1 San Diego, Californie (USA)Mise en place de référentiels partagés dans une grande organisation, exemple du VA aux USA Ecole d’été a Corte Juillet, 2007 Omar Bouhaddou VA et EDS San Diego, Californie (USA)

2 Les besoins Développer un dossier longitudinal du patient a travers tous les sites de soins Etendre l’aide a la décision a l’ensemble des données du dossier Faciliter l’échange de donnée standardisée entre les partenaires de santé Adhérer aux régulations fédérales et a un mandat du Congrès Américain A complete clinical picture of the patient is key to quality, safety, and the cost-effectiveness of patient care. Thus, there is an obvious business need to develop ONE NATIONAL EHR across all patient visits at all VA sites. We need to extend computerized decision support to all patient data, whether it is collected locally or at other sites. Since 40% of veterans receive care elsewhere, we also need to be able to share data between the VA and other healthcare partners, in particular the Department of Defense. Finally, we need to comply with regulations (HIPAA), standards (HITSP), and last but not least comply with a mandate from the US congress.

3 Standardisation des référentiels de terminologieLa standardisation des référentiels de terminologie réfère a l’identification, l’adoption, l’implémentation, la vérification, la maintenance, et l’adhérence aux normes Une stratégie clé pour le dossier partagé et l’interopérabilité

4 Plan Le système de santé du VADéveloppement des terminologies standards Mise en place dans un système opérationnel Maintenance des terminologies standards Bénéfices de la standardisation Leçons tirées Conclusion I will start with one slide background on the VA healthcare system. Then, I will describe how the VA has been adapting or developing terminology standards, how it is implementing them in its current clinical information system through a simple strategy that we believe is generic and can be used by others, how it is maintaining these standards, and finally describe some significant benefits that were made possible through the investment in terminology standardization.

5 Plan Le système de santé du VADéveloppement des terminologies standards Mise en place dans un système opérationnel Maintenance des terminologies standards Bénéfices de la standardisation Leçons tirées Conclusion I will start with one slide background on the VA healthcare system. Then, I will describe how the VA has been adapting or developing terminology standards, how it is implementing them in its current clinical information system through a simple strategy that we believe is generic and can be used by others, how it is maintaining these standards, and finally describe some significant benefits that were made possible through the investment in terminology standardization.

6 Veterans Affairs (VA) Un des plus grands système médical des Etats-Unis 1300 sites de soin dont 154 hôpitaux, 7.4 millions de patients enregistres, 60 millions de visites en 2006, employés, internes+étudiants par an Leader en qualité e.g., les critères de qualité des soins pour diabétiques sont meilleurs au VA que dans les systèmes privés The VA is the largest medical system in the US with 1300 care sites and 7.4 million enrolled patients. By many accounts, It is a leader in care quality. Independent studies have shown that based on multiple measures of quality outcomes, the VA compares favorably to commercial managed care. It is a leader in information system technology. Its VistA clinical information system is broadly adopted by clinician users who rely on it instead of the paper record. For instance, nearly 94% of the medication orders are e-prescribed. The quality of care and the efficiency of the information technology used to support it and measure it go hand in hand, as noted by the independent studies. Leader en technologies de l’information 128 implémentation de VistA: e.g., 107 million de prescriptions par an dont 93.6% sont électroniques

7 Plan Le système de santé du VADéveloppement des terminologies standards Mise en place au sein d’un système opérationnel Maintien des terminologies standards Bénéfices de la standardisation Leçons tirées Perspectives I will start with one slide background on the VA healthcare system. Then, I will describe how the VA has been adapting or developing terminology standards, how it is implementing them in its current clinical information system through a simple strategy that we believe is generic and can be used by others, how it is maintaining these standards, and finally describe some significant benefits that were made possible through the investment in terminology standardization.

