Building Intelligent Clinical Decision Support Systems .
Building Intelligent Clinical DecisionSupport Systems Using a LargeKnowledge BaseMingu Lee, Ph.D.Vice President, Samsung SDS AmericaDongwook Shin, Ph.D.President, TranformationCopyright 2011 Samsung SDS America, Inc. All rights reserved
Offerings DetailSamsung Group OverviewSamsung is the 19th most valuable brand in the world. Samsung Group’s total revenues of USD 173billion corresponds to rank 14th on Fortune Global 500.EmployeesSalesRevenueBrand ValueGlobalPresence277,000 around the globeUSD 173 BillionUSD 17.5 Billion (Global Ranking 19th)Over 60 CountriesRankGlobalRankingCompanyRevenues( million)12General Electric183,20713China National Petroleum181,123Samsung t Global Brands 2009, Interbrand)Sony76,945(Fortune Global 500, 2009)1
Offerings DetailSamsung AffiliatesSamsung is comprised of companies that are setting new standards in a wide range of businesses. Ofits 28 affiliated companies, Samsung Electronics, Samsung Life Insurance and Samsung C&T are listedin the Fortune Global 500, 2009.(Samsung Profile, ng C&T(Construction &Trading)Samsung LifeInsurance Ranked 40th, Fortune Global 500 USD 96 billion in revenues No.1 market share in the world: 164,600 employeesDRAM, SRAM, Flash Memory 179 offices in 61 countries No.2 in Mobile Phone World #1 Shipbuilder(based on order book) USD 9.7 billion inrevenues 12,500 employees 13 branch offices in 11countries Constructed world's talleststructure (Burj Khalifa, Dubai) USD 10.7 billion in revenues 7,200 employees 105 offices in 44 countries Largest insurancecompany in Korea USD 22.3 billion sales 6,377 employees 12 offices in 8 countries2
Offerings DetailSamsung SDS OverviewSamsung SDS is a global ICT service provider with over 11,000 employees and USD 3.2 billion inrevenues. With an average annual growth rate of 10% over the last four years, it is ranked as the 3rdlargest IT service company in Asia-Pacific region.Employees 11,678 Employees around the globe(1,824 Overseas)Revenue Growth(USD billion)SalesRevenue USD 3.2 BillionGlobalPresence 12 Offices in 12 Countries3.82 No. 1 Korean IT services providerwith the largest domestic marketshare (14.7%)Recognition 3rd largest IT services provider inAsia-Pacific region(Market Share: IT Services, Asia/Pacific, 2008, Gartner)2005 20062007 2008 2009* Includes 2010 projection* Numbers are as of May, 201032010(Projection)
Offerings DetailService Offering OverviewSmart Infrastructure Engineering(SIE)e-Government Citizen Services National Services Government Services Transportation ConstructionEnterprise Application Services Business and ICT Consulting ERP/SCM/PLM/MES/CRMMajorServiceOfferingsMobile Communication ServicesICT Infrastructure Data Center Network/Data Communication Cloud computingBusiness Process Outsourcing(BPO) Mobile Service Unified Communication Testing Services4
Offerings DetailSamsung SDS America Public Sector CapabilitiesMobile Mobile Groupware SAP Mobile BI Dashboard Oracle/Siebel Mobile CRM forPharmaceutical Sales Mobile Device Management Mobile Applications (Android OS)Portfolio Management Oracle Primavera EPA Enterprise Architecture Best Practices EPA Toxic Release Inventory White PaperERP DoD BTA ERP Samsung Electronics GlobalERP (Largest ERPimplementation in the world) Mobilizing ERP applicationsHealth IT HHS NIH Clinical DecisionSupport Systems (CDSS) –Natural Language ProcessingSystem IntegrationSecurity e-Procurement Systems(Vietnam and Costa Rica) Customs Services(Kazakhstan, DominicanRepublic, & Mongol) National Tax Service Networks(Sri Lanka) Kuwait Oil Facilities IntegratedSecurity System (US 300 M)5 EPA Portal and Web AccessManagement
