Secular trends were considered in the evaluation of universal hepatitis B vaccination in Taiwan

Objectives: The objective was to evaluate time-trend bias in the context of a series of studies reporting that a national hepatitis B virus vaccination program (launched in mid-1980s) substantially reduced childhood hepatocellular carcinoma (HCC) incidence. Study Design and Setting: We applied an age-period-cohort model to evaluate the relative importance of age, time-trend (period), and vaccination (cohort) effect, respectively, on the incidence and mortality rate of HCC in boys and girls in Taiwan from 1980 to 2009. Results: HCC incidence increased with age. The period effect analysis revealed that the incidence of HCC started to decrease in 1980s, leveled off in mid-1990s, and declined again in mid-2000s among boys. The period effect was flat among girls. Cohort effect analysis demonstrated that among boys, the incidence of HCC started to decrease by those born in 2000-2004, which was 15 years later than the first vaccinated cohort. Among girls, the incidence rate started to decline before the mass vaccination program was initiated. The analysis showed a decline in mortality for boys and girls born in 1980s. Conclusion: Time trend may play a more important role than the universal vaccination program in interpretation of the observed early decreasing trend in HCC in children, especially among boys. (C) 2015 Elsevier Inc. All rights reserved.