BACKGROUND: Age-period-cohort (APC) analysis could provide useful insight into the interpretation of time trends for disease rates. This study aimed to determine whether the patterns of age, period and cohort effects on epilepsy mortality in Taiwan were similar to those in the US and in England and Wales. METHODS: We employed a three-phase APC analysis method developed by Keyes and Li,(7) which conceptualizes the cohort effect as a partial interaction between age and period. The effects of age, period, and cohort were indicated by the relative risk (RR) of mortality rates. RESULTS: We found an increase in mortality rates in the elderly from 1981-85 to 2001-05 and no decline in mortality rates for children, adolescents and adults from 1991-95 to 2001-05. According to APC analysis, the RRs increased sharply during adolescence, and leveled off in adulthood, and then increased steeply in the elderly. The RR of period effects decreased from period 1971-1975 to period 1991-1995 and then increased. With the exception of cohort 1911-1915, no significant cohort effects were found. CONCLUSION: Despite the limitations of official published mortality data, we still found prominent age effects according to APC analysis, which were similar to the results of a previous study. However, the patterns of period and cohort effects in relation to epilepsy mortality in Taiwan differed from those in England and Wales and in the US.
BACKGROUND: Age-period-cohort (APC) analysis could provide useful insight into the interpretation of time trends for disease rates. This study aimed to determine whether the patterns of age, period and cohort effects on epilepsy mortality in Taiwan were similar to those in the US and in England and Wales. METHODS: We employed a three-phase APC analysis method developed by Keyes and Li,(7) which conceptualizes the cohort effect as a partial interaction between age and period. The effects of age, period, and cohort were indicated by the relative risk (RR) of mortality rates. RESULTS: We found an increase in mortality rates in the elderly from 1981-85 to 2001-05 and no decline in mortality rates for children, adolescents and adults from 1991-95 to 2001-05. According to APC analysis, the RRs increased sharply during adolescence, and leveled off in adulthood, and then increased steeply in the elderly. The RR of period effects decreased from period 1971-1975 to period 1991-1995 and then increased. With the exception of cohort 1911-1915, no significant cohort effects were found. CONCLUSION: Despite the limitations of official published mortality data, we still found prominent age effects according to APC analysis, which were similar to the results of a previous study. However, the patterns of period and cohort effects in relation to epilepsy mortality in Taiwan differed from those in England and Wales and in the US.