Please click here to obtain permission to reproduce this image.
Click on image to view larger version.
Figure 2 Bayesian analysis of the HOPE-2 trial. Triplots showing posterior (thick line) distributions derived from integrating evidence or likelihood (thin line) from the HOPE-2 trial and informative priors (dashed line) based on information derived from the NORVIT study (only data from comparison of combination therapy group vs. placebo are utilized to match the HOPE-2 trial treatment groups; see Table 3 for details) according to Bayes’ theorem (44). Posterior probabilities are estimated using priors for the primary composite end point of death from cardiovascular cause, myocardial infarction (MI), and stroke (log odds ratio µ = 0.206, standard deviation  = 0.115), death from any cause (µ = 0.183, = 0.153), myocardial infarction (µ = 0.219, = 0.122), and stroke (µ = –0.246, = 0.293). Probability of any effect size is calculated by computing area under the curve. The probabilities of benefit (Pb, log odds ratio <0) or harm (Ph, log odds ratio >0) are shown on the right of each plot. Superiority or inferiority is inferred at a posterior probability of benefit or harm >0.950.
|
