Background and objectives: To investigate the underlying reasons for variability in the incidence rate of amyotrophic lateral sclerosis (ALS) within the Irish population between the years 1996 and 2021.
Methods: The Irish ALS register was used to calculate the incidence and to subsequently extract age at diagnosis (age), year of diagnosis (period), and date of birth (cohort) for all incident patients within the study period (n = 2,771). An age-period-cohort (APC) model using partial least squares regression was constructed to examine each component separately and their respective contribution to the incidence while minimizing the well-known identifiability problem of APC effects. A dummy regression model consisting of 5 periods, 19 cohorts, and 16 age groups was used to examine nonlinear relationships within the data over time. The CIs for each of these were estimated using the jackknife method.
Results: The nonlinear model achieved R2 of 99.43% with 2-component extraction. Age variation was evident with those in the ages 65-79 years contributing significantly to the incidence (βmax = 0.0746, SE = 0.000410, CI 0.00665-0.00826). However, those aged 25-60 years contributed significantly less (βmin = -0.00393, SE = 0.000291, CI -0.00454 to -0.00340). Each successive period showed an increase in the regression model coefficient suggesting an increasing incidence over time, independent of the other factors examined-an increase of β from -0.00489 (SE = 0.000264, CI -0.00541 to -0.00437) to 0.00973 (SE = 0.000418, CI 0.0105-0.00891). A cohort effect was demonstrated showing that the contribution of those born between 1927 and 1951 contributed to a significantly greater degree than the other birth cohorts (βmax = 0.00577, SE = 0.000432, CI 0.00493-0.00662).
Discussion: Using the Irish population-based ALS Register, robust age, period, and cohort effects can be identified. The age effect may be accounted for by demographic shifts within the population. Changes in disease categorization, competing risks of death, and improved surveillance may account for period effects. The cohort effect may reflect lifestyle and environmental factors associated with the challenging economic circumstances in Ireland between 1927 and 1951. Age-period-cohort studies can help to account for changes in disease incidence and prevalence, providing additional insights into likely demographic and environmental factors that influence population-based disease risk.