Title: Electronic Medical Records: An Introductory Tutorial
1Electronic Medical Records An Introductory
Tutorial
- William Tierney, MD
- Atif Zafar, MD
- AHRQ PBRN Resource Center
2Outline of Presentation
- Introduction to EMRs (very basic information)
- Barriers to Adoption Some Problems with Data
Accessibility and Care Processes - EMRs for Clinical Research
- EMRs and HIPAA Security
- No Nonsense Guide to Selecting an EMR
- Examples of EMRs
- OpenSource
- Commercial
- Lessons Learned
3Introduction to EMRs
4Introduction to EMRs
- Why do we need Electronic Medical Records (EMRs)?
- Many problems with the current healthcare system
(underuse and overuse) - 30 of children receive excessive antibiotics for
otitis - 20-50 of surgical procedures are not necessary
- 50 of back pain x-rays not necessary
- 50 of elderly patients dont get a pneumovax
5Introduction to EMRs
- Why do we need EMRs?
- Clinical practice is a data intensive operation a
- Inadequate data communication causes medical
errors - Human cognition is good at pattern recognition
but not at remembering lists or evaluating
multiple business rules.
6Why do we need EMRs?
- Available 24 x 7
- Can be viewed by more than one user at a time
- Is available from remote locations
- To covering MDs
- Others with appropriate needs
- Data can nearly always be found
- Is legible
7Why do we need EMRs?
- Enhances Communication
- Between providers--clinical messaging
- Can tag EMR location with message
- Referrals
- Half of specialists didnt know what main
question was - A third of the time no information came back to
PCP
8Why do we need EMRs?
- Cost Savings
- Dictation cost savings
- 170/FTE/month
- Chart pull savings
- 217/FTE/month
- Savings accrue to practice, apply to all payers
9Why do we need EMRs?
- Assist with Decision Support
- Many domainscost and selection of
- Drugs
- 18 reduction found by Overhage
- Lab tests
- 10-15 reduction in cost for charges, last
result, probability of abnormal - Radiological studies
10Why do we need EMRs?
- Decision Support
- In inpatients, computerizing ordering decreased
- Serious medication errors by 55
- All medication errors by 81
- EMR can help by
- Structuring medication orders
- 34 error rate with paper vs. 6 with electronic
- Alerting about
- Allergies
- Duplicate medications
- Many other issues
11Introduction to EMRs
- Do EMRs make a difference?
- UNEQUIVOCALLY YES, BUT AT A COST!
- In multiple studies, EMRs have been shown to
- Shorten Length of Stay in a Hospital setting
- Decrease Adverse Drug Events (ADEs)
- Improve Readability, Consistency and Content of
the medical record - Improve Continuity of Care
- Reduce practice variation
- Most benefits come from Decision Support.
12EMR Use in the United States
- Even though the US Health Care system is the
costliest in the world, its performance ranks
37th in the world according to the WHO! - Only 5 of US primary care providers use EMRs
(Bates et. Al., JAMIA 2003), 7 of all physicians
(Wang, Bates, et. Al., American Journal of
Medicine, April 2003)
13EMR Use Around the World
- Use PCs Use EMR
- Australia 90 53
- Denmark 95 62
- Netherlands 95 88
- Sweden 95 90
- United Kingdom 95 58
- (c) 2001 Harris Interactive
14Breakdown by Function - 2002
- Australia UK
- Use EMR 90 99
- Of Those
- Prescrip 100 80
- Notes Unknown 45
- Reminders Unknown 70
- Clin Vocab 15 (ICPC) 100 (Read)
- Paperless Unknown 45
- 2B initiative by UK to get all physicians online
15What is an EMR?
- At their heart, EMRs are just a database
- This database hold many kinds of information
(coded and not coded) - This database is organized by date, time, pat ID
and contains - Patient registration data (name, contact info,
DOB, SSN, etc.) - Test results (laboratory, radiology, nuc med
etc.) - Medications (active, inactive) and Allergies
- Current list of diagnoses and problems
- Appointment Data
- Clinical Notes
- Billing Information
16What is an EMR?
