A K-Reversible Approach to Model Clinical Trajectories

Area of Science:

• Clinical informatics
• Computational biology
• Patient trajectory analysis

Background:

• Understanding patient health progression is crucial for timely clinical intervention.
• Identifying common clinical trajectories can improve patient outcomes and healthcare management.
• Mild traumatic brain injury (mTBI) patients often experience complex, varied recovery paths.

Purpose of the Study:

• To develop and validate an algorithm for clustering patient encounters and identifying common clinical trajectories.
• To demonstrate the effectiveness of the approach in a real-world dataset of mild traumatic brain injury (mTBI) patients.
• To provide clinicians with tools for better understanding and managing patient health paths.

Main Methods:

• Clustering of patient encounter data into similar groups.
• Development of an algorithm to generate automata representing common patient trajectories.
• Application of the method to a dataset of mTBI patients tracking symptoms like headaches, sleep disturbances, and PTSD.

Main Results:

• Successful clustering of patient data based on clinical encounters.
• Identification of distinct and common patient trajectories post-mTBI.
• Demonstration of the algorithm's capability to visualize and analyze complex health paths.

Conclusions:

• The proposed approach effectively identifies common clinical trajectories from patient data.
• This method offers valuable insights for managing mTBI and potentially other conditions.
• Automated trajectory analysis can enhance clinical decision-making and patient care.