Transcript 100

How electronic health records
may influence behavior
George Hripcsak, MD, MS
Department of Biomedical Informatics/
Medical Informatics Services
Columbia University & NewYork-Presbyterian
Promise of clinical decision support
• Long history of reminders
• McDonald, NEJM 1976
• Barnett, Med Care 1978
• Computerized orders
• Tierney, JAMA 1993
• Increase compliance with corollary orders
• Overhage, JAMIA 1997
• Reduce maximum dosing errors
• Teich, Archives Int Med 2000
• Improve prophylaxis
• Kucher, N Engl J Med 2005
Institute of Medicine
• To Err is Human: Building a Safer Health
System (1999)
• Crossing the Quality Chasm: A New
Health System for the 21st Century (2001)
Caveats
• Many positive studies from 4 institutions
• Chaudhry, Ann Int Med, 2006;144:E12-E22
• Unintended consequences of CPOE
• Koppel, JAMA 2005
• Increased mortality after CPOE
• Han, Pediatrics 2005
• CDSS improve process most of the time,
but outcomes are understudied
• Garg, JAMA, 2005
Documentation and Workflow
Will we repeat the hype cycle?
10/2/08
PGY1 Progress Note
S: No events o/n. CXR yesterday showed lung still reexpanded while on water seal. Pt participated in physical therapy yesterday, felt weak afterwards.
Still has transient cough.
O:
VS Tm 98.6 Tc 98.5 68-77 110-116/62-66 20-22 95%General NAD, sitting on edge of bed, with NC, appears improved
HEENT PERLA, EOMI, no JVD
CV RRR nml S1, S2
Pulm chest tube on R, dry crackles predominantly at the bases
Abd soft, nt, nd, + BS, no HSM
Ext trace ankle edema, no cords/calf tenderness
Labs: see webcis
ANA negative
RF negative
ESR 22
Hep B cAB/sAB positive, sAG negative, Hep C Ab negative
stool O and P- negative
Other Studies:
9/24 abd u/s Hepatomegaly. Increase in echogenicity and echotexture may be due to hepatic steatosis or a fibrotic process.
TTE: Moderately limited study due to poor acoustic penetration. The left ventricle is mildly hypertrophied with normal systolic function. The left atrium is
mildly dilated. The right ventricle is not optimally visuallized but overall right ventricular size and function are normal. No significant valvular abnormalities
are seen on limited views. The measured peak right ventricular systolic pressure is approximately 40mmHg.
A/P: 61 yo man with UIP vs. malignancy s/p VATS biopsy 2 wks ago at OSH, p/w worsening SOB found to have pneumothorax. Chest tube placed in
ER, PTX now resolved on CXR.
Pulm - likely HP, PTX s/p VATS biopsy and subsequent chest tube, now with reexpansion of lung. Hypersensitivity panel negative, though this does not
r/o hypersensitivity pneumonitis.
-f/u pulm recs
-decrease O2 to maintain O2 sat of 95%
-continue steroids
-appreciate thoracic surgery consult - chest tube now on waterseal
-PFTs when chest tube is out
-daily CXR
GI - LFT elevation, hepatomegaly of unclear source, hepatitis panels negative, TTE normal, LFTs have stabilized, relatively acute onset, possibly
reactivation of Hep B vs. parasitic infection
-appreciate GI consult - will repeat stool O and P/stool culture, f/u stronglyloidis and schistomiasis Ag, continue ivermectin, ANA, anti-sm Ab, quantitative
immunoglobulins, alfa 1-antitrypsin, Ceruloplasmin, and GGT
-MRCP if pt can have it with chest tube
Heme - eosinophila, likely 2/2 parasitic infection
-trend WBC count and eosinophila
-Ivermectin
FENGI
-Cardiac diet
PPX
-sub q heparin
FULL CODE
10/2/08
PGY1 Progress Note
S: No events o/n. CXR yesterday showed lung still reexpanded while on water seal. Pt participated in physical therapy yesterday, felt weak afterwards.
Still has transient cough.
O:
VS Tm 98.6 Tc 98.5 68-77 110-116/62-66 20-22 95%General NAD, sitting on edge of bed, with NC, appears improved
HEENT PERLA, EOMI, no JVD
CV RRR nml S1, S2
Pulm chest tube on R, dry crackles predominantly at the bases
Abd soft, nt, nd, + BS, no HSM
Ext trace ankle edema, no cords/calf tenderness
Labs: see webcis
ANA negative
RF negative
ESR 22
Hep B cAB/sAB positive, sAG negative, Hep C Ab negative
stool O and P- negative
Other Studies:
9/24 abd u/s Hepatomegaly. Increase in echogenicity and echotexture may be due to hepatic steatosis or a fibrotic process.
