Cornerstone I: Representing Knowledge

1
Cornerstone I Representing Knowledge

• From Data to Knowledge Through Concept-Oriented
  Terminologies
• James J. Cimino

2
The first step on the path to knowledge is
getting things by their right names.

• -Chinese saying

3
Overview

• What is data to knowledge?
• Knowledge representation choices
• Knowledge-based terminology efforts
• Medical Entities Dictionary
• Proof of concepts

4
What is data to knowledge?

• Start with patient data in the medical record
• Enhance knowledge by
• gaining a better understanding of the patient
• learning relevant knowledge
• bringing smart systems to bear to apply knowledge
• discovering new knowledge from health data

5
Knowledge Representation

• Terminology for representing symbols
• Format for arranging the symbols

6
Knowledge Representation Choices

• Guideline implementation

7
Guideline Implementation

• Starren and Xie, SCAMC, 1994
• National Cholesterol Education Panel Guideline

8
National Cholesterol Education Panel Guideline
Measure Cholesterol Assess Risk Factors
9
Guideline Implementation

• Starren and Xie, SCAMC, 1994
• National Cholesterol Education Panel Guideline

• Three representations
• PROLOG (first-order logic)

10
NCEP Guideline in PROLOG

• rule_j(PID)-
• check_lab(PID,hdl,HDL,_),!,
• HDL gt 35,
• total_risk(PID,Risk),!,
• Risk lt 2,
• check_lab(PID,cholesterol), C,_),
• C gt 200,
• C lt 239,
• print_rule_j.

11
Guideline Implementation

• Starren and Xie, SCAMC, 1994
• National Cholesterol Education Panel Guideline
• Three representations
• PROLOG (first-order logic)

• CLASSIC (frames)

12
NCEP Guideline in CLASSIC

• (CL-DEFINE-CONCEPT C-PATIENT
• (AND
• (ALL CHOL
• (AND INTEGER
• (MIN 200) (MAX 239)))))
• (CL-DEFINE-CONCEPT G-PATIENT
• (AND C-PATIENT LOW-RISK-PATIENT
• (ALL HDL (AND INTEGER (MIN 35)))))

13
Guideline Implementation

• Starren and Xie, SCAMC, 1994
• National Cholesterol Education Panel Guideline
• Three representations
• PROLOG (first-order logic)
• CLASSIC (frames)

• CLIPS (production rules)

14
NCEP Guideline in CLIPS

• (defrule C2G2J Rules to reach box J
• ?f1 lt- (calculated-patient (state c)
• (done no) (hdl ?hdl) (name ?name)
• (test (gt ?hdl 35))
• gt
• (printout Patient ?name needs treatment)

15
Guideline Implementation

• Starren and Xie, SCAMC, 1994
• National Cholesterol Education Panel Guideline
• Three representations
• PROLOG (first-order logic)
• CLASSIC (frames)
• CLIPS (production rules)

• All three representations proved adequate for
  encoding the guideline

16
Knowledge Representation Choices

• Guideline implementation

• Terminologic knowledge

17
Terminology Representation Choices

• Frame-based

18
Frame-Based Representation

• Serum Glucose Test
• is-a Lab Test
• Measures Glucose
• Specimen Serum
• Units mg/dl

19
Terminology Representation Choices
Terminology Representation Choices

• Frame-based

• Semantic network

20
Semantic Network Representation
Serum Glucose Test
21
Terminology Representation Choices
Terminology Representation Choices

• Frame-based
• Semantic network

• Conceptual graphs

22
Conceptual Graph Representation

• Serum Glucose Test -
• (is-a) -gt Lab Test
• (measures) -gt Glucose
• (specimen) -gt Serum

23
Terminology Representation Choices
Terminology Representation Choices

• Frame-based
• Semantic network
• Conceptual graphs

24
Knowledge Representation Choices

• Guideline implementation
• Terminologic knowledge

25
Knowledge Representation

• Terminology for representing symbols
• Format for arranging the symbols

• Terminology and format for representing
  terminologic knowledge

26
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991

27
Jochen Bernauer, SCAMC, 1991

• Conceptual graphs to model findings

28
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991

• Rector, Nolan and Glowinski, SCAMC, 1993

29
Rector, Nolan and Glowinski, SCAMC, 1993

• GALEN project
• conditions grammatically haveLocation bodyparts
• fractures sensibly haveLocation bones
• femurs sensiblyAndNecessarily haveDivision neck

