I have a set of data about the incidence of certain types of infections that occur in hospitals. The analysis has looked at the incidence trend for all aggregated results in a time series. This uses a GAM fitted spline to identify the incidence trend for aggregated data.
Over the past year the incidence trend has increased sharply for the aggregated data set. The corresponding incidence trends for each participating hospital do not look much different.
Is there a way to test a hypothesis that a specific individual hospital or hospitals drive the increase in the aggregated incidence?