Using RELMA Or… In Search of the Missing LOINC © 2007 Regenstrief Institute & James Case James T. Case MS, DVM, PhD Professor, Clinical Diagnostic Informatics California Animal Health and Food Safety Laboratory University of California, Davis Laboratory LOINC Meeting – June 2008

Acknowledgements RELMA Development Team John Hook, Mark Fisher, Ryan Phillips, Karen Ahmad, and more! Kathy Mercer Clem McDonald Dan Vreeman The Lab LOINC Committee Funding Support NLM, Regenstrief Institute, NCI, CDC © 2007 Regenstrief Institute & James Case

RELMA Development Team

John Hook, Mark Fisher, Ryan Phillips, Karen Ahmad, and more!

Kathy Mercer

Clem McDonald

Dan Vreeman

The Lab LOINC Committee

Funding Support

NLM, Regenstrief Institute, NCI, CDC

What will we cover today? Installing RELMA Setting personal preferences Loading a Local Observation File (LMOF) Preparing LMOF for Mapping Review of Map Screen Functions Setting Search Limits Mapping Local Terms to LOINC Viewing LOINC Term Details Proposing/Submitting New LOINC Terms Exporting/Printing Mapped Terms Mapping your own LMOF data

Installing RELMA

Setting personal preferences

Loading a Local Observation File (LMOF)

Preparing LMOF for Mapping

Review of Map Screen Functions

Setting Search Limits

Mapping Local Terms to LOINC

Viewing LOINC Term Details

Proposing/Submitting New LOINC Terms

Exporting/Printing Mapped Terms

Mapping your own LMOF data

Goals of Health Information Standards Interoperability – the ability to exchange information between organizations Comparability – the ability to ascertain the equivalence of data from different sources Data Quality – the measurement of completeness, accuracy and precision

Interoperability – the ability to exchange information between organizations

Comparability – the ability to ascertain the equivalence of data from different sources

Data Quality – the measurement of completeness, accuracy and precision

Levels of Interoperability Basic – allows messages to be exchanged between computer systems Word processing documents, spreadsheets Functional – describes the standard syntax (format) of the message Document templates, forms, data structures Message standards Semantic – requires use of standard vocabularies within the message

Basic – allows messages to be exchanged between computer systems

Word processing documents, spreadsheets

Functional – describes the standard syntax (format) of the message

Document templates, forms, data structures

Message standards

Semantic – requires use of standard vocabularies within the message

Comparability Meaning of the data is consistent when shared among different parties Common terminology required Should work in the background Not just words Codes – uniquely identifies terms Classification – groups related terms Vocabulary – specialized, precise terms that remove ambiguity

Meaning of the data is consistent when shared among different parties

Common terminology required

Should work in the background

Not just words

Codes – uniquely identifies terms

Classification – groups related terms

Vocabulary – specialized, precise terms that remove ambiguity

Potential Uses of Health Data are Constrained by Data Quality Factors Data quality issues Different for client communications vs. analysis Often constrained by external forces E.g. criteria for diagnosis often differs from the criteria for reporting Unidirectional effect of data consolidation Detailed -> General General -> Detailed

Data quality issues

Different for client communications vs. analysis

Often constrained by external forces

E.g. criteria for diagnosis often differs from the criteria for reporting

Unidirectional effect of data consolidation

Detailed -> General

General -> Detailed

The Problem “ In attempting to arrive at the truth, I have applied everywhere for information, but in scarcely an instance have I been able to obtain hospital records fit for any purpose of comparison. If they could be obtained, they would enable us to decide many other questions… They would show [subscribers] how their money was being spent [and] what amount of good was really being done with it…” Florence Nightingale - Notes on a Hospital, 1873

“ In attempting to arrive at the truth, I have applied everywhere for information, but in scarcely an instance have I been able to obtain hospital records fit for any purpose of comparison. If they could be obtained, they would enable us to decide many other questions… They would show [subscribers] how their money was being spent [and] what amount of good was really being done with it…”

Why LOINC? “ Within one laboratory, local jargon terms may be used which are usually well understood between colleagues, but would not be sufficiently widely known for communication with the outside world.” U. Forsum et al., Pure Appl. Chem 72:555-745, 2000 Properties and Units in the Clinical Laboratory Sciences Part VII. Properties and Units in Clinical Microbiology

“ Within one laboratory, local jargon terms may be used which are usually well understood between colleagues, but would not be sufficiently widely known for communication with the outside world.”

U. Forsum et al., Pure Appl. Chem 72:555-745, 2000 Properties and Units in the Clinical Laboratory Sciences Part VII. Properties and Units in Clinical Microbiology

Test comparisons Lab A Test Name: Lyme Disease Serology Measures: B. burgdorferi Ab IgG Method: ELISA Scale: quantitative e.g.: Titer 1:40 Lab B Test Name: Lyme Disease Antibody Measures: B. burgdorferi Ab IgM Method: Immune blot Scale: qualitative e.g.: Positive LOINC Code = 5062-5 LOINC Code = 6321-4 What you see in the order list

Lab A

Test Name: Lyme Disease Serology

Measures: B. burgdorferi Ab IgG

Method: ELISA

Scale: quantitative

e.g.: Titer 1:40

Lab B

Test Name: Lyme Disease Antibody

Measures: B. burgdorferi Ab IgM

Method: Immune blot

Scale: qualitative

e.g.: Positive

LOINC ® 101

Anatomy of a LOINC Term 5193-8 LOINC Code Hepatitis B virus surface Ab ACnc Pt Ser Qn EIA Component Property Measured Timing System Scale Method There are six major LOINC axes 5193-8:Hepatitis B virus surface Ab:ACnc:Pt:Ser:Qn:EIA

