? ? Medical errors and their timely identification remain an important problem in clinical practice. Electronic medical record repositories and electronic data processing offer an opportunity to identify such errors in time to prevent them or at least attenuate their harm. Typical computer-based error detection methods rely on the use of clinical knowledge, such as expert-derived rules, that is incorporated into the monitoring and alerting systems. Alerting that is based on knowledge is generally reliable; however, it is time-consuming and costly to extract and codify such knowledge, and as a consequence such systems are relatively narrow in their scope. We propose to develop and evaluate a data-based approach for detecting clinical outliers (anomalies) that is complementary to knowledge-based approaches. This new approach is based on comparing clinical actions, such as medications given and labs ordered, taken for the current patient to those actions taken for similar patients in the recent past, as recorded in a clinical database. If a clinical action for the current patient is highly unusual, then a cautionary alert is raised along with an explanation for why the action appears to be unusual. Key advantages of the new technique are that it works with minimal prior knowledge, and it may detect anomalies for which no rules have yet been written. Thus, this data-driven approach to clinical anomaly detection is expected to complement knowledge-based alerting methods. We propose to implement a data-driven anomaly detection method, and then evaluate it in a laboratory setting using retrospective data for the cohort of surgical cardiac patients. The project investigators comprise a multidisciplinary team with expertise in rule-based alerting in a hospital setting, clinical pharmacy, laboratory medicine, biomedical informatics, statistical machine learning, knowledge based systems, and clinical data repositories. ? ? ?
| Batal, Iyad; Cooper, Gregory; Fradkin, Dmitriy et al. (2016) An Efficient Pattern Mining Approach for Event Detection in Multivariate Temporal Data. Knowl Inf Syst 46:115-150 |
| Hauskrecht, Milos; Batal, Iyad; Valko, Michal et al. (2013) Outlier detection for patient monitoring and alerting. J Biomed Inform 46:47-55 |
| Batal, Iyad; Fradkin, Dmitriy; Harrison, James et al. (2012) Mining Recent Temporal Patterns for Event Detection in Multivariate Time Series Data. KDD 2012:280-288 |
| Batal, Iyad; Valizadegan, Hamed; Cooper, Gregory F et al. (2011) A Pattern Mining Approach for Classifying Multivariate Temporal Data. Proceedings (IEEE Int Conf Bioinformatics Biomed) 2011:358-365 |
| Nguyen, Quang; Valizadegan, Hamed; Seybert, Amy et al. (2011) Sample-efficient learning with auxiliary class-label information. AMIA Annu Symp Proc 2011:1004-12 |
| Valko, Michal; Kveton, Branislav; Valizadegan, Hamed et al. (2011) Conditional Anomaly Detection with Soft Harmonic Functions. Proc IEEE Int Conf Data Min 2011:735-743 |
| Valko, Michal; Hauskrecht, Milos (2010) Feature importance analysis for patient management decisions. Stud Health Technol Inform 160:861-5 |
| Batal, Iyad; Hauskrecht, Milos (2010) Mining Clinical Data using Minimal Predictive Rules. AMIA Annu Symp Proc 2010:31-5 |
| Visweswaran, Shyam; Mezger, James; Clermont, Gilles et al. (2010) Identifying Deviations from Usual Medical Care using a Statistical Approach. AMIA Annu Symp Proc 2010:827-31 |
| Hauskrecht, Milos; Valko, Michal; Batal, Iyad et al. (2010) Conditional outlier detection for clinical alerting. AMIA Annu Symp Proc 2010:286-90 |
Showing the most recent 10 out of 14 publications
