Yt = | |
α1 | Baseline level term |
+β1T1t | Baseline trend term (preslope) |
+α2I2t+β2 | T2t changes in terms after the beginning of the 1st multifaceted intervention |
+α3I3t+β3 | T3t changes in terms after the end of the 1st intervention |
+α4I4t+β4 | T4t changes in terms after the 2nd, less extensive intervention |
+ϵt | residual term |
Yt is the adjusted time series; ϵt are the residuals; α1 estimates the mean number of PPI consumptions per 10 000 patient-days before the first intervention; β1 estimates the baseline trend for PPI consumption; α2, α3 and α4 estimate, respectively, the change in the mean monthly consumption per 10 000 patient-days immediately after the first, the second and the third segment (ie, after the beginning of the first intervention, after the end of the first intervention, and after the second intervention); I2t, I3t and I4t are indicator variables which are set to null before the second, third and fourth segment (ie, before June 2003, July 2004 and February 2006), respectively; β2, β3 and β4 estimate the change in trend in the mean number of PPI consumptions per 10 000 patient-days that occurred after the first, second and third segment (compared with the monthly trend of the previous segment), respectively; T1t, T2t, T3t and T4t are continuous variables counting the number of months at time t from the start of the observation period, after the first segment (June 2003), second segment (July 2004) and third segment (February 2006), respectively.