Identifying variation in GP referral rates: an observational study of outpatient headache referrals.

Objective: This study provides a framework methodology for identifying GP surgeries with unexpected rates of referral to specialist services, using headache referrals to outpatient neurology as an example. Design: This is a retrospective observational study using routinely collected and open-source data. Setting and participants: Data was collected from a single consultant outpatient neurology clinic and 202 GP surgeries across seven CCGs in the Northwest of England. Primary Outcome: The number of headache referrals from each GP surgery during the study period of three and a quarter years was used as the primary outcome in a generalised linear model. The standardised residuals from this model were then used to identify GP surgeries that were likely to have referred unexpected patient numbers for headaches to an outpatient neurology clinic during the study period. Results: In the model using data from the CCG in which the outpatient neurology clinic is located we identified one GP surgery referring more headache patients than expected. The model showed that the clearest predictor of headache referrals was the number of referrals for other types of neurological disorders. In the model using data from all seven CCGs we identified four GP surgeries with unexpected numbers of referrals. This model showed that there were two predictors of headache referral, namely other neurology referrals and the distance from the clinic. Conclusion: We have developed a flexible methodology for identifying GP surgeries with unexpected numbers of referrals to specialist services. This methodology was demonstrated using headache referrals but could be adapted to any type of referral or geographical area.