Home Economics Applied Surrogate Endpoint Evaluation Methods with SAS and R
Data Analysis and Output
The macro %CONTCONTFULL produces default numerical and graphical output. The first part of the output, shown in Figure 12.3, consists of two descriptive plots, which show the distribution of the patients in the trials by treatment arm (Figure 12.3, left panel) and a scatter plot between the true and the surrogate endpoints (Figure 12.3, right panel).
As shown in Figure 12.4, for the ARMD study, individual and trial-level surrogacy are equal, respectively, to R2ndiv = 0.4866 (0.3814, 0.5919) and Rtrial = 0.7031 (0.5333, 0.8730). Both surrogacy measures indicate that the
Surrogacy measures with their 95% confidence intervals (C.I.).
change in visual acuity after 24 weeks is a surrogate of moderate value for the visual acuity at 52 weeks after starting the interferon treatment.
The results for the leave-one-out sensitivity analysis are presented in Figure 12.5. Note how the value of Д/На1 is dropping when center 13830 is left out (Rtriai13830^ = 0.6295). We can see in Figure 12.3 (left panel) that this center is the one with highest number of patients in both treatment arms. Leaving out this center has an impact on the estimated trial-level surrogacy, since it has the highest weight.
A table containing the surrogacy measures estimated using the leave-one- out sensitivity analysis is presented in Figure 12.6.
|< Prev||CONTENTS||Next >|