Physical Restraint and the Therapeutic Relationship Medium secure psychiatric services in the UK are designed to fulfil a therapeutic function in addition to those of security and custodianship (Mason et al., 2009). The development of strong nurse-patient therapeutic relationships is seen as essential in developing an effective therapeutic milieu in such environments (Thomas, Shattell, & Martin, 2002). However, in addition to such a therapeutic role, nurses who work in secure settings also have to deal with security issues such as compulsory detention, forced treatment, and risk to others (Mason, Lovell, & Coyle, 2008). It has been questioned whether staff who work in such services can fulfill both security and therapeutic roles, or whether these roles are incompatible. Physical restraint activities may be used in a forensic environment as part of routine care, but concerns have been raised about how restraint fits with ethical practice (Mohr, 2010). In addition to being the one who employs the restraint techniques, it is also usually nursing staff who make the decision of whether the patient is restrained. The emotional effect of being involved in restraint or nurses being seen as being part of the system of detention may damage any staff / patient therapeutic alliance (Steele, 1993;. Conacher, 1993). The literature on the impact of physical restraint on the therapeutic relationship is limited. Research has mainly focused on nursing staff views of physical restraint. For example, Lee, Gray, Gournay, Wright, Parr and Sayer (2003) found that nurses reported that physical restraint was useful in regaining control after an incident, but expressed feelings of ambivalence and concern regarding the after-effects of the restraint. Bigwood and Crowe (2008 found that nurses felt a conflict between their therapeutic role and the milieu of control in the service. They expressed feeling better about physically restraining a patient if they were able to undertake the restraint in a therapeutic way and if they thought that all other options had been explored. Perkins, Prosser, Riley and Whittington (2012) found that nurses
