A PopulationBased Laboratory Information Strategy

1
A Population-Based Laboratory Information Strategy

• Michael McNeely MD FRCPC
• Consultant in Medical Informatics, Victoria BC

2
Overview

• There will be an ever-increasing need for
  laboratory results to be knowledge-based to be
  interpreted, to guide treatment, and to smoothly
  integrate with the medical record.
• Canada Health Infoway is a government of Canada
  project whose goal is to have electronic medical
  records (EMR) for 80 of Canadas population by
  2010.  
• The Provincial Laboratory Information Solution is
  a BC project to provide a unified database of all
  laboratory results produced in the province.
  These two projects are at an early stage but
  eventually (phase III-IV) will incorporate
  knowledge support.
• The presentation will, by way of a review,
  discuss the potential for these initiatives to
  carry forward existing programs involving
  laboratory utilization control, risk management,
  chronic disease management, telepathology,
  epidemiology, genominformatics, and sample
  management.

3
Canada Health Infoway http//www.infoway-inforoute
.ca/

• What is Infoway?
• Canada Health Infoway Inc. invests with public
  sector partners across Canada to implement and
  reuse compatible health information systems that
  support a safer, more efficient healthcare
  system. Infoway is an independent, not-for-profit
  organization whose Members are Canada's
  14 federal, provincial and territorial Deputy
  Ministers of Health. Launched in 2001, Infoway
  and its public sector partners have over 100
  projects, either completed or underway,
  delivering electronic health record (EHR)
  solutions to Canadians solutions that bring
  tangible value to patients, providers and the
  healthcare system.

4
Canada Health Infoway http//www.infoway-inforoute
.ca/

• Mission
• To foster and accelerate the development and
  adoption of electronic health information systems
  with compatible standards and communications
  technologies on a pan-Canadian basis, with
  tangible benefits to Canadians.
• To build on existing initiatives and pursue
  collaborative relationships in pursuit of our
  mission.

5
Canada Health Infoway http//www.infoway-inforoute
.ca/

• Vision
• A high-quality, sustainable and effective
  Canadian healthcare system supported by an
  infostructure that provides residents of Canada
  and their healthcare providers with timely,
  appropriate and secure access to the right
  information when and where they enter into the
  healthcare system. Respect for privacy is
  fundamental to this vision.
• Goal
• To have an interoperable EHR in place across 50
  per cent of Canada (by population) by the end of
  2009.

6
Canada Health Infoway http//www.infoway-inforoute
.ca/

• Components of the HER
• Patient and provider registries 110 m
• Laboratory Results 150 m
• Medical Imaging 220 m
• Drugs 185 m
• Interoperable EHR 175 m
• Telehealth 150 m
• Public Health 100 m
• Innovation and adoption 60 m
• Infostructure 25 m

7
Evolution of EHR
8
Canada Health Infoway http//www.infoway-inforoute
.ca/

• Progress to date
• Standards adoption
• HL 7
• LOINC
• SNOMED CT
• Provincial Projects
• Ontario
• Others
• British Columbia

9
BC - The Provincial Strategyhttp//www.healthserv
ices.gov.bc.ca/cpa/publications/ehealth_framework.
pdf

• An HER provides each British Columbian with a
  secure and private lifetime record of their key
  health history and care within the health system.
• The record is available electronically to
  authorized health care providers and the
  individual anywhere, anytime, in support of
  high-quality care.
• For more information on the Electronic Health
  Record, please seehttp//healthnet.hnet.bc.ca/in
  dex.html

10
Provincial Laboratory Information Solution (PLIS)

• Planning and development activities to support
  Technology Transformation are being led by a
  dedicated PLIS Office within the PLCO, working
  with the Ministry of Healths Knowledge
  Management Branch. A joint PLCO/Ministry strategy
  which will lead to the creation of a Provincial
  Laboratory Information Solution (PLIS) for
  British Columbia.
• The overall guiding vision behind the creation of
  a Provincial Laboratory Information Solution
  (PLIS) for British Columbia is to provide access
  to clinical laboratory information (results,
  orders and decision support) to care providers at
  the point of care anywhere in British Columbia.
  PLIS is also a leading initiative within the
  Ministry of Health's broader E-Health strategy to
  develop the Electronic Health Record and support
  IT infrastructure for health care in BC.
• The Provincial Laboratory Information Solution
  (PLIS) will
• provide a standardized province-wide approach to
  presenting a patient's lab test results