8 Les Référentiels au VA: AvantVistA est une collection de plus de 100 logiciels écrits en Mumps, e.g., la pharmacie, le laboratoire, la liste des diagnostiques, la radiologie, etc. Chaque logiciel maintien ses propres référentiels en terminologie a sa propre façon Le VA est divise en 20 régions et chacune a une copie localisée et isolée de VistA Une telle décentralisation a beaucoup aidé a l’adoption du système par les cliniciens au point ou le dossier papier a presque disparu

9 Les Référentiels au VA: Aujourd’huiAujourd’hui la stratégie est différente, centrée sur un dossier complet et partagé du patient, au delà des institutions ou des logiciels Une équipe centrale désigne et maintien les nouveaux référentiels. Elle met a jour de façon synchronisée les 128 instances de VistA Les échanges de données avec d’autres institutions sont facilites par l’utilisation de normes nationales et internationales

10 Signes Vitaux – Avant StandardisationThere is a saying within the VA that, if you have seen one VA system, you have seen one VA system. Terminology Standardization is about to change that. Here’s an example of an OLD Vital Signs input screen at one VA site. There were 128 flavors of Vital Signs screen like the one shown here.

11 Signes Vitaux – Apres StandardisationAfter standardization, all 128 VistA sites use the same list of Vitals Signs and Qualifiers. The standard list can only be updated from a central Terminology Server, and local additions and changes are prohibited.. The Vital Signs terminology standard is derived from the LOINC standard.

12 Réactions AllergiquesAvant (rappelez vous qu’il y a 128 sites de soin!) Site 1 Site 2 Site 3 Site 4 Epistaxis Nose bleed Nosebleeds epistaxis Erythema Redness of skin zzSkin redness BP 90/58 Euphoria, xxxx euphoria Excessive salivation Drooling ptyalism drolling Excessive tear production Watering eyes Epiphora Onions-tearing This example illustrates Allergy Reactions terms BEFORE and AFTER data standardization. AFTER standardization, one standard list is presented to all users. Old local terms are preserved as search terms. No free text entry is permitted. The allergy reaction standard list was derived from SNOMED CT and it has been proposed as a draft national standard for Allergy Reactions. Après VA Standard Term VUID Synonyms Epistaxis 58852 Nose bleed*Nosebleeds*Bleeding from nose Erythema 86753 Redness of skin*Skin redness Euphoria 85242 Excessive salivation 46677 Drooling*Ptyalism Excessive tear production 89509 Watering eyes*Epiphora*Tearing

13 Titres de document Titres crées localement et difficiles a déchiffrer:AUDIO/REEVAL/HA CHECK (T)(CI) CH-SARP NOTE CIH/STAR II CONSULT DME CLINIC (T) IC/ID/V MH-CWT PATIENT CHECKLIST (CH) NURSING PRRTP NOTE SOCIAL WORK BOMH DISPOSITION NOTE (O) (T) UR 67CD (T)(K) A third example in the Document Titles domain shows locally created titles that are difficult to decipher. A new standard for document titles standard was created in collaboration between LOINC and the VA and is now part of the latest LOINC release. 120,000 titres de document actifs et utilisés

14 Titres de document HEART FAILURE CLINIC FIRST VISIT NOTELocal Note Title HEART FAILURE CLINIC FIRST VISIT NOTE SUBJECT MATTER DOMAIN SETTING SERVICE DOCUMENT TYPE 2452 clear, succinct clinically relevant titles CARDIOLOGY OUTPATIENT INITIAL EVALUATION NOTE Standard Note Title

15 Approche basée domaineAnalyser les données locales des 128 VAs Rechercher les normes nationales Recommander un standard Etablir un plan d’implémentation Nettoyer les données locales Assurer l’adhésion au standard Maintenir le standard This slide summarizes the VHA approach to terminology standardization. It is a domain-based approach and a domain sequence plan has been established reflecting the clinical and business priorities of the enterprise. For each domain, there is a Domain Action Team that is responsible for the analysis of VHA existing data and the identification of applicable national standards. The DAT recommends a standard to the VA Data Standardization Oversight Board for approval, following a balloting process similar to that used by standards development organizations like HL7. For each domain, the DAT also defines a starting strategy: some domains are standardized time forward only, like Vital Signs (meaning only newly recorded data is affected and past data is not) and other domains, like Allergy, are standardized from the oldest data available on record.

