how to get a handle on your patient identity challenges
TRANSCRIPT
Presenter Name: Michelle Schneider, Senior Solutions Engineer [email protected], 978-805-4143
How to Get a Handle on Your Patient Identity Challenges
Agenda
1. Patient Identity
2. EMPI – What is it
3. Where to begin
4. Managing an EMPI
5. Iatric Systems EMPI Solutions
Patient Identity?
Who is a patient and are we sure? • A patient is someone who is receiving or registered to receive
medical treatment. • Who is the someone and are we sure they are who they say:
• ID card (SSN) • Drivers license • Credit card
• Other verification methods – Equifax, Experian, TransUnion. • Identifies you are who you say you are…
Normally not the process of an EMPI to identify a specific person through additional means or checks.
Patient Identity across a delivery network • A single system has painstakingly identified a
single person. (if previous slide was true) • The patient is sent elsewhere for additional lab
work, another consult visit or for additional treatment.
• The delivery network (HIE, IDN) can now have
segregated information in multiple systems for a single identified “Patient”.
This is where trouble can occur
But, EMPI’s can help
EMPI
Situation: Multiple instances of the same or similar demographic information on a “patient” can be found in different registration systems, clinical systems, etc.
Remedy: • Software employs complex algorithms to determine matching and
non-matching characteristics. • Processes to manage the work flow of remediation of matched and
un-matched • Reporting, checks and balances
Deterministic Matching – Usually represents an exact match between two pieces of data. It can include some rules but all rules must be met every time. A large HIE will outgrow this quickly because there will be too many unmatched, duplicate records, searches/queries. Probabilistic Matching – A statistical approach to evaluate the probability that two records represent the same person. By associating algorithms and applying scores to outcomes, matches can be made with confidence when a threshold is met. Example:
Common Terminology
EMR 1 EMR 2 Deterministic Match? (In it’s purest form)
Probabilistic Match? (With algorithms in place)
First JOHN JON No Yes
Last SMITH SMYTHE No Yes
Middle R R Yes Yes
Addr 1 110 SOUTH MAIN ST.
110 S. MAIN STREET No Yes
City SOUTH WINDSOR S. WINDSOR No Yes
State CT CT Yes Yes
Zip 06074 06074 Yes Yes
DOB 19900808 19900808 Yes Yes
WILL THE SYSTEM MAKE THE MATCH? NO YES
False Positive – A match between two records that do not represent the same person False Negative– A match result that fails to match two records that represent the same person. The records are thought to relate to separate individuals. Example:
Common Terminology
Match Score
Partial Match/Nickname Score
Completely different
Missing
First 8 5 0 0
Last 15 10 -4 -3
Middle 4 1 -2 0
Gender 4 0 -4 0
Addr 1 12 8 0 0
City 6 4 -2 0
State 4 2 -5 0
Zip 6 0 0 0
DOB 8 4 -4 0
SSN 20 6 -10 0
Threshold for Matching 45
Common Terminology
Field EMR 1 EMR 2 Score
First JOHN GREGORY 0
Last SMITH SMITH 15
Middle R R 4
Gender M M 4
Addr 1 Evergreen Nursing Home
Evergreen Nursing Home
12
City S. WINDSOR S. WINDSOR 6
State CT CT 4
Zip 06074 06074 6
DOB 19300808 19320808 4
SSN 047-68-5412 047-62-2874 -10
WILL THE SYSTEM MAKE THE MATCH?
45=Yes
Field EMR 1 EMR 2 Score
First JOHN JON 5
Last SMITH SMITH 15
Middle R 0
Gender M M 4
Addr 1 110 SOUTH MAIN ST.
110 S. MAIN STREET
8
City SOUTH WINDSOR
S. WINDSOR 4
State CT CT 4
Zip 06074 06074 6
DOB 19900808 19900808 8
SSN 047-68-5412 027-86-5841 -10
WILL THE SYSTEM MAKE THE MATCH?
