OBJECTIVE To compare outpatient referral patterns in fundholding and non-fundholding practices before and after the NHS reforms in April 1991. METHODS Prospective collection of data on general practitioners' referrals to specialist outpatient clinics between June 1990 and January 1994 and detailed comparisons of three phases--October 1990 to March 1991 (phase 1), October 1991 to March 1992 (phase 2), and October 1993 to January 1994 (phase 3). METHODS 10 first wave fundholding practices and six non-fundholding practices in the Oxford region. METHODS Patients referred to consultant out-patient clinics. RESULTS NHS referral rates increased in fundholding practices in phase 2 and phase 3 of the study by 8.1/1000 patients a year (95% confidence interval 5.7 to 10.5), an increase of 7.5% from phase 1 (107.3/1000) to phase 3 (115.4/1000). Non-fundholders' rates increased significantly, by 25.3/1000 patients (22.5-28.1), an increase of 26.6% from phase 1 (95.0/1000) to phase 3 (120.3/1000). The fundholders' referral rates to private clinics decreased by 8.8%, whereas those from non-fundholding practices increased by 12.2%. The proportion of referrals going outside district boundaries did not change significantly. Three of the four practices entering the third and fourth wave of fundholding increased their referral rates significantly in the year before becoming fundholders. CONCLUSIONS No evidence existed that budgetary pressures caused first wave fundholders to reduce referral rates, although the method of budget allocation may have encouraged general practitioners to inflate their referral rates in the preparatory year. Despite investment in new practice based facilities, no evidence yet exists that fundholding encourages a shift away from specialist care.