16 Equipe Action Domaine Une équipe multidisciplinaire ArchitectesInformaticiens Utilisateurs Représentants des SDOs Experts de domaine Project manager This slide summarizes the VHA approach to terminology standardization. It is a domain-based approach and a domain sequence plan has been established reflecting the clinical and business priorities of the enterprise. For each domain, there is a Domain Action Team that is responsible for the analysis of VHA existing data and the identification of applicable national standards. The DAT recommends a standard to the VA Data Standardization Oversight Board for approval, following a balloting process similar to that used by standards development organizations like HL7. For each domain, the DAT also defines a starting strategy: some domains are standardized time forward only, like Vital Signs (meaning only newly recorded data is affected and past data is not) and other domains, like Allergy, are standardized from the oldest data available on record.

17 Les décisions importantesQuelles données du domaine standardiser? A partir de quelle date: futur ou passé? Si les « free-text » seront aussi convertis? Quel sera l’impact sur la qualité des soins et sur le dossier médical légal ? Si le standard sera implémenté directement ou indirectement (mapping)? This slide summarizes the VHA approach to terminology standardization. It is a domain-based approach and a domain sequence plan has been established reflecting the clinical and business priorities of the enterprise. For each domain, there is a Domain Action Team that is responsible for the analysis of VHA existing data and the identification of applicable national standards. The DAT recommends a standard to the VA Data Standardization Oversight Board for approval, following a balloting process similar to that used by standards development organizations like HL7. For each domain, the DAT also defines a starting strategy: some domains are standardized time forward only, like Vital Signs (meaning only newly recorded data is affected and past data is not) and other domains, like Allergy, are standardized from the oldest data available on record.

18 Exemples de décisions DOMAINE DEBUT STANDARD IMPLEMENTATIONSignes Vitaux Future seulement VA/LOINC Native Allergies Future + passe VA-DoD standard Titres de Document LOINC + VA extensions Mapping This slide summarizes the VHA approach to terminology standardization. It is a domain-based approach and a domain sequence plan has been established reflecting the clinical and business priorities of the enterprise. For each domain, there is a Domain Action Team that is responsible for the analysis of VHA existing data and the identification of applicable national standards. The DAT recommends a standard to the VA Data Standardization Oversight Board for approval, following a balloting process similar to that used by standards development organizations like HL7. For each domain, the DAT also defines a starting strategy: some domains are standardized time forward only, like Vital Signs (meaning only newly recorded data is affected and past data is not) and other domains, like Allergy, are standardized from the oldest data available on record.

19 Standardized VistA Files At-A-GlancePackage Files How Much? Standard Demographics 2 200 ALL VistA Files (key fields) HL7 Vitals 120.51 120.52 120.53 Point Forward Pharmacy Phase I 50.6 50.605 50.416 50.68 Non-inactive Medications RxNorm Allergy 120.8 120.86 120.85 UNIIs, UMLS, NDF-RT, SNOMED CT Lab Phase I (CH Subscript) 95.3 LOINC The next 4 slides show the files being standardized for each domain and the standard being used. NDF_RT – National Drug File Reference Terminology per CHI Standards LOINC - Logical Observation Identifiers, Names, and Codes HL7 - HL7 is not only a standard used in message format transmissions but it also includes data definitions that allow for the exchange of clinical data.

20 Standardized VistA Files At-A-GlancePackage Files How Much? Standard Lab Phase II (LDSI Files) 61.2 62 62.06 Point Forward SNOMED CT Orders 100 100.01 100.02 VistA Files (HDR required fields) Clinical Documents (TIU) (existing) 8925.1 8925.7 8925.5 8925.6 (new) 8926.2 8926.3 8926.4 8926.5 8926.6 Clinical LOINC Axes Lab Phase II – Tied to LEDI (July-December 07 timeframe) Nine Lab files will be standardized that support microbiology and pathology. Since LDSI (Lab Data Sharing Interchange) is the first project to use SNOMED CT, priority will be given to 4 files for the LDSI Project (Topography, Etiology, Collection Sample and Antimicrobial Susceptibility). (SNOMED CT may be used in HL7 messages which allows exchange of results of Lab tests between VA and DoD.) Topography (#61) Morphology (#61.1) Function (#61.3) Disease (#61.4) Procedure (#61.5) Occupation (#61.6) Etiology (#61.2) Collection sample (#62) Antimicrobial Susceptibility (#62.06) **Good news is SNOMED CT mapping is being completed in the 4 LDSI related files by Data Standardization and will be done through an automated process with minimal impact to sites, as opposed to the LOINC mapping. Orders – Phase I Clinical Documents (TIU) – mapping recently completed and data will be included in the HDR at a later date.

