Welcome! The Health Story Project Dictation to Clinical Data: Automating the Production of Structured and Encoded Documents

1
Welcome!The Health Story ProjectDictation to
Clinical Data Automating the Production of
Structured and Encoded Documents
Kim Stavrinaki s

• AHDI Conference, July 2009
• Nick van Terheyden, MD, Chief Medical Officer,
  MModal

2
Presentation Overview

• Background The Current Situation
• Enabling the EMR with the Missing Link
• A User Experience (GE/RISL)
• The Health Story Project
• Conclusion

3
Background

• The Current Situation

4
Electronic Health Record Universe

• Critical to the success of EHRs is to reconcile
  two opposing needs
• Enterprise need for structured and coded
  information capture
• Physicians practical need for a fast and easy
  method for creating clinical notes.

5
The Current Situation Structured

• Tedious manual process
• Time-consuming
• Documentation lacks expressiveness of natural
  language
• Lack of Flexibility
• Poor user interface
• Cost
• Fails to Meet Individual Physician Time vs.
  Benefit Test
• Cultural resistance
• Oblivious to HIM Requirements
• Incomplete and Inadequate Semantic Standards

Direct Data Entry Structured and encoded
information.
6
The Current Situation

• Transcription can be expensive
• Subject to longer turn-around times
• Clinical data lost, because documents are neither
  structured nor encoded
• Majority of attested information is only in the
  document
• Contains the detail and comprehensive scope of
  patient information
• Support human decision making
• Reimbursement is based on narrative documentation
• Retains current workflow, favored by physicians
• Interoperable
• Under utilized source of data for EMR

Dictation Fast and easy, expressive.
7
The Current Situation

• High cost of documentation
• Cost of ownership and physician time vs.
  transcription cost
• 60 of the data lost to the EHR
• Care process inefficiencies and impact on quality

8
Enabling the EMR

• The Missing Link in
• Information Capture in Healthcare

9
Data Entry Time

• The average physician spends 33 seconds dictating
  an establish office visit
• 92 of all office visits are established
• If the average physician sees 40 patients a day,
  total dictation time of 30 minutes plus time to
  search for the data.
• Using a traditional EHR application, the same
  number of patients would require 140 minutes of
  data entry time.
• Physicians are not willing to spend an additional
  90 minutes per day for data entry.

(40 X 92 x 33 seconds) (40 x 8 x 125) lt 30
minutes per day
Data and Chart courtesy Mark R. Anderson,
FHIMSS, CPHIMS, CEO, AC Group
10
Crossing the Chasm

• What if you could continue to use narrative and
  dictation and at the same time increase usage of
  the EMR and make more records available for the
  health information exchange?

11
Health Story Project Vision

• Comprehensive electronic clinical records that
  tell a patients complete health story
• All of the clinical information required for
• good patient care
• administration
• reporting and
• research
• will be readily available electronically,
  including information from narrative documents

12
Based on HL7 CDA

• Clinical Document Architecture Requirements
• Human readable document
• Must be presentable as a document
• Rendered version covers clinical information
  intended by the author
• Can contain machine-processable data
• Cross platform and application independent
• Can be transformed with style sheets

13
Adoption

• Incremental adoption overcomes the not me first
  dilemma
• Not dependent on recipients ability to receive
  or process
• Reverse adoption (can encode headers of existing
  documents)
• Non-proprietary
• Readable with any browser

14
Accessible Clinical Data
15
User ExperienceGE/RISL
Kim Stavrinakis Sr. Manager, Product
Definition, GE Healthcare

• The Missing Link in
• Information Capture in Healthcare

16
Key Workflows

• Self Editing
• real time read, proof, sign each exam
• batch mode - read multiple exams then sign via
  signature queue
• VR edits
• Option to send to Medical Editor during reporting
  process
• Batch Option dynamic combinations of workflow
  based on confidence models
• user based thresholds that determines how report
  is returned/reviewed to signature queue
• preliminary/draft to signature queue
• transcriptionist then preliminary to signature
  queue
• Transcriptionist Medical Editor workflow

17
Results Reporting Workflow
Data Center
Dictation Report in conversational speaking
When dictation is complete and EOL is pushed
Report is returned ready for edits
Dictating the Procedure
18
Results Reporting Workflow 2
Data Center
After final sign the report is processed in the
NLP engine for learning
Edit Mode using local capture tool voice in
selection between brackets
Voice in options for brackets, sign report, add
via voice more dictation in the sections, then
sign
19
Results Reporting Batch Mode
Report goes to Medical Editor or signature queue,
Radiologist moves on to next exam
Dictating the Procedure
When dictation is complete
20
Radiology Imaging of Lakeland Florida

• Radiology Imaging Specialists (RIS)
• physician-owned
• twenty board-certified radiologists
• many sub-specialized
• live since November 12, 2008

