Yves A' Lussier, Eneida Mendona, Dongwen Wang, Edward Shortliffe

1

• Yves A. Lussier, Eneida Mendonça, Dongwen Wang,
  Edward Shortliffe
• Laboratory for Knowledge Technology in
  Biomedicine
• Department of Biomedical Informatics
• Columbia University

2
Outline

• Current Problems
• Vigilens System
• Current Research

3
Problems with Current DSS

• Monitoring tightly coupled with repositories
• (DSS Specific to IT Implementation)
• Syntax rules
• Semantics vocabulary
• Modeling database structure
• Daunting Costs
• Short-lived up-to-date clinical guidelines
• Labor-intensive Development and Maintenance
• Knowledge-bases healthcare professionals
• Translation in DB queries IT professionals
• IT Processing-intensive
• Commonly impounds severely the DB performance and
  processor cycles

4
Current Notification Problems

• Complexity of Effective and Secure Notification
• Wireless devices (protocols, HIPAA)
• User Preferences
• Institution Policies
• 2-way Acknowledgment of Receipt Feedback
• Support of Guidelines beyond Events
• Continuous data feeds of biomedical devices

5
Vigilens Solutions Notification

• Server-based, DB schema independent
• Vigilens support notification rules integrating
  institution policies and user preferences
• Separates the pre-coordination that exists
  between the destination and the alert itself,
• Notification Intelligence is built into the
  message, enabling third-party notification
  servers (e.g. IBM Websphere Everyplace Connection
  Manager )
• The clinical event monitor system supports
  message priority and severity,

6
Vigilens Solutions Notification

• Vigilens support notification rules integrating
  institution policies and user preferences
• Separates the pre-coordination that exists
  between the destination and the alert itself,
• Notification Intelligence is built into the
  message, enabling third-party notification
  servers (e.g. IBM Websphere Everyplace Connection
  Manager )
• The clinical event monitor system supports
  message priority and severity,

7
Goals

• Demonstrate that an institution-independent
  decision support server is technically feasible
  and significantly improves healthcare

8
Background

• Faculty members, developers and trainees of the
  Dept. have previously established
  significantmilestones in decision support systems
  (MYCIN), event monitoring (Clinical Event
  Monitor), and knowledge representation (Arden
  Syntax, GLIF). The current Vigilens project
  builds on this experience and provides additional
  functionality to the monitoring processes of the
  NYP hospital.

9
Environment

• New York Presbyterian Hospital
• Department of Medical Informatics
• Clinical Event Monitor / Arden Syntax
• WebCIS
• MED
• Generic Repository
• Data Warehouse

10
Integration to Existing DMI Telehealth Systems

• 2002 - Vigilens
• 2001 - MIHEART
• 1999 - IdeaTel
• 1999 - PalmCIS
• 1996 - PatCIS
• 1994 - WebCIS
• 1990 Clinical Event Monitoring
• 1988 HIT / MED / Generic Repository

11
Vigilens Monitoring Architecture
Datagate Server
Ancillary Systems
VigiLens
Clinical Repository
12
(No Transcript)
13
Vigilens Notification Architecture
Devices
HTTP Server / Vigilens Components
SMS, WAP Push, Email, Voice
Wireless clients, Messaging Clients, WAP devices
Internet
DataTAC, Mobitex, CDPD, GPRS, SMS, PAP, SMTP,
SNPP, others
IP
NAS
JAVA based Middle Tier Server

• Manages user configuration
• Manages dynamic role monitoring
• Enables Policy definitions and executions
• Managers HIPPA compliance (encryption
  algorithms, etc)

Repository
Packet, Cellular, and some Private Networks
14
Benefits

• Integration of Care (Interoperability)
• Empowerment
• Electronic Vigilance
• Knowledge Deployment
• Cost Reduction of DSS
• Reusable and Sharable DSS
• Maintenance costs distributed
• Hardware costs shared by groups of organizations
• Rules development and maintenance distributed
  across a community of clinician, administrators,
  researchers and educators.

15
Software Process

• Evolutionary prototyping
• Staged Delivery

16
Ancillary Projects

• Biosurveillance Rx/Dx Quality Safety,
  Language Understanding
• Practice Guidelines for Individualized
  Medicine
• Novel Relational Database Monitoring
• Advanced Notification
• Execution of Clinical Practice Guidelines

17
Research Opportunities

• Fundamental MI Sciences
• Interoperability of DSS Schema
• DB Triggers-based DSS
• Cognitive and Behavioral Science
• Social cultural system-approaches to deployment
  and impact
• DSS communication modalities and
  user-configurable
• Clinical Informatics
• Clinical care
• Quality of healthcare
• Patient Safety
• Accountable performance-based medicine
• Clinical research
• Education of clinicians
• Administration of healthcare
• Public Health Informatics (Monitor of Monitored
  events)

18
Future Directions

• Clinical Practice Guidelines (CPG)
• GLIF Enabled
• Automated Knowledge Acquisition
• Non event-driven
• CPGs for Biodefense
• Individualized CPGs
• Increased Notifications Modalities

19
Acknowledgments

• Co-Investigators Edward Shortliffe, Eneida
  Mendonça, Stephen Johnson, Yves Lussier, Soumitra
  Sengupta, George Hripcsak, Justin Starren, James
  Cimino, Bruce Forman ,David Wajngurt
• Stephen Johnson, Robert Jenders and Soumitra
  Sengupta for inspiration
• Developers Jianrong Li, Raghunathan(Nathan)
  Kamasamudram, Dennis Chua, Hillary Feldman
• New York Presbyterian Hospital Collaborators of
  the Patient Safety and Quality Committees
• This project was partially supported by grant
  528753/PO P417322 from the National Aeronautics
  and Space Administration (HRSA/UVC) and the 1 D1B
  TM 00043-01 grant from the Health Resources
  Services Administration (HRSA/OAT). Additional
  support was provided NLM Medical Informatics
  Training Grant LM07079-07.

20
Contact Information

• Yves A. Lussier
• Phone [phone]
• email Lussier_at_dmi.columbia.edu
• http//www.dmi.columbia.edu/homepages/yal7001/Vig
  ilens.html