An efficient algorithm to assess multivariate surrogate endpoints in a causal inference framework

Multivariate surrogate endpoints can improve the efficiency of drug development process, but their evaluation raises many challenges. Recently, so-called individual causal association (ICA) has been introduced for validation purposes in causal-inference paradigm. The ICA is a function partially identifiable correlation matrix (R) and, hence, it cannot be estimated without making untestable assumptions. This issue addressed via simulation-based analysis. Essentially, assessed across set values non-identifiable entries R that lead to valid and this implemented using fast algorithm based on partial correlations (PC). Using theoretical arguments simulations, shown that, spite its computational efficiency, PC may spurious effect adding non-informative surrogates, i.e., surrogates convey no information treatment true endpoint, seemingly reduces range. To address this, modified (MPC) proposed. Based MPC removes nuisance increases efficiency.