Objectives To examine the role of microsystem characteristics in the translation of an evidence-based intervention (the Diabetes Prevention Initiative (DPI)) into practice in a community-health centre (CHC).

Design Case study.

Analysis Constant comparative method of qualitative analysis.

Setting Community-health centre in a mid-sized city in the USA.

Participants 27 administrators, clinicians and staff of a community-health centre implementing a DPI.

Main outcome measures Perceptions of microsystem characteristics that influence the implementation of this initiative.

Results Five characteristics of high-performing microsystems were reflected, but not maximised, in the implementation of the DPI. First, there was no universally shared definition of the desired purpose of the DPI. Second, investment in quality improvement (QI) was strong, yet sustainability remained a concern, since efforts were dependent upon external grant support. Third, lack of cohesiveness between the initiative planning team and the rest of the organisation served to both facilitate and constrain implementation. Fourth, administrators showed both support for new initiatives and a lack of strategic vision for QI. Fifth, this initiative substantially strained already-stretched role definitions.

Conclusions Translation of the DPI in this CHC was constrained by the lack of a cohesive QI infrastructure and incomplete alignment with characteristics of high-performing microsystems. The findings suggest an important role for microsystem characteristics in the process of implementing evidence-based interventions. Enhancing the level of microsystem performance of CHCs is essential to informing efforts to improve quality of care in this critical safety-net system.