Injury and ill health rates over time are examined using smoothing techniques which aim to reduce irregularities (random fluctuations) in the time series. A Generalised Additive Model (GAM), which is an extension of a Generalised Linear Model (GLM) with ‘smooth’ terms has been used, where the smoothed term is the year. The modelling was completed using the mixed generalised additive vehicle package for GAMs, supplied with the software R1.
A downward/upward trend is recorded when two conditions are met: a smooth model can be fitted to the data and the end points of the fitted model are statistically significantly different. Where no downward/upward trend is indicated this implies that one of these conditions has not been met.
1. Wood, S N (2006): Generalised Additive Models: An introduction to R.