Issue:
A major pharmaceutical client sought to create a data mining approach that would help them identify target populations of greatest need for their early development products.
Solution:
- We conducted a Classification and Regression Tree (CART) analysis for Rheumatoid Arthritis to identify high risk patients that would inform clinical and market access strategies.
- Using claims data, we identified key risk factors associated with high cost and/or low persistency associated with current therapies. Additionally, by varying the parameters of the CART models (‘pruning’ the regression trees), we came up with various risk scenarios, which provided valuable actionable information that was used by our client.
- The findings from this analysis were presented as a case study at a ISPOR workshop on phenotyping analytic techniques (ISPOR Annual Meeting, New Orleans, LA. 2013)
