Efforts to model the health effects of low-dose ionizing radiation (IR) have often focused on cancer. Meanwhile, significant evidence links IR and age-associated non-cancer diseases. Modeling of such complex processes, which are not currently well understood, is a challenging problem. In this paper we briefly overview recent successful attempts to model cancer on a population level and propose how those models may be adapted to include the impact of IR and to describe complex non-cancer diseases. We propose three classes of models which we believe are well suited for the analysis of the health effects in human populations exposed to low-dose IR. These models use biostatistical/epidemiological techniques and mathematical formulas describing the biological mechanisms of the impact of IR on human health. They can combine data from multiple sources and from distinct levels of biological/population organization. The proposed models are intrinsically multivariate and non-linear and capture the dynamic aspects of health change.