Transcript Session 1 HINF 371

Clinical Information Systems
HINF 371 - Medical Methodologies
Session 5
Objective

To review types of information
needed in decision making in a
clinical setting
Reading

Geisbuhler A and Miller RA (2000) Chapter 14: Computer
Assisted Clinical Decision Support, in Decision Making In
Health Care: Theory, Psychology and Applications, Cambridge
University Press, USA
Decision Support Needs
52 percent patient specific
information
 23 percent general knowledge that
could be found in a library, textbook,
or Medline
 26 percent synthesis of patient
information and medical knowledge

Patient Information

Computerized patient information






Better availability
Better readability
Clear and well organized displays of informatin
Ability display trends and patterns
Ability to select and organize subsets of information
Physicians most interested in




Which drugs
What happened in last hospitalization
What are the lab results
What is the status during the last visit to the physicians
Difficulties with Computerized
Patient Information

Getting data into the system – physicians are not
good at it





Transcription and interpretations of patients records
Voice recognition
Provision of immediate feedback as data entered
Getting information out of the system
Difficulties in implementation



Busy care providers are reluctant to use systems that
slow their work
25percent of difficulties are technology related
75 percent of difficulties are caused by inadequate social
engineering/change management
General Knowledge

Ability to perform literature searches using
Medline



Improved decisions
Reduce costs and length of stay
Other sources of information





Textbooks
Full-text journals
Practice guidelines
Printable educational material for patients
Drug information database
General Knowledge

Local policy and procedures
Formalized policy and procedures
 Operational rules
 Functioning of institution
 How to apply and break rules
 Resource inventories, operating hours,
names of people, turn around times for
lab tests, etc.

General Knowledge


Many medical errors are results of intrinsic
limits in physician’s ability to process
ongoing events
Attention focusing tools




Drug interactions (can be done through EMR or
Order Entry)
Reminders for tests, follow-ups, information
transfer, etc.
Results of past tests
Ranges of normal and likelihood of normal for
given tests
Patient Specific Support

Diagnostic decision support systems
Clinical algorithms – BC Nurseline or
Healthwise
http://bchealthguide.org/kbaltindex.asp
 Predictive Tools
http://groups.csail.mit.edu/medg/project
s/hdp/hdp-intro-tab.html

Patient Specific Support

Clinical Algorithms




Problems of scalability
Inability to deal with uncertainty or incorrectness
inherent in the nature of clinical data
Better used in non-intersecting subsets
Predictive tools




How representative are the patients who were used in
developing a given predictive tool
Which patients are likely to be misclassified and
subjected to unnecessary interventions
How severe are the adverse effects that occur in patients
receiving unnecessary interventions
How integrated the system to the workflow of physicians
within the architecture of clinical information systems

The clinical significance of a computer
based decision support tool resides in
the ability to augment the native skills
of the physician during clinical
practice, not its function in isolation
as an “omniscient oracle” Miller 1996
Is Medicine a science or an art?
Discussion