BACKGROUND: Although risk assessment charts have been proposed to identify patients at high cardiovascular risk, in everyday practice general practitioners (GPs) often use their knowledge of the patients to estimate the risk subjectively. DESIGN: A cross-sectional study aimed to describe how GPs perceive, qualify and grade cardiovascular risk in everyday practice. METHODS: General practitioners had to identify in a random sample of 10% of their contacts the first 20 consecutive patients perceived as being at cardiovascular risk. For each patient essential data were collected on clinical history, physical examination and laboratory tests, for the qualification of risk. At the end of the process GPs subjectively estimated the overall patient's level of risk. General practitioners grading was compared with the risk estimate from a reference chart. RESULTS: Over a mean time of 25 days 3120 patients perceived as being at cardiovascular risk were enrolled. According to the inclusion scheme each GP had contact with more than 200 patients at cardiovascular risk every month. Thirty percent of these patients had atherosclerotic diseases. Up to 72% of patients without any history of atherosclerotic diseases but perceived to be at risk could be classified according to a reference chart as being at moderate to very high risk. Comparing GPs' grading of risk with a chart estimate there was agreement in 42% of the cases. Major determinants of GPs' underestimation of risk were age, sex and smoking habits, while obesity and family history were independently associated with overestimation. CONCLUSIONS: On the basis of their perception GPs properly identify patients at cardiovascular risk in the majority of cases. General practitioners subjective grading of risk level only partially agreed with that given by a chart.