What is NOT part of a LOINC Name? The instrument used in testing Specific details about the specimen Priority (e.g. STAT) Where testing was done Who did the test Test interpretation Anything that is not an intrinsic part of the name of the result Other things that are carried in; The OBX segment An HL7 Version 3 Observation Object

The instrument used in testing

Specific details about the specimen

Priority (e.g. STAT)

Where testing was done

Who did the test

Test interpretation

Anything that is not an intrinsic part of the name of the result

Other things that are carried in;

The OBX segment

An HL7 Version 3 Observation Object

Component/Analyte The substance or entity that is measured, evaluated, or observed - Sodium - Glucose - Brucella sp. organism - Influenza A Virus antigen Cytomegalovirus Virus antibody Lipids.Total 5193-8: Hepatitis B virus surface Ab :ACnc:Pt:Ser:Qn:EIA

- Sodium

- Glucose

- Brucella sp. organism

- Influenza A Virus antigen

Cytomegalovirus Virus antibody

Lipids.Total

Component/Analyte Structure Formal name of Analyte (e.g. Calcium) Must specify any “subanalyte” e.g. Coronavirus Ag May have a subclass – separated by “.” e.g. Calcium.Free Challenge - e.g, 1H post 100 gm Glucose PO Two subparts separated by “post” <time delay> post <challenge type> Adjustments/corrections E.g. Adjusted to pH 7.4 Analyte Name^Challenge^Adjustments

Formal name of Analyte (e.g. Calcium)

Must specify any “subanalyte”

e.g. Coronavirus Ag

May have a subclass – separated by “.”

e.g. Calcium.Free

Challenge - e.g, 1H post 100 gm Glucose PO

Two subparts separated by “post”

<time delay> post <challenge type>

Adjustments/corrections

E.g. Adjusted to pH 7.4

Mass: Observations reported with mass (milligrams, grams, etc.) in the numerator of their units of measure Substance: Observations reported with moles or milliequivalents in the numerator of their units of measure Catalytic activity: Observations that report enzymatic activity Arbitrary: Results that report arbitrary units in the numerator of their units of measure Number: Counts Property* The characteristic or attribute of the analyte that is measured, evaluated, or observed the most difficult LOINC axis Major Categories 5193-8:Hepatitis B virus surface Ab: ACnc :Pt:Ser:Qn:EIA

Mass: Observations reported with mass (milligrams, grams, etc.) in the numerator of their units of measure

Substance: Observations reported with moles or milliequivalents in the numerator of their units of measure

Catalytic activity: Observations that report enzymatic activity

Arbitrary: Results that report arbitrary units in the numerator of their units of measure

Number: Counts

MCnc – mass concentration (mass/unit vol) MCnt – mass content (mass/unit mass) NCnc – number concentration (number/unit vol) TmStp – time CCnc – catalytic concentration (activity) Prid – presence or identity Imp – impression/interpretation Find – subjective or objective observation Type – “Kind-of” Property* the most difficult LOINC axis Combine Major Categories with Subtypes for Full Property 5193-8:Hepatitis B virus surface Ab: ACnc :Pt:Ser:Qn:EIA

MCnc – mass concentration (mass/unit vol)

MCnt – mass content (mass/unit mass)

NCnc – number concentration (number/unit vol)

TmStp – time

CCnc – catalytic concentration (activity)

Prid – presence or identity

Imp – impression/interpretation

Find – subjective or objective observation

Type – “Kind-of”

Common Issues with LOINC Properties Fraction = Part/whole Number fraction (NFr): % Eosinophils Substance fraction (SFr): % HGB which is A2 Ratio = Measures from the same system Mass concentration ratio - MCrto e.g., BUN/Creat in urine specimen Substance ratio-SCrto Urea/Creatinine expressed as mmol/L (SI units) Relative Ratio = Measures from different systems RelRto – multiple of the median RlTm – time from actual and normal control Fraction (proportion) vs. Ratio a/a+b vs. a/b

Fraction = Part/whole

Number fraction (NFr): % Eosinophils

Substance fraction (SFr): % HGB which is A2

Ratio = Measures from the same system

Mass concentration ratio - MCrto

e.g., BUN/Creat in urine specimen

Substance ratio-SCrto

Urea/Creatinine expressed as mmol/L (SI units)

Relative Ratio = Measures from different systems

RelRto – multiple of the median

RlTm – time from actual and normal control

Pt - at a point in time 12H - a twelve hour collection 24H - a twenty four hour collection Timing* *non-Pt timings are usually associated with Ratio Property The interval of time over which the observation or measurement was made

Pt - at a point in time

12H - a twelve hour collection

24H - a twenty four hour collection

System The system (context) or specimen type upon which the observation was made. Super System Second subpart (^). When not included, “patient” is the default. Used to indicated blood product unit (BPU), a bone marrow donor, or a fetus.

