Data Mining Electronic Health Records to Support Evidence-Based Clinical Decisions

摘要: This study investigated the extent of use data mining on electronic health records to support evidence-based clinical decisions, reasons why only few healthcare institutions integrate it in workflow, and resolutions increase its utilization actual practice. A literature review was conducted get examples studies where applications were used, particularly radiation oncology, critical care, in-hospital mortality prediction, pharmaceutically treated depression, visualizing event patterns, diabetes research. For each reviewed, objectives, methodology, procedure for integration decision system (CDSS) various issues analyzed documented. brief description required infrastructure including policies procedures ensure smooth deployment also documented whenever available. Clinical is used mostly gain new insights, do predictions, risk assessments, recommendations. Many find CDSSs a good learning environment. Issues preprocessing, class imbalance, feature engineering, performance evaluation mentioned. Other include need more active collaboration with stakeholders, access anonymized data, formulation interoperability standards, reevaluation results using from other institutions, finally addressing ethical legal issues. Although still infancy, experience early adopters have been promising thus should encourage