21 Standardized VistA Files At-A-GlancePackage Files How Much? Standard Immunization/ Skin Test 811.1 Point Forward CDC CVX Table/CPT VistA File (Skin Test) Encounters 45.6 45.82 VistA Files HL7 Problem List ICD 9  SNOMED CT CDC CVX Table – An HL7 external code set for vaccines that the CDC maintains. CPT codes are mapped to the corresponding CVX vaccine code for use in the HL7 data transmission standard by immunization registries. HL7 - HL7 is a standard used in message format transmissions. It also includes data definitions that allow for the exchange of clinical data.  HL7 specifies value sets for some data fields and for other data fields “user-defined” tables provide suggested values. STS standardization work is guided by the recommendations of the Consolidated Health Informatics (CHI) group. The CHI reviewed potential standards (such as HL7; X12N; ASTM E ; SNOMED; etc.) for all the domains and made recommendations for standards to be used for each domain. CHI recommended the HL7 standard be used for the Encounters Domain. The Encounters domain reviewed these HL7 defined value sets/tables for each of the files it standardized. If a HL7 specific value set was identified as a match for a particular Encounters file, then the DAT recommended it be adopted for use. Values in “user-defined” tables are only suggested by HL7 and other values may be specified by the user for use. The VA has created several “user-defined” tables to specify information needed for Ambulatory Care data reporting.

22 Standardized VistA Files At-A-GlancePackage Files How Much? Standard Pharmacy Phase II 50.64 50.607 50.606 50.67 50.95 50.608 50.609 56 51.2 Active Medications NDF, RxNorm Radiology 71 All LOINC, CPT code, SNOMED & NDF or NDF-RT 71.2 Point Forward VistA Files 71.6 On Hold Pharmacy Med Route 71.7 TBD 72 73.1 DICOM 75.2 78.1 SNOMED 78.3 79.2 Radiology is working on mapping strategies for files 71 ,72, 75.2 and 79.2.

23 Domaine allergies FREE TEXT CODED PCN Penicillin Sulfa drugsAce Inhibitor Meds ACE INHIBITORS ETHANOL ALCOHOL, ETHYL MONDAYS X This is a mapping file sample of all the possible categories of clean-up. This was kept in an Access Database. You can see here that PCN was cleaned up to the GENERIC DRUG file and given a name of Penicillin. You can see that was done similarly for the allergies mapped to other clean-up files. You will notice that free text terms such as LOVE were marked as entered in error. You will notice that the last entry NKA was also marked as entered in error but has a NAME of NKDA to activate some different programming logic. A Free text entry of “TYPE II DM” would not be included in file and would have been left as free text. However, the term was evaluated to determine that it would be free text. Structured Output is needed for the mapping file. Exact matches are required. Even an extra space would throw off the result.

24 Nombre de variantes distinctesVariations Nombre de variantes distinctes Occurrences totales Penicillin 1264 102292 Sulfa 904 48073 Aspirin 378 9974 Morphine 306 4278 Tuberculin, PPD 302 3161 Codeine 293 16695 Nitroglycerin 284 1599 Erythromycin 244 2484 Lisinopril 218 1131 Influenza 170 4357 Ibuprofen 165 3269 At the peak of the mapping variations there were 1,264 variations of Penicillin—this equaled more than 200,000 instances of the data with a sample of 600,000 free text entries. There were 904 variations of sulfa. This equaled more than 48,000 instances of data with a sample of 600,000 free text entries. The variations of the data began to decrease after this point but there were 24 mappings with more than 100 variations and the list here depicts some of those additional mappings.

25 Résultats par site This software has been released in May of Sites have 30 days to install the software. At the end of 30 days we will have the data from all sites but as of the time that this slide was pulled together we had 26 sites that had installed this software. With 26 sites having installed the software we have had the range from 28 to 100% of free text being changed to coded values. With 26 sites reporting we had approximately 85% of free text converted to a coded allergy record.