11
Provincial Laboratory Information Solution (PLIS)

• electronically distribute lab test results to
  ordering and/or copied physicians
• make historical lab test results from both public
  and private laboratories within the province
  available to physicians
• create an electronic lab test ordering system
  with decision support tools
• improve the ability to aggregate laboratory
  information in order to support both
  administrative and clinical decision-making
• provide a provincial capacity to measure and
  manage the provision and utilization of
  laboratory services
• contribute to the realization of the provincial
  Electronic Health Record (EHR)
• Through the use of technology and standards, the
  new system will ensure laboratory information is
  of a high quality, available to authorized health
  care providers and administrators throughout the
  province, part of each patient's provincial
  Electronic Health Record

12
Provincial Laboratory Information Solution (PLIS)

• Features
• Organizational structure
• Unique bid process Joint Services RFP
• Development
• Time frame
• FUTURE COMPONENTS OF INTEREST
• Data Mining
• Clinical Decision / Knowledge Support
• Telepathology

13
Data Mining

• Utilization Control
• Reduce unnecessary duplication of testing
• Ensure adherence to utilization protocols
• Facilitate data evaluation in order to design
  utilization strategies

14

• Chronic Disease Management
• Clinical Practice Guidelines
• Provide objective data for CPG development
• Outcomes analysis
• Follow-up of adherence
• Follow-up for outcomes studies
• Makes more elaborate CPGs possible
• Disease epidemiology
• Assist individual physicians patient tracking
  (e.g. lists of diabetics in a physicians
  practice).
• Provide physician reminders re chronic disease
  patient reviews
• Provide availability to a package of physician
  specific database searches on their own patients
  (e.g. a list of all registered diabetics in a
  given practice with statistics on their frequency
  of A1C testing compared to provincial norms).
• Patient reminders

15

• Special Disease Registries/Services
• Automated development of registries of diseases
  characterized by laboratory test results (e.g.
  hemoglobinopathies, hypercholesterolemia,
  diabetes, hemochromatosis, and many others as
  genetic testing expands)
• Specialized knowledge support tools and
  information for both physicians and patients
• Mailing list of physicians/patients to be
  informed when new information becomes available.
• Epidemiology
• Classic infectious disease epidemiology (but
  closer to real-time)
• Real-time epidemiology for epidemics (e.g. SARS)
  and bioterrorism
• Chronic disease epidemiology (non-infectious)
• Health Care System Management
• Outcomes data
• Utilization management
• Population trends
• Test usage and deployment of resources
• Physician ordering profiles

16
In 1982 I gave a talk on this very same
subject. I covered the following types of
automated interpretations.

• Level 1 Standard comment on every report of a
  specific test.
• Level 2 Result specific comment 1-test.
• Level 3 Result specific comment 2-or more
  tests, over time, or other clinical information
• Level 4 More sophisticated approaches.
• Now, in 2006 we havent managed Levels 1-3
  completely but were now looking at Level 4 and
  various projects may bring Level 4 to fruition
  within the next few years.

17
Canned Comments

• GOOD THINGS
• Demonstrated ability to change physician
  behaviour
• Demonstrated ability to enhance use of laboratory
  testing (e.g. ?utilization, ?diagnosis)
• CAUTIONS
• Limited clinical information
• Comment added whether needed or not
• Consume space on a paper report
• Paper report has a rigid format
• Some doctors feel threatened/insulted
• Patient overreaction (patient access)

18
Human Generated Comments

• Questions
• Are the interpretations part of the legal report?
• Should the interpretations be added to EMR?
• Who should be permitted to prepare such
  interpretations?
• Human generated reports have error rate of up to
  50 (Lim Clin Chem 2004)
• Marshall Challand (Ann Clin Biochem 2000)
• Variation amongst interpreters
• Communication style variable
• Clinical information available is not always
  appropriate to the test being interpreted
• Little feedback regarding usefulness
• Interpretations should be recipient specific

19
Laposata (Clin Chem 2004 50 471)

• Laposata has championed the need for
  human-generated, patient-specific narrative
  interpretations
• He has criticized the canned comment
• BUT he compares Apples and Oranges
• Laposata makes the case for why Knowledge Support
  is needed.