30
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993

• Campbell and Musen, SCAMC, 1993

31
Campbell and Musen, SCAMC, 1993

• Conceptual graphs and SNOMED
• Pain Chest Radiation to Left Arm

Pain -
(located in) -gt Chest (radiating to) -gt
Arm -gt (with laterality) -gt Left
32
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993
• Campbell and Musen, SCAMC, 1993

• Lindberg, Humphreys, McCray, Methods 1993

33
Lindberg, Humphreys, McCray, Methods 1993

• Unified Medical Language System

Concept
Lexical group
String
String
34
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993
• Campbell and Musen, SCAMC, 1993
• Lindberg, Humphreys, McCray, Methods 1993

• Rocha, Huff, et al., CBM, 1994

35
Rocha, Huff, et al., CBM, 1994

• VOSER
• A server architecture for managing terminologic
  knowledege

36
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993
• Campbell and Musen, SCAMC, 1993
• Lindberg, Humphreys, McCray, Methods 1993
• Rocha, Huff, et al., CBM, 1994

• Campbell, Cohn, Chute, et al., SCAMC 1996

37
Campbell, Cohn, Chute, et al., SCAMC 1996

• Convergent Medical Terminology
• SNOMED/Kaiser/Mayo
• Galapagos

38
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993
• Campbell and Musen, SCAMC, 1993
• Lindberg, Humphreys, McCray, Methods 1993
• Rocha, Huff, et al., CBM, 1994
• Campbell, Cohn, Chute, et al., SCAMC 1996

• Brown, ONeil and Price, Methods, 1997

39
Brown, ONeil and Price, Methods, 1997

• Read Codes
• Representation with GALEN model

40
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993
• Campbell and Musen, SCAMC, 1993
• Lindberg, Humphreys, McCray, Methods 1993
• Rocha, Huff, et al., CBM, 1994
• Campbell, Cohn, Chute, et al., SCAMC 1996
• Brown, ONeil and Price, Methods, 1997

• Spackman, Campbell, and Côte, SCAMC 1997

41
Spackman, Campbell, and Côte, SCAMC 1997

• SNOMED RT (Reference Terminology)
• Convergent Medical Terminology
• Description Logic Format

42
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993
• Campbell and Musen, SCAMC, 1993
• Lindberg, Humphreys, McCray, Methods 1993
• Rocha, Huff, et al., CBM, 1994
• Campbell, Cohn, Chute, et al., SCAMC 1996
• Brown, ONeil and Price, Methods, 1997
• Spackman, Campbell, and Côte, SCAMC 1997

• Huff, Rocha, McDonald, et al., JAMIA 1998

43
Huff, Rocha, McDonald, et al., JAMIA 1998

• Logical Observations, Identfiers, Names and Codes
  (LOINC)
• 4764-5 GLUCOSE3H POST 100 G GLUCOSE PO SCNC
  PT SER/PLAS QN

44
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993
• Campbell and Musen, SCAMC, 1993
• Lindberg, Humphreys, McCray, Methods 1993
• Rocha, Huff, et al., CBM, 1994
• Campbell, Cohn, Chute, et al., SCAMC 1996
• Brown, ONeil and Price, Methods, 1997
• Spackman, Campbell, and Côte, SCAMC 1997
• Huff, Rocha, McDonald, et al., JAMIA 1998

• Pharmacy system knowledge base vendors

45
Pharmacy System Knowledge Base Vendors
Country-Specific Packaged Product
Ingredient
Manufactured Components
Composite Trademark Drug
46
Knowledge-Based Terminology Efforts

• Jochen Bernauer, SCAMC, 1991
• Rector, Nolan and Glowinski, SCAMC, 1993
• Campbell and Musen, SCAMC, 1993
• Lindberg, Humphreys, McCray, Methods 1993
• Rocha, Huff, et al., CBM, 1994
• Campbell, Cohn, Chute, et al., SCAMC 1996
• Brown, ONeil and Price, Methods, 1997
• Spackman, Campbell, and Côte, SCAMC 1997
• Huff, Rocha, McDonald, et al., JAMIA 1998
• Pharmacy system knowledge base vendors