21

• You didnt change the radiologists work, and
  that is what made it easy on me.
• David Marichal, CIO, Radiology and Imaging Spec.
  of Lakeland, FL

22
Conversational Documentation

• transformation of dictation directly into
  structured clinical documents while encoding data
  depending on the care givers and organizations
  needs

23
Results

• VOC
• flexibility is key
• full-time rads 70 Medical Editor workflow/30
  self-edit
• part-time radiologists can use it in batch
  digital dictation mode
• radiologist love not having to dictate accession
  , name, signs/symptoms, etc
• quality of the engine is very good
• self-edit for stat exams has reduced of calls
  from the hospital

24
The Health Story Project and Meaningful Clinical
Documents
Kim Stavrinakis Sr. Manager, Product
Definition, GE Healthcare

• The Missing Link in
• Information Capture in Healthcare

25
Meaningful Clinical Documents vs. Text

• Structured and encoded clinical content enables
• pre-signature alerts,
• decision support,
• best documentation practices,
• multiple output formats,
• multi-media reporting,
• data mining
• Implements HL7 CDA4CDT standard compliant
  document types
• Increases quality of documentation

26
Health Story Document Types

• Implementation Guides
• Completed
• History Physical
• Consultation
• Operative Report
• DICOM Imaging Reports
• Upcoming
• Discharge Summary in progress through HL7
• Billing and Reimbursement Requirements
• Progress Notes
• .PDF work with Adobe

27
Project Members

• Founders
• Promoters
• Participants

28
Our Advocacy To Date

• Participation in public comment periods
• NCVHS Hearing on Meaningful Use
• HHS Request for Input on Meaningful Use
• HITSP Request for Input on ARRA
• Comments are posted on our site
• www.healthstory.com

29
Our Advocacy Messages

• Dictation is the documentation method of choice
  for 85 of physician providers
• Standardization of dictated notes is an
  achievable step for providers Standards are
  available today
• The current EHR systems certification process
  does not include requirements for integration
  with dictated notes per available standards
• The current draft definition of meaningful use
  focuses on recording clinical documentation in
  the EHR through data entry

30
Our Advocacy Requests

• Actions Requested
• Require certified EHR systems to accept
  interfaced data from dictation/transcription
  process per available Healthstory standards
• Modify the definition of meaningful use to
  recognize use of certified EHR systems with the
  above capabilities
• Assist in spreading the word about this avenue
  for getting important information into the EHR
  that allows physicians to continue dictating and
  that provides patients with comprehensive
  electronic records

31
Conclusion
32
Crossing the ChasmBabel Must Go

• Medical text typed from dictation
• has no meaning
• black marks on a page
• info must be tagged as discrete data
• elements in order to assign meaning
• Clinical documentation uses wide variety of terms
  with same meaning.
• and terms that sound the same that have different
  meanings..
• authors have a wide variety of styles, accents,
  methods of dictation

33
Health Story

• Captures meaningful clinical documents
• Is the bridge between
• free form narrative and expressive notes, and
• fully structured clinical data
• Improves the quality of clinical documentation
• Generates semantically interoperable clinical
  data that will
• solve the fundamental challenges with EMRs -
  allowing clinical decision support, alerts,
  decision support, data mining
• enable interoperability, reporting, patient
  safety initiatives, PQRI (pay for performance),
  PSI (patient safety indicators) and improve
  billing data capture

34
Impact

• Allows providers to maintain preferred workflow
  and documentation methods
• Increases the value and usability of narrative
  documents
• Accelerates the implementation of interoperable
  electronic health records
• Allows reuse of information

35
Getting Involved

• Become an Ambassador
• We need a grass roots effort to help spread the
  word Support our advocacy messages
• You can help educate your employers, clients,
  etc. about Health Story
• Joint the Effort
• Varying membership levels, including individuals
• Volunteer for a Project
• Currently developing data standards for discharge
  summary
• Participate in HL7 ballots on project draft
  standards
• Encourage Implementation
• E.g. Include requirements for standards in
  transcription RFPs

36
Membership Options and Benefits
37
QA
Kim Stavrinakis Sr. Manager, Product
Definition, GE Healthcare
38
Where You Can Find Me
Nick van Terheyden, MD, CMO, MModal Twitter htt
p//twitter.com/drnic1 Technorati http//technora
ti.com/people/technorati/nvt1 RSSSpeech
Understanding http//speechunderstanding.blogspot.
com/feeds/posts/default MyBlogLog http//www.mybl
oglog.com/buzz/members/nvt LinkedIn http//www.l
inkedin.com/in/nickvt Plaxo http//nvt.myplaxo.c
om FaceBook http//profile.to/drnick Digg http
//digg.com/users/nvt1 Delicious http//deliciou
s.com/nvt1 E-Mail nvt_at_mmodal.com GrandCentral (
301) 355-0877