44=No
False Negative
False Positive
James Sir Brown, Male, 01/01/1980, 101 Home, Leakey, Texas 78006, 830-555-1345
Jim Brown, Male, 01/10/1980, 830-555-1345, 403-55-1356
Facility ABC:123456 DOS: 01/20/1999
Facility XYZ:345789 DOS: 09/27/2009
Information sent from different facilities on the same person
EMPI:1101101
Can now aggregate data and use for other purposes Analytics, Pop Health, etc
Diagram of Best Practice Workflow
James Brown, M, 121 House, Fredericksburg, Texas 78213, 01/10/1980
Facility JFK:AE345672 DOS: 09/27/2015
Facility ABC:123456 James Sir Brown, Male, 01/01/1980, 101 Home, Leakey, Texas 78006, 830-555-1345
Facility XYZ:345789 Jim Brown, Male, 01/10/1980, 830-555-1345, 403-55-1356
Facility JFK:AE345672 James Brown, M, 121 House, Fredericksburg, Texas 78213, 01/10/1980
EMPI 1101101 James 830-555-1345, 403-55-1356 Brown, M, 121 House, Fredericksburg, Texas 78213, 01/10/1980 Sir
Benefits of an EMPI for Workflow
• Remove duplicates from existing systems. Eliminates duplicate charts within a facility and combines all relevant clinical information under 1 chart (carved in stone or on paper). HIM/Registration
• Assist in billing under the correct insurance information in the event of multiple MRN numbers. Accounting
• Truer picture of patients under a provider’s care or multiple providers at multiple sites. Providers
• Ability to aggregate data under multiple sources of information in the event of an HIE or large disparate integrated network. Providers/CCD/MU
Benefits of an EMPI for Workflow
• Benefit to patients all results should be viewable in patient portal. Patients
• Decrease cost of doing business when the number of patient lives represented is true. Budget
• Analytics are more accurate because the patient count is now correct. HIM/Analytics
Problems Associated with EMPI Issues
• Disparate and incomplete clinical information on patients across an enterprise.
• Billing duplications
• Incorrectly matching records within a facility or across an enterprise. For example: • Mislabels due to incorrect link • Father/son confusion due to Jr./Sr. information • Confusion between individuals with similar information from
the same address (assisted living site)
• Incorrectly NOT matching records within a facility or across an enterprise • Deterministic matching can present a risk
OK, I’ve got concerns. Where do I begin?
• Where is all the data coming from? • Interfaces:
• ADT • EMR • Physician Practices • Ancillary Orgs –Labs, Rads, SNF
• Flat Files • Order Messages • CCD
• Have you ever converted from another EMR? • What do those records look like? • Are the MRNs formatted differently?
OK, I’ve got concerns. Where do I begin?
• Get samples of data types and compare • Examples
• requirements from one ADT to another • How do unknown SSNs get processed?
• Understand the rules from each source. • Gather documentation from all sources
• Multiple MRNs for one person • Understand merge process in all sources
OK, I’ve got concerns. Where do I begin?
• Know your duplicate rate • Potential Duplicates
• Close but not quite right • How often does this happen? • Why does this happen?
• Simple name differences • Nickname • Missing information from a source • SSN • Baby Boy/Girl
• Any quick wins?
• Reeducation of registration staff • Require certain fields from particular source • Establish procedures – SSN, Baby naming, Multiples
OK, I’ve got concerns. Where do I begin?
