A quantitative approach to the spread of variance in translational research using Monte Carlo simulation

Abstract

The translation of promising preclinical research into successful trials often fails. One contributing factor is the “Princess and the Pea” problem, which refers to how an initially significant effect size dissipates as research transitions to more complex systems. This work aimed to quantify the effects of spreading variability on sample size requirements. Sample size estimates were performed by Monte Carlo simulation. To simulate the process of progressing from preclinical to clinical studies, nested sigmoidal dose–response transformations with modifiable input parameter variability were used. The results demonstrated that adding variabilty to the dose–response parameters substantially increases sample size requirements compared to standared calculations. Increasing the number of consecutive studies further increases the sample size. These results quantitatively demonstrate how the spread of variability in translational research, which is not typically accounted for, can result in drastic increases in the sample size required to maintain a desired study power.