One of the key challenges in the development of personalized medicine is the enormous genetic and phenotypic heterogeneity of individuals. Most of the currently used classification criteria depend on information derived from population averages. Personalized predictions for disease progression in an individual are very difficult to make. Concepts from physics and ecology may provide new approaches to get insight into the nonlinear behaviour of biological systems and to develop new predictive and interventive tools for medical applications. Systems biology approaches, e.g. the development of thermodynamically inspired dynamic network biomarkers that make use of not only absolute values but also on fluctuations and correllations will be discussed and examples of succesfull applications presented.