Under the term "big data", the secondary use of available IT-based clinical routine data promises to easily support quality management and research. Available clinical data can, for example, be used to measure care quality, as this is an important aspect in the evaluation of health interventions. Different indicators are available for measuring care quality, e.g. nursing sensitive outcome indicators. However, a specific methodology for using clinical routine data to measure quality of nursing care is not available at the moment. We present first ideas for a model to use clinical routine data to derive nursing outcome indicators measuring quality of care. In particular, we are interested in the question of which care indicators can be extracted to predict care outcome.