Qn - quantitative Continuous numeric (real, integer, ratio) Optional operator (>,  ,  , <) Ord - ordinal a ranked set of possible values (1+, 2+, 3+) Nom - nominal an unranked collection of possible values a taxonomy (e.g list of bacteria) Nar - narrative free text narrative (e.g., visit note) Scale

Qn - quantitative

Continuous numeric (real, integer, ratio)

Optional operator (>,  ,  , <)

Ord - ordinal

a ranked set of possible values (1+, 2+, 3+)

Nom - nominal

an unranked collection of possible values

a taxonomy (e.g list of bacteria)

Nar - narrative

free text narrative (e.g., visit note)

Scale (Special) OrdQn – Ordinal or Quantitative Primarily used for antimicrobial observations e.g. MIC reported as resistant, intermediate, susceptible or as the mm diameter of the inhibition zone Use is discouraged Multi – structured text “globs” e.g. chromatography output Use is discouraged Doc – Clinical documents Set – Clinical attachments (headers)

OrdQn – Ordinal or Quantitative

Primarily used for antimicrobial observations e.g. MIC reported as resistant, intermediate, susceptible or as the mm diameter of the inhibition zone

Use is discouraged

Multi – structured text “globs”

e.g. chromatography output

Use is discouraged

Doc – Clinical documents

Set – Clinical attachments (headers)

Method Methods only needed if interpretation affected Different normal ranges Test Sensitivity Listed only at the generic level Agglutination Immunoassay Probe with target amplification

Methods only needed if interpretation affected

Different normal ranges

Test Sensitivity

Listed only at the generic level

Agglutination

Immunoassay

Probe with target amplification

Mapping Terms to LOINC Things to Remember The thing ordered is not always the thing measured: Blood Culture – live organism(s) identified VDRL – Treponema pallidum Ab Urinalysis – lots of different things The question (what am I measuring?) is not the answer (e.g. Pos) You are mapping the question, not the answer! You must know the specifics of the component being tested for (what is this test actually measuring?)

The thing ordered is not always the thing measured:

Blood Culture – live organism(s) identified

VDRL – Treponema pallidum Ab

Urinalysis – lots of different things

The question (what am I measuring?) is not the answer (e.g. Pos)

You are mapping the question, not the answer!

You must know the specifics of the component being tested for (what is this test actually measuring?)

RELMA ® Regenstrief LOINC Mapping Assistant

Purpose of Tutorial “Eliminate” need to read the User’s Manual Become familiar with RELMA features Provide some insight into mapping tips/pitfalls Help begin the mapping process for your institution Note: Most screen shots from version 3.24(early release)

“Eliminate” need to read the User’s Manual

Become familiar with RELMA features

Provide some insight into mapping tips/pitfalls

Help begin the mapping process for your institution

Note: Most screen shots from version 3.24(early release)

New in this Version (rel. 3.24) Determine which panels have any of or all of a group of LOINCs. LOINC Details new display type, &quot;Custom Display&quot;. Custom Export - Select columns to send to the Clipboard, Excel, a CSV file, or email program. Exporting mapped local terms to a delimited file support LOINC parts Component, Property, Time, System, Scale, and Method of mapped terms. Row numbers now displayed on restriction trees and multi-axial tree search results. Grids allow switch between word-wrapped and truncated cell contents. Truncated cells display an ellipsis (...)

Determine which panels have any of or all of a group of LOINCs.

LOINC Details new display type, &quot;Custom Display&quot;.

Custom Export - Select columns to send to the Clipboard, Excel, a CSV file, or email program.

Exporting mapped local terms to a delimited file support LOINC parts

Component, Property, Time, System, Scale, and Method of mapped terms.

Row numbers now displayed on restriction trees and multi-axial tree search results.

Grids allow switch between word-wrapped and truncated cell contents.

Truncated cells display an ellipsis (...)

New in this Version (rel. 3.24) Context menu now has the &quot;Propose a LOINC&quot; action. When adding or editing a local term, can use the context menu. When displaying LOINC/Part Details, you can type a LOINC/Part number into the LOINC/Part number box, hit Enter, and RELMA will display that LOINC/Part. Then, that LOINC/Part is added to the end of the list of displayed LOINCs/Parts. The Change Submission form now lists the name of the local term file, the number of local terms in the file, the number of mapped terms, the number of unmapped terms, and the percentage of mapped terms. When RELMA is set to ask for a comment at each mapping, if no comment is supplied the mapping is not performed.

Context menu now has the &quot;Propose a LOINC&quot; action.

When adding or editing a local term, can use the context menu.

When displaying LOINC/Part Details, you can type a LOINC/Part number into the LOINC/Part number box, hit Enter, and RELMA will display that LOINC/Part.

Then, that LOINC/Part is added to the end of the list of displayed LOINCs/Parts.

The Change Submission form now lists the name of the local term file, the number of local terms in the file, the number of mapped terms, the number of unmapped terms, and the percentage of mapped terms.

When RELMA is set to ask for a comment at each mapping, if no comment is supplied the mapping is not performed.

New in this Version (rel. 3.24) Units from the original submission are included as part of the example units on the LOINC Details. On the LOINC Details display, externally defined answers are now explicitly shown as externally defined. Importing HL7 switches to the local term file of the HL7 import. Switching preferred language from the User Preferences form immediately loads the language. Support for searching the LOINC database in multiple languages. The HIPAA HL7 attachments displays have been modifed to display on the &quot;Panels, Forms & Surveys&quot; form.

Units from the original submission are included as part of the example units on the LOINC Details.

On the LOINC Details display, externally defined answers are now explicitly shown as externally defined.

Importing HL7 switches to the local term file of the HL7 import.

Switching preferred language from the User Preferences form immediately loads the language.

Support for searching the LOINC database in multiple languages.

The HIPAA HL7 attachments displays have been modifed to display on the &quot;Panels, Forms & Surveys&quot; form.

