This project aims to introduce a clinical decision support system, CORA (Cardiac Outcomes Risk Assessment) for advanced heart failure (HF) a devastating disease affecting over 5 million Americans, and a leading cause of Medicare hospital admissions. The project specifically aims to moderate the deployment ventricular assist devices (VADs) which have been recently approved as destination (DT) such that patients most likely to benefit will be prioritized, while those better served with palliative care will be spare from unnecessary risks of debilitating side effects. The proposed system will leverage proven prognostic models based on advanced data mining and machine learning algorithms, and be packaged as a user-friendly computer application that would be used at key decision points in the course of treatment. It would provide both physician and patient with prognostic assessment based on the fusion of: patient-specific data, expert knowledge, and retrospective meta-analysis. This four-year project will develop a multi-platform software application that integrates modules for: forecasting, decision support, communication, decision making and visualization of data. This will be achieved through completion of four specific aims: (SA1) Expand existing inference (prognostic) algorithms and develop expert decision models to compute personalized estimate of life expectancy and risk of adverse events; (SA2) Create a computer application to assist patients to navigate their treatment options as they reflect their personal preferences and beliefs; (SA3) Incorporate these prognostic and decision models into user-friendly user interfaces for physicians, patients and their family caregivers; and (SA4) Conduct a multi- center study with respect to both qualitative and quantitative metrics of efficacy related to mortality, morbidity and quality of life. Training data capitalizing upon the NIH-funded Interagency Registry for Mechanical Circulatory Support (INTERMACS) as well as recently published guidance documents by the American Heart Association (AHA) and International Society of Heart & Lung Transplantation (ISHLT.) This project will involve a multi-disciplinary group of specialties including: bioengineering, advanced heart failure cardiology, cardiac surgery, computer science (machine learning/data mining), medical ethics, and palliative care. Carnegie Mellon University, one of the Nation's leading institutions for Computer Science will coordinate this project to assure a productive collaboration with four clinical sites: Allegheny Health Network (Pittsburgh), U. Colorado Denver, Duke University, and Harvard/Brigham & Women's hospital. By achieve consensus amongst stakeholders, we hope to assure the acceptance of CORA into practice and ultimately improve the efficiency, efficacy, and cost- effectiveness of the treatment of end-stage heart failure.

Public Health Relevance

Congestive heart failure (CHF) is a debilitating disease affecting over 5 million Americans, and a leading cause of Medicare hospital admissions. Treatment can be very complicated, and for some, may involve advanced technologies like heart-assist devices and stents, and for others palliative care. This project will develop a software application for patients and their doctors to work together in choosing the best treatment strategy.

Agency
National Institute of Health (NIH)
Institute
National Heart, Lung, and Blood Institute (NHLBI)
Type
Research Project (R01)
Project #
1R01HL122639-01A1
Application #
8818809
Study Section
Biomedical Computing and Health Informatics Study Section (BCHI)
Program Officer
Baldwin, Tim
Project Start
2015-01-15
Project End
2018-12-31
Budget Start
2015-01-15
Budget End
2015-12-31
Support Year
1
Fiscal Year
2015
Total Cost
$920,133
Indirect Cost
$203,650
Name
Carnegie-Mellon University
Department
Biomedical Engineering
Type
Schools of Engineering
DUNS #
052184116
City
Pittsburgh
State
PA
Country
United States
Zip Code
15213
Lohmueller, Lisa C; Kanwar, Manreet K; Bailey, Stephen et al. (2018) Retrospective Evaluation of Bayesian Risk Models of LVAD Mortality at a Single Implant Center. Front Med (Lausanne) 5:277
Kanwar, Manreet K; Lohmueller, Lisa C; Kormos, Robert L et al. (2018) A Bayesian Model to Predict Survival After Left Ventricular Assist Device Implantation. JACC Heart Fail 6:771-779
Jo, Yohan; Yoder, Michael Miller; Jang, Hyeju et al. (2017) Modeling Dialogue Acts with Content Word Filtering and Speaker Preferences. Proc Conf Empir Methods Nat Lang Process 2017:2169-2179
Kanwar, Manreet K; Lohmueller, Lisa C; Kormos, Robert L et al. (2017) Low Accuracy of the HeartMate Risk Score for Predicting Mortality Using the INTERMACS Registry Data. ASAIO J 63:251-256
Yang, Qian; Zimmerman, John; Steinfeld, Aaron et al. (2016) Investigating the Heart Pump Implant Decision Process: Opportunities for Decision Support Tools to Help. ACM Trans Comput Hum Interact 2016:4477-4488
Loghmanpour, Natasha A; Kormos, Robert L; Kanwar, Manreet K et al. (2016) A Bayesian Model to Predict Right Ventricular Failure Following Left Ventricular Assist Device Therapy. JACC Heart Fail 4:711-21
Loghmanpour, Natasha A; Kanwar, Manreet K; Druzdzel, Marek J et al. (2015) A new Bayesian network-based risk stratification model for prediction of short-term and long-term LVAD mortality. ASAIO J 61:313-23