James J' Cimino MD, Jianhua Li MD,

1
Theoretical,
and Practical
Empirical,
Approaches to Resolving the
Unmet Information Needs of
Clinical Information System Users

• James J. Cimino MD, Jianhua Li MD,
• Suzanne Bakken RN DNSc, Vimla L. Patel PhD
• Department of (Bio)Medical Informatics
• Columbia University
• New York, New York, USA

2
Information Needs of CIS Users

• Unmet information needs in practice
• Unmet needs lead to deferred or uninformed
  decisions (errors!)
• Information needs of CIS users not studied

3
Why Care about CIS Users?

• Needs occur when learning new information
• Needs occur during decision making
• Simple setting suggests predictable needs
• Setting is conducive to resolving needs
• Retrieval may be accomplished automatically
  (infobuttons)

4
A Brief History of Infobuttons

• 1991 Medline Button translated ICD9 to MeSH
  and performed Medline searches
• 1994 World Wide Web
• 1994 First CPMC Web-based CIS
• 1996 WebCIS
• 1996 Infobuttons link microbiology and
  medication data to PubMed and Micromedex
• 1997 Infobuttons for patients (PatCIS)

5
Theoretical Approach

• Theory Information needs are situated
• Hypothesis needs arising during use of a CIS are
  predictable based on context
• IF detect the context THEN predict need
• IF predict need THEN address need
• IF address need THEN reduce unmet needs
• IF reduce unmet needs THEN reduce errors

6
Context-Dependent Information Needs
7
Empirical Approach

• Direct observation of CIS users
• Think aloud about information needs
• Data capture video (screen) and audio
• Transcription articulation of needs
• Coding characterize needs and contexts
• Summariation needs and contexts

8
Portable Usability Lab
Users Workstation
75 foot cable
Video Monitor
Converter Controller
Microphone
Video Converter
Headphones
VCR
Cassette Recorder
9
(No Transcript)
10
Practical Approach

• Context passed to Infobutton Manager (IM)
• IM matches context to information needs
• Search strategies are instantiated
• Questions (links) are returned to user

11
Matching Context to Need
C o n t e x t
12
Context Parameters

• Patient age, gender
• User physician, nurse, student, patient,
• Setting lab reports, medications,
• Concept of interest lab test, drug,

13
Infobutton Parameters

• Question type fixed, fill-in-the-blank
• Source type simple query, additional data
• URL
• http//www.guideline.gov/SEARCH/search_results.asp
  ? sSearch_stringltgtbSort_by_datequery_id
  bDebugresults10viewsearch_results
• Question
• What are guidelines for evaluation of ltgt?

14
(No Transcript)
15
lta href"Javascriptget_info('35702')"gtltimg
src"info.gif"gtlt/agt
function get_info(mc) var infowin
window.open("","INFO") infowin.focus()
document.INFO.info_med.valuemc
document.INFO.submit()
ltform method"POST" nameINFO
action"wc_infomanage.cgi"gt ltinput type"HIDDEN"
name"info_med" value""gt ltinput type"HIDDEN"
name"info_context" value"LabDetail"gt ltinput
type"HIDDEN" name"info_institute"
value"CPMC"gt ltinput type"HIDDEN" name"DOB"
value"1951-05-26"gt ltinput type"HIDDEN"
name"SEX" value"F"gt ltinput type"HIDDEN"
name"USER" value"ciminoj"gt lt/formgt
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
Evaluation Plan

• Baseline observation
• Creation of Infobutton Manager tables
• Resources used
• Unmet information needs
• Log of access to online resources
• Introduction of infobuttons
• Usability studies
• Log of access to online resources (infobuttons)
• Post-intervention observation
• Use of infobuttons and other resources
• Unmet information needs

20
Status Report

• Infobutton Manager built
• Web-based table management tools built
• Log files being collected
• Three nursing stations, CCU and clinic
• 5 days of data
• 15 hours of audio/video
• Needs coded along six dimensions
• patient/institution
• administrative/health knowledge
• foreground/background
• explicit/implicit
• human/computer/paper
• success/failure/deferred

21
Discussion

• Empirical method is detecting needs
• Summarizarion is next step
• Technical approach is feasible
• Evaluation should detect impact
• Information needs may be universal
• Our approach is institution independent

22
Context Parameters

• Patient age, gender
• User physician, nurse, student, patient,
• Setting lab reports, medications,
• Concept of interest lab test, drug,

• Institution CPMC, RMRS, LDS,

23
ltform method"POST" action"wc_infomanage.cgi"gt lti
nput type"HIDDEN" name"info_med"
value35702"gt ltinput type"HIDDEN"
name"info_context" value"LabDetail"gt ltinput
type"HIDDEN" name"info_institute"
value"CPMC"gt ltinput type"HIDDEN" name"DOB"
value"1951-05-26"gt ltinput type"HIDDEN"
name"SEX" value"F"gt ltinput type"HIDDEN"
name"USER" value"ciminoj"gt lt/formgt
24
ltform method"POST" action"wc_infomanage.cgi"gt lti
nput type"HIDDEN" name"info_med"
valueDIGOXIN.."gt ltinput type"HIDDEN"
name"info_context" value"LabDetail"gt ltinput
type"HIDDEN" name"info_institute"
valueRMRS"gt ltinput type"HIDDEN" name"DOB"
value"1951-05-26"gt ltinput type"HIDDEN"
name"SEX" value"F"gt ltinput type"HIDDEN"
name"USER" value"ciminoj"gt lt/formgt
25
Conclusions

• Theoretical, empirical and practical approaches
  are working synergistically
• Satisfaction of CIS users needs is feasible
• Other institutions can share if they have
• Web-based CIS
• Users with information needs

26
Acknowledgments

• National Library of Medicine grant R01-LM07593
• Research assistants
• Anne-Marie Ramierez
• Mark Graham
• Leanne Currie