This is consistent with previous studies reporting that falls are a common problem after stroke (Stolze et al 2004, Lamb et al 2003, Ramnemark et al 1998). Our data may also be an underestimate as we used retrospective recall rather than monthly calendars, which are the gold standard for falls data. The high proportion of fallers is likely to be a reflection

of poor recovery in terms of walking speed. A recent study by Tiedemann and colleagues (2008) suggested that a walking speed of less than 1 m/s was a predictor of multiple falls in community dwelling older persons. Using this criterion, 94% of our entire sample was at risk of multiple Procaspase activation falls. There are several limitations to our study. First, as in most clinical trials of complex interventions, we were unable to blind therapists, and patients cannot be blinded, creating a potential source of bias. In addition, the high levels of disability and co-morbidities resulted in an incomplete dataset, eg, cognitive and language impairments often meant that it was not possible for questionnaires to be completed. In conclusion, Fluorouracil cost analysis of the secondary outcomes

of the MOBILISE trial, measured six months after entry to the study, demonstrates that treadmill walking with body weight support results in a greater walking capacity and higher perception of walking ability six months after commencement of training compared with overground walking. There is no evidence to suggest that treadmill walking with body weight support has a deleterious effect on walking quality. Clinicians should therefore feel confident about implementing this intervention. eAddenda: Appendix 1, Table 3 available at jop.physiotherapy.asn.au Ethics: Sydney University Human Research Ethics

Committee (08-2002/2916), Melbourne University Human Research ethics Committee (HREC No. 050881), Human Research Ethics committees from the following sites: Kingston Centre (Research Project Application No. 06018B), Sydney South West Area Health Service (Project no. 2007/066), South Eastern Sydney & Illawarra Area Health Service: Eastern section (Ref no. 98/043) / Southern section (Ref no. 02/79Ada), Royal Rehabilitation Centre Sydney (Research project 02/08) and Sydney West CYTH4 Area Health Service (Reference no. 2004/8/4.9 (1923)) approved this study. All participants gave informed consent before data collection began. Competing interests: None declared. Support: This study was supported by a University of Sydney Sesquicentenary Grant and an NHMRC (Australia) Project Grant (no. 402679). Over 60 people assisted in this project and we would like to thank and acknowledge the physiotherapy staff of Prince of Wales Hospital, St George Hospital, Blacktown and Mount Druitt Hospitals, Bankstown Hospital, Royal Ryde Rehabilitation Centre, and the Kingston Centre.