The Future of All Things LIS

1
The Future of All Things LIS
Ulysses J. Balis, MD Past-President of the
Association for Pathology Informatics Associate
Professor and Director, Clinical
Informatics Co-Director, Division of Pathology
Informatics Department of Pathology University of
Michigan Health System Ann Arbor,
Michigan ulysses_at_umich.edu
Michael McNeely, MD President of the Association
for Pathology Informatics Medical Director -
Provincial Laboratory Information Solution
(PLIS) 4481 Shore Way, Victoria V8N3V1 (250)
477-7758 mmcneely_at_islandnet.com
APIII 1030 am 1200 noon October 22,
2008 Session B2
2
Anatomic Pathology

• Historical Overview of AP
• Present Gap Analysis
• Well-founded extrapolation of APs manifest
  future state

3
Common Themes Between AP and CP as compelling
Merger Use Cases

• Both realms may be viewed as production
  environments, managing a plurality of assets,
  both tangible and informational
• Both realms have multi-step processes,
  susceptible to error
• Both can benefit from automation, just-in-time
  principles and single piece workflow
• Both have shared vulnerabilities in their present
  value stream (process steps out of scope for
  extant workflow)

4
Then
AP data stewardship approach circa 1972
5
Contemporary AP-LIS Capabilities a brief gap
analysis

• Moderate, but not comprehensive, use of
  computational support
• Tracking largely manual (paper-based)
• Validation Methodologies and computational
  support of such lags behind CP
• Lack of LIMs-type concepts
• Uncontrolled inventory
• Loss of control of data once it is forwarded to
  external systems
• Lack of Standards
• Lack of flexible reporting of rich data formats
• Lack of Integrative reporting
• Lack of degree of development of Middleware
  concepts as in CP
• Fewer AP-LIS informatics Specialists than in CP
• Lack of recognition that AP workflow transcends
  the Pathology Department boundaries
• Current workflow and reporting approaches still
  represent a significant cumulative risk for error
• Specimen identification
• Correct reporting
• Successful clinical handoff

Now
6
Asset Tracking

• Although firmly entrenched in CP workflow, AP
  use-cases and solutions are just beginning to
  appear, ostensibly due to the lead time that was
  required for vendors to provide suitable
  technical solutions
• This delay in emergence was similarly potentiated
  by the relative paucity of AP-oriented
  Informaticians

7
Comprehensive Asset Tracking
Evolution of Enabling Technologies Many of the
present-state batch-based practices in use are in
place as a logical result of historical
limitations with respect to labeling and asset
tracking technology. In early test-deployments
of multilayer direct thermal label stock (A)
seven major classes of so-called high performance
formulations failed following a thirty minute
xylene immersion. Newer formulations, such as
General Data Stainer Shield labels (B), are able
to withstand both xylene immersion as well as
exposure to the far majority of routine
histochemical staining protocols, thus providing
for the first time the opportunity to leverage
just-in-time single piece workflow in the
histology suite. Similarly in the case of
cassette printing technology (C), first and
second generation based printing systems, which
were dependent on thermal carbon or ink jet
transfer (upper row), lacked both spatial
resolution and consistency to provide for a
consistently machine-readable barcodes, whereas
newer platforms, based on laser-ablative
reduction of a black polymeric coating, provide
precise and highly consistent barcodes of very
high resolution (lower row). Two dimensional
barcodes, which have intrinsic error correction,
are made possible by this higher rendered
resolution. While it is possible to leverage
laser-based printing approaches for direct slide
annotation (D), the expense associated with the
current generation of these printers makes them
unapproachable for just-in-time generation of
annotated slides at individual workstations. For
this reason, the label-based approach, which
allows for printing of slides at every cutting
station (see figure to right), is a superior
solution for maintaining single-piece workflow.
8
Real-time tracking of workflow
Single Piece Workflow as an Overall Process
Map When applying single-piece workflow
solutions to the entire value stream process, it
becomes possible to concurrently drive all
sources of error to near zero incidence. An
essential key of this transformation is
eliminating all manual or clerical steps, an
rather, replacing them with barcode/information
technology-driven solutions.
9
(No Transcript)
10

• Per-user workstations
• minimal bench footprint impact
• ergonomically positioned
• fast response

• Stainer Shield Labels (General Data Corporation)
• Solvent resistant
• Fade resistant
• Direct thermal and thermal transfer technologies

11
No manual labeling
12
In addition to full demographics, slides contain
routing codes as well as histology staff
identifiers.
13
Effect of Single Piece Workflow with use of
Middleware Barcode Tracking upon Productivity A
six month average estimate of time to section
each slide is depicted by the solid line.
Significant variability is noted. Following the
deployment of just-in-time slide label printing
(dotted line) there is both a significant
reduction in overall processing time per slide
(plt0.00001) and concurrently, a decreasing trend
line (near statistical significance). More
importantly, conversion to a barcode-based
approach for this workflow step essentially
removed all inter-slide label transposition
errors.
14
AP Test Validation Methodologies

• Question to what level of rigor do we subject
  our new AP-lab tests (IHC as an exemplar)?
• ROC analysis?
• Is such support for either the analysis or the
  documentation of such in the AP-LIS?
• What tools are in place for sliding-window
  analysis of the ongoing operational confidence
  limits of such tests?

