TY - ABST
T1 - Age-, Period- and Cohort modelling of Type 1 Diabetes Incidence in Europe among children age 0-14 years
AU - Ibfelt, Else Helene
AU - Carstensen, Bendix
AU - Dahlquist, G G
AU - Neu, A
AU - Rosenbauer, J
AU - Skrivarhaug, Torild
AU - Bratina, Natasa
AU - Rami-Merhar, Birgit
AU - Jarosz-Chobot, P
AU - Cinek, O
AU - Marčiulionytė, Dalė
AU - Ionescu-Tirgoviste, C
AU - deBeaufort, Carine E
AU - McNally , R
AU - Gillespie, K
AU - Gorus, Frans K
AU - Castell, C
AU - Konrad, Daniel
AU - Kocova, M
AU - Svensson, J
AU - Patterson, Christopher C
PY - 2019/10
Y1 - 2019/10
N2 - AimDescribe incidence trends in type 1 diabetes (T1D) for centers participating in EURODIAB 1989 to 2013 using fine modelling age-period and cohort models (APC-models).Methods19 centers contributed with data in one-year age classes of type 1 Diabetes cases age 0-14 years. Cases were validated using capture-recapture methods. There were 57,725 individual cases from 19 centres from 16 countries. We used age-period-cohort models with natural splines for the age, period and cohort effects, and calculated the overall drift for each country. Further a likelihood ratio test was used to test for non-linearity of period and cohort effects. We fitted a final age-cohort model for all centres with separate age-effect for boys and girls and separate cohort effect for each center.ResultsThere was no clear common cohort effect across centers, and the non-linear period-effects were small. The overall result was an increase in incidence with the highest increase observed in countries with former lowest incidence. During the entire period Sweden had the highest incidence and Macedonia had the lowest incidence. The male:female ratio changed with age, with a higher rate in boys in the youngest (< 5 years of age) and in the oldest (> 12 years of age). The peak incidence was at age 12 years for girls and at age 13 years for boys.ConclusionThe modelling in 1-year classes allowed us to show the changing male:female ratio with age. A common cohort effect across Europe could not be confirmed, but we found that centres with the former lowest incidence rates seemed to have the largest increase over time.
AB - AimDescribe incidence trends in type 1 diabetes (T1D) for centers participating in EURODIAB 1989 to 2013 using fine modelling age-period and cohort models (APC-models).Methods19 centers contributed with data in one-year age classes of type 1 Diabetes cases age 0-14 years. Cases were validated using capture-recapture methods. There were 57,725 individual cases from 19 centres from 16 countries. We used age-period-cohort models with natural splines for the age, period and cohort effects, and calculated the overall drift for each country. Further a likelihood ratio test was used to test for non-linearity of period and cohort effects. We fitted a final age-cohort model for all centres with separate age-effect for boys and girls and separate cohort effect for each center.ResultsThere was no clear common cohort effect across centers, and the non-linear period-effects were small. The overall result was an increase in incidence with the highest increase observed in countries with former lowest incidence. During the entire period Sweden had the highest incidence and Macedonia had the lowest incidence. The male:female ratio changed with age, with a higher rate in boys in the youngest (< 5 years of age) and in the oldest (> 12 years of age). The peak incidence was at age 12 years for girls and at age 13 years for boys.ConclusionThe modelling in 1-year classes allowed us to show the changing male:female ratio with age. A common cohort effect across Europe could not be confirmed, but we found that centres with the former lowest incidence rates seemed to have the largest increase over time.
M3 - Conference abstract for conference
T2 - ISPAD 2019 45th Annual Conference
Y2 - 30 October 2019 through 2 November 2019
ER -