Study	Reason for exclusion
Ada 2003	The intervention included more than 50% training on the treadmill, so that overground gait training was not a majority of the treatment
Andersen 2000	Participants had a stroke within previous 6 months
Caldwell 2000	Not a RCT (case series with 1 or 2 participants per group)
Ceceli 1996	The intervention included high‐technology equipment not readily available in a community PT setting
Chen 2006	The intervention included numerous components and overground gait training was only one of them, less than 50% of the total
DePaul 2007	Some participants had a stroke within previous 6 months All groups participated in walking exercises
Dickstein 1986	Participants had a stroke within previous 6 months
Duncan 2003	Participants had a stroke within previous 6 months
Duncan 2007	Not a RCT (protocol for a RCT; intervention used body weight supported treadmill training, not overground gait training)
Eich 2004	Participants had a stroke within previous 6 months All groups participated in walking exercises
Flansbjer 2008	Experimental intervention consisted of strength training and stretching; did not include overground gait training
Ford 2007	Not a RCT (no control group)
GAPS 2004	Participants had a stroke within previous 6 months All groups participated in walking exercises
Green 2004	This study combined the results of two other studies
Hesse 1994	Not a RCT (single group repeated measures design)
Hesse 1995	Not a RCT (single group repeated measures design)
Hesse 1998	Not a RCT (single group repeated measures design)
Hesse 1999	Not a RCT (single group repeated measures design)
Hesse 2001	Not a RCT (survey design)
Hesse 2003	Not a RCT (review)
Hesse 2005	Participants had a stroke within previous 6 months All groups participated in walking exercises
Horn 2005	Not a RCT (cohort design)
Ivey 2007	Experimental intervention was a fitness programme and did not include overground gait training Measured physiological variables such as insulin response and glucose tolerance
Kalra 1993	Intervention was placement into stroke unit or general hospital setting, not a specific therapeutic protocol
Kilbreath 2006	Not a RCT (repeated measures cohort design)
Kilbreath 2006b	Intervention included high‐technology equipment not readily available in a community PT setting
Kollen 2005	Not a RCT (cohort design)
Kwakkel 1999	Participants had a stroke within previous 6 months
Kwakkel 2002	Participants had a stroke within previous 6 months
Langhammer 2000	Participants had a stroke within previous 6 months All groups participated in walking exercises
Langhammer 2007	All groups participated in walking exercises The experimental treatment was focused on high intensity exercise for endurance, strength, and balance rather than on overground gait training per se The control group received functional task‐oriented training during the first 3 months and standard PT treatment for the remaining 9 months as needed
Langhammer 2008	Participants had acute stroke
Latham 2005	Not a RCT (survey design)
Laufer 2001	Participants had a stroke within previous 6 months The control group participated in walking exercises
Lee 2008	Intervention consisted of progressive resistance training or aerobic cycling, not overground gait training
Lennon 2006	Participant had acute stroke No control group
Liston 2000	Not a RCT (single group cross‐over trial)
Logigian 1983	Included only participants with acute stroke The control group participated in walking exercises
Luft 2008	All groups participated in walking exercises
Lynch 2007	Intervention consisted of sensory retraining, not overground gait training
Macko 2005	All groups participated in walking exercises
Mandel 1990	Intervention for experimental group included high‐technology equipment not available in a typical community PT setting
Mauritz 2004	Not a RCT (review paper)
Mayr 2007	All but one participant had a stroke less than 6 months prior to the study beginning Both groups received physical therapy gait training
McClellan 2004	The experimental treatment was primarily a home exercise programme
Mead 2007	The experimental treatment focused on progressive endurance and resistance training and did not involve overground gait training
Montoya 1994	Intervention for experimental group included high technology equipment not available in a typical community PT setting
Morris 1992	Intervention for experimental group included high technology equipment not available in a typical community PT setting
Olney 1991	Not a RCT (review paper)
Olney 2006	The experimental treatment provided strength and conditioning training rather than a overground gait training The control group participated in walking exercises at home
Peurala 2007	Not a RCT (cohort design)
Plummer 2007	Not a RCT (feasibility study with 7 participants)
Pohl 2002	Participants had a stroke within previous 6 months All groups participated in walking exercises
Pohl 2007	Participants had a stroke within previous 6 months All groups participated in walking exercises
Pyoria 2007	The control group received overground physical therapy treatment and thus included a substantial amount of gait training
Rimmer 2000	The experimental treatment provided strength and conditioning training rather than providing overground gait training
Rodriquez 1996	Not a RCT (case series)
Roerdink 2007	Not a RCT (compared stroke patients to healthy controls)
Schauer 2003	The intervention included high‐technology equipment not readily available in a community PT setting
Smith 1981	Participants had a stroke within previous 6 months The experimental treatment compared placement into a stroke unit or a general medical unit
Smith 1982	Participants had a stroke within previous 6 months The experimental treatment compared placement into a stroke unit or a general medical unit
Stephenson 2000	Not a RCT (survey study)
Stern 1970	Participants had a stroke within previous 6 months
Stibrant Sunnerhagen 2007	Experimental intervention focused on muscle strengthening and muscle endurance using a circuit training methods, and did not involve overground gait training Measured physiological variables like oxygen consumption and muscle enzymes
Sullivan 2007	All 3 interventions included high‐technology equipment (body weight supported treadmill training) not readily available in a community PT setting
Sutbeyaz 2007	Both groups performed ankle dorsiflexion movements, not overground gait training
Tangeman 1990	Not a RCT (single group repeated measures design)
Teixeira 1999	The experimental treatment was a strength and conditioning programme rather overground gait training
Teixeira 2001	Not a RCT (single group repeated measures design)
Thaut 1997	Participants had a stroke within previous 6 months Neither group acted as a non‐gait control group
Treger 2004	The experimental treatment was a home‐based walking programme rather than gait training
van Vliet 2005	Participants had a stroke within previous 6 months
Waagfjord 1990	Not a RCT (case series)
Wagenaar 1990	Not a RCT (case series)
Weng 2006	Neither group acted as a non‐gait control group
Werner 1996	The experimental treatment included 50% OT and 50% PT Of the PT services, gait training accounted for only a portion Thus, overall, gait training did not account for 50% of the experimental treatment
Werner 2007	Not a RCT (cohort design)
Wolfe 2000	The experimental treatment was home‐based intervention by a rehabilitation team Fewer than 50% of the participants received gait training or physical therapy services
Yang 2005	All groups participated in walking exercises
Yavuzer 2007	Participants had a stroke within previous 6 months, and the experimental treatment did not consist of overground gait training
Yen 2008	No non‐gait control group Both groups received general physical therapy
Young 1992	Participants had a stroke within previous 6 months

OT: occupational therapy 
 PT: physical therapy 
 RCT: randomised controlled trial