This paper analyses the 'failure' of a patient safety intervention. Our study was part of a randomised controlled trial (RCT) of bed and bedside chair pressure sensors linked to radio pagers to prevent bedside falls in older people admitted to hospital. We use agential realism within science and technology studies to examine the fall and its prevention as a situated phenomenon of knowledge that is made and unmade through intra-actions between environment, culture, humans and technologies. We show that neither the intervention (the pressure sensor system), nor the outcome (fall prevention) could be disentangled from the broader sociomaterial context of the ward, the patients, the nurses and (especially) their work through the RCT. We argue that the RCT design, by virtue of its unacknowledged assumptions, played a part in creating the negative findings. The study also raises wider questions about the kind of subjectivities, agencies and power relations these entanglements might effect and (re)produce in the hospital ward.We are using a performative perspective situated in science and technology studies (Barad 2007, Latour 2005, Mol 2002. The main insight of this tradition is that technologies are not neutral objects, and that they have an active existence, and meanings, within the sociomaterial