Offerings DetailMobile EMR by Samsung SDSMobile Electronic Medical Record (EMR) provides healthcare professionalsinstant access to patient information at the point of care.Features Search patient recordsBenefitsBenefits Compact 7” tablet form factor View clinical charts/ analyticssignificantly enhances Make annotationsmobility View lab results and findings View and compare images(e.g., X-ray film, CT Scan,Ultrasound, etc.) Always connected – datamaintained securely on server Improve quality of servicethru access to more accurate,timely information Process automation thrumobility significantlyimproves data quality and Remotely manage devicesenhances efficiency of(i.e., remote lock/wipe of data,hospital staffblack-list/white-list apps)6
Offerings DetailProject Overview U.S. Department of Health and Human Services(HHS) National Institutes of Health (NIH) NationalLibrary of Medicine (NLM) Clinical Decision Support Systems (CDSS) withArtificial intelligence (AI) approach to discoverAdverse Drug Events accurately Performed by Samsung SDS America &Tranformation7
Offerings DetailPotential ScenarioWhile his doctor is out-of-town, an elderly asthma patient whohas developed severe knee pain sees another physician in hisdoctor’s office. An EMR provided documentation of the last visit,including recent laboratory results and a list of the patient’smedications. This information easily brought the doctor up to dateon the patient’s condition. The doctor entered an order formedicine for the knee pain into the system, printed out a (legible)prescription for the patient, and sent him on his way.Unfortunately, within 2 months, the patient wound up in theemergency room with a bleeding ulcer caused by interaction ofthe pain medicine with the patient’s asthma medicine.- Referred From Agency for Healthcare Research and Quality (AHRQ)8
Offerings DetailTechnical Components I. DailyMed: Drug information (package insert):– A website Operated by NLM to publish drug labels– Contents provided by the U.S. Food and Drug Administration (FDA) II. Natural Language Programming (NLP): Convert drug informationin DailyMed to the corresponding logical forms III. Unified Medical Language System (UMLS): A medical ontologydeveloped by NLM– MetaMap: A tool to find UMLS concepts from sentences IV. OpenCyc: A large knowledge base system with logical inferenceV. Google Health: Personal Health Record (PHR) systemVI. Adverse Drug Event Detection ServicesVII. Android Mobile App & UI9
Offerings DetailConceptual Architecture10
Offerings DetailConceptual ArchitectureOpenCycReview the accuracyConversionusing NLPAssertion in logicformDailyMedConversionAssertion frompatient infoInferenceUMLSGoogle Healthbased PHRAlerting to patient/physician11
Offerings DetailAdvil information from DailyMedADVIL (ibuprofen) capsule, liquid filled[Wyeth Consumer Healthcare]WARNINGSStomach bleeding warning:This product contains a nonsteroidal anti-inflammatory drug (NSAID), whichmay cause stomach bleeding. The chance is higher if you:are age 60 or olderhave had stomach ulcers or bleeding problemstake a blood thinning (anticoagulant) or steroid drugtake other drugs containing an NSAID [aspirin, ibuprofen, naproxen, or others]have 3 or more alcoholic drinks every day while using this producttake more or for a longer time than directedI. Drug Information Services12
Offerings DetailPre-Processing of DailyMed Using XML format of DailyMed SPL– Process “Warning”, “Boxed warning”, “Indication”, “Contraindication”,“Precaution” and “Adverse reaction” part– These parts can be identified by the corresponding LOINC code section code code "34071-1" codeSystem "2.16.840.1.113883.6.1" displayName "WARNINGSSECTION"/ text paragraph Stomach bleeding warning: This product contains an NSAID, which maycause severe stomach bleeding. The chance is higher if you /paragraph list item are age 60 or older /item item have had stomach ulcers or bleeding problems /item item take a blood thinning (anticoagulant) or steroid drug /item item take other drugs containing prescription or nonprescription NSAIDs[aspirin, ibuprofen, naproxen, or others] /item /list /text I. Drug Information Services13
