The Implications of Complex Systems Science on Clinical Trials Practice

Abstract

Although the application of Complex Systems Science methods is fairly well-established in several Physical and Life Sciences (e.g. network analysis, multi-level simulation in Systems Biology and Nanotechnology), this perspective has far deeper implications that should be transforming the way we understand, report and analyse our experimental data. On the whole, experiments (including medical trials) tend to be reported in a “hypothesis supported/refuted” format, where some metric or statistical test is used to support a hypothesis given in cause-effect terms, for example, treatment A is shown to be effective against X because most of the individuals given A improved with respect to X while those who were not given A did not. While this is indeed valid and useful, we also seem to be missing the opportunity to understand more about what makes A effective. The Complex Systems paradigm suggests that non-linearity and apparently categorical responses (e.g. effective against X vs not effective against X) arise because systems under different conditions will be driven to different attractors, yet medical trials and biochemical experiments are often reported assuming a simple mechanistic view of the world.