CHRiSS: Cardiac Health Risk Stratification System

描述(由申请者提供)：概述本项目旨在制作和测试一个用于终末期心力衰竭的临床决策支持应用程序，名为CRISS：心脏健康风险分层系统。它将为进展性心力衰竭患者设计，他们最终可能成为脑室辅助装置(VAD)治疗的候选者。该软件将建立在现有的机器学习和数据挖掘技术的基础上，这些技术由PI和他的同事开发，旨在预测VAD植入后90天的死亡率。这一阶段的工作将需要为患者和临床医生开发图形用户界面 将个性化的预后信息整合到功能性决策支持工具中。推理算法也将扩展到包括不良事件和重新入院的风险。在这个第一阶段的可行性项目中，CRISS将在西宾夕法尼亚阿勒格尼卫生系统(WPAHS)进行编程和Beta测试。这将需要半结构化访谈来收集相关的专家知识，并挖掘电子病历(EMR)以根据特定机构的数据校准预测算法。这一第一阶段项目的成功完成将导致软件得到充分验证的多中心试验。该项目的最终成功将导致一个软件应用程序，该应用程序将优化VAD治疗的好处 公共卫生，同时通过减少令人无法接受的高不良事件和不必要的再次住院来降低成本。

DESCRIPTION (provided by applicant): SUMMARY This project aims to produce and beta-test a clinical decision support application for end- stage heart failure, titled CHRiSS: Cardiac Health Risk Stratification System. It will be designed for patients with progressive heart failure who may eventually become candidates for ventricular assist device (VAD) therapy. The software will be build upon existing machine learning and data mining technology, developed by the PI and colleagues, designed to predict 90-day mortality following VAD implantation. This Phase- 1 effort will entail development of graphic user interfaces for both patients and clinicians to incorporate personalized prognostic information into a functional decision-support utility. The inference algorithm will also be expanded to encompass adverse events, and risk of readmission. In this Phase-1 feasibility project, CHRiSS will be programmed and beta-tested at West Penn Allegheny Health System (WPAHS). This will entail semi- structured interviews to collect relevant expert knowledge, and mining the electronic medical records (EMR) to calibrate the prognostic algorithm to the institution-specific data. Successful completion of this Phase I project will lead to a multi-center trial in which the software is fully validated. Ultimate succes of this project will result in a software application that will optimize the benefit of VAD therapy to public health, while reducing cost by reducing the unacceptably high rate of adverse events and unnecessary readmissions.