Failures and Successes: How they have changed us

1
Failures and Successes How they have changed us
improve-it Toronto 2004 Nov 11

• Tom Rosenal MD
• Medical Director, Electronic Health Record
  Calgary Health Region

2
Rural
Community
U of C
FMC
PLC
Staff Physicians Beds Budget
22,000 2,200 2,200 2 B
ACH
RVH
3
Calgary Region EHR Clinical Focus Areas Goals
Quality
Safety
Access

• Chronic Disease Management
• Practice Variation
• Preventive Care

• Medication Errors
• Dx and Tx Errors
• Infection Prevention
• Blood Transfusion Errors

• Scheduling Waitlist Mgmt
• Continuity of Care
• Unnecessary Duplication

4
What is EHR Architectural Model
Active now or in lt 6 mos
Inpatient outpt
Sunrise Clinical Manager Results Orders Clin
Documentation CDS
Clinibase ADT
Parts active growing
Chronic Disease Mgmt IS Provincial EHR
Pharmaceutical Info Netwk Lab results History
Transcribed Reports POSP EMRs Results Delivery
to EMRs
Active in 2 years
EMPI Enterprise Master Person Index
5
Calgary Current Clinical Information System
status

• Intranet, single email system (Netscape), all
  sites
• Acute Care TDS 7000 results all orders, 2 sites
• Chronic Disease Management IS Orion, 2 programs
• Adult ICU, GE Quantitative Sentinel, all urban
  adult sites
• Emergency, HASS, all urban sites
• Anesthesia PICIS charting, all urban sites
• DI Agfa PACS, Cerner (Quadris), all urban sites
• Lab Cerner Classic, urban region, inpt
  community
• EKG MUSE all urban sites
• Data warehouse Oracle
• HL7 Interface engine e-Gate

6
Calgary In-progress Clinical Information System
status

• OR PICIS scheduling mat mgmt, all urban sites
• Registration Scheduling Clinibase, urban region
• Acute care Sunrise Clinical Manager results all
  orders documentation clinical decision
  support, all urban sites
• Community Care IS

7
Overview

• What is a failure
• The dimensions of failure
• A model of causes of failure
• Our experience in Calgary
• Measuring failure success
• How we are translating past experience into
  current approaches

8
Failure Literature Review
All happy families are happy alike, all unhappy
families are unhappy in their own way. Anna
Karenina 1874, Leo Tolstoy
The Ten Commandments of Effective CDS JAMIA 2003
David Bates et al
The Ten Plagues Exodus
Evaluating computerised health information
systems hard lessons still to be learnt BMJ 2003
Littlejohns P, Wyatt JC, Garvican L
9
Failure
Lies along a spectrum of degree Multidimensional
Depends on measure

• What
• Who
• Where
• How
• When

• Goals, IT, adoption, impact
• Which part of organization, which user
• Which site, department
• Survey, database dump, interview
• Immediate post-implement, later

10
Critical Care at the Bedside
11
Levels of success of a clinical information system

• System as planned is implemented
• System stable
• Used by intended providers
• Process improvement
• Outcomes improvement

12
Special considerations

• Health Care
• Life limb
• 24X7
• Physicians are key users but not employees
• Core function is care, not information management

• Calgary Health region
• Large entity
• Complexity community to acute
• Heterogeneous systems, experience, expectations
• Projects may take 5 years

13
IT Projects outcomes

• 25
• 46
• 26

• Outcome
• Failed to be implemented
• Challenged (Change scope, miss deadline, over
  budget)
• Successful

Standish Group 1998 Ric Irving in Health Tech
Canada Sep 2004
What if Could not achieve goal of reduced cost,
so established new goal of gt80 rate of physician
direct order entry
14
Clinical Info System Challenges

• Twain Change is inevitable progress is optional
• Capacity for change
• Time needed to implement
• Timing of work redesign
• Effectively engaging clinical community
• Balancing priorities
• Addressing needs of all clinicians
• Aligning across Region
• Aligning provincially federally
• Balancing access privacy
• Funding
• Balancing integration with best-of-breed
• Betting on the clinical information system
  industry
• Technologic

15
Determinants of failure

• Organization (Formal informal)
• Structure
• Leadership
• Culture
• People (Knowledge, skills)
• Clinicians
• Informaticians
• Project managers
• Technologists
• Goals expectations
• Users
• Administration
• Technology
• Funding

