Peurala 2009.

Methods	RCT Method of randomisation: sealed envelopes (stratified according to ability to walk) Blinding of outcome assessors: no Adverse events: 2 in treatment group A, 3 in control group Deaths: 1 in control group Dropouts: 5 in treatment group A, 1 in treatment group B, 3 in control group ITT analysis: not stated
Participants	Country: Finland 56 participants (22 in treatment group A, 21 in treatment group B, 13 in control group) Non‐ambulatory at start of study Mean age: 68 years Inclusion criteria: first supratentorial stroke or no significant disturbance from an earlier stroke (Modified Rankin Scale 0 to 2), acute phase after stroke with a maximum duration of 10 days, FAC 0 to 3, voluntary movement in the leg of the affected side, Barthel Index 25 to 75 points, age 18 to 85 years, no unstable cardiovascular disease, body mass index < 32, no severe malposition of joints, no severe cognitive or communicative disorders Exclusion criteria: not stated
Interventions	Between June 2003 and December 2004, random allocation to 2 arms took place (2 walking exercise groups) Gait training with Gait Trainer device (GT‐Group) Overground walking training (WALK‐Group) All participants received 55 minutes daily gait‐oriented physiotherapy and additional gait training for 15 sessions over 3 weeks (each session lasting maximum of 20 minutes of walking). Between January 2005 and February 2007, random allocation to 3 arms took place (3 walking exercise groups) Gait training with Gait Trainer device (GT‐Group) Overground walking training (WALK‐Group) Control group (CT‐Group) All participants received 55 minutes daily gait‐oriented physiotherapy and additional gait training for 15 sessions over 3 weeks (each session lasting maximum of 20 minutes of walking). However, CT‐Group received 1 or 2 physiotherapy sessions daily but not at the same intensity as in the other groups.
Outcomes	Outcomes were recorded at baseline and after 3 weeks and 6 months: FAC 10‐metre walk test 6‐minute walk test Modified Motor Assessment Scale Rivermead Motor Assessment Scale Rivermead Mobility Index
Notes	Because we were interested in the effects of automated electromechanical‐ and robotic‐assisted gait‐training devices for improving walking after stroke, we combined the results of the CT‐Group and the WALK‐Group as one group, which we compared with the results from the GT‐Group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	An investigator not involved in the study randomly assigned participants to groups using concealed envelopes.
Allocation concealment (selection bias)	Low risk	Allocation was performed by an independent person not otherwise involved with the participants.
Blinding of outcome assessment (detection bias) All outcomes	High risk	No
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Unclear whether reasons for missing outcome data are unlikely to be related to true outcome