26 Domaine titre de documentLocal Note Title HEART FAILURE CLINIC FIRST VISIT NOTE SUBJECT MATTER DOMAIN SETTING SERVICE DOCUMENT TYPE 2452 clear, succinct clinically relevant titles CARDIOLOGY OUTPATIENT INITIAL EVALUATION NOTE Standard Note Title

27 Domaine titre de documentThere were 156,000 local note titles before and now there are 2,452 standard note titles. T The structure of the titles now allows the content to be clear. If you are a clinician looking for a Cardiology note that note is now easier to find regardless of which medical center you are working at. It is easier because you no longer need to know the name of the local titles to identify the relevant content in a note.

28 Degré de standardisation et taille du domaineDegree of Standardization Size of Domain (normalized)

29 Plan Le système de santé du VADéveloppement des terminologies standards Mise en place au sein d’un système opérationnel Maintenance des terminologies standards Bénéfices de la standardisation Leçons tirées Conclusion I will start with one slide background on the VA healthcare system. Then, I will describe how the VA has been adapting or developing terminology standards, how it is implementing them in its current clinical information system through a simple strategy that we believe is generic and can be used by others, how it is maintaining these standards, and finally describe some significant benefits that were made possible through the investment in terminology standardization.

30 Implémentation – 1 / 4 Equipe d’implémentationNettoyage des données antécédentes (si nécessaire) Matches reçoivent VHA Unique IDs (VUIDs) et un statu “actif” Non-matches aussi reçoivent VUIDs, mais un statu “inactif” Les termes similaires sur plusieurs sites de soins reçoivent les mêmes VUIDs Les fichiers locaux sont verrouillés et la gestion des changements est assurée par l’équipe centrale Once the DAT has approved a new standard, the Domain Implementation Team is responsible for the implementation of the new standard. The standard is implemented at the point of care as opposed to a back office mapping method. First, legacy data is cleaned up if necessary. Then, local terms that match concepts in the standards list receive VHA Unique Identifiers or VUIDs and an “active” status. This means that these terms can be selected by a user. Local terms that do not match any standard concept also receive VUIDs, but the status is set to “Inactive.” These terms are non-selectable post standardization. Similar terms across 1 or more sites receive identical VUIDs. Finally, local files are locked, and future changes to these files are managed through a centralized Terminology Server.

31 Implémentation – 2 / 4 Etape 1: Les termes standards sont identifiés et un VUID leur est assigne Allergen Name VUID Activation Status STRAWBERRIES BLACKBERRIES Active Status Effective Date 1 01/20/2005 Step 1 in the implementation process is to create the standard list and assign a VUID to each term, as illustrated in this simple example. Note that the activation status for these terms is set to ‘1’ and that each status has also an effective date. This way, a term can be active at one point in time, then inactive at another point, and then active again at a third point. Active Status Effective Date 1 01/20/2005

32 Implémentation – 3 / 4 Etape 2: Le fichier local est nettoyé pour refléter la liste standard Local ID Term 1 STRAWBERRIES 2 strawbery 3 Blackberries Local ID Term VUID Status Effective Date 1 STRAWBERRIES 1/20/2005 2 strawberries 2/15/2005 3 BLACKBERRIES Step 2 of the process: At each VA site, the local file is cleaned up to reflect the standard list. New fields are added (marked in blue). Matches receive an active status and non-matches receive an inactive status. Since in the Allergy domain, the domain action team decision was to standardize all data including legacy data, all local terms receive a VUID.

33 Implémentation – 4 / 4 Etape 3: les logiciels d’accès sont ‘filtrés’Application Filter: active terms only Local ID Term VUID Status Effective Date 1 STRAWBERRIES 1/20/2005 2 strawberries 2/15/2005 3 BLACKBERRIES Step 3 and final of the process: The existing access routines to the terminology files are changed to filter in only active terms. In other words, after the standard is implemented, only the active terms will show up on the user pick list. The applications themselves have not changed. The data that is stored in the local patient record also remain unchanged. Patient File (local ID)

