Electronic Medical Records:
key implementation issues
C.T. Lin MD
Senior Medical Director, Informatics University of Colorado Hospital
January, 2008 For CLSC 6800 HLTH 6071
Outline
Paper Medical Records
– The burning platform
Electronic Medical Records
– What and why
– EMR Current design challenges – Translation vs. Transformation
• Historical record of care • Communication/continuity • Preventive Care • Quality assurance • Legal record • Financial record • Research
• can’t find the chart
• can’t find the result in the chart • can’t read the chart
• can’t easily collate the data • can’t compare across patients • no analytic capacity
Paper disadvantages
Paper
Paper
Medical
Medical
Records
Records
Paper records: Legibility?
One
One
day
day
’
’
s worth of
s worth of
papers to
papers to
be
by ctl
Quick!
Quick!
RecallRecall: Find all the patients : Find all the patients who take Vioxx!
who take Vioxx!
P4P
P4P: Prove that we give : Prove that we give Aspirin to all our CAD
Aspirin to all our CAD
patients!
patients!
NQF
NQF: Have all our diabetes : Have all our diabetes
patients had a Pneumovax?
patients had a Pneumovax?
Vaccine
Vaccine: Call all high risk : Call all high risk
patients to get flu vaccine
patients to get flu vaccine
now!
now!
Screening
Screening: We have free : We have free
PFT screening next week!
PFT screening next week!
Who would benefit?
Who would benefit?
Paper records:
Paper records:
the
the
burning platform
burning platform
•
•Legibility suspectLegibility suspect •
•Costly to maintainCostly to maintain •
•Not disasterNot disaster--proofproof
•
•CanCan’’t qualify fort qualify for
pay
pay--forfor--performanceperformance •
•Population and quality Population and quality studies impractical
studies impractical
Electronic Medical
Records
Functional components of EMR
• Integrated view of patient data • Clinical decision support • Clinician order entry
• Access to knowledge resources • Integrated communication support
Why must we have an EMR?
• Because the current system is inadequate
• Because expert bodies recommend it • Because the government says so • Because insurers are going to require it • Because patients are going to demand it for better safety and improved service
“…
information technology
must
play a central role in the redesign of
the health care system if a
substantial improvement in quality
is to be achieved over the coming
decade.”
– Institute of Medicine, 2001
EMR design challenges
• Human-machine interface (user-friendly?) • Data acquisition (how to get data in?) • Coded data (how useful is the stored data?) • Technology adoption (who will use it?)
• Donald Norman
– The Design of Everyday Things – The Invisible Computer • Edward Tufte
– Visual Display of Quantitative Information – Envisioning Information
– Visual Explanations
– The Cognitive Style of PowerPoint • William Cole:
– Semantics of Graphic Data Display
The Human-Machine Interface
• Paper record of scribbled notes? (fast!) • Scan handwritten notes? (DHH) • Typing of narrative text? (VA) • Dictation? (UCH)
• Voice recognition (Dragon: narrative) • Templates, pick lists (Coded data)
Data Acquisition
• Narrative
– Easily recorded as written or spoken – No limits on nuances (choked on spaghetti?) – (Difficult to analyze patterns across time and
across patients)
• Coded
– Standardized definitions – No ambiguity
– Powerful for aggregate analysis – (Difficult to enter data)
Narrative vs. Coded data
• A finite enumerated set of terms intended to convey information unambiguously. (note that the English language fails all criteria).
• Current dictionaries have standardized definitions, but not all terms unique, not always reproducible (SNOMED, ICD9).
(dyspnea = shortness of breath = respiratory distress = labored breathing = panting = winded = SOB = breathless = exacerbation of asthma)
• A senior resident wants to do a study on ER pain management in sickle cell anemia
• Full paper system very awkward (1+ year effort):
– Obtain ER registry for the past 5 years – Obtain a list of 200,000 patients,
– Manually select those aged under 20, African-American, with hematocrits of less than 20%.
• Electronic registration system simplifies selection (1 month effort)
• Coded EMR (1 minute effort)
Using coded data for research
• A clinician wants help with appropriate billing of a Medicare patient
– Attend weeks of training for Evaluation and Management Coding
– Or, by electronically coding data, allow computer support of E/M coding as note is completed
Technology adoption
The 80 - 20 rule
• In my opinion, the success of a project is perhaps 80 percent dependent on the development of the social and political interaction skills of the developer and 20 percent or less on the implementation of the hardware and software technology!
