The interesting article about GP referral rates by Rowlands and Lewis ('Risky business', page 28, 31 May) makes clear that, as with many other subjects, we need a rounded view of what is happening before drawing conclusions.

They imply, correctly I believe, that we need to be careful about making value judgements on high or low-referring GPs. There may be valid differences because of the needs of their patients, the skills of the GP, or the services available within the practice.

It may also be that the GP is using other treatment modes - for example, for musculoskeletal conditions we have compared GP usage of non-steroidal drugs, physiotherapy, and orthopaedic referrals. They presumably used a referral rate per 1,000 population as a comparator. But other measures would help us to flesh out the picture - for example, a referral rate per 1000 GP consultations, and differences between GPs in the did not attend rate of conversion to a waiting list for surgical specialties.

Steve Elliott Commissioning manager Derbyshire Dales and South Derbyshire primary care group Repton Derbyshire