47
Medical Entities Dictionary (MED)

• New York Presbyterian Hospital
• 60,000 concepts (procs, results, drugs, probs)
• 208,242 synonyms
• 84,677 hierarchical links
• 113,906 semantic links
• 238,040 other attributes
• 66,404 translations (ICD9-CM, LOINC, MeSH, UMLS)

48
Central Controlled Terminology
49
MED Data Structures

• Semantic network

50
MED Semantic Network
Medical Entity
Plasma Glucose
51
MED Data Structures

• Semantic network

• MUMPS global

52
MED MUMPS Global

• med(1600) ltSERUM GLUCOSE MEASUREMENTgt
• med(1600,1) ltC0202041gt
• . . ,4) lt32703,50000gt
• . . ,5) ltgt
• . . ,6) ltSerum Glucose Measurementgt
• . . ,7) ltgt
• . . ,8) lt1724gt
• . . ,12) ltGLUCgt
• . . ,14) lt169gt
• . . ,16) lt31987gt
• . . ,17) ltmg/dlgt
• . . ,20) ltC000006gt
• . . ,23) lt1178gt
• . . ,50) ltSerum Glucosegt
• . . ,138) lt40444,40445,40446,59165gt
• . . ,156) ltMCNCgt
• . . ,161) ltQNgt

53
MED Data Structures

• Semantic network
• MUMPS global

• DB2

54
MED DB2 Tables
55
MED Data Structures

• Semantic network
• MUMPS global
• DB2

• Unix

56
MED UNIX Data Structure

• 1600SERUM GLUCOSE MEASUREMENT 1C020241432703
  45000012GLUC17mg/dl........

57
MED Data Structures

• Semantic network
• MUMPS global
• DB2
• UNIX

58
Proof of Concepts

• Merging data and application knowledge

59
Merging Data and Application Knowledge

• Class-based, reusable lab summaries

Chem20 Display
Serum Glucose Test
Fingerstick Glucose Test
Plasma Glucose Test
60
DOP Summary
61
WebCIS Summary
62
Merging Data and Application Knowledge

• Class-based, reusable lab summaries

Chem20 Display
Serum Glucose Test
Fingerstick Glucose Test
Plasma Glucose Test

• Expert system for application maintenance

63
Proof of Concepts

• Merging data and application knowledge

• Smarter retrievals from the record

64
Smarter Retrievals from the Record

• Repository stores events and results
• Clinical problems at a different level of
  granularity
• Re-use knowledge to map from problems to clinical
  data
• Produce problem-specific views of the medical
  record

65
Concept-oriented (Heart)
Radiology 2/28/96 Head CT
Lab 12/28/96 Sickle Cell Test
Admission 3/14/96 Stroke
Lab 1/1/99 Blood Type Test
Radiology 2/1/97 Knee X Ray
Admission 2/14/98 Angina
Discharge 1/15/99 CHF
Radiology 2/23/99 Chest X Ray
Lab 1/1/99 Cardiac Enzyme Test
66
(No Transcript)
67
(No Transcript)
68
(No Transcript)
69
(No Transcript)
70
Proof of Concepts

• Merging data and application knowledge
• Smarter retrievals from the record

• Just-in-Time education

71
Just-in-time Education

• Medline button
• Infobuttons

72
(No Transcript)
73
(No Transcript)
74
(No Transcript)
75
(No Transcript)
76
(No Transcript)
77
(No Transcript)
78
(No Transcript)
79
(No Transcript)
80
(No Transcript)
81
(No Transcript)
82
Just-in-time Education

• Medline button
• Infobuttons

• Text-to-Web

83
(No Transcript)
84
(No Transcript)
85
(No Transcript)
86
(No Transcript)
87
(No Transcript)
88
(No Transcript)
89
(No Transcript)
90
(No Transcript)
91
(No Transcript)
92
(No Transcript)
93
Just-in-time Education

• Medline button
• Infobuttons
• Text-to-Web

94
Proof of Concepts

• Merging data and application knowledge
• Smarter retrievals from the record
• Just-in-Time education