34 Modélisation VHA Terminology The cloned concepts LOINCtake on a new identity as VHA Terminology concepts. SNOMED CT Iraqi Freedom Syndrome 989 UMLS Standardization would not be possible without a central Terminology Services environment. The first component of this environment is the Terminology Modeling or Authoring environment. In one section, we import and represent the SDO code sets such as LOINC, SNOMED CT, etc. These are our reference terminologies. On the other section, we clone from these national standards all concepts that our clinical applications need. This is our interface terminology. When a concept is missing from the national standard (like the “Iraqi Freedom Syndrome,” for example) we create it within the VHA standard and then submit it for addition to the appropriate national standard. ICD-9-CM VHA Terminology CPT 4

35 Modélisation The terminology structure of the SDO code sets or the VHA Terminology space is the same. It is the classic concept-based representation where a concept or unit of thought is associated with 1 or more designations or expressions, has defining properties, and is related to other concepts through many relationships. Concepts are grouped into subsets to facilitate the development of pick lists for given applications. We evolved this model from studying several operational systems including Kaiser, IHC, SNOMED, NLM …

36 Exemple Relationship has child has child Designation Property 4021641Vaccines Toxoids Relationship has child has child VA Class Type: Major Immunilogical Agents Designation Property NDFRT Level: VA Class IM000 Drug Class Antihistamines Antimicrobials Immunological Agents Vitamins Subset

37 Déploiement des terminologiesTerminology Modeling Standards Terminology Deployment HL7 National Interface Engine Local HL7 VistA 2 VistA 1 After the terminology is modeled in the Enterprise Terminology Modeling environment, it is ready to be deployed to the field. The Terminology Deployment Server tracks and automatically captures all changes that are made in the modeling environment and packages them into subsets that are ready for deployment. These subsets are deployed as HL7 messaging to 1 or many or all 128 VistA instances. An enterprise Interface Engine routes the HL7 messages to their destinations. At each VistA site, a core application, called the Master File Server, receives the messages and converts the information they contain into updates to MUMPS terminology files using the VUID as the key index. This HL7/MFS approach provides a rapid mechanism for updating all VA sites with the same terminology content. MFS Allergy Application Allergy Terms User Vitals Terms Vitals Application

38 Plan Le système de santé du VADéveloppement des terminologies standards Mise en place au sein d’un système opérationnel Maintenance des terminologies standards Bénéfices de la standardisation Leçons tirées Conclusion

39 Maintenance La médecine n’est pas statique … les standards non plus.Une fois un domaine standardisé, les révisions locales sont interdites, plus de “free-text” Un processus appelé “New Term Rapid Turnaround” ou NTRT permet aux sites de demander des changements We’re standardized, now what? Healthcare is not static … neither are standards. After a domain has been standardized, free text and local additions to standardized reference files are no longer permitted. This makes it necessary to implement a process by which clinicians and others can request additions/changes to the enterprise standard. We refer to this process as New Term Rapid Turnaround or NTRT.

40 Maintenance NTRT web site Terminology Modeling Deployment ServerStandards Deployment Server HL7 NTRT web site National Vitria IE Local VistA 2 VistA 1 HL7 If clinician users can’t find the term they need to document a patient observation, they can access the New Term Rapid Turnaround or NTRT Web site to request a change to the standard. They fill out a form with details about the term needed. Each domain has a specific form. Then, the request is automatically forwarded to the domain data stewards who process it, provide feedback to the user requestor, and finally deploy the new term to all VA sites. Urgent updates are processed in a matter of hours and regular updates are done in a matter of days. MFS Allergy Application Allergy Terms User Vitals Terms Vitals Application

41 NTRT: allergie au ‘durian’For example, a provider wants to document that his patient is allergic to “durian”, which is a big, green thorny fruit from South East Asia. He types “durian” into the search box and clicks “search.”

42 Nouvelle allergie au ‘durian’Cannot find term: Please try your search again (try using only the first 3 letters) or go to Tools Menu to Request New Term Something similar to this dialogue box would pop up. The user is encouraged to search again or to request a new term by selecting “Request New Term” on the Tools Menu.

43 Lien au NTRT Request New TermWhen he selects “Request New Term”, he will be linked to the NTRT website.

44 Site web NTRT This web form will automatically authenticate the user and be populated with the term for which he searched. The user would then specify certain aspects of the term. In this case, he would indicate that durian is a “reactant” of type “food.” The user should then provide any clarifying information about this term so that ERT can create it as specified. This information might be related names or synonyms along with any comments that clarify the meaning or usage of the term. The user would then submit the request to ERT.

