For a seven year period (1985-91) clinical and epidemiological data were prospectively collected on children aged < 10 years with microbiologically confirmed invasive Haemophilus influenzae type b infection in the Oxford region to study the epidemiology of the disease and determine the potential impact of early primary immunisation in infants. Computer records of primary immunisations given to these cases were retrospectively analysed and, where necessary, hospital and general practitioner records were searched to determine the immunisation history. Over the seven year period, 416 cases of invasive H influenzae type b disease were reported. Widescale immunisation against H influenzae type b began in 1991 as part of a regional trial. The estimated annual incidence for invasive disease between 1985 and 1990 was 35.5 cases per 100,000 children aged less than 5 years; for H influenzae type b meningitis it was 25.1 per 100,000 children aged less than 5 years. The cumulative risks for invasive disease and meningitis by the fifth birthday were one in 560 and one in 800 respectively. The majority of disease (71%) occurred in children less than 2 years of age with the peak monthly incidences at 6 and 7 months of age. The overall mortality was 4.3% and 50% of these deaths occurred suddenly. Most (91%) of the children had received at least one primary immunisation against diphtheria, tetanus, and pertussis before H influenzae type b infection and there was only one case of parental refusal of immunisation. None had received H influenzae type b immunisation. Given a vaccine uptake of 90% by 5 months of age it is estimated that at least 82% of the H influenzae type b infections could have been prevented. Extrapolated nationally, 1150 cases of infection and 50 deaths could be prevented each year by routine primary immunisation.