This research explores the relationships between diagnostic and procedural codes in a medical setting, with the objective of developing a general classification and predictive model. We describe the inherent relationship between codes in the context of a general data model, but note that the model is somewhat tenuous and requires extensions through other data analytic / data mining techniques. One of those techniques is decision tree induction, which is described briefly as a possible supplement to the initial code-to-code patterns. The paper concludes with implications for future research, including the investigation of additional analytic techniques and the extension of the model into other domains where problem-solving has been codified.