Manual and automated mapping assistant Same free use as LOINC (see license) Comes with the LOINC files and indexes Developed in Visual Basic (No API) RELMA tools transform local words in local file User creates file of local term/name and codes Assigns LOINC term to local test/battery code RELMA Functions

Manual and automated mapping assistant

Same free use as LOINC (see license)

Comes with the LOINC files and indexes

Developed in Visual Basic (No API)

RELMA tools transform local words in local file

User creates file of local term/name and codes

Assigns LOINC term to local test/battery code

Highlights Browseable Hierarchies Display search results in tree view Enhanced “details” view of terms/parts References Descriptions Sample units Empirically-derived common test list Special features for panels/forms Enhanced export/copy-paste options

Browseable Hierarchies

Display search results in tree view

Enhanced “details” view of terms/parts

References

Descriptions

Sample units

Empirically-derived common test list

Special features for panels/forms

Enhanced export/copy-paste options

For a given test, users can select LOINC records containing: A particular category of test (e.g., microbiology) Any selected word in the input term Additional keyed in word(s) Wild cards of ? and * Any intersection of the above RELMA

For a given test, users can select LOINC records containing:

A particular category of test (e.g., microbiology)

Any selected word in the input term

Additional keyed in word(s)

Wild cards of ? and *

Any intersection of the above

RELMA Logic Not  #<word A> Or  <word A> | <word B > And  <word B> <word A>

Not  #<word A>

Or  <word A> | <word B >

And  <word B> <word A>

For a given test: User can select LOINC records containing: A set of words connected by ANDs or Ors Wild cards of ? and * Selected classes, systems, components, Any intersection of the above Note: RELMA assumes exact match on word unless user adds terminal “*” to indicate wild card. RELMA Search logic

For a given test:

User can select LOINC records containing:

A set of words connected by ANDs or Ors

Wild cards of ? and *

Selected classes, systems, components,

Any intersection of the above

Note:

RELMA assumes exact match on word unless user adds terminal “*” to indicate wild card.

Additional RELMA Features User Specified Search Limits Selectable trees for: Class System (specimen) Component Method

User Specified Search Limits

Selectable trees for:

Class

System (specimen)

Component

Method

Installing RELMA®

Installation Steps Make sure you have enough free disk space! At least 250Mb are needed From CD Start – Run – D:RELMASetup Specify installation directory Life will be easier if you accept the default Two database files installed RELMA.MDB – LOINC Terms Database LMOF3.MDB – Local Master Observation File Two sample files copied (C:Program FilesRELMA) Run from Start – (All) Programs – Regenstrief – RELMA

Make sure you have enough free disk space!

At least 250Mb are needed

From CD

Start – Run – D:RELMASetup

Specify installation directory

Life will be easier if you accept the default

Two database files installed

RELMA.MDB – LOINC Terms Database

LMOF3.MDB – Local Master Observation File

Two sample files copied (C:Program FilesRELMA)

Run from Start – (All) Programs – Regenstrief – RELMA

Running RELMA®

Copyright Screen © 2006 Regenstrief Institute & James Case

Main Menu/Welcome Screen © 2006 Regenstrief Institute & James Case Click File, Set Preferred Language

Select Linguistic Variant(s)

LOINC Terms with Spanish Linguistic Variant Spanish!

Welcome Screen

Setting User Preferences

Select User Preferences From the File Menu Set User Preferences

Set User Preference Dialog © 2006 Regenstrief Institute & James Case User Name or Initials File Locations Startup Screen Preference RELMA Color Scheme

RELMA Color Schemes Predefined Schemes Create, Save, Delete custom color schemes Customize Colors

User Map Screen Preferences Click Map Screen Tab

Importing Local Terms into RELMA

4 Ways to Load LMOF Files Direct entry into LMOF from within RELMA (painful) Create an Access table that mimics the LMOF structure (less painful but tedious) Appendix A: RELMA Manual Create a delimited ASCII file from your local test catalog (good choice) Load directly from HL7 v2.x messages (best? choice) Pulls data from OBR and OBX segments

Direct entry into LMOF from within RELMA (painful)

Create an Access table that mimics the LMOF structure (less painful but tedious)

Appendix A: RELMA Manual

Create a delimited ASCII file from your local test catalog (good choice)

Load directly from HL7 v2.x messages (best? choice)

Pulls data from OBR and OBX segments

Constructing a Local Dataset Create extract of your test catalog with: Battery/Panel Code Battery/Panel Description or Name Local Code Test Description or Name Include Method if Important Units Example Values Laboratory Section

Create extract of your test catalog with:

Battery/Panel Code

Battery/Panel Description or Name

Local Code

Test Description or Name

Include Method if Important

Units

Example Values

Laboratory Section

Creating Delimited ASCII File RELMA can’t parse free text Need to create separate fields Can use any of these delimiters Tab, Semicolon, Comma, Space Can define your own (“|” character recommended) Fields can be in any order Minimum required fields Local Code Local Description Units (highly recommended)

RELMA can’t parse free text

Need to create separate fields

Can use any of these delimiters

Tab, Semicolon, Comma, Space

Can define your own (“|” character recommended)

Fields can be in any order

Minimum required fields

Local Code

Local Description

Units (highly recommended)

Entering New Local Terms into RELMA

Manual Enter/Edit © 2006 Regenstrief Institute & James Case View/Add/Edit Local Terms

Edit Term Dialog Add New Test Print Preview List Text alignment option Custom Export Format

Add Local Term Form Click to Add to Current Working File MICRO BUBPLAG BUBONIC PLAGUE – RRT PCR

Print Preview Local Term File

Text Alignment Truncated vs. Wrapped Truncated terms end with ellipsis

Custom Export Options Output Options

Exported as Excel File

Steps to Import Local Terms Click Import Local Terms Button from Main Menu a) File>Import Local Terms from Delimited File Locate your local terms text file Sample file loaded into C:Program FilesRELMA Name your Working Set LMOF database can contain multiple work sets Define default section (Optional) Identify file delimiter Assign fields to LMOF attributes Ignore fields you don’t need Combine fields if needed Check “Case-sensitive” if needed Click Import .

