Generalized additive models (GAMs) investigating the shape of the relationship between CFR of 28 human infectious diseases and age.

For each model, we report the results for the linear (regression model analysis) and the non-linear (smoothing model analysis) components. The sign of the estimate indicates whether the linear trend was positive or negative. The models were run using a binomial distribution of errors (number of deaths/number of cases). Diseases are ordered by increasing values of CFR (deaths/cases). The column Replicates indicates the number of replicated datasets per disease. After a sequential Bonferroni correction, the non-linear trend for Western equine encephalitis was not statistically significant.