summary. The long-term efficacy of a childhood hepatitis B vaccination programme was evaluated. A total of 112 newborn babies of hepatitis B carrier mothers were given hepatitis B immune globulin (HBIG) and a 10-μg three-dose regimen of plasma-derived vaccine administered at a conventional (0, 1, 6 months), delayed (2, 3, 8 months) or accelerated (0, 1, 2 months) schedule. The vaccinees were followed up to determine their anti-HBs status over a 16-year period. Upon completion of the vaccination schedules, 92.6% developed antibody against surface antigen (anti-HBs) seroconverion, the rate of which fell to 33.3% at year 16. The three schedules were equally effective in preventing chronic infection, with a protective efficacy of 88.9% from hepatitis B surface antigen (HBsAg) carriage, compared with historical control. Vaccinees on the delayed schedule had a slightly higher seroconversion rate over years, and were better able to maintain an anti-HBs level of ≥ 100 iu/L. Overall, a quarter demonstrated evidence of exposure to the virus, being positive for antibody against core antigen or HBsAg, or mounting a rise in anti-HBs during the follow-up period. We conclude that a three-dose hepatitis B vaccination regimen is generally effective in protecting newborns of hepatitis B carrier mothers from infection and chronic carriage. Booster is not needed even after 16 years of monitoring.