- So if an EMR is just a database, how is it
different from other databases, and why is it so
useful? - Value Added
- A Clinical Knowledge Heirarchy (term dictionary)
- How do clinical concepts work together
- Ex Digoxin toxicity can occur with hypokalemia
- A List of Current Clinical Recommendations
- A List of Appropriate Medication Indications,
Doses, Adverse Effects and Interactions and Cost
Estimates - Costs, Indications and Utility of Tests
17What is an EMR?
- What are some typical EMR Components
- Lab System Contains all lab tests ordered and
their results and stored as coded results (LOINC
etc.) in many systems - Radiology System Stores test reports
- Pharmacy SystemList of current medications,
inactive meds and when they were last dispensed
or ordered - Billing System A list of diagnostic codes used
for billing (ICD9, CPT, etc.) - Registration System Names, Contact Info,
Personal Info, etc. for patients
18What is an EMR?
- Additionally, many EMRs have
- An Order Entry System (where physicians enter
orders, prescriptions, notes etc. online) - A Decision Support System
- Often linked to the order entry system to provide
guidance at the point of care - Contains databases for clinical knowledge,
guidelines, list of medication indications, doses
etc.
19What is an EMR?
- The spectrum of EMRs
- EMRs target specific user bases, from solo
office-based practices to large, multispecialty
tertiary care centers - Many features are thus directed at managing
workflows specifically to these user bases - For example, large commercial EMRs unbundle
services such as clinical documentation, results
display etc. while office systems typically
integrate all of these under the same interface.
20How do Clinicians Interact with EMRs
Order Entry/Results Reporting
21Different Types of EMRs
- EMRs dont necessarily need to be expensive and
complicated or require that a computer be used to
enter data - Can have hybrid computer/paper based approaches
- Ex In the CHICA System, paper is used to
interact with an electronic data repository - Standardized paper forms are printed and then
scanned - Characters are recognized and the electronic data
so generated interacts with the data repository
22Different Types of EMRs
- At Indiana University, pediatric clinics use this
system - A data repository was developed using Microsoft
SQL Server - A clinical guideline system was written in Arden
Syntax - An optical character recognition system called
Cardiff Teleforms is used to process handwritten
numerical data on preprinted scanned forms - The data so generated is stored in the database
and dynamic reminders are generated for the
physician - These are printed on the clinic computer
- The entire operation takes lt 2-3 minutes!
23Different Types of EMRs
- The Mosoriot Medical Record System
- Indiana University has an HIV Effort in Kenya
- A Simple MS Access based database holds all
patient records (3 years worth!) - Provides forms for data entry, standard term
dictionary, medication listings, registration
system, clinical documentation system etc. - Created by 1 programmer over 2-3 weeks!
- Highly effective, easy to maintain, inexpensive!
24Data Sources
- So how can EMRs populate their databases?
- Data can come from many many sources
- Admission/Discharge/Billing
- Anesthesia Systems
- Cytology Systems
- Diagnostic Imaging Management Systems
- EKG Carts
- Endoscopy Systems
- ER Systems
25Data Sources
- More Data Sources
- Home Care Systems
- ICU Monitoring Systems
- IV Fluid Infusion Control Systems
- Laboratory Systems
- Nurse Triage
- Order Entry Systems
- Pharmacy Systems (Inpatient/Outpatient)
- Pulmonary Function Systems
26Data Sources
- More Data Sources
- Radiology systems
- Risk Management systems
- Registration Systems
- Scheduling and Clinic Charge Systems
- Transcription Systems
- Unit Dose Dispensing machines
- Ventilator Management systems
27Data Sources
- So if there are so many data sources available
and so many people are interested in using EMRs,
why are they not more prevalent?
28The Challenges of EMR Implementation
29Problems with Electronic Data
- For the last 30 years the medical informatics
community has struggled with how to architect the
vessel that will hold patient data - Problem is that they have focused on the wrong
problem! - We dont just want to create a system that
permits entry of data electronically, we want to
create a system that can acquire this data
automatically from other electronic data
repositories and make it available at the time of
service.
