DISCUSSION
Demonstrated with the identified studies, staff training was an essential part of a program for reducing physical restraint. Noticeably, for the purpose of enhancing the effectiveness in restraint reducing, Stewart et al. (2010) suggested staff training should be implemented together with other strategies, including changing local policy, review procedure, crisis management, the PRN medication, patient education, violence awareness, and risk assessment. In terms of staff training, lecture learning and seminar were two of the most widely employed approaches. However, staff training emphasized cultivating nurses' capability of providing health education in specific topic which required professional knowledge in mental health nursing, such as emotion management, aggressive behavior management, and stress coping strategy (Pellfolk, Gustafson, Bucht, & Karlsson, 2010). In addition to nurses' capability, strengthening the awareness of risk and reflections of clinical dynamic scenario were other indispensable elements in training programme; and seminar and group discussion were main approaches to achieve these goals (Putkonen et al., 2013). This review has following limitations. Firstly, the identified studies were randomized controlled trials and quasi-experimental studies. However, the evidence of quasi-experimental studies provided practical but not convincing argument in relation to the proposed PICO question. To summarize the identified evidence above, in relation to the randomized controlled trials, despite the rigor in randomization, the design of every trial were reasonable and acceptable. The measurements of patient and staff at baseline and follow-up were absent. However it was preferable but unrealistic to collect the characteristic of participants as it consumes great resources. On the other hand, the trials concerning sta ff training should be conducted at organizational level, meaning all the members within an organization would take part in its implementation.
