1. MDX queries were executed on an OLAP cube to analyze the CHSI healthcare data and gain business insights. Key findings included the relationship between formulary type and charge quantity, most profitable years, medications, and care settings.
2. Data mining models including decision trees, neural networks, and regression were built to predict medication discontinuation reasons. The most important variable for prediction was found to be infusion time.
3. Model performance was evaluated using lift charts, and the models were found to fit the data reasonably well with similar predictive strength.