30Problems with the Data Sources
- Too many repositories or islands of systems
- Difficult to bridge and combine in useful ways
- Contain different data at different levels of
granularity - Each uses a different code to identify the same
information. - Many institutions do not capture all of the data
of interest to clinicians. - Labs are sent to external reference laboratories
- Patients fill their scripts at community
pharmacies - As a result many implementations do not lead to
satisfactory achievement of the intended quality
assurance goals
31Problems with Data Sources
- Another problem is that there are many many care
providing sites in the United States - Hospitals 5000
- Nursing Homes 19000
- Pharmacies 59722
- Physician offices 200000
- Laboratories 63000
- Emergency Rooms 4856
- Hospice Care 2800
- Home Care agencies 4258
- All of these sites generate data that are not
necessarily compatible.
32Problems with Electronic Data
- Thus, the problem is not one of creating database
fields de novo, it is one of merging existing
fields from many different sources in meaningful
ways - When commercial and other EMR vendors create
proprietary, closed, systems, with custom
database architectures, they often worsen the
problem and make it harder to populate the
database with useful information, inexpensively
33(1) The Role of Standards
- Fortunately, most of the informatics community
has realized that the solution to the problem of
merging data lies in the implementation of
Standards for Data Communication. - These standards permit data to be easily
translated from one database system to another
34(1) Standards
- There are many many standards, each for a
different purpose - Lab Data Communication
- General Clinical Messaging
- Radiology Image Transmittals
- Diagnostic Coding
- Procedure Coding
- Need to distinguish between coding standards and
messaging standards.
35(1) Standards
- HL7 (Health Level 7)
- Most widely used standard
- General clinical messaging standard
- Communicates structured data
- Fields for
- Diagnostic Results
- Notes
- Referrals
- Scheduling Information
- Nursing Notes
- Problems
- Clinical Trials data
36(1) Standards
- Health Level 7
- 2000 hospitals, the CDC and most referral labs.
- Also used in Canada, Australia, New Zealand,
Japan and extensively in Europe - Bridges many systems, including laboratory,
dictation, pharmacy, electronic patient records,
performance databases, data repositories (cancer
registries) etc. - Web Site
- http//www.mcis.duke.edu/standards/HL7/h17.htm
37(1) Standards
- LOINC
- Logical Observations and Indicators Names and
Codes - A coding standard that is used for LAB data
- Used for representing laboratory observations and
common clinical measurements - At least 5 large commercial labs (Corning,
MetPath, LabCorp, ARUP Labs and Life Chem) have
adopted LOINC
38(1) Standards
- DICOM
- Another messaging standard
- Standard of choice for transmitting diagnostic
images - Closely supported by all of the imaging vendors
and is working with the HL7 group - Web site
- http//www.xray.hmc.psu.edu/dicom/dicom_home.html
39(1) Some other coding standards
- ICD9/10 Used to code diagnoses
- CPT Used to code procedure data
- ISO - Used to code units of measure
- UMDNS Device classification standard
- NDC Drug entities classification
- SNOMED organism names, topologies, symptoms and
pathology - HOI Outcomes variables
- UMLS Metathesaurus for clinical nomenclature
- Arden Syntax Clinical knowledge
40(2) Patient Identification
- How do we ensure that the information belongs to
the correct person? - Patients move and change addresses/tels
- Patients change names or use aliases
- Patients sometimes have multiple SSNs
- There are differences in patient, provider and
place of service identifiers among data sources
41(2) Patient Identification
- Solutions to this problem do exist but at a local
institutional level at the moment - Our institution uses a combination of mothers
maiden name, SSN and DOB to uniquely identify the
patient - The Kassebaum-Kennedy Bill (PL 104-191) will make
this into a national effort and standardize
patient and provider identifiers
42(3) Physician Data Capture
- The ultimate EMR promises to capture whatever
data is needed to perform any EMR task outcomes
analysis, utilization review, profiling and cost
estimation. - This prospect excites many CEOs and CIOs
- However, much of the data needed for such
functionalities comes from physicians (disease
severity and clinical findings) and most of this
data is recorded as un-coded free text.
