To assess the impact of patient mix on general practitioners' work-load within a list patient system, a "key" was constructed for computing expected use of general practitioners' services by patients on a doctor's list. The key was based on empirically observed variations in consultation rates for different age and sex-groups. Computations were made for 90 general practitioners in Trondheim. Expected patient weight varied in accordance with the proportion of women and of old patients on the doctor's list. The mean expected use of consultations was approx. 6% higher for patients on female doctor's lists relative to patients on male doctors' lists. The difference between the patients of male and female doctors respectively was far less than previously expected. This is explained by the fact that female doctors - compared with their male colleagues - usually have more children and fewer old people on their list of patients.