BACKGROUND: The effectiveness of an evidence-based health care intervention depends on it being delivered consistently to appropriate patients. Brief alcohol intervention is known to be effective at reducing excessive drinking and its concomitant health and social problems. However, a recent implementation trial reported partial delivery of brief alcohol intervention by general practitioners (GPs) which is likely to have reduced its impact. AIM: To investigate patient-practitioner characteristics influencing brief alcohol intervention in primary care. DESIGN OF STUDY: Cross-sectional analysis of 12,814 completed Alcohol Use Disorders Identification Test (AUDIT) screening questionnaires. SETTING: Eighty-four GPs who had implemented a brief alcohol intervention programme in a previous trial based in the Northeast of England. METHOD: GPs were requested to screen all adults (aged over 16 years) presenting to their surgery and follow a structured protocol to give a brief intervention (five minutes of advice plus an information booklet) to all 'risk' drinkers. Anonymized carbon copies of the screening questionnaire were collected from all practices after a three-month implementation period. RESULTS: Although AUDIT identified 4080 'risk' drinkers, only 2043 (50%) received brief intervention. Risk drinkers that were most likely to receive brief intervention were males (58%), unemployed (61%), and technically-trained patients (55%). Risk drinkers that were least likely to receive brief intervention were females (44%), students (38%), and university educated patients (46%). Logistic regression modelling showed that patients' risk status was the most influential predictor of brief intervention. Also, GPs' experience of relevant training and longer average practice consultations predicted brief intervention. However, personal characteristics relating to patients and GPs also predicted brief intervention in routine practice. CONCLUSION: Interpersonal factors relating to patients and practitioners contributed to the selective provision of brief alcohol intervention in primary care. Ways should be found to remedy this situation or the impact of this evidence-based intervention may be reduced when implemented in routine practice.