Abstract
Referral rates and urgency of referral to hospital by six pairs of trainers and trainees were compared over one month in 1986. From 5846 consultations, 327 resulted in referral to hospital. Overall, the referral rates was low at 5.6 per 100 consultations; however, there was considerable variation in individual referral rates with trainees tending to follow their trainer's referral patterns, despite statistical adjustment for sharing the same practice. This paper critically examines the wisdom of using general practitioner referral rates for the purpose of resource allocation and education.