TTE: Moderately limited study due to poor acoustic penetration. The left ventricle is mildly hypertrophied with normal systolic function. The left atrium is
mildly dilated. The right ventricle is not optimally visuallized but overall right ventricular size and function are normal. No significant valvular abnormalities
are seen on limited views. The measured peak right ventricular systolic pressure is approximately 40mmHg.
A/P: 61 yo man with UIP vs. malignancy s/p VATS biopsy 2 wks ago at OSH, p/w worsening SOB found to have pneumothorax. Chest tube placed in
ER, PTX now resolved on CXR.
Pulm - likely HP, PTX s/p VATS biopsy and subsequent chest tube, now with reexpansion of lung. Hypersensitivity panel negative, though this does not
r/o hypersensitivity pneumonitis.
-f/u pulm recs
-decrease O2 to maintain O2 sat of 95%
-continue steroids
-appreciate thoracic surgery consult - chest tube now on waterseal
-PFTs when chest tube is out
-daily CXR
GI - LFT elevation, hepatomegaly of unclear source, hepatitis panels negative, TTE normal, LFTs have stabilized, relatively acute onset, possibly
reactivation of Hep B vs. parasitic infection
-appreciate GI consult - will repeat stool O and P/stool culture, f/u stronglyloidis and schistomiasis Ag, continue ivermectin, ANA, anti-sm Ab, quantitative
immunoglobulins, alfa 1-antitrypsin, Ceruloplasmin, and GGT
-MRCP if pt can have it with chest tube
Heme - eosinophila, likely 2/2 parasitic infection
-trend WBC count and eosinophila
-Ivermectin
FENGI
-Cardiac diet
PPX
-sub q heparin
FULL CODE
Cut and paste
• Once entered, a mistake lasts forever
… 36 year old man … 27 year old woman …
• Doctors are telling us not everything needs
to be restated every time
Sublanguage
• Misspellings and interesting abbreviations
– text messaging
s/p LURT 1998 c/b 1A rejection 7/07 back on HD
pHtn 2/2 ASD w L->R shunt p/w abd pain x 3
• Doctors are telling us data entry and
review must be made more efficient
Medicine resident daily progress note:
Events overnight
Medicine resident daily progress note:
Subjective
Medicine resident daily progress note:
Vital sign flowsheet
Medicine resident daily progress note:
Vital signs by physician
Medicine resident daily progress note:
Medications
Medicine resident daily progress note:
Physical exam
Medicine resident daily progress note:
Laboratory
Medicine resident daily progress note:
Radiology
Medicine resident daily progress note:
EKG and telemetry
Medicine resident daily progress note:
Assessment
Medicine resident daily progress note:
Problem list
Medicine resident daily progress note:
Plan
Proposed addition for compliance
Inform
Pt edu
Smoke
Pain
PERRLA
Structured data entry
general
o fatigue
 fever or chills
o lumps or masses
eyes
 wear glasses/contacts
 visual changes
o eye pain
o itchy/watery eyes
nose and throat
o bloody nose
 congestion/runny nose
o sore throat
 hoarseness
gastrointestinal
o dysphagia (trouble swallowing)
 heartburn
o nausea and vomiting
o abdominal pain
o jaundice
o diarrhea
 constipation
cardiovascular
 chest pain/tightness
o palpitations
o fainting spells
 edema or fluid retention
ears
o hearing aids
o earache
o tinnitus (ringing in ears)
 ear drainage
 recurrent infections
respiratory
o shortness of breath
 cough/congestion
o wheezing
 productive of sputum/phlegm
o hemoptysis (coughing up blood)
dermatology
o skin lesions/skin cancer
 rash
…
Cost per click
• $16M nationally per checkbox
– # doctors, # notes per year, time on checkbox
• Should do cost benefit
Weekly Notes Written in Eclipsys XA:
Inpatient Providers
17,991
8,227
October 2007
October 2008
“I don’t read notes anymore; I just write
them.
There is no information in them. I do look
at vital signs, labs, and resident signout
notes.”