43(3) Physician Data Capture
- In order for physician generated data to be
useful it needs to be in coded form so that
algorithmic assertions can be made - The problem of coding free text data is of
paramount importance and information systems
designers have struggled with this as long as the
field of medical informatics has been in existence
44(3) Physician Data Capture
- One approach we could take would be to translate
existing free text dictations into coded,
computer readable information, but - Human coding is error prone and expensive and is
at too high a level of granularity to be useful - Decades of research into computer based coding
has still not yielded satisfactory results - Or the physician could code the data themselves
by entering structured notes but - This is costly in terms of time as it requires
the user to map the terms into computer
understandable words at a level of granularity
which is useful
45(3) Physician Data Capture
- Commercial EMR vendors bypass the problem and
provide every mode of data entry possible - Direct keyboard entry
- Dictation with human transcription
- Voice Recognition
- Structured Data Entry
- Paper based data collection
- Web/PDA/Mobile devices
- Problem is that we dont know which one is the
most efficient so users have to think with their
feet
46(3) Physician Data Capture
- We did a study at Indiana University comparing
voice recognition with typing and
dictation/transcription and found that (at least
for 1 user) - Voice recognition almost doubled the note size as
compared with typing - It took longer to use voice recognition by 1.3
min as compared with typed notes - Voice recognition was 30 fold less accurate than
dictation/transcription - During proofreading, the user missed 30 of
errors - 1.2 of errors changed the intended meaning
- Dictated note turnaround time was from 2-5 days!
47(3) Physician Data Capture
- Managers and quality analysts want data that is
often never captured - Formal functional status
- Detailed Guideline criteria
- And we dont even know how much of this kind of
information is needed? - For some disorders (angiography and knee surgery)
data sets have been developed but we do not know
the operating characteristics or predictive value
of the data elements? - How do we define and collect the soft data
elements?
48(3) Physician Data Capture
- We do have some instruments for some disorders
(CAGE, Hamilton Scale, SF12/36 etc.) - But we lack them for many other clinical entities
and for much of specialty clinical care - And checklist based symptom questionnaires as
opposed to validated instruments elicit many more
symptoms than open ended questions, so which of
these are really important?
49(3) Physician Data Capture
- Coding of all medical information is unnecessary
- So where do we draw the line?
- (how much should be coded and how much can be
stored as free text) in order to maximize the
utility of the information. - The other issue is with longevity of clinical
notes. - How often do you use a note from 2 years ago?
- How long do we need to keep the EMR data?
50(4) Cost
- Cost is perhaps the biggest barrier to
implementation - Unfortunately there are few studies that have
looked at the long term ROI with EMRs - Most existing studies have been done by the
system vendors and so the data should be examined
with a cautious note - However, the data that is available suggests that
the ROI is excellent!
51(4) EMR Cost Analysis Studies
- Several studies are worth mentioning
- (1) Renner et. Al. looked at implementing an EMR
in 1996 in 40 primary care practices - Its net present value (1996 dollars) was about
280,000 based on a 5-year model - They found that reducing the cost of medications
and preventing ADEs was of the greatest benefit
in primary care
52(4) EMR Cost Analysis Studies
- (2) Wang, Bates et. Al. looked at the cost of
implementing a full EMR in primary care as
compared with paper based chart systems - Primary outcome was the cost benefit per provider
over a 5-year period - Used average statistics from their institution
(Partners Healthcare, Boston), expert opinion and
national data to estimate costs - System Costs (13,100 initial, 3100 each year
HW) - Induced Costs (11,200 in year 1)
53(4) EMR Cost Analysis Studies
- (2) Wang, Bates et al.
- Benefits resulted from costs averted (/year)
- Transcription savings (2700)
- Reduction in need for chart pulls (5/chart
pulled) - Drug cost savings and prevention of ADEs (2200)
- Laboratory and Radiology cost savings (10,700)
- Charge capture improvement (7700)
- Decrease in Billing Errors (7600)
- All benefits finally being realized in year 4
54(4) EMR Cost Analysis Studies
- (2) Wang, Bates et. Al.
- Resulted in present value of net benefit (2002
dollars) to be 86,400/provider in year 5 - Breaking down by EMR feature they got
- Light EMR (net loss of 18,200/doc in year 5)
- Online patient charts only
- Medium EMR (net benefit of 44,600/doc in year 5)
- Adds an Electronic Prescribing Module
- Full EMR (net benefit of 86,400/doc in year 5)
- Adds Lab, Radiology and Charge Capture systems
55(4) EMR Cost Analysis Studies
- (2) Wang, Bates et. Al.
