Patient diagnosis and treatment involves factors that occur at different times. In order to identify as early as possible those factors associated with excessive length of stay in a hospital, time staging was added to induced decision trees. Time-staged induced decision trees uncovered useful patterns in data from the earliest time period, which represents a patient’s medical history, even though linear regression did not identify any significant variables in this data.
Murphy, Catherine, "Induced Decision Trees for Temporal Medical Data" (1998). AMCIS 1998 Proceedings. 66.