Web-Based Clinical Information Systems: Architecture, Perils and Promises

1
Web-Based Clinical Information Systems
Architecture, Perils and Promises

• James J. Cimino, M.D.
• Department of Medical Informatics
• Columbia University

2
Web-Based Clinical Information Systems

• Clinical Data Access
• Information Resource Access
• Integrating Information Resources

3
Clinical Data Access Architecture
Clinical Information System
Web Server
Clinician
4
Clinical Data from a Web Server
Web Server
CGI
Static Documents
Clinical Information System
CGI
Multiple Views of Patient Data
Clinician
5
Clinical info from web server (screens)

• Static Documents - Reports
• Viewers - scanned in written documents
• CGI - button list
• Viewers - tabular data
• Viewers - graphical data
• Viewers - ecg
• Viewers - xray
• Viewers - pathology
• Viewers - Image Engine

6
Clinical Data Access Promises

• Universal access
• Platform-independence

7
Clinical Data Access Perils

• Security
• Hypertext Paradigm

8
Perils Security

• Internet security
• Web browser security

9
Web Browser Security
ltENTERgt
User ID ciminoj
ltENTERgt
Password
ltENTERgt
ltFORWARDgt
ltBACKgt
ltFORWARDgt
ltBACKgt
ltBACKgt
10
Perils Hypertext Paradigm

• Cognitive Overload
• Disorientation
• Mismatch with clinical application paradigms

11
Cognitive Overload
Select Function
Choose Document
Select data type
Select graph type
12
Disorientation
Step 1
Step 2
Step 3
Select Function
Choose Document
Select data type
Select graph type
13
Paradigm Mismatch
Modify Patient List
New Patient List
ltENTERgt
ltENTERgt
Add Patient
ltBACKgt
14
Information Resource AccessArchitecture
Clinical Information System
Web Server
Internet
Clinician
Local Information Source
Remote Information Source
15
Information Resource AccessArchitecture (screen
shots)

• Text Databases - Chorus
• Search engines - Utah
• Expert systems - DXplain

16
Information Resource Access Promise

• Improved access to knowledge
• Economies of scale

17
Information Resource Access Perils

• Quality
• Availability
• Permission
• Disorientation
• Cognitive Overload

18
Integrating Informatics Islands
19
Integrating Information Resources Architecture
Vocabulary Server
Clinical Information System
Web Server
Internet
Clinician
Information Source
Information Source
20
Vocabulary Services
Intravascular Glucose Tests
Glucose
Serum Glucose
Plasma Glucose
Glycosuria
Blood Glucose
Abnormalities of Blood Glucose
Glucose Preparations
Hyperglycemia
Hypoglycemia
Robitussin
Ringer's Lactate
21
Integrating Information Resources (screen shots)

• lab summary - could be from multiple sources
• Drug to pdr example
• chest x-ray to information sources (text, pics)
• lab to medline
• lab to dxplain - gopher
• lab to dxplain - front end
• lab to dxplain - api
• lab to cholesterol

22
ltFORM METHOD"POST" ACTION"http//www.cpmc.columb
ia.edu/triadbin/chguide/0!177!255!!!M!65!!201202
!"gtltINPUT TYPE"submit" NAME"ACTION"
VALUE"GUIDELINE SERVER"gtlt/FORMgt
Integration of CGI's

• CAD No
• LDL 177
• DIET Unknown
• CHOLESTEROL 255
• HDL Unknown
• GENDER Male
• AGE 65 years
• PREMATURE MENOPAUSE Unknown
• RISK FACTORS Family History of CAD
  (201) Smoker (202)

23
Integrating Information Resources Promises

• Improved access
• Increased usability
• Increased usefulness

24
Integrating Information Resources Perils

• Lack of context
• Local variability
• Translation errors
• Lack of interchange standards
• Lack of vocabulary standards
• Liability

25
Challenges

• Data servers
• Vocabulary
• Resource interfaces
• User information needs

26
Challenges Data servers

• Access to repositories
• Security
• Authentication
• Confidentiality

27
Challenges Vocabulary

• Controlled vocabulary
• Translation
• Vocabulary services

28
Challenges Resource Interfaces

• Content
• Interface standards
• Licensing
• Liability

29
Challenges User Information Needs

• User profiles
• Problem orientation
• Cognitive processes

30
Conclusions

• Improve access to patient data
• Integration of information sources
• http//www.cpmc.columbia.edu/cisdemo