15
Lack of LIMs-type concepts in AP Systems

• LIMs Laboratory information management systems
• Term associated with research labs, pharma, CROs
  and reference facilities
• Ability to track not only specimen inventory, but
  also that of
• Reagents (lot numbers, batch validation data,
  systematic drift)
• Initial test validation and ongoing test
  performance data
• Ability to generate detailed worklist and
  rework-list data, based upon possible sentinel
  event data (failed reagent (IHC, molecular
  assays)
• Data views and management reports based on
  ongoing ROC data
• AP systems, in the absence of such capabilities,
  have considerable vulnerabilities for
• Test failure
• Failure to detect test failure
• Failure to seamlessly recover from such failures,
  and in an automated fashion
• AP Systems represent, consequently, only a
  partial leveraging of the full potential of
  information technology and process tracking.

16
Lack of Standards for AP Information Exchange

• What is the current state of the art in AP
  consult case exchange?
• Paper-based
• Postal based
• No common format for electronic data interchange
• Not even a common barcode format of set of
  barcode data element definitions
• Several timely developments
• Lab Infotech Summit 2009 LITS-Interop
• CLSI Barcode Specimen Label Standards Initiative

17
Lack of Flexible Reporting for Rich Data Formats
in AP

• HL7 2.x remains largely text based
• Column width limits and loss of tabular
  formatting
• Pagination horror stories
• Loss of rich content

18
Lack of Integrative Reporting in AP

• Artificial segregation of datasets based on
  historical compartmentalization of the lab,
  despite the reality that (1) both effective case
  triage (within the lab) and (2) clinical
  interpretation would be well-served by
  integration of said content.
• Exemplar Hematopathology reporting
• Separate hematology, flow histopathology/immunohis
  tochemistry interpretation
• Not insignificant possibility that the resultant
  plurality of reports will not exhibit internal
  agreement ? clinical management risk

19
Lack of Middleware Concepts in AP

• Inability to add functionality independent of
  core vendor participation / development
• Stasis in the development of new workflow and
  patient safety initiatives
• Delay in the realization of tracking / JIT
  technologies
• Difficulty in integrating multiple repositories
  for aggregate management of data

20
Fewer AP-LIS Informatics Specialists Than for CP

• Less specialty activity and expertise fewer
  solutions rendered over time, as compared to
  degree of data management sophistication and
  automation in the Clinical Lab
• Will slowly rectify itself as more Informatics
  Fellowships emerge (as these typically emphasize
  AP, as well as CP, concepts)

21
Closed Loop Clinical Communication of AP Results

• Printing the report / sending the report to a
  downstream interface is no longer a minimum
  standard, if we are to provide true patient care.
• Need to close the loop with the clinician NOT
  with the EMR
• Patients DIE every day due to this lack of
  capability AP Middleware allows this feature
  instantiation without the need to resort to
  vendor upgrades or platform exchanges
• Dr. Jeff Myers will lecture on this area
  extensively a little later this meeting.

22
Rare Event Detection Enhancement of Patient
Safety

• Real-time tracking of ever aspect of asset
  inventory and workflow progression
• Use of rules engines to identify when a low
  incidence event is not detected as humans are
  just pathetic (at best) at executing exception
  workflow practices based upon low incidence
  events. Moreover, the threat of punishment (or
  even bodily harm) does not work, in terms of
  alleviating subsequent errors.

23
Trends with Central Hospital I.T. Strategic
Planning

• Realization that central repositories designed
  for clinical decision support are increasingly
  difficult to support.
• Realization that ancillary departments (LIS, RIS,
  OR, ED, etc.) are optimally empowered to control
  both content and display of their data.
• Identification of monolithic repositories as
  being ideal for research support and not
  necessarily for clinical support.

24
Present State

• Effort expended to interface instruments to the
  LIS and the LIS to the HIS represent an
  increasing percentage of overall operational
  effort of the generally fixed Lab IT resource.
• Hospital migration of HIS platforms can impose a
  sudden and daunting resource requirement on lab
  IT resources to design, implement and test the
  myriad of extant interfaces required by such a
  new system
• The current language of interface HL7 2.x
  represents a formalized exchange construct that
  requires precisely constructed specification of
  the sender/receiver pair to create a functional
  data channel this is labor-intensive and
  requires specialized programming expertise.

25
Some Estimates of Effort

• Typical LIS implementation
• 10 instrument interfaces
• 2 HIS interfaces
• 40 programmer hours/interface/year as a
  conservative support metric
• 800 hours effort/site-year
• Millions of programmer hours of interface
  maintenance and customization effort annually.

