Annually, it is estimated that more than 795,000 Americans experience a stroke. Stroke remains a major cause of death in the United States, behind only cardiovascular disease and cancer;and is the leading cause of long- term disability, with only 25% of adults recovering to normal health from this condition. The severity of neurological damage due to an acute stroke is mitigated by the early restoration of blood flow to the affected area;and more people are now surviving strokes through earlier intervention with thrombolytic agents and interventional clot retrieval devices. Unfortunately, the rapid development of new drugs and devices in this area has made it difficult to provide treatment guidance for a given patient, and metrics for comparing outcomes between treatment groups are lacking. This proposal focuses on the creation of an observational database that is subsequently used to support an influence diagram for acute stroke treatment. The database is predicated on the specification of a unified in- formation model for stroke and its treatment: thus far, no comprehensive framework has been established to incorporate these variables using standardized representations and/or controlled vocabularies. We take ad- vantage of current data at our institution;and a forthcoming national dataset through the American Heart Association (AHA) to instantiate this database. From this database, an influence diagram is established, using in- formation on patient presentation, medical history, imaging, and available treatment drugs/devices to compute an optimal treatment decision maximizing outcomes and other considerations (e.g., quality of life). Methods to handle missing data via imputation and propensity scores are used to help compute the required conditional probability tables underlying the influence diagram. A spectrum of utility functions will be considered over the course of this effort to explore the trade-offs between health outcomes and cost. Using the database and influence diagram, case-based similarity retrieval is supported, enabling discovery of related past cases to review treatment decisions and outcomes;and as a means to define and compare subgroups (and their outcomes). A graphical user interface (GUI) for querying the influence diagram and performing case-based retrieval will be developed. Evaluation focuses on assessing the performance of the influence diagram relative to known out- comes and compared to other conventional statistical models (e.g., logistic regression, decision trees);and the overall impact of the system to impact decision-making. This R01 leverages nationally-recognized clinical expertise in acute stroke treatment from the UCLA Stroke Center;as well as a longstanding medical informatics research group. The ultimate result of this effort will be a set of informatics-driven modeling tools and a database for acute stroke treatment.

Public Health Relevance

Stroke remains a major cause of death in the United States, behind only cardiovascular disease and cancer;and is the leading cause of long-term disability, with only 25% of adults recovering to normal health from this condition. Recent progress in the development of new thrombolytic agents and interventional clot retrieval de- vices is helping more people survive strokes by minimizing neurological damage;however, physicians have few guidelines for selecting the drugs/devices that will optimize a given individual's outcomes. The focus of this research is the creation of an influence diagram that will provide insight into acute stroke, establishing a tool to aid medical decision making regarding the different treatment options.

National Institute of Health (NIH)
National Institute of Neurological Disorders and Stroke (NINDS)
Research Project (R01)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1-SBIB-Q (80))
Program Officer
Odenkirchen, Joanne
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of California Los Angeles
Schools of Medicine
Los Angeles
United States
Zip Code
Petousis, Panayiotis; Han, Simon X; Aberle, Denise et al. (2016) Prediction of lung cancer incidence on the low-dose computed tomography arm of the National Lung Screening Trial: A dynamic Bayesian network. Artif Intell Med 72:42-55
Meng, Frank; Morioka, Craig (2015) Automating the generation of lexical patterns for processing free text in clinical documents. J Am Med Inform Assoc 22:980-6
Ho, King Chung; Speier, William; El-Saden, Suzie et al. (2014) Predicting discharge mortality after acute ischemic stroke using balanced data. AMIA Annu Symp Proc 2014:1787-96
Bui, Alex A T; Hsu, William; Arnold, Corey et al. (2013) Imaging-based observational databases for clinical problem solving: the role of informatics. J Am Med Inform Assoc 20:1053-8
Love, Alexa; Arnold, Corey W; El-Saden, Suzie et al. (2013) Unifying acute stroke treatment guidelines for a Bayesian belief network. Stud Health Technol Inform 192:1012