• Expert systems

95
Expert Systems

• Hripcsak, et al., Ann. Int. Med., 1995

96
Hripcsak, et al., Ann. Int. Med., 1995

• Identify chest x-ray reports suspicious for 6
  clinical conditions to trigger alerts
• Method Sens Spec
• Laypersons 22-47 97-99
• Radiologists 73-98 96-99
• Internists 68-98 97-99
• Keyword 51-79 79-92
• NLP/MED/Rule-based 81 98

97
Expert Systems

• Hripcsak, et al., Ann. Int. Med., 1995

• Clinical decision support system

98
Clinical Decision Support System

• Data monitor runs rules against incoming reports
• Tuberculosis cultures come back 4-8 weeks later
• One day, hundreds of TB alerts came in

99
What Happened to the Tuberculosis Alert?
?
Medical Logic Module
No Growth to Date
No Growth
100
How We Outsmarted the Lab
?
Medical Logic Module
No Growth to Date
No Growth
101
Expert Systems

• Hripcsak, et al., Ann. Int. Med., 1995
• Clinical decision support system

• DXplain Button

102
DXplain Button

• Elhanan, et al., SCAMC 1997
• Convert of test results to clinical findings

Serum Cholesterol Test

• Pass findings to DXplain

103
(No Transcript)
104
(No Transcript)
105
(No Transcript)
106
(No Transcript)
107
Expert Systems

• Hripcsak, et al., Ann. Int. Med., 1995
• Clinical decision support system
• DXplain Button

108
Proof of Concepts

• Merging data and application knowledge
• Smarter retrievals from the record
• Just-in-Time education
• Expert systems

• Data mining

109
Data Mining

• Wilcox and Hripcsak, SCAMC 1997

110
Wilcox and Hripcsak, SCAMC 1997
111
Data Mining

• Wilcox and Hripcsak, SCAMC 1997

• Wilcox and Hripcsak, SCAMC 1998

112
Wilcox and Hripcsak, SCAMC 1998

• Compare traditional coding methods with NLP to
  identify conditions in a set of patient records
  (x-ray reports)
• Method Sens Spec
• Laypersons 36 86
• Expert-coded cases 27-37 95-98
• ICD-9-coded cases 12-29 86-90
• Physicians 85 98
• NLP/MED/Rule-based 81 98

113
Data Mining

• Wilcox and Hripcsak, SCAMC 1997
• Wilcox and Hripcsak, SCAMC 1998

114
Proof of Concepts

• Merging data and application knowledge
• Smarter retrievals from the record
• Just-in-Time education
• Expert systems
• Data mining

• Database maintenance and use

115
Database Maintenance and Use

• Tables, columns, events all modeled in the MED
• Allows linkage of data model to controlled
  terminology
• Terminologies can be reused
• Impact of terminology changes on data model can
  be tracked

116
Proof of Concepts

• Merging data and application knowledge
• Smarter retrievals from the record
• Just-in-Time education
• Expert systems
• Data mining
• Database maintenance and use

• Terminology maintenance and use

117
Terminology Maintenance and Use

• Integrating terminologies from merging hospitals
• Automated update of medication terminology
• Detection of errors and inconsistencies

118
Proof of Concepts

• Merging data and application knowledge
• Smarter retrievals from the record
• Just-in-Time education
• Expert systems
• Data mining
• Database maintenance and use
• Terminology maintenance and use

119
Is it Worth the Trouble?

• Meed
• noun
• 1 archaic an earned reward or wage
• 2 a fitting return or recompense
• Date before 12th century
• Etymology from Old English
• MED

120
Summary

• Putting knowledge in your terminology gets you
• Better ways to get knowledge out of your EMR
• Better ways to get knowledge out of resources
• Better ways to use other knowledge bases
• Bettter ways to use terminology
• Better ways to manage applications
• Better ways to manage data and terminology
• Representation scheme is less important
• Desiderata for controlled terminology

121
Desiderata

• Desirable qualities for terminology

122
Desiderata

• Desirable qualities for terminology

Go placidly amid the noise and haste, and
remember what peace there may be in
silence. Id rather be sailing