45 Traitement de la requête NTRTUne fois la requête soumise, le message suivant apparait sur l’écran: Thank You Your request has been submitted. Once it has been processed (48-72 hours), you will receive a CPRS GUI notification as a reminder to resume documentation. Note: The allergy has not been documented in the patient’s record. If it is clinically important please include it in a text note. A message similar to this one will then be displayed to the user to explain that he will receive feedback on the request within 72 hours. It will also notify him that nothing was documented in the patient’s record at this time and if it is clinically important that it should be documented in a text note.

46 Notification The requestor will receive feedback as a CPRS GUI notification. EIGHT, INPATIENT Moderate New Term “Durian” Created

47 Signes vitaux examplesRefusés Intraoccular Pressure, Pain : nécessitent un changement du logiciel Le terme demandé existe déjà Acceptés Laryngeal Mask Airway and Reservoir Cannula ajoutés comme adjectifs de Pulse Oximetry/O2 Saturation Temporal ajouté a Temperature Hip ajouté a Circumference/Girth

48 Signes vitaux: sommaire d’activitésDenied entries have not been accepted because the content represents an enhancement request, a duplicate entry or an established business rule for the domain is in conflict with the request. The domain specific review teams review these requests and make a decision about the best recommendation possible for the clinician to record this allergy record. Accepted entries are ones that the Domain specific review groups agrees will add value to the standard. The accepted terms are deployed on the next weeks deployment.

49 Allergies: sommaire d’activitésDenied entries have not been accepted because the entry belongs in the National Drug File, does not fit into the established business rules for the content needed for the standard file, there is already a duplicate request,or there is an equivalent term in the standard that can be used for recording this allergy. The domain specific review teams review these requests and make a decision about the best recommendation possible for the clinician to record this allergy record. Accepted entries are ones that the Domain specific review groups agrees will add value to the standard. The accepted terms are deployed on the next weeks deployment.

50 TIU Examples Rejetés Acceptés Ophthalmology Consent Pharmacist NoteAsthma Outpatient Consult VHA Transplant Referral Acceptés Rescinded DNR Polytrauma Nursing Note Radiology Preprocedure Note Skin Test Note Ophthalmology Consent (only 5 total consents in standard, should use Consent or Procedure Consent instead) Pharmacist Note (Pharmacist not appropriate as a Role unless Pharmacy cannot be used as a Subject Matter Domain) Asthma Outpatient Consult (we want to avoid using specific diagnoses within titles as much as possible) VHA Transplant Referral (no way to fit gracefully into standard currently)

51 Titres: sommaire d’activitésClinical Documents (Aka TIU) Started 10/16/2006. The volume of Clinical Document requests has been the largest volume of NTRT requests received to date with a total of 723 submitted requests for TIU in only 7 months time. You will notice the note on the graph above indicated that this was the approximate date for mapping compliance at the sites. You will notice that the numbers of new requests have tapered off significantly after this milestone was reached according to current counts. Mapping Compliance Date

52 Plan Le système de santé du VADéveloppement des terminologies standards Mise en place au sein d’un système opérationnel Maintenance des terminologies standards Bénéfices de la standardisation Leçons tirées Conclusion I will start with one slide background on the VA healthcare system. Then, I will describe how the VA has been adapting or developing terminology standards, how it is implementing them in its current clinical information system through a simple strategy that we believe is generic and can be used by others, how it is maintaining these standards, and finally describe some significant benefits that were made possible through the investment in terminology standardization.

53 Bénéfices de la standardisationUn dossier complet du patient Une aide a la décision élargie L’interopérabilité des données La sante publique et la bio-surveillance sont facilitées La qualité, la sécurité, le cout, et l’efficacité du processus de soin sont améliorés

54 Dossier complet du patientStandards & Terminology Services National Reference Terms (HL7 Engine) VistA 1 VistA 2 Patient Data (local ID) Patient Data (local ID) Local Patient Data (HL7 Engine) Some benefits are realized immediately after terminology is standardized. For example, the VA has created a national patient data base or Health Data Repository aggregating data from all its 128 VistA sites. Local data from each of the 128 VA care sites is pushed to the national database as HL7 messages. The triggers that push the data from the local site to the national database associate to each data element the corresponding VUID from the terminology file. As a result, whereas patient data is identified with local IDs in the local database, it is identified with the national VUIDs in the HDR, thus making the data comparable and sharable. Department of Defense HDR (VUID) National