Click Import Local Terms Button from Main Menu

a) File>Import Local Terms from Delimited File

Locate your local terms text file

Sample file loaded into C:Program FilesRELMA

Name your Working Set

LMOF database can contain multiple work sets

Define default section (Optional)

Identify file delimiter

Assign fields to LMOF attributes

Ignore fields you don’t need

Combine fields if needed

Check “Case-sensitive” if needed

Click Import .

Example Delimited File © 2006 Regenstrief Institute & James Case

Importing Local Files © 2006 Regenstrief Institute & James Case Select Import Local Terms

Open File in RELMA © 2006 Regenstrief Institute & James Case Name your working set. RELMA allows multiple sets in LMOF database Select your delimiter Import Button

Open File in RELMA © 2006 Regenstrief Institute & James Case

Fields Segregated © 2006 Regenstrief Institute & James Case Assign LMOF Attribute

Ready to Import © 2006 Regenstrief Institute & James Case Now you can click this button!

Alternative (best) Way Use large set of HL7 messages Automatically make dataset of: OBR ID OBR description OBX ID OBX description Sample of results with Real values Units Abnormal flags Normal ranges

Use large set of HL7 messages

Automatically make dataset of:

OBR ID

OBR description

OBX ID

OBX description

Sample of results with

Real values

Units

Abnormal flags

Normal ranges

Generate Local term file from HL7 messages Select Import Terms from HL7 File

Select File to Import Select HL7 File

Specify HL7 File Name and Sample Size Name the Local Term File (working set)

HL7 Messages Data Statistics © 2006 Regenstrief Institute & James Case

New Terms added to LMOF New Battery Code Sample Data

Import the two sample files provided Import_Sample_OBR.txt Contains battery code and description Import_Sample_OBX.txt Contains test code and description Create 2 working sets Lenny L’OINC says: “ Load the Sample Files Into RELMA now!”

Import the two sample files provided

Import_Sample_OBR.txt

Contains battery code and description

Import_Sample_OBX.txt

Contains test code and description

Create 2 working sets

Cleaning your data

Preparing your Data for Mapping Improve mapping success by: Expanding abbreviations Standardizing colloquial terms Ignoring “administrative” terms Standardizing time references Can be done prior to importing Better to use tools built into RELMA

Improve mapping success by:

Expanding abbreviations

Standardizing colloquial terms

Ignoring “administrative” terms

Standardizing time references

Can be done prior to importing

Better to use tools built into RELMA

Preparing your Data for Mapping Find local Words not in LOINC Tools>Concepts in Local Term File Unknown to RELMA Creates a list of local terms in your file not found in LOINC Only needs to be run once per working set Can be printed or saved as a file Autosearches LOINC terminology Global Replace, Interactive Replace, Ignore

Find local Words not in LOINC

Tools>Concepts in Local Term File Unknown to RELMA

Creates a list of local terms in your file not found in LOINC

Only needs to be run once per working set

Can be printed or saved as a file

Autosearches LOINC terminology

Global Replace, Interactive Replace, Ignore

RELMA Cleaning Tools Use this tool to find unknown terms

Scanning Local Terms

Replace Local Terms © 2006 Regenstrief Institute & James Case Assign LMOF Attribute Replacement Options

RELMA Cleaning Tools Use this tool to make global substitutions

Term Substitution Save this global substitution for only this working set

Mapping Local Terms

Mapping Local Terms Select “ Map Local Terms to LOINC ” from Welcome Screen Select your Working Set to Map File>Change Local term File Select the subset of terms to work with: All Mapped Unmapped Set your Search Limits Set Search Limits Button

Select “ Map Local Terms to LOINC ” from Welcome Screen

Select your Working Set to Map

File>Change Local term File

Select the subset of terms to work with:

All

Mapped

Unmapped

Set your Search Limits

Set Search Limits Button

Standard Mapping Window Features Resizeable grid Reorder Grid columns View details of LOINC term Sort by column Click column Custom Sort Print or export results grid Spell check squiggly line to signify words not known to RELMA

Resizeable grid

Reorder Grid columns

View details of LOINC term

Sort by column

Click column

Custom Sort

Print or export results grid Spell check squiggly line to signify words not known to RELMA

Mapping Screen View or edit local term

View Local Term Details © 2006 Regenstrief Institute & James Case

Standard Mapping Screen Begin a search

Initial Mapping Results Click to show words used in search EEK! What’ll I do?!