- Conclusions An Ambulatory EMR
- Resulted in net benefits across a range of
assumptions, which increase as more features are
added and as the time horizon lengthens - Most benefit was derived from reductions in drug
expenditures, improved test utilization, improved
charge capture and reduced billing errors - The greater the portion of capitated patients the
greater the return, although benefits also accrue
for fee-for-service patients (but mostly to
payers and not health care institutions) - Limitation Did not consider malpractice
reduction, increased provider productivity or
decreased staffing requirements. - Intangible benefits Improved Quality and
Decreased Errors
56(5) Other Barriers
- (1) Physician reluctance and fear that their
productivity may decline (which it does) - (2) Unreliability of EMR Vendors in a volatile IT
economy. Lack of adequate IT support from the
vendors - (3) Concerns over data security
57Summary Barriers to EMR Use
- Too many data sources, no simple way to
coordinate and connect them except to use
standards which are still evolving - Unique patient identification still a problem
esp in large tertiary care centers - Physician data capture inefficient and expensive
58Summary Barriers to EMR Use
- Startup costs can be prohibitive but long term
benefits are clearly evident form pilot studies - Physician reluctance a major barrier to use
- Concerns over security still an issue, eg HIPAA
- System vendors are transient and fail to provide
adequate support
59EMRs for Clinical Research
60EMRs for Research
- So what EMR functions do we need in order to
effectively do clinical research? - Answer Depends on what you want to do
- However, to be able to ask questions of your
practice, you need - Registration data (Registration system)
- Diagnoses (Billing data)
- Medications (Pharmacy data)
- Labs and other Test Results (Lab/Radiology data)
- AND
- A system to query these databases intelligently
61EMRs for Research
- You dont necessarily need a decision support or
order entry system but if you want to intervene,
you may want to include these systems as well
62EMRs for Research
- Note that the registration, billing, pharmacy and
lab/radiology data usually (but not always)
exists, outside of the context of any specific
EMR system - These are just data repositories which need to be
tapped into and queried - So you need a system to access and query these
databases, independent of any electronic medical
record system.
63EMRs for Research
- Alternatively, you could build a master
repository which acquires and stores this
information and permits intelligent queries to be
performed - This is exactly what we did in Kenya in the
Mosoriot Medical Record System, although data is
still hand-entered. Eventually it will be
downloaded using HL7 messages.
64Mosoriot Medical Record System
- An example of an EMR that is inexpensive and
functional and supports both clinical care and
research in rural Kenya - Built in 2-3 weeks by 1 programmer using
Microsoft Access - Consists of
- Data dictionary tables which define test names,
medications, diagnoses etc. - Forms which are used for data entry
- Has tables for registration data, billing data,
medication lists, lab and test results - Currently running on Tablet PC devices in Kenya
65Research Workflow Model
66EMR Features Conducive to Research
- Reliance on Standards (HL7, LOINC, ICD9, CPT)
- Easy access to data repository, i.e. database
structure is well documented - Built-in Practice Profile Management systems
- Built-in decision support and order entry
functionality - Able to export data in a standard format (CSV,
MDB etc.)
67HIPAA Security
68Introduction
- HIPAA Heath Information Portability and
Accountability Act - Final Security Rule Published in the Federal
Register on February 20, 2003 (effective 60 days) - http//www.cms.hhs.gov/hipaa/hipaa2/regulations/se
curity/default.asp - Designation 45 CFR 160, 162, 164
- Compliance Dates April 20, 2005
- Covered Entities 24 months after effective
date - Small Health Plans 36 months after effective
date
69HIPAA Security
- Some excellent links
- http//privacy.med.miami.edu/glossary/gt_security_
rule.htm - http//www.hipaadvisory.com/tech/wireless.htm
- http//www.hipaadvisory.com/regs/securityoverview.
