BACKGROUND: Increasing consultation rates have implications for the organisation of health services, the quality of care, and understanding the decision to consult. Most quantitative studies have concentrated on very high attenders--not those attending five or more times a year, who are responsible for most (60%) consultations--and have assessed neither the role of lifestyle nor patients' attitudes. AIMS: To assess associations with higher than average attendance (five or more times ayear). DESIGN OF STUDY: Postal questionnaire sent to a random sample. SETTING: Four thousand adults (one per household) from six general practices. METHOD: Data were analysed to identify predictors significantly associated with higher than average attendance. RESULTS: The response rate was 74%. Sef-reported attendance agreed with the notes (r = 0.80, likelihood ratio for a positive test = 9.4). Higher attendance was independently predicted by the severity of ill health (COOP score = 0-7, 8-9, and 10+; adjusted odds ratios= 1, 1.72, 1.91 respectively; test for trend P<0.001) and the number of reported medical problems (COOP score = 0, 1, 2, and 3+ respectively; adjusted ORs = 1, 2.05, 2.31, 4.29; P<0.001). After controlling for sociodemographic variables, medical problems, the severity of physical ill health, and other confounders, high attendance was more likely in those with medically unexplained somatic symptoms (0, 1-2, 3-5, and 6+ symptoms respectively, ORs = 1, 1.15, 1.48, and 1.62; P<0.001); health anxiety (Whitely Index = 0, 1-5, 6-7 and 8+ respectively, ORs = 1, 1.22, 1.77, and 2.78; P<0.001); and poor perceived health ('very good', 'good', 'poor' respectively, ORs = 1, 1.61, and 2.93; P<0.001). Attendance was less likely in those with negative attitudes to repeated surgery use (OR = 0.61, 95% CI = 0.47-0.78), or to doctors (Negdoc scale <18, 18-20, and 21+ respectively; ORs = 1, 0.87, 0.67; P<0.001), in those usually trying the pharmacy first (OR = 0.61, 95% CI 0.48-0.78), and those consuming alcohol (0, 1, 2, 3+ units/day respectively; ORs = 1, 0.62, 0.41, 0.29; P<0.001). Anxiety or depression predicted perceived health, unexplained symptoms, and health anxiety. CONCLUSION: Strategies to manage somatic symptoms, health anxiety, dealing with the causes of--or treating--anxiety and depression, and encouraging use of the pharmacy have the potential both to help patients manage symptoms and in the decision to consult. Sensitivity to the psychological factors contributing to the decision to consult should help doctors achieve a better shared understanding with their patients and help inform appropriate treatment strategies.