Characterizing the Distribution of Multiple Negative Control Outcome Risk Differences

Multiple NCOs are frequently analyzed simultaneously to assess unmeasured confounding, but the statistical distribution of NCO risk differences under different confounding scenarios had not been characterized.

Target RWE researchers presented an oral analysis characterizing the distribution of multiple NCO risk differences under varying confounding levels, providing a statistical framework for interpreting multi-NCO results and distinguishing true confounding signals from chance variation.

Providing a statistical framework for interpreting multiple NCO results enables Target RWE and pharma partners to make more calibrated decisions about whether NCO findings indicate actionable bias.

Multi-NCO risk difference distribution characterization for calibrated confounding interpretation in RWE programs, Statistical framework for NCO gating decisions in regulatory-grade comparative effectiveness studies