htm
70HIPAA Security
- Security should not be confused with Privacy or
Confidentiality - Privacy The rights of an individual to control
his/her personal information without risk of
divulging or misuse by others against his or her
wishes - Confidentiality only becomes an issue when the
individuals personal information has been
received by another entity. Confidentiality is
then a means of protecting this information - Security refers to the spectrum of physical,
technical and administrative safeguards used for
this protection
71HIPAA Security
- Addresses 3 tiers of protection
- Administrative Safeguards
- Physical Safeguards
- Technical Safeguards
72Administrative Safeguards
- Institutional level
- Develop security management process where
potential threats to PHI are determined - Provide training to all employees about HIPAA
- Provides appropriate level of authorization based
on a protocol for granting access - Violations should be clearly documented and
investigated - A disaster recovery plan should be in place
73Physical Safeguards
- Applies to 3 elements of the PHI data storage
infrastruture - Facility where PHI data is stored
- Workstations on which it is stored
- Media on which it is stored
74Physical Safeguards
- Require that the facility have access control
- Contingency plans need to be in place in case an
intruder gains access - Workstation security measures be in place
- Automatic logoff
- Screen is placed away from potential viewers
- PDAs should be password protected
- Devices and media should be appropriately
disposed of in case they are no longer needed and
data should be erased properly
75Technical Safeguards
- Applies to how information is stored, verified,
accessed and transmitted/received - Access and audit controls
- Emergency access to information when needed
- Automatic logoff is enforced
- Data is encrypted and decrypted during
transmission - Verify integrity of the storage and transmission
(digital signatures)
76Am I HIPAA Compliant?
- Questions to ask yourself and your institution
77Questions to ask your institution
- 1. Was a security audit done and if so what are
the results? - 2. Did I get the appropriate HIPAA training and
do I have a certificate to prove this? - 3. Are there procedures in place to grant access
to PHI to authorized users? - 4. What are the procedures in place in case of
disaster, data loss or data theft? Are Backups
made frequently?
78Facility, Workstation, Media
- 1. What are the procedures in place to safeguard
the facility from intruders? Are there
contingency plans for dealing with intruders,
data theft or other event? - 2. How do protect the safety of workstations? Are
they password protected? - 3. Can bystanders view the screens on which PHI
may potentially be displayed?
79Facility, Workstation, Media
- 4. Is an automatic logoff mechanism enforced?
What time limits are provided before this occurs? - 5. What types of data are stored on PDA devices
and if PHI is stored is it password protected or
encrypted? - 6. What procedures are used when disposing of,
reusing or archiving data on hard disks, CDs,
floppys and Zip disks? Are PHI data erased
properly if the disks are to be disposed of or
reused?
80Data Level
- 1. Are there audit mechanisms for checking who is
accessing the PHI data and is this done on a
regular basis by authorized personnel? - 2. Are there procedures in place to grant
emergency access to information if needed? - 3. Is data integrity checked when the data is
transmitted or received? (digital signatures,
digital certificates, checksums etc.) - 4. Is the data encrypted and decrypted during the
transmission process?
81HIPAA Wireless Security
82Before you Begin
- Do I really need to be wireless of can I get by
with a wired connection? - Is space limitation a problem?
- Is mobility absolutely necessary?
- Do I have the permission of my institution to
install wireless networks? - Do I have adequate IT support to do this?
8311 Steps to Wireless Security
- Wireless is inherently unsecure
- Many Many ways of hacking into wireless networks
- Technology base is there to make it secure
- Some simple steps can be taken to maximize the
security of your wireless network
8411 Steps to Wireless Security
- 1. Change the default SSID (network name) on the
router so that your name/location is kept secret - 2. Disable the SSID broadcast, if your router
supports it. This will prevent hackers from
seeing you - 3. Change the administrators password on your
router.
8511 Steps to Wireless Security
- 4. Turn on the highest level of security
supported by your hardware (i.e. Wireless
Equivalent Privacy WEP, which is older or WPA
which is the latest and most secure) - 5. Make sure you have the latest firmware
updates. Implement MAC (media access control),
which specifies exactly which WLAN PC cards can
access the network and excludes others
8611 Steps to Wireless Security
- 6. Place the Wireless Access Point (WAP) towards
the middle of the building, keeping the zone of
potential access within the building. - 7. Do your own security audit. Use Network
Stumbler (www.netstumbler.com) on your Tablet PC,
laptop of PDA and walk around the perimeter of
your building to see where and what a would-be
hacker may see
8711 Steps to Wireless Security
- 8. If you have a limited number of wireless
clients (Tablet PCs), provide them with static
IP addresses, and disable DHCP on your router.