20
Knowledge Support

• a.k.a. Clinical Decision Support
• Two forms
• Static PubMed, Lab Tests On-Line, ARUP
• Dynamic or CARTKS (Context Appropriate Real Time
  Knowledge Support)
• Specific Interpretations

21
The Case for Knowledge Support / Clinical
Decision Support

• Hundreds of publications have demonstrated its
  potential usefulness
• Several publications have pointed out potential
  problems but none has undercut the basic premise.
• Clinical Practice Guidelines
• Ever increasing numbers
• Poorly applied ( 25 adherence)
• Limited complexity

22

• It is likely that when electronic knowledge
  support tools become a standard feature of
  medical practice, the protocol and CPG approach
  will be maximized.
• McNeely Clinics of Laboratory Medicine 2002 22
  1-10
• It is so apparent that computerization will
  enhance the application of CPGs that it may be
  unethical to continue to perform trials to answer
  this question.
• Ellson and Connolly JAMA 1998 279 989.
• To be widely accepted by practicing clinicians,
  computerized support systems for decision making
  must be integrated into the clinical work flow.
  They must present the right information, in the
  right format, at the right time, without
  requiring special effort.
• James BC NEJM 1999 340 1202.

23
Ripple-Down Rules

• Developed by Paul Compton and Gordon Edwards of
  St. Vincents Hospital, Sydney AU
• Original system PIERS
• Now marketed by Pacific Knowledge Systems
  http//www.pks.com.au/ as LabWizard
• Rule-Based but no knowledge engineer

24
Ripple-Down Rules
Lab Completes Test
Knowledge Base Inference Engine
Verified Result Combination?
No
Yes
LIS Reports Result And Interpretation
Result Combo Interpreted
Integrator
25
LabWizard (example)
26
BloodLink
Clin Chem 2002 48 605.
Marc van Wijk MD PhD Delft, The Netherlands
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
BloodLink Evaluation
Test reduction of 19.6
50 GPs Two Groups 1-Year
32
Laboratory Advisory System

• Chang E, McNeely MDD, Gamble K. Strategies for
  choosing the next test in an expert system.
  Proceedings of the congress on medical
  informatics. AAMSI 1984 2198-202.
• McNeely MDD, Smith B. An interactive expert
  system for the ordering and interpretation of
  laboratory tests to enhance diagnosis and control
  utilization. Canadian Medical Informatics.
  May/June 199516-19.
• Smith BJ and McNeely MDD. The Influence of an
  Expert System for Test Ordering and
  Interpretation on Laboratory Investigations.
  Clinical Chemistry 1999 45(8) 1168-1175.
• Clinical-Laboratory.com Old Marlebone Rd,
  London, England

33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
(No Transcript)
37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
Results of a trial
45
The LAS Study Conclusion

• The development of test ordering strategies can
  be enhanced.
• The interpretation of the test results can be
  enhanced.
• A statistical database of diagnosis, clinical
  information, test orders, and results can be
  readily derived. Such information is unique and
  is available for optimizing and developing
  testing strategies and for laboratory management.

46
The LAS study conclusion (cont)

• An appropriate search of the database would
  enable clinician-targeted education and
  utilization feedback to be derived.
• Examination of the database at the time of
  ordering would enable the development of a module
  to identify unnecessary, duplicate testing.

47
Contextualized Report Dr. Jonathan Kay
(Oxford)Drs. Bruce Friedman and Jules Berman
Lab Medicine 2006 37 121.
48
Smith, John H. Male
46 yoa 23957988-1 Dr. Louis
Pasteur DOS June 7, 2006
Test Name
Result
Reference Interval
Alkaline Phosphatase
128 ?
20 105 U/L
49
Smith, John H. Male
46 yoa 23957988-1 Dr. Louis
Pasteur DOS June 7, 2006
Test Name
Result
Reference Interval
ALERT !! Patient is taking Chlorpromazine which
is known to cause Cholestasis with increased Alk
Phos.
Alkaline Phosphatase
128 ?
20 105 U/L

• Analytical Information Alkaline Phosphatase
• Laboratory validation studies
• Method reference
• Instrument validation studies
• Proficiency testing record
• Complete Bibliography Click here ?

