Bar-coding in AP - Pathology Informatics 2015

advertisement
Bar-coding in AP: OmniTrax
as a Full Middleware Solution
Rodney Schmidt, MD, PhD
Professor of Pathology, Director of
Medical Informatics (Pathology)
University of Washington, Seattle
Today’s Story
Lessons from OmniTrax
– Lean processes and workflow
– Deeper understanding of barcoding
• Different levels of barcoding with different benefits
– Measures of benefits
• Quality and efficiency
• Workflow dependent!
– Current capabilities
Trade-offs using a middleware solution
Need for a bar-code standard
Disclosure
• Bar-coding software developed at UW (OmniTrax and
OmniImage) has been licensed by UW to Pathway
Pathology Consultants for PowerPath end-users.
• Dr. Schmidt and his team have a revenue-sharing
agreement with UW.
• Dr. Schmidt has a consulting agreement with ThermoFisher for educational talks.
• No other financial relationships with hardware or
software manufacturers.
Why barcode?
• Expensive
– $23k/gross station
– $10k/cutting station
– Software
• Workspaces change
– Wiring, networking
• Time investment
• Processes change
– Material handling
– QA
• Jobs change
– Workflow
– Change management
• Pathologists affected!
– Software fast
– Workspaces slow
– Financing slow
Who needs the hassle?!
Why barcode?
• Error reduction and patient safety
– Errors labeling things
– 1/300 (manual) to < 1/10,000,000 (datamatrix)
• Reduced medical-legal liability
• Custodial responsibility & inventory control
• Self-interested reasons
– Helps you do your job faster
– Reduced time wasted on error resolution
– Indirect efficiencies because of better knowledge
about where things are
What is Bar-coding?
• Labeling
– Putting barcodes on things
– Technically easy, cheap (some methods)
• Tracking
– Location updates; inventory control
– Added work; needs software; modest cost
• Driving
– Using barcodes to expedite workflow
– Disruptive technology; expensive; LIS
interoperability
Bringing Bar-coding to AP
• Track slides (2005)
– Eliminate the “lost slide” problem
– Ease conference prep
• Specimen labels (2006)
– Tissue discards and tracking
– Drive gross photography
• Block creation and labeling (2008)
– Automated JIT production of barcoded blocks
– Gross room QA process and tracking
• Slide creation and labeling (2008)
– Automated JIT creation of barcoded slides
– Facilitate workflow and QA
• Eliminate all manual labeling (and errors)
• Facilitate workflow – JIT information display
Achieved Benefits
• Marked reduction in labeling errors
• Improved inventory control (i.e. knowledge of
where things are)
• Direct savings of ~ 3 FTE
• Indirect savings of >> 0.5 FTE
• Improved image collection and management
(paperwork, gross, micro, EMs, IF, etc)
• Increased job satisfaction
Bar-coding Options
• Buy LIS-specific
– Available? Capable?
• Buy 3rd party solution (middleware)
– Available? Capable?
• Build LIS-specific middleware
– Can be quick. Investment.
• Build LIS-agnostic middleware
– Most complex; most control
Design Principles
• No scanning without benefit
– User acceptance; minimal training
• No manual data entry
– Eliminate human errors
• Use barcodes to drive workflow
– Efficiency
• Make nothing until it’s needed
– Eliminate handling and error opportunities
• No assumptions – only trust scan events
– Quality timestamps, locations, personnel
• Leverage LIS
• LIS-agnostic design
Material identification (2005)
• Handwritten
specimen
labels
• Manual, offline cassette
labeling
• Hand-written
slide labels
Primary labeling errors (2004)
1000
900
800
700
600
500
400
300
200
100
0
?
Blocks
Recorded
Slides
Actual
Accession number is re-entered into a
standalone cassette imprinter
Targets – Gross Room
• Foolproof labeling
– No human labeling/data entry
• Reduced dependence on
support staff
– Off-hours availability
– Redirection of support
personnel
• Reduced waste of cassettes
• Grossing step at least as
fast as current
• (Record timestamps)
The unsupervised Resident!
Targets - Accession
Receive specimen
and enter data into
the LIS
Generate a bar
coded label for the
specimen and
laboratory request
form.
Minimum extra
keystrokes (one)
Classic Grossing
Workflow
Accession specimens
Label specimens
Label cassettes
Group with specimens
Move to staging area
*
*
*
Move to gross bench
Lay out cassettes
Fill cassettes
*
*
Request more cassettes
Store excess with specs
Handling steps
Rack filled cassettes
Reconcile with LIS
*
Transport for
processing
*
Possible errors
* QA steps
Just-in-Time Printing
Accession specimens
Bar-code specimens
Fewer handling
steps
Fewer (1) error
opportunities
Scan/print cassettes
Lay out cassettes
Fill cassettes
*
*
Fewer QA
processes
Courtesy General Data
Rack filled cassettes
*
Transport for
processing
*
Q&E Benefits
“Classic”
“Just-in-Time”
Handling steps
11
5
Error opportunities
9
1
Manual QA steps
7
4
Primary labeling errors
988/yr (est.);
(1.2%)
2 in 3 mo (initial);
0 in next 7 mo;
(0.003%)
Cassette wastage
~25/d (~7%)
~0
Grossing efficiency
--
At least as fast
Support staff
--
0.75+ FTE saved
Histology – Embedding
• Target
– View critical
information about block
and specimen
– Efficient workflow
• Block scan:
– Embedding instructions
– Number of pieces of
tissue
– Specimen info
– (Record timestamps)
Histology – Cutting
• Targets
– Present critical information
(block, specimen)
– Eliminate manual slide labeling
– Block/slide verification
– Multiple workflows
– No clutter
– Efficient
• Touch-screens; no keyboards
• Block scan:
– JIT slide printing/labeling
– Info display
• Slide scan:
– Block/slide match
Cutting - Benefits
• Elimination of
hand labeling
• Much faster than
manual labeling
for blocks with
many slides
• Fewer block/slide
mismatches
• Overall throughput
increased ~10%
Slide Life Cycle
Histology
Pathology Offices
Sendouts
Faculty signout
File
Histology work
order completes
with scanning
Pull for
conference
Ship
Resident
review
Deliver
Slides – Benefits
• Less staff time looking for slides
• Faster to find last location than make a phone
call
• Fewer arguments over whether slides were
delivered
• Fewer recuts?
