Ashburn 2007.

Methods	Parallel group design. Stratified by NHS using blocks of size four. Random allocation by telephoning the medical statistics group at University of Southampton. Participants were informed of their allocation by telephone. Data analysed on an intention‐to‐treat basis. Treated as outpatients 7 times a week for a 6‐week period, for a total period of 42 hours. Assessed at baseline, 8 weeks, and 6 months.
Participants	70 participants in the exercise group and 72 in the control group. 6 dropouts in the exercise group and 8 in the control group. Participants' mean age 72.7 years (exercise), 71.6 years (control); male/female 38/32 (exercise), 48/24 (control); Hoehn and Yahr stage 3.14 (exercise), 3.11 (control); duration of PD 7.7 years (exercise), 9 years (control). Inclusion criteria: confirmed diagnosis of Parkinson's disease, independently mobile, living at home in the community, experienced more than one fall in the previous 12 months, passed a screening for gross cognitive impairment (Mini‐Mental State). Exclusion criteria: unable to participate in assessments because of pain, acute medical condition, in receipt of or soon to receive treatment.
Interventions	Exercise: personalised home‐based exercise and strategy programme. After assessment, treatment goals were established with participants, and exercises from the exercise menu were taught. Participants were visited weekly at home by a physiotherapist for approximately 1 hour. 6 levels of exercise progression comprised muscle strengthening, range of movement, balance training, and walking. Strategies of falls prevention and movement initiation and compensation taught by physiotherapist. Participants were asked to complete the exercises daily for max of 1 hour and to keep a record. Phoned monthly to encourage exercises. Control: usual care, contact with local Parkinson's disease nurse. Drug therapy was not described.
Outcomes	Self‐reported falls diary. Functional reach. Timed up and go test. Chair stand test. Berg balance test. Euroqol‐5d, QoL thermometer. Self‐assessment Parkinson's disease disability scale. Tests were carried out midway between drug doses.
Notes	At 6 months, 34% in the control group were participating in extra rehabilitation compared with 25% in the exercise group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Eligibility Criteria	Low risk
Randomisation Method	Low risk	Block randomisation (block size 4).
Concealment of Allocation	Low risk	Telephone call to central office.
Similarity at Baseline	Low risk
Withdrawals Described	Low risk	6% at 8 weeks and 8% at 6 months.
Intention To Treat Analysis	Low risk	Analysis was on an intention‐to‐treat basis.
Cointerventions Constant	Unclear risk	Drug therapy was not described.
Credible Placebo	Low risk	Controls had contact with Parkinson's disease nurse.
Blinded Assessors	Unclear risk	Assessor remained blind to group allocation but reported being aware of the allocation of 18 exercise and 11 control participants at 8 weeks, and 25 exercise and 14 control participants at 6 months.