Initial Mapping Results Use term checkbox Number of LOINC terms containing keyword Battery terms included in search

Revised Mapping Results Number of matching records found Match units selected by default

Re-revised Mapping Results Number of matching records found Match units unselected

Standard Mapping Screen Clear all input fields

Cleared Mapping Screen Enter keywords here CANINE DISTEMPER VIRUS IF

Ad hoc term search

Standard Mapping Screen Navigate through the local terms Show entire local term file

Show all terms Displays printed report format

© 2006 Regenstrief Institute & James Case

Additional Search Screen Options Navigate to viewed screens View Mode: Grid (default) or Tree (new)

Tree View Replaces “Group” View in prior versions Still in beta Results displayed hierarchically Defined by the multi-axial hierarchy in search restrictions Map to a term in tree by clicking Map button of double clicking term Only rows that have LOINC Codes

Replaces “Group” View in prior versions

Still in beta

Results displayed hierarchically

Defined by the multi-axial hierarchy in search restrictions

Map to a term in tree by clicking Map button of double clicking term

Only rows that have LOINC Codes

Tree View Cannot map to these Can map to these

Assigning a LOINC Map Highlight correct term Click “Map” Button (or doubleclick)

Adding Mapping Comments If comments option is selected, prompt appears Check with Frank in Chemistry

Assigning a LOINC Map LOINC Term Assigned

View LOINC Term Details Right clicking on a LOINC term brings up a Task Menu View details for a specific LOINC Term

LOINC Term Details Can scroll through returned subset of terms Change to expanded details view Change text size Can scroll down a single formatted page

Expanded Details View

Expanded Details View

Using Wildcards in Searches “*” replaces any number of characters GLUC* *COSE GLU*SE “?” replaces a single character GLUCOS? ?LUCOSE GLUC?SE

“*” replaces any number of characters

GLUC*

*COSE

GLU*SE

“?” replaces a single character

GLUCOS?

?LUCOSE

GLUC?SE

Other Search Qualifiers “ #” represents the “NOT” operator #GLUCOSE – terms without the Word Glucose “ |” is the “OR” operator GLUCOSE | LACTOSE Can have multiple “OR”s per box Can Mix OR and NOT and Wildcards

“ #” represents the “NOT” operator

#GLUCOSE – terms without the Word Glucose

“ |” is the “OR” operator

GLUCOSE | LACTOSE

Can have multiple “OR”s per box

Can Mix OR and NOT and Wildcards

Limiting Searches

Setting Search Limits Narrows search to specific subset of LOINC terms Reduces number of candidate terms Limits can be applied to all components Component attribute can be further restricted by number of words Tree structure allows for hierarchical constraints

Narrows search to specific subset of LOINC terms

Reduces number of candidate terms

Limits can be applied to all components

Component attribute can be further restricted by number of words

Tree structure allows for hierarchical constraints

RELMA - Search Constraints Controls general search constraints including: Limit to LOINC terms compatible with submitted units Force match with any specimen contained in name Methodless terms only ( now enhanced ) Limit to components with N or fewer words in their name Pop up search timing statistics after each search Use carefully or search may not be successful (Note parallel control switches at bottom of screen)

Controls general search constraints including:

Limit to LOINC terms compatible with submitted units

Force match with any specimen contained in name

Methodless terms only ( now enhanced )

Limit to components with N or fewer words in their name

Pop up search timing statistics after each search

Use carefully or search may not be successful

(Note parallel control switches at bottom of screen)

Setting Search Limits Click Hierarchy & Search Limits Button

Setting search limits Can also set most of them by toggling buttons at the bottom of the screen Toggle Buttons

Can also set most of them by toggling buttons at the bottom of the screen

Search Constraints Predefined, general search constraints Attribute trees Local Unit Constraint: Now defaults to “ON”

Terms Consistent with Units Only terms consistent with mmol/L appear

Setting Search Limits Specimen constraints

Terms Consistent with Specimen CSF Enter default Specimen

Terms Consistent with Specimen Only CSF Terms are returned

Setting Search Limits Methodless Terms Restriction Override Methodless Terms Restriction

Limit to Methodless Terms

Limit to Methodless Terms Only Methodless Terms Appear

More on Methodless Terms Some LOINC categories do not have methodless terms Checking methodless only will remove these from view on results grid Checking additional box allows these to be seen

Some LOINC categories do not have methodless terms

Checking methodless only will remove these from view on results grid

Checking additional box allows these to be seen

Methodless Only Unchecked All terms returned

Methodless Only Checked Only Methodless terms returned

Conditional Methodless More terms returned

Setting Search Limits Limit to Lab Tests Only (No Clinical LOINC Terms)

Setting Search Limits Limit to lab tests that comprise 99.8% of INPC volume

Setting Search Limits Limit Number of Words in the Component Attribute

Limit Number of Component Words Without Limit Applied: 698 terms returned

Limit Number of Component Words With Limit Applied: 64 terms returned Restriction appears on Toggle Bar (cannot toggle off)

Setting Search Limits Include Deprecated LOINC Codes with Returned Terms

Include Deprecated LOINC Terms Deprecated LOINC Terms appear as Strikethrough Text with a “Do Not” Symbol

Favor Property Restriction Favor one type of property over others

Order/Observation Restriction Order or Observation preference

LOINC Hierarchies – Class Tree Three top-level branches Tree Navigation Buttons

Before Class Restriction Large number of candidate terms Many terms have class of “CHAL”

Search Trees chal Restrict eligible tests to non-challenge chemistry tests only

Following Class Restriction Only Non-Challenge Chemistry tests returned

Improvements to Trees Continued reorganization to provide a Specimen hierarchy

Improved Hierarchy for Component Show the LOINC Codes associated with these components

Show Associated LOINCs Click on details for more information about the selection

Details Screen

Tree Export Tools Save tree data to a file Export tree data to Excel Send Tree data as email

RELMA - Tree Features All trees operate the same way Shows terms spelled out Can expand and collapse parts or all of tree. Tree is string searchable Search can be based on one or more branches of a tree with or without other criteria Tabs are marked with an asterisk if criteria selected

All trees operate the same way

Shows terms spelled out

Can expand and collapse parts or all of tree.

