OBJECTIVE--To determine the incidence of haemorrhagic disease of the newborn in the British Isles, study risk factors, and examine the effect of vitamin K prophylaxis. DESIGN--Prospective survey of all possible cases of haemorrhagic disease of the newborn as reported by consultant paediatricians using the monthly notification cards of the British Paediatric Surveillance Unit and a follow up questionnaire for each case to validate the diagnosis and accrue further data. SETTING--Britain (England, Scotland, and Wales) and Ireland (Northern Ireland and the Irish Republic) during December 1987 to March 1990. PATIENTS--27 infants classified as having confirmed (n = 25) or probable (n = 2) haemorrhagic disease of the newborn. RESULTS--24 of the 27 infants were solely breast fed. 10 suffered intracranial haemorrhage; two of these died and there was clinical concern about the remainder. 20 infants had received no vitamin K prophylaxis, and seven had received oral prophylaxis. Relative risk ratios for these groups compared with babies who had received intramuscular vitamin K were 81:1 and 13:1 respectively. Six infants had hepatitis (alpha 1 antitrypsin deficiency in four), unsuspected until presentation with haemorrhagic disease of the newborn, of whom four had received oral prophylaxis. One other baby had prolonged jaundice. One mother had taken phenytoin during pregnancy. CONCLUSIONS--All newborn infants should receive vitamin K prophylaxis. Intramuscular vitamin K is more effective than oral prophylactic regimens currently used in the British Isles.