Background Age-Period-Cohort (APC) models are frequently used in epidemiological studies to investigate temporal trends in demographic rates such as incidence and mortality. Here we apply the method of Partial Least Squares Regression (PLSR) to study APC trends in incidence of Amyotrophic Lateral Sclerosis (ALS) in Ireland from 1996–2013. ALS is a rare progressive neurodegenerative disease with an unknown cause. The crude incidence of ALS for those aged 18 or over is estimated to be approximately 2.7 per 100,000 person-years in Europe. A number of studies have suggested that the incidence of ALS may be changing. Incidence data for ALS in Ireland was compiled using a population based register and age-specific population estimates were taken from national records.
Methods A long with standing problem in conducting APC analysis is the issue of perfect collinearity amongst the variables age, period and cohort. An identifiability issue arises due to the intrinsic mathematical relation: age + cohort = period. This presents a methodological challenge, as standard regeression techniques cannot be applied. In recent years, PLSR has been applied to analyse APC effects in blood pressure, obesity and overall mortality, thereby providing simultaneous joint estimates of all three variables.
Results Our linear PLSR model found that the risk of ALS increased with age, as already well documented in the literature. In addition, a small negative effect for birth cohort with incidence of ALS was found, showing increased risk for older cohorts. Further analysis to investigate non-linear effects also revealed three of the oldest birth cohorts have a risk of ALS substantially higher than other cohorts.
Conclusion Through PLSR, which separates the joint effects of age, period and cohort in a single model, we have identified a number of birth cohorts at increased risk of developing ALS. Epidemiological investigations such as this may inform future research studies into the potential risks associated with ALS and other neurodegenrative diseases.