Objective: To better understand the reasons for lack of impact of sending feedback on prescribing pattern to general practitioners (GPs).
Methods: Semi-structured interviews with GPs who had all participated in intervention studies addressing effects of prescriber feedback. Interviews were audiotaped and fully transcribed. Transcripts were studied repeatedly and coded into categories in order to produce meaningful patterns.
Results: None of the GPs believed they altered prescribing practice after they received prescriber feedback. Unsolicited prescriber feedback was perceived as violating the GPs' autonomy. The GPs wanted to decide for themselves what data should be sent to them. Aggregated data were difficult to interpret, and GPs did not regard it as a problem that their practice pattern deviated from that of other practices. There was a mistrust of the validity of data, and the GPs wanted to be able to identify the patients with a need for optimised therapy directly from the information provided in the prescriber feedback. In addition they wanted advice on how to optimise therapy.
Conclusion: Postal prescriber feedback (not revealing the patients' identities) is not effective because it does not motivate GPs to change nor does it address the barriers to change. Prescriber feedback requested by the GPs may be more effective, however, if it includes identities of inappropriately treated patients combined with relevant advice on how to optimise prescribing.