In France, 97% of children under age of 3 turn to a GP at least once a year. The GP/child encounter benefits from a particular status -no gate keeper constraint, baby extra fee- even if neither the nature nor the economic burden of this activity have been yet studied. OBJECTIVES: This study aimed to: 1/ analyse the nature of the activity for children in General Practice 2/ find out the determinants of the prescription decision. METHODS: We got data from a representative sample of 922 French GPs (BKL-Thalès panel). Information was directly collected on the basis of 60 consecutives visits per GP, through a patient management software. We only picked children visits. In addition to standard GPs and patients characteristics, other data were also collected on the child: health insurance status, status in the GP's practice: regular / new) and on the visit (length, reasons for encounter, prescription of drugs…). We use the Hierarchical Linear Models to identify the determinants of the prescription at the respective level of children and GPs. RESULTS: Activity is highly concentrated: Upper Respiratory Tract Infections represent 54.8% of the 6652 visits. Obviously URTI favours prescription decision by 3.96(OR) [p<0.0001] and more generally an acute pathology by 3.26(OR) [p<0.0001]. GPs prescribe on average 2 medicines per visit (even if 25.1% did not lead to any prescription) but surprisingly male prescribe much more than female: 2.01(OR) [p<0.0062]. Long visits (over 15 minutes) lead more often to a drugs prescription: 1.80(OR) [p<0.0001]. Furthermore, a child under age of 6 is more likely to receive a drugs prescription 1.66(OR) [p<0.0001]. CONCLUSION: The diagnoses and the child age clearly explain drugs prescription and unexpectedly GPs sex takes significantly part in the decision of prescription.