Background: Much of European primary care is computerised and many groups of practices pool data for research. Technology is making pooled general practice data widely available beyond the domain within which it is collected.
Objective: To explore the barriers and opportunities to exploiting routinely collected general practice data for research.
Method: Workshop, led by primary care and informatics academics experienced at working with clinical data from large databases, involving 23 delegates from eight countries. Email comments about the write-up from participants.
Outputs: The components of an effective process are: the input of those who have a detailed understanding of the context in which the data were recorded; an assessment of the validity of these data and any denominator used; creation of anonymised unique identifiers for each patient which can be decoded within the contributing practices; data must be traceable back to the patient record from which it was extracted; archiving of the queries, the look-up tables of any coding systems used and the ethical constraints which govern the use of the data.
Conclusions: Explicit statements are needed to explain the source, context of recording, validity check and processing method of any routinely collected data used in research. Data lacking detailed methodological descriptors should not be published.