Tree is string searchable

Search can be based on one or more branches of a tree with or without other criteria

Tabs are marked with an asterisk if criteria selected

Navigating through the Mapping Process

Limit effort to one lab section at a time and focus expertise Chemistry and hematology will be easiest Consider prioritizing by frequency LOINC Mapping Tactics

Limit effort to one lab section at a time and focus expertise

Chemistry and hematology will be easiest

Consider prioritizing by frequency

Vreeman DJ, Finnell JT, Overhage JM. A Rationale for Parsimonious Laboratory Term Mapping by Frequency. AMIA Annu Symp Proc . 2007;:771-775.

More LOINC Mapping Tactics Try using method-less terms first Specific methods can be transmitted in: OBX–17 (v2.x), Observation.methodCode (v3.0) Examine local units or real results to verify correct properties Properties are rarely distinguishable in tests You don’t have to do it all at one sitting Use the “Unmapped” function to return where you left off With every release - Update previous mappings to identify deprecated terms

Try using method-less terms first

Specific methods can be transmitted in:

OBX–17 (v2.x), Observation.methodCode (v3.0)

Examine local units or real results to verify correct properties

Properties are rarely distinguishable in tests

You don’t have to do it all at one sitting

Use the “Unmapped” function to return where you left off

With every release - Update previous mappings to identify deprecated terms

Common Mapping Issues Locally Defined Test Name Ambiguity Reuse of local test code “ Analyte-free” Local Test Names Incongruent Value sets (Scale ambiguity) Result vs. Interpretation Available LOINC Terms too Specific Available LOINC Terms too General Panel vs. Discrete Test Common in Microbiology

Locally Defined Test Name Ambiguity

Reuse of local test code

“ Analyte-free” Local Test Names

Incongruent Value sets (Scale ambiguity)

Result vs. Interpretation

Available LOINC Terms too Specific

Available LOINC Terms too General

Panel vs. Discrete Test

Common in Microbiology

Using your Mapped Terms Print results of LOINC Mapping Export to File

Print results of LOINC Mapping

Export to File

Intelligent Mapper Run in batch mode to find N- closest terms Can then use this output to do final mapping Again units are VERY important Can pick all limits available to regular mapping With Radiology can use CPT to help Intelligent Mapper Radiology References Vreeman DJ, McDonald CJ. A Comparison of Intelligent Mapper and Document Similarity Scores for Mapping Local Radiology Terms to LOINC. Proc AMIA Symp. 2006;809-813. Vreeman DJ and McDonald CJ. Automated Mapping of Local Radiology Terms to LOINC. Proc AMIA Symp. 2005;769-773.

Run in batch mode to find N- closest terms

Can then use this output to do final mapping

Again units are VERY important

Can pick all limits available to regular mapping

With Radiology can use CPT to help

Intelligent Mapper Radiology References

Vreeman DJ, McDonald CJ. A Comparison of Intelligent Mapper and Document Similarity Scores for Mapping Local Radiology Terms to LOINC. Proc AMIA Symp. 2006;809-813.

Vreeman DJ and McDonald CJ. Automated Mapping of Local Radiology Terms to LOINC. Proc AMIA Symp. 2005;769-773.

Picking the Intelligent Mapper

Intelligent Mapper Start Screen

Intelligent Mapper - Mapping Screen

Exporting Mapped Terms Export Terms to Delimited File

Export Terms Dialog Check Fields you wish to export Select Delimiter

Output File © 2006 Regenstrief Institute & James Case Mapped terms include LOINC code

Other RELMA ® Features

Panels, Forms and Survey Review New Menu Function: Review Panels, Forms and Surveys

Government Forms Form Name Component LOINC terms

Laboratory panels Expand category to see panels Double-click panel name to see components

Expanded Panel Detail

Expanded Clinical Panel © 2006 Regenstrief Institute & James Case

Clinical Panels New Feature: Find where a term has been used

LOINC Term Usage 35089-2, 35090-0

LOINC Term Panel Usage

Other Survey Instruments

HIPAA Attachments View Various HIPAA Attachments

HIPPA Attachment Tree

Expanded Attachment Data

Can’t find the term you want?

Search Hints and Tips Keywords with zero frequency are ignored May need to rephrase – use synonym Some causes for no returned terms Too many keywords in search – uncheck some Limits applied that don’t make sense E.g. Method-less tests plus Method tree set to EIA Did not find and revise words not in RELMA Local units not in RELMA Units are GREAT discriminators You may have tests that need to be added to LOINC

Keywords with zero frequency are ignored

May need to rephrase – use synonym

Some causes for no returned terms

Too many keywords in search – uncheck some

Limits applied that don’t make sense

E.g. Method-less tests plus Method tree set to EIA

Did not find and revise words not in RELMA

Local units not in RELMA

Units are GREAT discriminators

You may have tests that need to be added to LOINC

Proposing New LOINC Terms

Proposing New Terms Make sure the term is really needed Think of other names for the same concept Avoid detailed methods or localizations Is the distinction really important? Supply sufficient annotation to justify the new term Package inserts, sample reports (email to Regenstrief) The more the better! Construct new terms within RELMA File>Propose a LOINC Linked to the “trees” to allow browsing Can review and submit them to LOINC from within RELMA

Make sure the term is really needed

Think of other names for the same concept

Avoid detailed methods or localizations

Is the distinction really important?

Supply sufficient annotation to justify the new term

Package inserts, sample reports (email to Regenstrief)

The more the better!

