Data collected through electronic health records is increasingly used for clinical research purposes. Common research tasks like identifying treatment cohorts based on similar treatment histories, assessing adherence to protocol based care, or determining clinical ‘best practices’ can be difficult given the complex array of treatment choices and the longitudinal nature of patient care. We present a temporal sequence alignment strategy to find patients with similar treatment histories starting from their initial regimen. The algorithm relies on a user defined threshold heuristic to further reduce the search space in large clinical databases. It also uses an ontology based scoring schema to measure the distance between two clinical treatment histories. We validate the algorithm with a search for patients who are placed on a guideline recommended alternative regimen for HIV after failing initial ideal therapy. Our approach can be used to summarize patterns of care as well as predict outcomes of care.
展开▼