• Data quality – Strive for complete, accurate and consistent data across your exchange. • Registration • Interfaces • CCD • All sources
• Think proactively – personal device integration
• Do you have an EMPI? • No –As you add sources, you increase your risk
• It will be difficult to add sources and maintain this model
• Yes – If your duplicate rate is high and false matches are frequent, consider analysis and remediation
Managing an EMPI Best practices include focus on: • Remediation efforts – auto link only process vs. remediation
processing • Resolve today’s issues and errors • Clean the EMPI • Keep it Clean
• Staffing – don’t under-estimate your staffing when trying to manage the remediation of a large EMPI. • Education – end users • Policy/Procedure – create, implement, audit • Algorithm research, testing, tweaking • Reaching out to patients – Are you who we think you are? • Interface management, modifications, testing • Merge/Unmerge management
Managing an EMPI Best practices include focus on:
• Data governance across the enterprise: consider who has the final say in what is a link if remediation is occurring
• Process and reporting – Understanding your data and what systems provide what information. • SSN – do 2 people share the same SSN but are not the same
person? • Hospital re-admits or drug seekers at multiple locations? • Report of “significant” changes
• First name • Gender • Date of birth • SSN – that was not all 9’s for example
• Establish consistency
Iatric Systems Solutions Data analysis – ARGO DPA tool to review data from a source or sources of patient records.
• Number of duplicates within a source, across sources (free report)
• Data quality (baby, reused SSN’s, default data) included in Free Analysis
• Can provide a linked file back to client for a fee. Remediation support – managing a work effort to clean up duplicate records and link records within the EMPI.
• Human resources • One time and/or ongoing
Iatric Systems Solutions Analytics Consulting, EMPI & 3rd Party Integration Practice
• 3rd Party Integration
• EMPI - Initiate, NextGate, Argo, InterSystems, MirthMatch, etc. • Consulting • Implementation • Data remediation • Matching/algorithm/rules consulting • Analytics/BI technical consulting
• Migration / Conversion-based technical work
Mirth Cloverleaf Rhapsody
Corepoint Openlink Etc.
Iatric Systems
Deliverables 1. FREE EMPI Analysis
• Duplicate Profile Analysis – Performed by Iatric Systems and Powered by Argo
2. Comprehensive reports including: a) Data review of sample files from multiple sources b) Match rate and accuracy by EMR and Overlap c) Match record distribution i.e. EMR vs. Network d) False positives/False negatives e) Population analysis and review i.e. Baby girl, high number
of same last name f) Field level analysis – every field i.e. first name, dob, zip
code – evaluated for accuracy, error rate and false positive risk
Open Discussion
Sorular Вопросы
Questions Frågor Fragen Ερωτήσεις
שאלות Pytania 質問 सवाल Domande
Preguntas Cwestiynau
Algorithm – A complex set of steps to create an outcome. (Felligi Suntor (1969) is the base algorithm that probabilistic algorithm’s employ with vendors adding additional parsing routines, data attribute availability, etc to enhance). Attribute – a variable that identifies a type of information to determine likeness on same values (ie..First Name, Last Name, DOB, zipcode, etc). Tasks – an item that may require remediation. Can be a duplicate (2 records with different MRN numbers from the same source of information), Overlays (a single MRN that may have been overwritten by someone else's information). Can also include potential links and potential duplicates. Weighting – A numeric value given to specific attributes that when combined with other attributes creates an overall score. Scoring – Adding of multiple attributes scores to give a total value.
Common Terminology
Thresholds – a set score that when achieved a set of records can be linked automatically (auto match) or in which a task can be created for manual review (potential match). Enterprise Master Patient Index (EMPI) – an indexed system of patient records across an enterprise of participating data sources that can represent a single entity of a patient record based on patient demographics. Link or Match – creating an association between 2 or more records across the same or multiple sources of information. Merge – creating a “surviving” record and an “obsolete” record from multiple records containing multiple charts in a single registration system. Not usually performed within an HIE. Entity – multiple sources of demographic information tied together with a single EMPI number thus creating an entity (single best record, golden record, etc)
Common Terminology
I Can Help!
Michelle Schneider Senior Solutions Engineer
Iatric Systems, Inc.
Phone: (978) 805-4143 E-mail: [email protected]
Connect with me on LinkedIn: MichelleSchneider
Read our Blog and sign up for the topics that interest you: new.iatric.com/blog-home
Follow Us
new.iatric.com/blog-home
Contact Us
Iatric Systems Account Executive [email protected]
EMPI – Professional Services Contact Us