• Improved job satisfaction
– ** Saved me 30 min the first day! **
• Overall savings > 2.0 FTE!
Slides Benefits
FTE Savings
Histology
+0.5
FTE
Reduced time hunting for
mis-delivered slides
+0.5
FTE
Auto completion of outstanding orders
when slide is scanned
Office staff +.5-1
FTE
+.25
FTE
Reduced time for conference
preparation
Increased efficiency regarding send
outs
Barcodes Enable…
• Imaging
–
–
–
–
Gross photos
Photomics
Documents
EM/IF
• Specimen
management
– Discards
– Locations
• Winscribe automation
• HPV workflow
– Reflex testing
– Digene/Luminex
Targets - Specimens
• Discards
– Accurate
– Efficient
– Documented
• Track location
• Drive photography
Specimen Discard
Workflow
– Device scans
specimen barcode
– Handheld device
queries AP-LIS
• If case signout
occurred <2wks prior
• If case signout
occurred >2wks prior
• If note on Req Data
tab, caution light and
note display
Barcoding Benefits
• Direct personnel (FTE)
– 2.0
– 0.75
– 0.1
– 0.1
– TBD
Slide delivery and tracking
Cassette printing
Specimen discards
Document scanning
Fluorescence image import
~$150,000/yr assuming $50,000/FTE
Barcoding Benefits
• Indirect personnel (FTE)
– 0.5
– TBD
– TBD
Scanned consult document availability1
Scanned Req forms
Slide location info (e.g. Pathologists)
• Reduced loss of materials
– Slide/Block tracking
– Specimen discards
1Schmidt,
RA, et al. Am J Clin Pathol 126:678-83, 2006
Barcoding Benefits
• Error Reduction
– Elimination of all manual labeling steps!
– Reduced labeling errors
• Specimens
• Blocks
– ~988/yr to near 0
– “How did you manage to do that?!”
•
•
•
•
Slides
Gross photos
Scanned documents
Photomicrographs
OmniTrax – What’s new?
• Interface model for interacting with LIS
• More customers
– OHSU
– NYU
• HPV workflow implemented
• Gross/Histo enhancements
• (Cytology support)
• (Immunostainer interfaces)
– Leica Bond 3
– BioCare intelliPATH
• (Archives tracking port)
• (Slide tracking port)
Middleware
Software that bridges a human to one or more major systems
Advantages
• Leverage the power of core
systems
• Deliver niche functionality
• Avoid duplication of core
functions
Disadvantages
• Ongoing interoperability
• If you build your own:
Independence and control
• LIS data model poor
–
–
–
–
–
Open hardware options
Portability between LISs
Short bug/fix cycle
Implement functions you need
Tune and refine prn
– LIS upgrades
– Might change LISs
• Negotiate interfaces
– Extract data
– Write data
– Too simple
– Missing concepts
• If you build your own:
Ongoing support obligation
Basic Architecture
OmniTrax
LIS
Agent
UI/
app
UI/
app
Agent
Business objects
Database
UI/
app
QA Reports
Local Extensions
OmniTrax
UI/
app
UI/
app
UI/
app
UI/
app
Web app
LIS
Agent
LIS
Agent
Agent
Business objects
Database
IIS
Web app
Reports
Reports
Growth and Complexity
as of Sept 7, 2010
•
Lab Framework Client DLL –
22,850 lines (about 460 printed pages)
•
OmniTrax Server –
11,554 lines (about 235 pages)
•
Agent –
4199 lines (85 pages)
•
Gross Room Manager –
4754 lines (97 pages)
•
Histology Manager –
5133 lines (104 pages)
•
That’s equivalent to:
Version 1: 22 tables
– Les Miserables
– All three Lord of the Rings books
Version 4: 48 tables
Need for a Standard
Problems
1. Multiple barcodes from diff. facilities on same item
2. No “assigning authority” in barcode
Interpreted differently by different software
3. Some proprietary uses
APIII focus group suggestions (2008)
1. The barcode should contain only an identifier (e.g.
“license plate”); software determines use
2. The barcode should contain something equivalent
to an “assigning authority”.
ID|application|installation
12356789|OmniTrax|UWPath98195
Why barcode?
Expensive
Workspaces change
Process changes
… true, but reasonable ROI
… it might be time
… new processes are better
Jobs change
… but more valuable activity
Pathologists affected
Time investment
… in good ways
… pays off!
Better lab efficiency
Error/liability reduction
Inventory control
Resident autonomy
Gateway to more functions
Acknowledgements
•
•
•
•
•
Phil Nguyen
Kevin Fleming
Rosy Changchien
Chris Magnusson
Victor Tobias
• General Data
• Thermo-Fisher
• Accu-Place
•
•
•
•
•
•
Dr. Erin Grimm
Dan Luff
Steve Rath
Pam Selz
Kim Simmons
All the Techs and
Office Folks!
Download