Introduction: Despite considerable efforts directed at developing international evidence based guidelines to improve clinical management, adoption of evidence based practices can be poor in low-income settings including Kenya. Studies in Africa rarely consider the implementation and change processes as influenced by the structural and organizational context in which clinicians are embedded nor how these can influence performance. This thesis builds on existing literature and theory on behavioural change, clinician-managers’ identity construction processes and contextualized leadership processes by examining these and their effect on guideline adoption in the complex contexts of Kenyan county hospitals. Methods: Methodologically I explored these issues through qualitative ethnographic approaches using in-depth interviews, focus group discussions and non-participant observations. I analyzed data inductively and deductively borrowing from the grounded theory approach to develop plausible explanations of collated data and observations. Results: Early work indicated limited attention to local dissemination of the new guidelines and poor leadership in implementing Evidence Based Medicine (EBM) as key barriers. However, specially introduced ‘study facilitators’ as part of an intervention study emerged as leaders of change often acting as role models, friendly supervisors and peer educators to facilitate EBM implementation. Further work reviewing literature on the roles of clinical mid-level managers (MLMs; department leaders) in improving service delivery emphasized the importance of ‘soft skills’ e.g. building interpersonal relationships, mentoring, coaching and effective communication skills. Subsequent in-depth empiric work on identity transitions of these clinical MLMs indicated that ‘identity work’, drawing on competing professional and managerial institutional logics resulted into ‘willing’, ‘ambivalent’ and ‘reluctant’ hybrids. Distributed leadership by hybrids was undermined by existing hierarchical professional autonomy and cadre delineations between nurses and doctors in the public county hospitals we studied. Discussion: The thesis describes both a set of work and a research journey. My initial work was predominantly based on applying the Theory of Planned Behaviour to explain behaviour of front-line health workers. However, it quickly became clear that this provided only a partial understanding of guideline adoption within a hospital overlooking the pivotal role of clinical team leaders / in influencing change. There emerged valuable lessons for current Kenyan leadership and management development programmes which are likely to be transferable to other African health systems. Particular recommendations from this work are the importance of a focus on the soft-skills of those stepping into clinical hybrid manager roles and considering the ‘practical norms’ of Kenyan public hospitals in understanding the gap between desired official institutional norms and health workers actual behavioural practices.