Medication reconciliation
Review of eight medical centers:
• ED enters meds on paper, review and edit on floor, no other med list
allowed in chart; await better software
• Nurse enters meds on paper, doctor reviews; await better software
• Nurse enters meds on paper, doctor reviews, doctor attests electronically c
hard stop on meds at 6 hours; await better software
• Nurse enters meds electronically (some from insurer), prints for doctor; await
better software
•
•
•
•
Pharm tech enters meds electronically, prints for doctor; await better software
Pharm tech enters meds electronically
All paper; await better software
Failed attempts at nurse and doctor entry; await better software
NYPH:
• Doctor (or nurse) enters meds electronically, doctor attests c hard stop at 18
hours; look forward to better software
Reconciliation and attestation
Unit
Patient
s
Med Rec Orders
#
#
Home Med List
%
#
Both Med and
Orders
%
#
%
Special
9
6
67%
6
67%
6
67%
ICU
10
10
100%
10
100%
10
100%
ICU
10
10
100%
10
100%
10
100%
ICU
9
9
100%
9
100%
9
100%
ICU
15
15
100%
14
93%
14
93%
Medsurg
32
31
97%
32
100%
31
97%
Medsurg
33
33
100%
32
97%
32
97%
Medsurg
20
19
95%
19
95%
19
95%
Medsurg
24
24
100%
24
100%
24
100%
Medsurg
18
18
100%
18
100%
18
100%
Medsurg
38
37
97%
35
92%
35
92%
Medsurg
23
20
87%
20
87%
20
87%
Medsurg
15
12
80%
12
80%
12
80%
TOTAL
591
570
96%
565
96%
564
95%
...
Lessons
• Quality initiatives improve quality, not
EHRs
– Why home-grown systems succeed
– EHR is an infrastructure, not an intervention
Lessons
• Focus with clear goals
– If the goal is only Leapfrog, that is all that will
be achieved
Lessons
• Slow, iterative process
– What does not kill the patient makes the
system stronger
Lessons
• Culture and buy in
– May get away with strong arm
Lessons
• Research
– Basic research: we don’t yet know how to do
this right
– HSR: evidence-based EHRs or at least better
art
Focused initiatives
• Focused initiatives with clear goals
• Measure process and outcomes
• Discharge summary writer
DSUM Writer vs. Dictation
(focused intervention)
% DSUMs entered via DSUM Writer
1/04 - 2/07
CUMC
100%
90%
80%
70%
%
60%
50%
% via DSUM Writer
40%
30%
20%
10%
0%
2004
2005
2006
Year
2007
DSUM Writer vs. Dictation
Turn-Around Time to Finalized DSUM
Days Saved by DSUM Writer
25
15
Days Saved by DSUM Writer
10
5
Month
Nov-07
Sep-07
Jul-07
May-07
Mar-07
Jan-07
Nov-06
Sep-06
Jul-06
May-06
Mar-06
Jan-06
Nov-05
0
Sep-05
Days Saved
20
Estimated Financial Impact of DSUM Writer - 2007
$1,000,000.00
$900,000.00
$800,000.00
Dollars
$700,000.00
Estimated Cumulative Dictation
Savings
$600,000.00
Development & Maintenance
Costs
$500,000.00
$400,000.00
$300,000.00
$200,000.00
$100,000.00
$0.00
Month
Next generation documentation
• What would really support both individual and
team care
• Past medical history as a central resource
• vs. cut and paste
• Document only current thoughts & actions
• review everything else
• Merge intern progress and signout notes
• Improved user interface technology
• natural language processing, speech
Data entry technology
• Natural language processing
– Convert narrative text to encoded form
– Natural interface for MD
– Computable for use in databases
Clinical data warehouse
2,500,000 patient records
62,000,000 laboratory test batteries
6,000,000 clinical notes: discharge summary,
admission, progress, signout, and visit notes
34,000,000 narrative reports from 40 ancillary
departments, including radiology, pathology,
cardiology, pulmonary
20,000,000 inpatient orders, outpatient orders
Flowsheeted nursing documentation c VS
Micro-consults
• Anticoagulation
– evidence for dosing, genetics, contraindications
– variable practice
• Reminders are insufficient
• Order a micro-consult
–
–
–
–
–
Advise on dosing based on EHR (automated to human review)
Primary MD gets order set
Consult tracks in a registry (with automated surveillance)
Escalate to consult as needed
Bill for micro-consult?