Offerings DetailPre-Processing of DailyMed Extracting Text and conversion into MEDLINE format– Combine the text inside paragraph tag into one sentenceconnecting sentences with “and” paragraph Stomach bleeding warning: This product contains an NSAID, which may cause severestomach bleeding. The chance is higher if you /paragraph list item are age 60 or older /item item have had stomach ulcers or bleeding problems /item item take a blood thinning (anticoagulant) or steroid drug /item item take other drugs containing prescription or nonprescription NSAIDs [aspirin,ibuprofen, naproxen, or others] /item /list PMID- 1f01c10a-9434-91a4-2ee4-352315a6b610 34071-1 advil ibuprofenAB - Stomach bleeding warning: This product contains an NSAID, which may cause severe stomachbleeding and the chance is higher if you are age 60 or older;I. Drug Information Services14
Offerings DetailUnified Medical Language System UMLS is a medical ontology developed in NationalLibrary of Medicine– Having more than 1 million concepts– Concepts are connected with several relations: ISA,ASSOCIATED WITH, AFFECTS Relations defined in UMLS can connect conceptswith specific relations, otherwise unrelated.II. UMLS Ontology Services15
Offerings DetailUsing MetaMap1000 C1514468:product [Entity]966 C0332256:Containing [Functional Concept]1000 C0003211:Non-steroidal anti-inflammatory drug (Anti-Inflammatory Agents, Non-Steroidal)[Pharmacologic Substance]1000 C1524003:Cause (Science of Etiology) [Conceptual Entity]972 C0235325:Gastric bleeding (Gastric hemorrhage) [Disease or Syndrome]Phrase: "if“861 C0001792:age (Elderly) [Population Group]861 C0470240:60 [Quantitative Concept]Phrase: "older,“Phrase: "or“Phrase: "have“1000 C0038358:Stomach Ulcers (Gastric ulcer) [Anatomical Abnormality, Disease or Syndrome]Phrase: "or“1000 C1515187:Take [Activity]983 C0260264:Steroid drugs [Steroid]II. UMLS Ontology Services16
Offerings DetailDailyMed & MetaMap & NLPDESCRIPTIONADVAIR DISKUS 100/50, ADVAIR DISKUS 250/50, and ADVAIR DISKUS 500/50 arecombinations of fluticasone propionate and salmeterol xinafoate.One active component of ADVAIR DISKUS is fluticasone propionate, acorticosteroid having the chemical name S-(fluoromethyl) ndrosta-1,4-diene-17β-carbothioate, 17propionate and the following chemical structure:CONTAINS(“Advair”, “Fluticasone propionate”).ISA(“Fluticasone propionate”, “Corticosteroid”).III. NLP Services17
Offerings DetailCertainty Factor Expert knowledge is often expressed withcertainty factor.– Using words, “most probable”, “probably”, “probablynot”, “improbable”.– In DailyMed, certainty factors are expressed as:“chance is higher”, “more likely”, “low chance”, “lesslikely”. “chance is higher”, “more likely”- CAN CAUSE(A, B, “high”). “low chance”, “less likely” - CAN CAUSE(A,B, “low”). No certainty factor - CAN CAUSE(A,B, “middle”).III. NLP Services18
Offerings DetailUsing NLP with certainty factorSYNONYM(“Advil”, “Ibuprofen”).CONTAINS(“Advil”, “NSAID”).CAN CAUSE( Patient “Gastric bleeding”, “middle”) IFTAKES( Patient, “NSAID”).CAN CAUSE( Patient “Gastric bleeding”, “high”) IFTAKES( Patient, “NSAID”) and IS OLDER( Patient, 60).CAN CAUSE( Patient “Gastric bleeding”, “high”) IFTAKES( Patient, “NSAID”) and TAKES( Patient, “Steroid drug”).III. NLP Services19
Offerings DetailPutting things together for InferenceSYNONYM(“Advil”, “Ibuprofen”).CONTAINS(“Advil”, “NSAID”).CAN CAUSE( Patient “Gastric bleeding”, “middle”) IF TAKES( Patient, “NSAID”).CAN CAUSE( Patient “Gastric bleeding”, “high”) IF TAKES( Patient, “NSAID”) andage( Patient) 60.CAN CAUSE( Patient “Gastric bleeding”, “high”) IF TAKES( Patient, “NSAID”) andTAKES( Patient, “Steroid drug”).CONTAINS (“Advair”, “Fluticasone propionate”).ISA(“Fluticasone propionate”, “Corticosteroid”).ISA(“Corticosteroid”, “Steroid drugs”).Inference reasoningtriggered by prescriptionof AdvilAdvil can cause gastric bleeding for the patient who takes Advair and the risk ishigh.III. NLP Services20
