It has been suggested that complex human diseases can be understood by studying the effects of perturbations on the functioning of biochemical reaction networks (BRN) describing the intracellular processes. In such models, the inter-individual genetic and epigenetic differences impacting response to therapy could be taken into account at the level of network mechanistic details or parameters. In this context, we propose a novel computational approach based on tropical geometry to identify perturbations resulting in changes in qualitative dynamical properties of BRNs i.e. critical transitions. In situations where the underlying mechanistic model is unknown, we propose a fully data driven approach based on state space modelling approach to identify such critical transitions. We demonstrate the challenges and recent advances in this topic by analysing well studied BRNs as well as clinical multivariate time series data obtained from the Intensive Care Units.