant	Unclear risk	Allowed variation in levodopa therapy.
Credible Placebo	Low risk	Falls prevention advice given in both arms.
Blinded Assessors	Low risk	Assessors were blinded to group allocation.