In May 1991 an immunization programme against Haemophilus influenzae type b (Hib) infection began within the Oxford region. We validate a deterministic mathematical model of Hib by comparison with the incidence of disease in the Oxford region, 1985–97. The comparison of model results with observed outcome allows an exploration of some of the poorly understood properties of the immunization programme. Model results and observed incidence are consistent with a vaccine that blocks the acquisition of carriage. Similarly, the data suggest that factors other than experience of Hib carriage are likely to have generated acquired immunity to Hib disease prior to the introduction of vaccination. Hence it is unlikely that waning of vaccine-derived protection will result in a resurgence of disease. The inclusion in the immunization schedule of a booster dose, as used in other countries, would have provided very little extra benefit.