This ensures that only authorized machines can
see your network.
8811 Steps to Wireless Securit
- 9. If we are in an enterprise setting, use VPNs
(Virtual Private Networks). You can isolate your
WLAN from the wired network using products such
as the Netgear FVM318 or the SonicWall SOHO TZW.
Then you can use the VPN to tunnel directly into
the wired network securely
8911 Steps to Wireless Security
- 10. Avoid using public hotspots, areas that are
insecure and open for general use. - 11. Turn off file and print sharing on your
Tablet PCs. Most devices do not prevent
client-to-client traffic, so people sitting
across the street from you can be looking at your
shared directory remotely.
90Guide to Selecting and Deploying an EMR
91Selecting an EMR
- Award winning EMRs
- CPRI Davies Award Winners (1995-2000)
- Emphasis on successful implementation, not on
technology that is behind the design - Functional Requirements
- Integrate data from multiple sources
- Provide decision support
- Used by caregivers as primary source of
information - Must enhance care, not just replace paper
- So who are there award winners and what are their
strategies for success?
92Davies Award Winners
- Intermountain Healthcare System, Salt Lake City
- Columbia Presbyterian Medical Center
- Department of Veteran Affairs CPRS (now
open-source) - Brigham and Womens Hospital
- Kaiser Permanente, Cleveland OH
- Regenstrief Medical Record System
- North Mississippi Health Services
- Kaiser Permanente, Portland OR
- Northwest Memorial Hospital, Chicago
- Kaiser Permanente, Rocky Mtn. Region
- Harvard Vanguard System
93Davies Award Winners
- Common Strategies and Attitudes towards
implementing EMRs
94Common Strategies
- Vision of healthcare as an information business
- Sustained leadership (5 years )
- Run by project leaders and not CIOs
- Most projects had physician champions
- EMRs subjected not to a cost benefit ROI analysis
but to an unremitting pressure to show value
95Common Strategies
- Customer Service, Customer Service!
- Frequent, sustained, end-user orientations and
feedback with demonstrated responsiveness to
feedback! - Weekly Regenstrief Pizza Meetings
- Kaiser physician focus groups
- Northwestern weekly feedback with supplements
- System developers were also the salespeople,
troubleshooters, coaches and colleagues!
96Common Strategies
- Plans in place for system evaluation and change
management - All winners had to re-engineer some workflow
process dont automate a manual process that
occurs commonly but does not work! - Incremental deployment dont rush things
- Each increment overcame a specific barrier to
care - Systems were viewed as tools to enable care
process improvement and were not an end to
themselves
97Common Strategies
- All resulted in a decreased reliance on
paper-based sources of information - Decision Support, Decision Support, Decision
Support -gt provides the largest value added
compared to a paper system - Focus on standards based data architecture rather
than specific applications to do this or that - FAST RESPONSE TIME!
- Flexible enough to adapt to organizational change
98So what can I do to implement an EMR in my
practice?
99Can I implement an EMR?
- Depends on your size and your budget
- Solo practice -gt yes, definitely
- Multispecialty group (2-100) -gt probably (cost is
around 4-20K per provider) - Multispecialty, multisite groups maybe
- Tertiary care centers with scattered secondary
care sites -gt probably need to be brave and
wealthy!
100What EMR should I choose?
- Do not start in product selection mode
- Begin by identifying the practice processes that
you wish to improve first - Then search for the functions you need
- Problem List Medications
- Clinical Encounters Lab/Xray/Pathology
- Telephone Calls Referrals
- Preventive Care Managed care
101Which EMR should I choose?
- Anticipate primary and secondary users
- Primary
- Clinical decision making,
- Documentation
- Support for Billing
- Secondary
- Provider profiling and service utilization
- Quality report cards and outcomes analysis
- Regulatory reporting and justification for studies
102What if I have a limited budget?
