Strengths of realistic evaluation:
Our approach deviated in some respects from the approach recommended by Pawson et al. 139. For example, Pawson et al. 139 do not advocate a comprehensive literature search and data extraction, while I used rigorous search strategy reviewed by a health-sciences librarian and extracted the data by organized literature matrix developed according to the needed information to answers the research questions. Moreover, data analysis relied on well-known frameworks such as using the Role Domains Model for Knowledge Brokering 90 to report the roles performed by OL and KB, using EPOC taxonomy 99 to report types of KT interventions associated with OL and KB interventions, and lastly using TICD checklist 100 and ICF 101 to report types of outcomes of OL and KB interventions. A comprehensive search strategy and theoretical frameworks increase the quality of this research.
Limitations of realistic evaluation:
This review was limited to interventions delivered to rehabilitation practitioners, so results of this review cannot be taken as evidence for the other health care disciplines. In addition, relevant stakeholders should have been involved to reflect on formulating the research questions, provide input throughout the research process, and assist in disseminating research results; however, stakeholders have not been involved in this research as this work was done by one person (DG) because of the comprehensive exam rules. Notwithstanding its limitations, this research provided an essential piece of knowledge to the researchers, practitioners, and administrators working in rehabilitation sectors.
The current research is the first realist evaluation of human agents in rehabilitation. This research was conducted based on previous research recommendations, which demanded to better understand human agents’ context, mechanism, and expected outcomes 41. This current research describes the available evidence about contexts in which human agents have a better impact, mechanisms through which human agents can properly function, and outcomes addressed in the literature. In general, it was relatively difficult to reach a definite conclusion about which features can lead to better outcomes because of data heterogeneity. However, our research emphasized that human agents who were adequately skillful, embedded in the organization, and properly identified to play this role (context); and who played the required roles with a suitable mode of delivery and parameter (mechanism), had greater potential to achieve targeted goals (outcome). Moreover, this research highlights a number of challenges and provides directions to inform future research to enrich evidence about human agents’ intervention as promising methods to promote the use of research evidence in practice. This research informs many stakeholders such as researcher who plan to conduct KT trials using human agents’ interventions were informed how they can maximize human agents’ impact; employers who intend to employ human agents in rehabilitation settings were informed about the possible roles and activities that human agents can perform as well as recommended parameter of OL and KB interventions and expected outcomes; and human agents themselves were informed about how can they properly do their job to achieved their target during their work period. In general, this research can improve human agents’ utilization in rehabilitation.