Orthogonalized Regression for Causal Inference: A Pharmacoepidemiologist's Dream?

Author: Nuvan Rathnayaka

Challenge

Orthogonalized regression methods had emerged in statistics and machine learning but had not been evaluated from a pharmacoepidemiologist's perspective for applicability to standard drug effectiveness questions.

Solution

Target RWE researchers presented an oral critical evaluation of orthogonalized regression for causal inference in pharmacoepidemiology settings, assessing its advantages, limitations, and practical conditions under which it outperforms or underperforms standard propensity score methods.

Impact

Critically evaluating emerging causal inference methods positions Target RWE as a rigorous and discerning methodologist, building pharma partner and regulatory confidence in the company's analytic decision-making.