55 Interopérabilité : VA – DoDDecision Support Decision Support VA systems DoD systems HDR CHDR CHDR CDR This slide shows the flow of information in the data exchange program currently running between the VA and the Department of Defense which was made possible post-standardization. Starting at a specific VistA site, a clinician prescribes “Abacavir sulfate 300mg Tab” This order is recorded locally with a the local ID or IEN=76. Local order checks are run against this prescription, if applicable. Then, a trigger pushes this data element to the national HDR and, at the same time, associates to it the national VUID= , which is available in the new field added to the local terminology file. Once in HDR the data is available to all VA sites in case the same patient presents for a visit at another VA site. If the patient is an Active Dual Consumer (meaning the patient receives care at both a VA and a DoD facility), the ABACAVIR prescription is also pushed to CHDR, which is the gateway for computable data exchange between the VA and the DoD. The VA CHDR first structures the information from HDR into a commonly agreed upon HL7 message structure. Then, the key coded elements of the message are translated into a common mediation terminology. The two agencies agreed to exchange Pharmacy data through RxNorm, Drug Allergens through UMLS, and Allergy Reactions through SNOMED CT. After translation, the VA CHDR sends the message across the firewall to the DoD CHDR. The same process then takes place at the DoD side in reverse. First, the DoD CHDR translates the key encoded elements from RxCUI to its own national ID, the NCID. Then, it restructures the information for storage into its national Clinical Data Repository or CDR. Finally, when the patient shows up at a DoD site, all data, including VA data, is pulled from the CDR and made available to the clinician, just as if it was recorded locally during a previous visit. Mediation Terminologies: RxNorm, UMLS, SNOMED CT, LOINC

56 Aide a la décision élargiePhase I (aujourd’hui): « order checks » sur toutes les données de tous les VA et originaires du DoD Drug-drug interactions Drug-allergy checks Duplicative drug therapy checks Phase II (futur): laboratoire, liste des diagnostiques

57 Plan Le système de santé du VADéveloppement des terminologies standards Mise en place au sein d’un système opérationnel Maintenance des terminologies standards Bénéfices de la standardisation Leçons tirées Conclusion I will start with one slide background on the VA healthcare system. Then, I will describe how the VA has been adapting or developing terminology standards, how it is implementing them in its current clinical information system through a simple strategy that we believe is generic and can be used by others, how it is maintaining these standards, and finally describe some significant benefits that were made possible through the investment in terminology standardization.

58 Leçons tirées Les données appartiennent a l’institution et non a un logiciel (e.g., Allergies) ou a la clinique d’origine Une maintenance centralisée est nécessaire et demande des outils adaptés (e.g., serveur de terminologies) Une coordination étroite est nécessaire entre développeurs, SDOs, terminologistes, utilisateurs, etc. - le travail des SDO est essentiel! Standardisation de la terminologie exige un soutien complet de l’organisation, pas seulement ITC Le soutien du gouvernement est irremplaçable (e.g., des objectifs concrets, un coordinateur pour faciliter les collaborations, et l’élimination des barrières cout – achat de la License SNOMED CT)

59 Conclusion VA a démontré un model générique qui permet l’implémentation de référentiels standards avec un minimum d’impact sur les systèmes opérationnels Des bénéfices importants peuvent être réalisés y compris Un dossier complet du patient Une aide a la décision sur l’ensemble des données du patient Une interopérabilité sémantique est possible In conclusion, hopefully you’ll leave today optimistic about terminology standardization. As demonstrated by the VA, standardization is possible today without having to re-write your existing systems. The VA terminology standards are based on national standard recommended by HITSP and CHI and are public domain and freely available. It is hard work but the payoffs are worth it. Once terminology standardization is in place, then the institution can build a consolidated EHR, can engage into semantic interoperability with other healthcare partners, and can apply decision support to all the patient data, where ever it was collected, thus providing a more complete clinical picture of the patient and resulting in efficient quality healthcare.

60 fin Thank you.