16
E.g. 1 - Old ICU (12 beds) e-Charting System 1982

• 1 junior
• 0
• Vendor
• High
• Easy to use
• Good data extractable
• Primitive text terminals, PDP-11
• Tight, no operating

• Organization Structure
• Leadership
• Culture
• People (Knowledge, skills)
• Clinicians
• Informaticians
• Project managers
• Technologists
• Goals expectations
• Users
• Administration
• Technology
• Funding

Result Never implemented Staff remained optimistic
Learned Technology needs minimal capability Need
internal staff to support users
17
E.g. 2 - New ICU (50 beds) e-Charting System 1996

• Highly technology friendly docs
• 1 mid, 1 junior
• 1
• 1
• High
• Paperless charting
• Good data extractable
• Fit except first regional HL7 lab feed
• Tight, from donors

• Organization Structure
• Leadership
• Culture
• People (Knowledge, skills)
• Clinicians
• Informaticians
• Project managers
• Technologists
• Goals expectations
• Users
• Administration
• Technology
• Funding

Result Best of breed 4 sites Staff very satisfied
Learned Department needs to link to
hospital Easier to develop than measure impact
18
E.g. 3 Anesthesia (80K cases/yr) e-Charting

• Localized 1 site
• Those experienced too little involved
• 0
• 0
• 1 internal, Vendor
• Modest
• Easy to use
• Good data extractable
• Good but abandon users after1 site

• Organization Structure
• Leadership
• Culture
• People (Knowledge, skills)
• Clinicians
• Informaticians
• Project managers
• Technologists
• Goals expectations
• Users
• Administration
• Technology
• Funding

Result Adoption good only first site Polarized
community, angry
Learned Need to design regionally from
start Ensure funding can address fixes, enhance
19
E.g. 4 TDS 4-7000 1-2 Hospitals 1988-2002

• Very strong
• Start users min involved
• Low at start, high at end
• Start 0, end 2
• Start 0, End 1
• gt6 internal, many vendor
• Modest
• Easier to get results to place orders
• Save
• Good but config 1 site, then abandon
• Never enough

• Organization Structure
• Leadership
• Culture
• People (Knowledge, skills)
• Clinicians
• Informaticians
• Project managers
• Technologists
• Goals expectations
• Users
• Administration
• Technology
• Funding

Result Adoption fairly good Some users became
very sophisticated
Learned Install painful part at end Match project
to expectations
20
TDS 7000 Common lab tests
21
TDS 7000 - Variable Dose 3 day - Prednisone
22
TDS 7000 Early Result
23
Riding the PCIS Rollercoaster
24
Usability 2000
25
TDS How a project can be a failure AND a success

• Bad
• Many users hate
• Factor making some physicians quit hospital
• Some charting not reviewed
• Bad press 1992, 2000
• Butt of jokes FRED, Tedious
• Blamed even when network is the problem
• Paper usage dramatically increased
• Project team hypersensitive

• Good
• Some users love
• Many providers understand results, order entry
  and documentation needs
• Physician order entry a norm
• Orders support evidence-based care (IV heparin)
• Real-time data to support some QI projects
• Project team knowledgeable

26
Some past evaluations in Calgary Health Region

• TDS 7000
• Survey users of pre- 3 post-implementation
• Usability studies pre- post-stabilization
  project
• Proportion direct physician order entry
• ICU GE Quantitative sentinel
• Survey user satisfaction
• Fit/gap re QI needs
• Sunrise Clinical manager
• Baseline TDS usage
• Interview provider satisfaction with engagement
  process

27
Learning from our failures successes
28
Our current approach

• Ensure executive commitment
• Know thyself (thy culture)
• Anticipate impact of system on relationships
  power
• Make goals achievable, balance expectation
  management with generating excitement
• Perform baseline measure perform continuous
  multifaceted formative evaluation
• Adopt project management methodologies people
• Achieve user engagement the right clinical
  users
• For broad rollouts, involve users from the start
• Describe accommodate workflow
• Prepare users
• Ensure transition project to operations
• What does not kill you, makes you strong
• Serious business but have a sense of humor

29
Pachyderm shift
30
Discussion