Charles Hurmiz, Manager, Software Engineering

1
CRIS A Configurable Shared Database For
ResearchAn OverviewAcademic Health Centers
Clinical Research ForumMarch 29, 2005

• Charles Hurmiz, Manager, Software Engineering
• Clinical Research Informatics
• Clinical Research Administration

2
History of Research Databases

• Separate Database Systems
• Build one Database for each study or research
  need
• Data was disjointed and difficult to bring
  together

B
C
A
3
History of Research Databases

• Open Database Connectivity (ODBC) Tools
• ODBC tools made physical separation less
  important
• ODBC tools facilitated data sharing
• Separate data bases still suffered from lack of
  data standardization.

B
ODBC
ODBC
C
A
ODBC
4
History of Research Databases

• CRIS
• Single architecture houses all research
  information
• Rapidly deploy study databases
• Resolves data consistency issues

5
CRIS Features

• Web Based
• Zero footprint (nothing is installed on the
  client machine)
• Requires only access to an internet browser
• Platform Independent
• Operates on Windows, Macintosh, Unix
• Any platform that has access to an internet
  browser
• Centralized Data Storage
• All data is stored in a single, consistent
  database architecture
• Configure vs. Code
• Adding additional data collection applications is
  a configuration point in software
• Adding additional interfaces to foreign systems
  is through configuration

6
Metadata Driven

• Application Configuration
• Meta data is captured in multiple layers
• Different layers define various aspects of
  configuration
• Security Model
• Implements a robust role based security model
• Security can be configured from within or through
  interface to central security i.e. LDAP

7
Technologies Employed

• Open Source Software
• JBOSS Application Server
• Cocoon Framework
• Written In Java
• J2EE
• Oracle Database
• Object Oriented Schema
• Storage of Graph trees

8
Application Architecture
9
Data Collection vs. Data Analysis

• Data Collection
• Generally looking at single record at a time
• Creating Source data
• Referencing data from foreign sources
• Creating relationships between instances of data
• Data Analysis
• Generally looking at multiple records at a time
• Selecting data based on defined criteria
• Selecting data on ad hoc criteria

10
Traditional Approach to Data Warehousing
11
CRIS Approach to Data Warehousing
12
CRIS Overview

• CRIS is developed to be flexible
• Highly customizable
• Metadata driven
• The flexibility of CRIS allows it to accommodate
  any type of data
• Clinical Research Data
• Approved protocol and IRB studies
• Other Research Data
• Tissue Bank
• Non-Research Non-Clinical Data
• Financial
• Inventory
• Other