- Again, think of using selected modules to enhance
parts of your practice - Clinical Note Systems
- Prescription Writer
- Use one or more of the OpenSource EMRs
- Need some level of IT expertise to deploy
- No real support available from the developers
103Examples of OpenSource EMRs
- a. OpenEMR (http//www.synitech.com/openemr/
lthttp//www.idltechnology.com/products/openemr/ind
ex.phpgt) - b. Care2002 (lthttp//sourceforge.net/projects/care
2002/gt) - c. Open Infrastructure for Outcomes UCLA
(lthttp//sourceforge.net/projects/open-outcomes/gt)
- d. PatientRunner (lthttp//sourceforge.net/projects
/patientrunner/gt) mental health record system - e. OpenSDE (lthttp//sourceforge.net/projects/opens
de/gt) structured note entry system - f. MedSurvey (lthttp//sourceforge.net/projects/med
survey/gt) clinical information system for
Windows PCs - g. OpenEMed (lthttp//sourceforge.net/projects/open
med/gt) Java based EMR - h. Hardhats (VAs VISTA software) yes this IS
open source now and available to EVERYONE
(lthttp//www.hardhats.org/gt), (lthttp//sourceforge
.net/projects/hardhats/gt)
104EMRs for Primary Care Practice
- Recent survey done by the journal Family Practice
Management (2001) - Surveyed 28 vendors
- Price structure highly variable
- Found that the market is highly volatile and some
vendors went out of business or merged with
others during the time of the survey
105EMRs for Primary Care Practice
- Five star systems
- ChartWare
- HealthProbe Patient Information Manager
- EpicCare
106EMRs for Primary Care Practice
- Four Star Systems
- Logician
- NextGen
- Pearl
- Physician Practice Solutions
- PowerMed EMR
- Practice Partner Patient Records
- QD Clinical
107EMRs for Primary Care Practice
- Four Star Systems
- SOAPWare
- Welford Chart Notes
- Clinical Works Module (ASP)
- NextGen (ASP)
- Physician Practice Solution (ASP)
- topsChart (ASP)
108EMRs for Primary Care Practice
- 4 physician practices
- ENTITY, Logician, NextGen, ClinicalWorks
- 10 physician practices
- EpicCare, PEARL, Physician Practice Solution
- All others can serve practices of any size
109EMRs for Primary Care Practice
- Most allow ICD9 and CPT codes
- Many allow access from the web
- Most allow multiple modes of data entry
(keyboard, mouse, touch-screen, light-pen, voice
recognition etc.) - Most permit integration of hospital data with a
primary care database
110Integration with Handhelds
- Some EMRs allow data access from PDAs and other
handheld or laptop devices - ChartWare - O-HEAP
- DOCUMENTOR - Partner
- ENTITY - PowerMed
- EpicCare - SOAPWare
- MedicWare - Welford ChartNotes
- NextGen - ClinicalWorks
- topsCHART
111Other EMR Surveys/Resources
- HealthCare Informatics 2004 Resource Guide
- Comprehensive listing of EMRs, features, costs,
contact information etc. - 50 per copy
- Order from
- http//www.healthcare-informatics.com
112Some Lessons Learned
113Lessons learned the hard way
- Well-designed renowned vendor products meet about
80 of your needs -gt where will the other 20
come from? - Poorly designed systems will be quickly abandoned
by time-pressured end-users - Caveat Emptor Total Solution, Turnkey
solution, esp if a proprietary black box
114Lessons learned the hard way
- Clinical/Administrative information is inherently
structured. Capturing it in unstructured ways
(images) is a costly mistake - Data acquisition costs may be more expensive than
operational expense (I.e. keyboard entry time
more costly than provider input)
115Lessons learned the hard way
- Users will accept a tradeoff if there is a clear
payback in functionality - Attitudes towards computer use are not age
dependent - Be the 10th customer to a vendor, never the
first! - Beware of vendors who say we can do that what
is it?
116Lessons learned the hard way
- The most important information a vendor will give
you is the address of 2-3 sites where their
system is currently in use
117Acknowledgements
- David Bates, MD
- Daniel Masys, MD
118DISCUSSION AND QUESTIONS