Whether one looks for data defects in an Electronic Data Capturing (EDC) System or explores water on Mars, the methods a researcher is applying are nowadays the same.
As it is very expensive to send a shuttle to our neighbor planet Mars, NASA has decided to apply more data analytics to look for promising regions, where water on Mars can be. Researchers try remotely unlocking mysteries of water on Mars by using data mining and statistical learning. A project “Automated Identification and Characterization of Land forms on Mars” funded by NASA applies data mining and machine learning mechanisms to oversee complex patterns across Mars’ surface.
Similarly, Risk-based Monitoring Analytics (RBMA) is combining techniques from data-mining, statistics and machine learning, providing novel data analysis methods for early risk identification. From a data-mining point of view, clinical data is very complex, heterogeneous, diffuse and, often, unstructured. Therefore, it requires special techniques to detect critical patterns in it.
Each Key Risk Indicator (KRI) is usually a complex module developed by a team of mathematicians, statisticians, and medical staff. Validated by a retrospective approach. It guarantees that a risk will be found in the ocean of unstructured information, found early enough to be mitigated without complications for a pharmaceutical company. A good KRI applies a similar approach as used by NASA: dataset preparation, segmentation, learning, detection.
Summing up, RBM requires design and development of statistical-learning tools, which can use the historical information from a wide range of data available and convert it to knowledge.