Public health surveillance is the ongoing systematic collection, analysis, interpretation, and dissemination of health data for the planning, implementation, and evaluation of public health action. To achieve effective public health interventions, it is pivotal to analyse and interpret the vast amounts of data collected by surveillance systems to enable good understanding of all factors having an impact on health. For example, we can consider child protection, which is an important public health issue. Often relatively extensive data exist on families in official statistics, research reports, social services reports, school and medical records, etc. However, these data are dispersed and hard if not impossible to relate and compare. This leads to numerous interventions that are conducted without adequate knowledge of target families that are to benefit from these interventions. Inadequate knowledge also leads to lack of intervention where it is needed, such as undetected cases of child abuse. In many cases the basic data that are needed for intervention decisions exist, but are not available to decision makers due to inadequate communication and lack of data integration, analysis and interpretation. Chronic condition management is another area where extensive disparate data exist from statistics and various health services and intervention agencies. In this area too there are numerous organizations offering services and a great need to better coordinate these services to achieve better outcomes for patients and also to reduce soaring costs of the healthcare system. We propose DIONE (Decision Intelligence for Organizations in Networked Environments), a decision support system that uses a complex systems approach to offer a means to integrate and maximally exploit all available data to optimize intervention decisions. It will transform dispersed data into relevant information for decision makers. The proposed system has its place on top of existing research and data collection, rather than duplicating any existing research, and will optimize utilization of existing data and research results. DIONE will enable more focussed health and social service interventions to encourage and help people to manage life stages and events, in view of achieving the best possible health and behaviour outcomes. The system will use data mining and analysis techniques such as predictive analytics, text analytics, data mining, machine learning, and Bayesian statistics to integrate and exploit all available quantitative and qualitative data from heterogeneous data sources, such as statistics, social services, schools, medical services, etc., to identify risk factors and people most at risk. This will enable more focussed health and behaviour interventions. The problem with risk factors for health outcomes is that they are in general hard or impossible to observe directly. The proposed data mining techniques will be calibrated with synthetic data, obtained from agent-based simulation or micro-simulation models of the populations to be stud...