Construct new terms within RELMA

File>Propose a LOINC

Linked to the “trees” to allow browsing

Can review and submit them to LOINC from within RELMA

Propose new LOINC terms Select Propose a LOINC

Propose new LOINC terms Navigate through proposed terms Create new term, save current term or Exit form

Preview Proposed Terms Review your new terms

Preview Proposed Terms All proposed terms fully editable Choose whether to send or postpone X X X X

Use Existing Terms as Template Select your closest match Click the Propose Term button

Use Existing Terms as Template All required components filled in

Using LOINC and SNOMED Together

Where do LOINC and SNOMED Fit? LOINC represents the question: Is there any Botulism toxin in my specimen? (33708-9) Organisms identified in specimen? (634-6) SNOMED represents the answer: Negative (SCTID 260385009) E. coli O157:H7 (SCTID 103429008) Remember:

LOINC represents the question:

Is there any Botulism toxin in my specimen? (33708-9)

Organisms identified in specimen? (634-6)

SNOMED represents the answer:

Negative (SCTID 260385009)

E. coli O157:H7 (SCTID 103429008)

Where do LOINC and SNOMED Fit? In an HL7 message, LOINC may be used: In OBR-4 (Universal Service Identifier) In OBX-3 (Observation Identifier SNOMED may be used: In OBX-5 (where nominal values are needed) Almost anyplace else in an HL7 message where coded values are needed

In an HL7 message, LOINC may be used:

In OBR-4 (Universal Service Identifier)

In OBX-3 (Observation Identifier

SNOMED may be used:

In OBX-5 (where nominal values are needed)

Almost anyplace else in an HL7 message where coded values are needed

OBX: With a Coded Value OBX||CE|6609-2^Listeria ID^LN||36094007^L. monocytogenes^SCT The code is from SNOMED The code is from LOINC OBX-5: Data A code for L. monocytogenes A code that identifies the data type in OBX-5 as a coded element OBX-3:A code that identifies the data in OBX-5 (Listeria culture)

HL7 2.5 Example (ER-7 format) MSH|^~&| LABGL1|| DMCRES|| 199812300100|| ORU^ R01| LABGL1199510221838581| P| 2.3||| NE| NE PID||| 6910828^ Y^ C8|| Newman^ Alfred^ E|| 19720812| M|| W| 25 Centscheap Ave^^ Whatmeworry^ UT^ 85201^^ P||( 555) 777- 6666|( 444) 677- 7777|| M|| 773789090 OBR|| 110801^ LABGL| 387209373^ DMCRES|634-6^Bacteria XXX Aerobe Cult ^LN||| 199812292128||||||||Stool| IN2973^ Schadow^ Gunther^^^^ MD^ UPIN|||||||||||||||| CA20837^ Spinosa^ John^^^^ MD^ UPIN OBX|| CE| 634-6^Bacteria XXX Aerobe Cult^ LN||50136005^Salmonella typhimurium ^SCT|||||| F||| 199812292128|| CA20837 OBR|| 110801^ LABGL| 387209373^ DMCRES| 29567-9^Bacterial Susceptibility Panel^ LN||| 199812300934||||||||Bacterial isolate| IN2973^ Schadow^ Gunther^^^^ MD^ UPIN||||||||| Salmonella typhimurium ||||||| CA20837^ Spinosa^ John^^^^ MD^ UPIN OBX|| CE|23631-5^Trimethoprim/Sulfasoxazole^ LN||264841006^Intermediate ^SCT|||||| F||| 199812300934|| CA20837 OBX|| CE|18967-7^Penicillin^LN||30714006^Resistant^SCT|||||| F||| 199812300934|| CA20837 OBX|| CE|18928-2^Gentamicin^ LN||131196009^Susceptible^SCT|||||| F||| 199812300934|| CA20837 First Order OBR (Culture and ID) First Result Value (SNOMED) First Result Description OBX (LOINC) Second Order OBR (Susceptibility)

MSH|^~&| LABGL1|| DMCRES|| 199812300100|| ORU^ R01| LABGL1199510221838581| P| 2.3||| NE| NE

PID||| 6910828^ Y^ C8|| Newman^ Alfred^ E|| 19720812| M|| W| 25 Centscheap Ave^^

Whatmeworry^ UT^ 85201^^ P||( 555) 777- 6666|( 444) 677- 7777|| M|| 773789090

OBR|| 110801^ LABGL| 387209373^ DMCRES|634-6^Bacteria XXX Aerobe Cult

^LN||| 199812292128||||||||Stool| IN2973^ Schadow^ Gunther^^^^ MD^ UPIN|||||||||||||||| CA20837^ Spinosa^ John^^^^ MD^ UPIN

OBX|| CE| 634-6^Bacteria XXX Aerobe Cult^ LN||50136005^Salmonella typhimurium ^SCT|||||| F||| 199812292128|| CA20837

OBR|| 110801^ LABGL| 387209373^ DMCRES| 29567-9^Bacterial Susceptibility Panel^ LN||| 199812300934||||||||Bacterial isolate| IN2973^ Schadow^ Gunther^^^^ MD^ UPIN||||||||| Salmonella typhimurium ||||||| CA20837^ Spinosa^ John^^^^ MD^ UPIN

OBX|| CE|23631-5^Trimethoprim/Sulfasoxazole^ LN||264841006^Intermediate ^SCT|||||| F||| 199812300934|| CA20837

OBX|| CE|18967-7^Penicillin^LN||30714006^Resistant^SCT|||||| F||| 199812300934|| CA20837

OBX|| CE|18928-2^Gentamicin^ LN||131196009^Susceptible^SCT|||||| F||| 199812300934|| CA20837

Q and A Session

Exercise Map your own data

Need RELMA? http://tinyurl.com/2vrms7