--Reed Gardner, LDS Hospital
A formula for adoption
+ Executive support and clear vision
+ Physician champion(s) at executive and clinic levels + Alignment of incentives for individual docs
+ Adequate analyst support + Adequate time
+ Robust hardware and software performance! = Successful implementation
A formula for adoption
Executive support and clear vision
Organizational behavior change At least a cost neutral ROI Transcription cost reduction
Marketability of patient safety (e-RX) Necessary for future quality initiatives Necessary for Patient Centered IT
Transcription Cost Reductions in Clinics Completing Allscripts Deployment
$0.00 $10,000.00 $20,000.00 $30,000.00 $40,000.00 $50,000.00 $60,000.00 $70,000.00 $80,000.00 $90,000.00 04/0 5/20 05 05/05/ 2005 06/0 5/20 05 07/0 5/20 05 08/0 5/20 05 09/05/ 2005 10/05/ 2005 11/0 5/20 05 12/0 1/20 05 01/0 1/20 06 02/01/ 2006 03/0 1/20 06 04/0 6/20 06 05/0 6/20 06 06/06/ 200 6 07/0 7/20 06 08/08 /200 6 09/0 9/20 06 10/10 /200 6 11/11/ 2006 12/0 6/200 6 01/06/ 200 7 02/0 6/20 07 03/06 /200 7 04/0 4/20 07 05/0 5/200 7 06/0 6/20 07 Monthly Transcription $
Adv Reproductive Dermatology Endocrinology Interventional Pain Rheumatology Ophthalmology Spine Center Integrative Med UFM AFWilliams UFM Boulder
Urology Spine Pain Rheum Derm Endo IntegrativeAdv Repro UM Denver UFM Boulder UFM Park Mdw UM Anschutz UFM Westminster Greater than 80% Greater than 80%
reduction in monthly cost,
reduction in monthly cost,
over $1M saved annually
A formula for adoption
Physician champions at executive and clinic levels
Chief Medical Information Officer Seek out clinics with internal champions
(Not necessarily the geek, but the respected clinician instead. Rare that one person is both)
Photo apl
A formula for adoption
Align incentives for individual docs
Feds paying for quality (PQRI) University version:
Pay internal bonus based on use of EMR Physician billing improves by 14% Reduce workload by pre-loading data Generate useful reports (recall, quality)
A formula for adoption
Analyst supportand/or Enough Time
You can bring a clinic live with 2 analysts and 40 physicians (20 faculty, 20 residents)… in about a year.
We have 3 teams of 4 analysts + 1 training specialist to implement 3 clinics every 10 weeks
Protect physician time for training, initial use Seeing 50-60 patients a day, and learning EMR is unrealistic
The impossible triangle
Successful implementations are like keeping all three sides of a triangle facing up. You can have 2 out of 3, but you’ll never get all 3.
cheap good
fast
A formula for adoption
Robust hardware and software
(performance)!
The goal: sub-second response
Counting clicks vs. fast systems
Tablet PC medical assistant story
“The tapper” story
Database server fiasco
A formula for adoption
Miscellaneous thoughts:
Include your patients in a media campaign (pardon our dust, we’re making improvements)
The attitude and efforts of 1 person can make all the difference in success or failure of a clinic
QuickTime™ and a decompressor are needed to see this picture.
A formula for adoption
+ Executive support and clear vision
+ Physician champion(s) at executive and clinic levels + Alignment of incentives for individual docs
+ Adequate analyst support + Adequate time
+ Robust hardware and software performance! = A successful implementation must have all these
elements but should have at least one of these
in SPADES.
Assembling an EMR
So, you have 2 EMR systems.Well, can’t you just hook the 2 systems up? What’s the big deal?
EMR Systems Map
IDX Visit Management, Patient Billing, & Scheduling 3M Clinical Workstation: integrated viewer, clinics + hospital
Provider Portal (MedXplore -> McKesson) Patient Portal
Lab RIS Path Dictation
IDX ADT Allscripts: Deployed 20 of 40 Clinics Em er g enc y McKesson
RN docu, Bar Code, Inpt Pharmacy, CPOE
Viewer Interactive system Infra-structure GI p ro ce d CV proced OB GY N Ps yc h Pe ri -Op Onc ology Tr ans p la nt
EMR:
Substitution
versus
Transformation
EMR: Substitution
• Its faster to create a new patient chart
• EMR charts can’t be misplaced
• Notes are LEGIBLE
• Prescriptions are LEGIBLE
• No more sticky notes
• 2 people can use the chart at a time
Transformation: Safety
• Electronic documentation instantly
available, legible, longitudinal
• Safer prescribing (drug interactions,
allergy check, formulary check)
• Vioxx recall: Patients identified,
Transformation: Quality
• Flu vaccine: identify highest risk
patients to immunize first
• Diabetes: Track patients with
highest Hemoglobin A1c’s
• Heart disease: monitor use of
ACE-I, Aspirin, beta-blockers
Transformation: Patients
• Patients and physicians both
contribute to a shared medical record
• Patients collaborate with providers to
set their own treatment goals
• Patient can access, or give access,
to their record anywhere in the world
Transformation: no EMR
• Conversations between physicians,
nurses, and patients undergo:
– Voice recognition
– Natural language processing
– Integrated, non-interruptive alerts based on latest evidence
Serious Creativity, de Bono, 1992
Substitution vs. Transformation
Like improvements in transportation, EMR developments generally are incremental at first…
Substitution vs. Transformation
It was hard to foresee what the carriage couldbecome…
Substitution vs. Transformation
And in the broader sense, what transportation could become
Substitution vs. Transformation
And we will always dream…