Limited evidence suggests that delayed prescribing may influence future consultation behaviour. To assess the effects of antibiotic prescribing strategy on reconsultation in the year following presentation with acute lower respiratory tract infection (LRTI). Balanced factorial randomised trial. Primary care. Eight hundred and seven subjects, aged>or=3 years, had acute illness presenting with cough as the main symptom, plus at least one symptom or sign from sputum, chest pain, dyspnoea or wheeze. The subjects were randomised to one of three prescribing strategies (antibiotics, delayed antibiotic, no antibiotic) and a leaflet. Prior antibiotic use and reconsultation were assessed by medical record review. Patients who had been prescribed antibiotic for cough in the previous 2 years were much more likely to reconsult (incidence rate ratio [IRR]=2.55, 95% confidence interval [CI]=1.62 to 4.01) and use of a delayed prescription strategy is associated with reduced reconsultation in this group. In those with prior antibiotic exposure, there was a 34% reduction in consultation rate in the no antibiotic group (IRR=0.66, 0.30 to 1.44, P=0.295) and a 78% reduction for the delayed antibiotic group (IRR=0.22, 0.10 to 0.49, P<0.001) when compared with those given immediate antibiotics. This effect was not observed in patients who had not been prescribed antibiotics in the previous 2 years; there was no reduction in consultations in the no antibiotic group (IRR=1.23, 0.79 to 1.92, P=0.358) or the delayed antibiotic group (1.19, 0.78 to 1.80, P=0.426). There was an increase in consultation rate with an information leaflet (IRR=1.27, 0.86 to 1.87, P=0.229). Past attendance with cough, or past attendance with other respiratory illness and smoking, also predicted reconsultation with cough. Delayed antibiotic prescribing for LRTI appears effective in modifying reconsultation behaviour, particularly in those with a prior history of antibiotic prescription for LRTI.