Data collection, integration and normalization presented to DIMACS

1
Data collection, integration and normalization
presented to DIMACS

• Gil Delgado
• October 17, 2002

2
Discussion

• Challenges and barriers as well as processes and
  methods for collection, integration, and
  harmonization of clinical data.
• Clinical Data as a resource for Disease
  Surveillance, BioTerrorism detection, and Adverse
  Events Reporting.

3
Agenda

• Challenges and Barriers
• Overview of clinical data collection, integration
  and harmonization
• Processes
• Methods
• Demonstration of Health Data System

4
Challenges and Barriers

• Social, Political, and Ethical
• HIPAA
• Incentives
• Resources
• Available Infrastructure
• Vocabularies and Codes
• Appropriate Filtering

5
Collection, Integration and Harmonization Overview

• Enterprise Application Integration
• Interface Engines
• HL7
• Enterprise Master Patient Indexing
• De-Identification
• Vocabulary Harmonization
• Standard vs. Proprietary
• Mapping

6
Processes and Methods

• Rationals Unified Process (RUP)
• Message-based Processing
• Error Reporting
• Change Management
• Patient Matching
• Vocabulary Mapping
• Quality Assurance
• Performance Metrics and Management

7
Deployment Architecture
8
Commercial Research Architecture
9
Public Health Surveillance Architecture
10
Demo

• Health Data System

11
Appendix

• The following three slides present more details
  about the data that Emergint provides for
  research.
• These slides discuss the value of the data, from
  how it is collected, what is collected, how it is
  harmonized, and how it is made available.

12
Appendix Value Creation

• Emergint captures the same data directly used in
  the delivery of clinical care, rather then
  secondary data derived from clinical care, such
  as claims data and chart abstraction
• Patient data are multi-encounter and
  longitudinal, de-identified in compliance with
  HIPAA, continually updated and linked across
  multiple inpatient and outpatient providers

13
Appendix Value Creation (cont.)

• Data consists of demographic (age, sex, weight,
  etc.), diagnosis, procedures and results
  (laboratory results, medications, transcriptions
  including history and physical and discharge
  summary)
• Data are electronically captured and maintained
  in original terminology and transported in a
  common format (XML) to retain maximum data
  integrity

14
Appendix Value Creation (cont.)

• Data are comparable across sources or original
  terminology by enhancing data with
  standards-based concepts for indexing and
  retrieving data
• Common search tools and customized report
  formatting provide electronic access to the data