Background: Primary care is recognised as a medical specialty and its unique information needs justify the existence of its own health informatics sub-specialty: primary care informatics (PCI). A challenge for PCI is how to raise the standard of computerised medical records so that meaningful conclusions can be drawn from them. In the UK the Primary Care Data Quality (PCDQ) programme has eight years experience of using feedback in an educational context to improve data quality.

Objective: This literature review set out to define the characteristics of a feedback process most likely to achieve change; the principles of which could be applied to PCDQ or to other data quality initiatives.

Method: A literature review of the major medical bibliographical databases, and the websites and working groups of the international medical informatics associations. Results: There are generalisable lessons for primary care derived from the literature about implementing best evidence, feedback and the theory of diffusion of innovation. The principles identified are: (1) Engage and support local innovators – i.e. those most likely to adopt change, demonstrate the evidence-base for the intervention and the form of feedback most acceptable to them (2) Model the clinical context in which quality improvement is required; (3) Develop an understanding of the health system, its culture and management system; and, (4) Identify and address technical issues relating to computer use and coding.

Conclusions: Feedback is most effective when: clinically relevant, educationally orientated, given by peers, and sensitive to the socio-technical context.