50
Smith, John H. Male
46 yoa 23957988-1 Dr. Louis
Pasteur DOS June 7, 2006
Test Name
Result
Reference Interval
Alkaline Phosphatase
128 ?
20 105 U/L

• Analytical Information Alkaline Phosphatase
• Laboratory validation studies
• Method reference
• Instrument validation studies
• Proficiency testing record

51
Smith, John H. Male
46 yoa 23957988-1 Dr. Louis
Pasteur DOS June 7, 2006
Test Name
Result
Reference Interval
Alkaline Phosphatase
128 ?
20 105 U/L

• Reference Interval Alkaline Phosphatase
• Literature Reference
• In-house studies
• Notes
• Graphical Presentation

52
Smith, John H. Male
46 yoa 23957988-1 Dr. Louis
Pasteur DOS June 7, 2006
Test Name
Result
Reference Interval
Alkaline Phosphatase
128 ?
20 105 U/L

• Reference Interval Alkaline Phosphatase
• Literature Reference
• In-house studies
• Notes
• Graphical Presentation

53
Smith, John H. Male
46 yoa 23957988-1 Dr. Louis
Pasteur DOS June 7, 2006
Test Name
Result
Reference Interval
Alkaline Phosphatase
128 ?
20 105 U/L

• Interpretation Alkaline Phosphatase
• Causes of an increased Alkaline Phosphatase
• Causes of an decreased Alkaline Phosphatase
• Specific Interpretation of this result
• Request a personal consultation on this result

54
Genoinformatics

• Screening tests
• PCR Testing
• Proteomics
• Physician Understanding
• Patient Information - Counselling
• Family Studies
• Long-term
• Follow-up
• New Knowledge

55
Knowledge Assembly

• Major problem is the creation/assembly of
  Knowledge Support tools (e.g. 1 rule per hour or
  committee)
• Must have AUTOMATED knowledge assembly
• Must have generic Inference Engines
• Must rely on the integrative intelligence of the
  user

56
Inference Engine
IF
Alkaline Phosphatase gt ULN
AND
Age gt 70
THEN
3
Consider Pagets Disease
57
Knowledge Assembly
2ND most important slide

• Facts
• From Electronic Medical Record
• Added at time of ordering
• Added during interpretation
• Rules
• Grunt approach
• Formal Committees (worldwide?)
• CPGs
• Wikipedia format
• Medical Literature
• Database

?
Constructed
?
Automated
58
Wikipedia
59
Medical Literature
60
Data Mining
61
Telepathology

• Goals of Province-wide program VISION
• Organizational structure
• Overview
• Standards

62
(No Transcript)
63
Evolution Step
Infrastructure
STDS
Accred Rules
Privacy

• Feasibility trials of SF
• Thinking about V
• Education

PLIS PACS
Existing
Record of image (not image)

• Routine SF
• Some trial virtual

Linkage to API LIS
Licensing ?
HL7SNOWMED.CT LOINC

• Mature use of S F
• Some HA use V routinely for limited APS trial
  HA-HA

Intra DeptStorage
DICOM2009
ImageRepository PACS

• Mature SF
• Routine, limited ApV, routine HA-HA

• Additional Aps more common use

64
Conclusion
Most important slide

• The EHR 2009-2010 Clinical Decision Support
  If Decision Support is expected in 2 - 4 years
  then planning MUST start now.
• If Knowledge Support is to be meaningful then
  building the Knowledge Bases must begin soon
  but, we will need to know how they will be
  executed and what the Inference Engine will look
  like.
• If Laboratory Professiolnals expect to be
  involved in the interpretation of the results
  they produce they must get involved in the
  development of the Decision Support modules or
  risk being disintermediated.

65
Conclusion