Constructing a practical medical data analysis system for real-life scenario remains challenging. While the development of various medical data analysis algorithms is rapidly advancing, it has not been matched by the development of a feasible, practical application framework for their implementation. In this paper, we introduce a medical data analysis framework based on natural language QA interface to structured medical data sources. We demonstrate that the complexities of performing medical data analyses can be simplified into question answering framework. Utilizing standard medical vocabularies, two hypothetical clinical questions are interpreted and the required analytics are automatically performed to produce textual and non-textual data for clinicians.