*Discount rates are 6% per year for costs and 1.5% per year for benefits. All event rates (normalised to 1 patient) and all costs (normalised to 1 patient) are at five years.

Each value of the relative risk refers to adjunctive drug (considered in each individual row of table) v placebo. Contradiction arises because relative risks (obtained from either clinical studies or meta-analyses) compare drug under examination (adjunctive drug) with placebo, whereas cost effectiveness assessment in same row compares combination with absence of adjunctive drug v combination with presence of adjunctive drug (with no comparison v placebo). Hence, comparators in relative risk assessment (adjunctive drug v placebo) differ from those used for cost effectiveness comparison (absence v presence of adjunctive drug). Contradiction is only apparent because relative risks are applied in a chain of subsequent multiplications throughout the table. For example, if cost effectiveness comparison of aspirin v aspirin+diuretic and b blocker appears with relative risk of placebo v diuretic and b blocker (RR=0.83), there is no error because relative risk of aspirin v placebo (RR=0.75) has already been incorporated as multiplicative factor in previous row and its presence affects all subsequent rows (including comparison of aspirin v aspirin+diuretic and b blocker).