• “Mega-reminders”
Primary Care Information
Project (PCIP)
Public Health’s Role in Health Information Technology:
The New York City Model
Farzad Mostashari
Mat Kendall
New York City Department of Health and Mental Hygiene
[email protected]
PCIP
• 3000 Medicaid providers in NYC
• The following storyline illustrates the
TCNY Clinical Decision Support System in
action
Jane Doe, a 48 year-old woman is cared
for by her family practitioner, Dr. James
Bear.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
Dr. Bear wants to
find out how he is
performing
compared to other
physicians in his
practice in controlling
high blood pressure
for his patients.
• Dr. Bear
queries the EHR
to identify which
of his patients
have diabetes
and an HbA1C >
7.
Using the QUALITY MEASURE REPORTS FUNCTION, Dr. Bear is inspired by the performance of his
peers in managing the blood pressure (BP) of their hypertensive patients; only one-third of his
hypertensive patients have achieved good BP control.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
Dr. Bear wants to
improve his score on
BP control and
queries the EHR to
identify patients with
poorly controlled
hypertension
Using the ENHANCED REGISTRY FUNCTION, Dr. Bear identifies five patients with high blood
pressure who do not have an appointment scheduled, and reaches out to each patient; he generates a
letter scheduling a follow-up visit with patient Jane Doe.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
• Jane Doe receives
the letter and makes
a f/u appointment
• During the visit, Dr.
Bear’s assistant
takes her history and
vitals
• Jane mentions that
she has had a few
weeks of excessive
thirst and fatigue
Jane’s blood pressure is elevated (150/90) and highlighted in red by the AUTOMATIC
VISUAL ALERT FUNCTION. Dr. Bear can trend her BP over time.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
• Based on Jane’s
chief complaint of
excessive thirst, Dr.
Bear performs a
fingerstick test and
confirms his
suspicion that Jane
has diabetes
• Dr. Bear enters a
diagnosis of
diabetes into the
EHR
Based on Jane’s new diagnosis of diabetes, the CLINICAL DECISION SUPPORT
FUNCTION identifies four preventive care services that should be performed. This
list of services is automatically populated in the CDSS panel.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
Dr. Bear agrees
that these tests
are appropriate
and should be
performed
Dr. Bear uses the QUICK ORDER FUNCTION to order an HbA1C test for Jane, as
well as a flu vaccine; the alerts disappear from the panel once they are ordered.
Dr. Bear may also choose to suppress alerts, if he deems them unnecessary.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
Dr. Bear also
selects the “LDL
control (high
risk)” alert, which
displays the
order set for high
LDL levels
The 1st part of the COMPREHENSIVE ORDER SET displays a selected list
of recommended medications (brand & generic) for lipid control.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
Dr. Bear views
other order sets
for high LDL
levels
The 2nd part of the COMPREHENSIVE ORDER SET displays a selection of
recommended labs, immunizations, follow-up appointments, referrals as well as
printable physician and patient education materials.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
• Dr. Bear wonders
if he should change
Jane’s medication
regimen to better
control her lipids and
wants know what
medications have
been filled by her in
the past 90 days
• Jane has signed a
consent form to give
the provider access
to her medication
history
Since Jane is a Medicaid patient, Dr. Bear can use the eMedNY FUNCTION to view
her 90-day medication history. He notices that Jane has not filled her lipid medication
(simvastatin) for the past three months; she admits that she has stopped taking them
because she wondered if her tiredness might have been due to these pills.
1. Measure Reports
2. Enhanced Registry
3. Automatic Visual Alerts
4. CDSS
5. Quick Orders
6. Comprehensive Order Sets
7. eMedNY
8. CIR and School Health
• While she’s there,
Jane asks Dr. Bear
for a school health
form for her 5 yearold (Tim) who is
entering day care.
Tim
male
01/01/03
Mother
• Dr. Bear generates
a preloaded NYC
School Health form
populated with Tim’s
information for Jane
to take with her.
Tim’s information has already been automatically uploaded to the CITYWIDE
IMMUNIZATION REGISTRY. The CIR will maintain a complete record of Tim’s
immunizations which can be accessed by other providers as needed.
Quality initiatives improve quality,
not EHRs
• Partnership among clinical leadership,
quality, IS
Focus with clear goals
Slow, iterative process
Culture and buy in
• Need to pair bottom-up initiative with
top-down, evidence-based approach
Research
• There is something there
• We need to find it