Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Clinical Element Models (CEMs) SHARP F2F Meeting Mayo Clinic June 21, 2010 Stanley M Huff, MD #1 A Simple Model #2 Use of detailed clinical models in SHARP • Guide for data normalization widgets • Target for structured output from NLP • Logical structure for data payload in NHIN Connect services • Reference for data that participates in the phenotype logic and queries #3 Model Classes Created • Patient, Employee, Provider, Organization, ContactParty, PatientContact (visit), ServiceDeliveryLocation, AdmitDiagnosis • HealthIssue (Problem), Allergy, Intolerance, Document • Order – OrderLab, OrderLabMicro, OrderBloodProduct – OrderMedAmb, OrderMedCont, OrderMedInt, OrderMedPCA, OrderMedReg – OrderNutrition, OrderRadiology, OrderNursing, OrderRepiratory, OrderTherapies • LabObs, MicroLabObs, Assert, Eval, Meas, Proc • Qualifiers, Modifiers (Subject), Attributions, Panels #4 Model Subtypes Created • Number of models created - 4384 – Laboratory models – 2933 – Evaluations – 210 – Measurements – 353 – Assertions – 143 – Procedures – 87 – Qualifiers, Modifiers, and Components • Statuses – 26 • Date/times – 27 • Others – 400+ – Panels – 79 #5 Access to the models • Send me an email and I will send you a zip file –[email protected] • Web browser –www.clinicalelement.com –Works best with Mozilla Firefox browser #6 What if there is no model? Site #1 Dry Weight: 70 kg Site #2 Weight: 70 kg Dry Wet Ideal #7 Relational database implications Patient Identifier Date and Time Observation Type Observation Value Units 123456789 7/4/2005 Dry Weight 70 kg 123456789 7/19/2005 Current Weight 73 kg Patient Identifier Date and Time Observation Type Weight type Observation Value Units 123456789 7/4/2005 Weight Dry 70 kg 123456789 7/19/2005 Weight Current 73 kg How would you calculate the desired weight loss during the hospital stay? #8 Model Centered Data Representation SNOMED LOINC FDB RxNorm ICD-10 CPT SNOMED LOINC FDB RxNorm ICD-10 CPT Context Specific Mapping Tables Internal Terminology (ECIDS) Models Models and Concepts ECIS Thesaurus Mayo Thesaurus IH Thesaurus LexGrid Terminology Server #9 We assume that the model is used in association with a terminology server. # 10 Model and Terminology Model MedicationOrder ::= SET { drug Drug, dose Decimal, route DrugRoute, frequency DrugFrequency, startTime DateTime, endTime DateTime, orderedBy Clinician, orderNumber OrderNumber} Instance data MedicationOrder { drug PenVK, dose 250, route Oral, frequency Q6H, startTime 09/01/95 10:01, endTime 09/11/95 23:59, orderedBy Don Jones, M.D., orderNumber A234567 } If the medicationOrder.drug is_a “antibiotic” then notify the infection control officer. Concept Semantic Network Drugs Antibiotics Penicillins Pen VK Analgesics Cephalosporins Amoxicillin Cardiovascular Aminoglycosides Nafcillin # 12 Denormalized Semantic Network Drugs Drugs Drugs Antibiotics Antibiotics Antibiotics Penicillins Penicillins Penicillins has-child has-child has-child has-child has-child has-child has-child has-child has-child Antibiotics Analgesics Cardiovascular Penicillins Cephalosporins Aminoglycosides Pen VK Amoxicillin Nafcillin Drugs Drugs Drugs Drugs Drugs has-member has-member has-member has-member has-member Antibiotics Penicillins Pen VK Amoxicillin Nafcillin # 13 Mods and Quals of the Value Choice • Mods - Component CE’s which change the meaning of the Value Choice. • Quals - Component CE’s which give more information about the Value Choice. # 14 A Panel containing 2 Observations # 15 The use of Qualifiers # 16 The use of Modifiers # 17 XML Model with Term Binding The name of this model Binding to a single <cetype name="BloodPressurePanel" kind="panel"> “observable” concept <key code="BloodPressurePanel_KEY_ECID" /> <item name="systolicBloodPressureMeas" type="SystolicBloodPressureMeas" card="0-1" /> <item name="diastolicBloodPressureMeas" type="DiastolicBloodPressureMeas" card="0-1" /> <item name="meanArterialPressureMeas" type="MeanArterialPressureMeas" card="0-1" /> <qual name="methodDevice" type="MethodDevice" card="0-1" /> <qual name="bodyLocationPrecoord" type="BodyLocationPrecoord" card="0-1" /> <qual name="bodyPosition" type="BodyPosition" card="0-1" /> <qual name="relativeTemporalContext" type="RelativeTemporalContext" card="0-M" /> <qual name="patientPrecondition" type="PatientPrecondition" card="0-M" /> <mod name="subject" type="Subject" card="0-1" /> <att name="observed" type="Observed" card="0-1" /> <att name="reportedReceived" type="ReportedReceived" card="0-1" /> <att name="verified" type="Verified" card="0-1" /> … </cetype> # 18 Binding to a “domain” (value set) Path to the coded element <constraint path="qual.methodDevice.data.cwe.domain" value="BloodPressureMeasurementDevice_DOMAIN_ECID" /> The name of the terminology “domain” that the element is “bound” to <constraint path="qual.bodyLocationPrecoord.data.cwe.domain" value="BloodPressureBodyLocationPrecoord_DOMAIN_ECID" /> # 19 Compiler XML Template Java Class “In Memory” Form HTML CEML Source File CE Translator UML? openEHR Archetype? HL7 RIM Static Models? OWL? # 20 Decomposition Mapping Precoordinated Model (User Interface Model) SystolicBPRightArmSittingObs data SystolicBPRightArmSitting 138 mmHg Post coordinated Model (Storage Model) SystolicBP SystolicBPObs data 138 mmHg quals BodyLocation BodyLocation data Right Arm PatientPosition PatientPosition data Sitting # 21 How much data in a single record? • “Chest pain made worse by exercise” – Two events, but very close association – Normally would go into a single finding • “Ate a meal at a restaurant and 30 minutes later he felt nauseated, and then an hour later he began vomiting blood.” – Discrete events with known time and potential causal relationships – May need to be represented by multiple associated findings • Semantic links are used to represent relationships between distinct event instances # 22 Representation of Semantic Links InstanceId 1 (123) Nausea Relationship followed-by InstanceId 2 (987) Vomiting • Semantic links can also have certainty and attribution – Certainty – Attribution (who or what asserted the relationship, when, and why?) # 23 Area 6 Discussion and Planning # 24 Terminology Services Detailed CEMs) Model (including Using NHIN for transmitting data And the Internet for managing Content Internet Terminology, Models, Logic, NLP Semantics, etc. Normalized Data Instances NHIN Canonical EMR+ Normalized Data Instances Normalized Data Instances ETL EDW Staging NLP Widgets Normalization Widgets ETL + Rules EMR 1a Patient Billing Imaging EMR 2a Providr Claims Sched Lab Facility Rx …. Facility a NLP Widgets Analytic Health Repository Decision Support CER HTA QI CDS Normalization Widgets EMR 1b Patient Billing Imaging EMR 2b Providr Claims Sched Lab Facility Rx …. Facility b Discussion • Evaluation projects – Sharing data through NHIN Connect and/or NHIN Direct • What, who, when, where? – Comparison of data processed through SHARP to data in existing Mayo and Intermountain data trust, EDW, AHR • What, who, when, where? – Others? • Evaluation of NLP outputs and value? Focus on a specific domain: X-rays, operative notes, progress notes, sleep studies? • Questions – – – – What is the target set of normalization widgets that we want to build? Can we do the evaluations on de-identified data? Do we need patient consent to do the evaluations? Others? # 26