DataBabes : a case study in feto-maternal clinical data mining

![CDATA[This paper aims to test whether data mining and supporting technology components can provide Feto-Maternal Medicine (FMM) Clinicians with an improved environment for patient data analysis. Currently Online Transaction processing (OLTP) systems support the treatment and care of mothers and unborn children, thus generating large volumes of clinical data which is not ideally structured for complex analysis. Other application domains, such as financial and anti-terrorism, are well advanced in the use of data mining to aid complex data analysis. Health care exploitation of data mining has largely focussed on organisational management rather than patient data analysis. Clinicians within the FMM Unit at Liverpool Hospital held hypothesis regarding the correlations between multiple factors within the OLTP system data. For example, the relationship between pregnancy outcome and the gauge of needle and type of needle used for transabdominal Chorionic Villus Sampling (CVS). To address the research aim, the DataBabes architecture was defined and prototyped in the FMM Unit at Liverpool Hospital within the South Western Sydney Area Health Service (SWSAHS). The CRISP-DM process and supporting technology components, as implemented in DataBabes, were found to provide clinicians with an improved data analysis capability. Of particular interest to clinicians was the C5.0 data mining algorithm as they could see the predictive power of the algorithm given a set of known factors. There are many future research directions from this research, perhaps the most challenging lies in addressing how data mining can be best exploited in generating valid medical knowledge that is acceptable for evidence based medicine practice.]]