26
The Imbalance of Infrastructure to Required
Support
Ideal Oversight Model
Hospital I.T. Continuum
Clinical Laboratory Continuum
Demarcation of Intended Scope of
Influence/oversight
27
The Imbalance of Infrastructure to Required
Support
Diminishing Central IT Expertise
LIS
HIS
Hospital I.T. Continuum
Clinical Laboratory Continuum
Demarcation of Intended Scope of
Influence/oversight
28
The Imbalance of Infrastructure to Required
Support
Diminishing Central IT Domain Expertise with Lab
IT Compensation
LIS
HIS
Hospital I.T. Continuum
Clinical Laboratory Continuum
Demarcation of Intended Scope of
Influence/oversight
29
The Pending Imbalance
Growing domain reporting complexity of
ancillary specialty information systems

Growing Gap in representation model
Complexity
Fixed Hospital IT expenditures as a percentage of
the overall budget
2013
2011
2007
2009
LIS Data Model complexity exceeds HIS receiver
capabilities
30
HL7 Receiver
HL7 Receiver
Just-in-time Aggregation
Multiple unique instances of customized HL7
interfaces which are not necessarily centrally
managed or organized for consistency in
deployment style
Just-in-time delivery
HL7 Receiver
Repository
Future State
Present State
31
What is Federation?
Conventional Data Model
LIS
User
Centralized HIS
RIS
OR
other
E.D.
Conventional HL7 interface
32
What is Federation?
Conventional Data Model
Chem
User
Centralized LIS
AP
Coag
Flow
Hematology
Real-time interfaces
33
What is Federation?(at the hospital level)
Potential Revised Data Model
LIS
RIS
OR
E.D.
other
Web-based Just-in-time aggregation
User
Single SQL (or ODBC) query Concurrently vectored
to each Participatory Single Source of Truth
SSOT shifted to the appropriate domain-specific
stewards of data from the HIS domain
Participatory SQL servers
34
What is Federation?(at the hospital level)
Potential Revised Data Model
LIS
RIS
OR
E.D.
other
Web-based Just-in-time aggregation
User
Single SQL (or ODBC) query Concurrently vectored
to each Participatory Single Source of Truth

• Consequences of shifting to a SQL-based SSOT
  model
• Data only represented once in overall enterprise
  model
• Reduction in number of interfaces requiring
  support
• Potential to transfer classes information other
  than text
• Reduction in support responsibilities of central
  hospital IT.

Participatory SQL servers
35
What is Federation?
Lab Areas
1
2
3
other
User
Web-based Just-in-time Aggregation, analysis
and report generation tools
Service Oriented Architecture (SOA), Single SQL
(or ODBC) query Concurrently vectored to
each Participatory Single Source of Truth
LIS
SSOT shifted to the appropriate domain-specific
stewards of data from the central LIS domain
Participatory SQL servers
36
Anticipated Growth Areas for Merger AP/CP
Information System Design and Deployment

• Integrated reporting of crossover result types
• Hematopathology
• Indexed / multiplexed assays
• Integrated LIMs capabilities
• Integrated LEAN / Six Sigma tools
• Integrated ROC validation capabilities
• Best-of-breed enablement made possible by common
  Service-oriented architectures, data buses and
  effective use of federation
• Use of HIS Resulting Federation Models
• Integrated CRM tools
• Portals
• Just-in-time approaches applied to all aspects of
  workflow
• Redefinition of the boundaries of the LIS (TLIS)

37
Integrated Reporting of AP/CP Crossover Result
Types

• Report models based on clinical management needs
  and not that which is convenient to the labs
  workflow
• Ability to effectively identify discordant
  primary data, prior to releasing clinically
  confusing or even misleading diagnostic
  information, thus preventing mismanagement of
  patient treatment

38
Integrated LIMs Capabilities

• Ability to identify and correct failed
  high-throughput testing
• Validation of methodologies
• Documentation
• ROC analysis

39
Use of HIS Resulting /Federation Models

• Ability to take control of both content and
  formatting of resultant data
• Removal of referential Integrity problems that
  currently plague complex enterprise systems with
  multiple interfaces

40
Best-of-Breed Enablement

• With use of federation it becomes possible to
  extend extant systems in a direct and
  straightforward way, without dependence on vendor
  expertise and resources
• Allows site to employ quick-follow strategies,
  uncoupled from vendor development timelines

41
Integrated CRM tools and portals

• Will become requisite as the standard of practice
  will mandate that both Patient portals and
  physician portals be made available
• Will allow for integration with the emerging
  plethora of personal health record repository
  services

42
The Complexities Brought On by High-throughput
Testing

• Reagent inventory must be tracked just as
  specimens are
• Possible total-batch or sub batch failure
• Results are subject to periodic reinterpretation
• Need for a closed-loop model with the ordering
  clinician to convey new susceptibility
  interpretations, odds ratios and prognostic
  information

43
Final Thoughts on the Pending Merger

• It WILL happen
• Solutions will impart quality, flexibility and
  efficiency to our existing workflow
• AP will benefit to an even greater extent than CP
• Conventional lines of demarcation between
  historically segregated work centers will
  diminish in importance.

44
Open Discussion and Questions