The aim of this thesis is to study whether the Swedish state’s efforts to govern the professional practice of social workers (SWs) by knowledge to achieve an evidence-based practice, address the difficulties that SWs encounter in their practice. The Swedish state’s ambition is to govern the practice of social work by ideas and methods originating from medicine. The thesis therefore compares, through four different studies, the professional practice of SWs with general practitioners (GPs) as the most comparable sub-profession in medicine. The research questions that the thesis seeks to answer are: What situations are perceived as problematic and non-problematic by SWs and GPs? How does the professionals’ relationship with clients affect the performative aspects of these occupations’ enactment of their practice? How do SW and GPs experience the state’s efforts to govern their professional practice by knowledge? What problems can arise when the Swedish state builds its knowledge governance of social work practice on a comparison with medicine, even though these professional fields exhibit a number of important differences? The empirical material in the first and second studies consists of 52 narratives, written by SWs and GPs on problematic and non-problematic situations. The third study is based on vignette-based focus group interviews with 25 SWs and GPs. And the fourth study is based on a literature review, consisting of 54 articles about SWs’ and GPs’ practice. A first finding is that professionals rarely describe lack of knowledge or difficulties with knowledge use as problematic. A second finding is that non-problematic situations are connected to professionals’ control of the intervention process. SWs gained control of the relationship with the client either by use of coercive means or by the client’s active co-operation. GPs gained control of the intervention process by the use of professional knowledge. A third finding is that an understanding of professional practice as a linear process consisting of diagnosis, inference and treatment reflects GPs’ practice but not all aspects of SWs’ practice. A fourth finding is that both professions hold ambivalent positions towards evidence-based risk reductions technologies. The responses towards organisational risk reductions technologies differ in that GPs are sceptical, whilst SWs take a more pragmatic view. A main conclusion is that the Swedish state’s efforts to govern the performative aspects of SWs’ practice by knowledge, runs the risk of becoming a insufficient strategy since they tend to exhibit two main blind spots. The first is that these efforts tend to ignore that the relationship with the client has a crucial affect on and conditions SWs’ and in problematic situations also GPs’ knowledge use. The second is that the efforts also tend to ignore that control of the intervention process is not always maintained through control of knowledge use. For SWs control of the relationship with clients represents another mode of professional control of the intervention process. / Hur använder socialarbetare och allmänläkare kunskap i praktiken?, FAS/FORTE dnr 2009-0251