Offerings DetailNLP Conversion Rules Rule for contraindication– Sample sentence: Aldactone is contraindicated for patients with anuria,acute renal insufficiency, significant impairment of renal excretory function,or hyperkalemia. Pattern:– [phsu] - [contraindicate] - [patient] - [with] -[dsyn or patf] Generate logical statements from this pattern– CONTRAINDICATE( X, phsu) IF HAS DISORDER( X, dysn or patf).– CONTRAINDICATE( X, "bronchospasm“, “middle”) IF HAS DISORDER ( X,"anuria") or or HAS DISORDER ( X, “hyperkalemia")– phsu : Pharmacological Substance– dsyn: Disease and Syndrome– patf: Pathologic FunctionIII. NLP Services21
Offerings DetailNLP Conversion Rules (Cont’d) Rule for allergy– Sample sentence: Ibuprofen may cause a severe allergic reaction, especiallyin people allergic to aspirin. Pattern:– [phsu or orch]-[cause] -[dsyn]-[in]-[people]-[allergic]-[to]-[phsu or orch] Generate logical statements from this pattern– CAN CAUSE( X, "allergic reaction“, “high”) IF TAKES( X, "ibuprofen") andALLERGIC TO( X, “aspirin”)certainty factorespeciallyIII. NLP Services22
Offerings DetailPhysician involvement in NLP NLP conversion is a complicated process andneeds expert’s intervention and feedback.– All the rules are reviewed by physicians who haveexpertise in medical informatics. The result of NLP conversion is also reviewed byexperts and thoroughly tested withcomprehensive test set.III. NLP Services23
Offerings DetailMaintaining Microtheories Microtheory is a set of rules and knowledgetreated independently from one another inOpenCyc.– Each microtheory can have knowledge conflicting withone another.– DailyMed information is converted into rules andmaintained as a big microtheory.– Each patient information is translated in a separatemicrotheory and interact with DailyMed microtheory.IV. Knowledge Mgmt Services24
Offerings DetailGoogle Personal Health Record (PHR)Google Health will retire onJanuary 1st of 2012.V. PHR Services25
Offerings DetailDetailed Medical Information in Google PHRProfile IDDemographicsProblem ListMedication ListAllergy ListV. PHR Services26
Offerings DetailXML output from Google PHRProfile IDDemographicsProblem ListMedication ListAllergy ListLab Test ResultsProceduresV. PHR Services27
Offerings DetailInference flowCONTAINS(“Advil”, “Non-steroidal anti-inflammatory drug”).CAN CAUSE( Patient, “Gastric bleeding”, “middle”) IF TAKES( Patient,“NSAID”).CAN CAUSE( Patient, “Gastric bleeding”, “high”) IF TAKES( Patient, “NSAID”)and IS OLDER( Patient, 60).CAN CAUSE( Patient, “Gastric bleeding”, “high”) IF TAKES( Patient, “NSAID”)and TAKES( Patient, “Steroid drugs”).COMPONENT(“Advair”, “Fluticasone propionate”).ISA(“Fluticasone propionate”, “Corticosteroid”).ISA(“Corticosteroid”, “Steroid drugs”).TAKES(“John”, “Advair”).AGE(“John”, 62).TAKES(“John”, “Advil”).:- CAN CAUSE(“John”, X, Y).InferenceDoctor prescribes “Advil”CAN CAUSE(“John”, “Gastric bleeding”, “high”).John’s Medical RecordAge: 62Medication: AdvairVI. ADE Services28
Offerings DetailCurrent WorkflowGoogle HealthDownloadJohn’s dataProcessOpenCycJava Interface(step 2)Assert to OpenCyc:(# take # John # Advair)")(#take # John # Advil)AssertJohn’s data(step 2)John calls our applicationInput query to OpenCyc:"(# canCause # John ?X ?Y)"QueryTo OpenCyc(step 3)Upload resultJohn can have knee pain ifhe takes Advil and the risk is high.(step 5)John gets the resultJava code getting output:while (iterator.hasNext()){item ame());}Get inferenceresultfrom OpenCycCyc Microtheory 1Assert to OpenCyc:(# take # John # Advair)")(#take # John # Advil)Cyc Microtheory 2(# implies((# take ?Patient # Steroid)# and(# take ?Patient # Advair))(# canCause ?Patient # kneePain # high))(step 4)(step 1)* All the data transfer format is ASCII and synchronousVI. ADE Services29Translation donein advance usingNLPDailyMed
Offerings DetailUser InterfaceVII. User Interface30
Mobile Communication Services . Offerings Detail Samsung SDS America Public Sector Capabilities Mobile ERP Health IT Mobile Groupware SAP Mobile BI Dashboard Oracle/Siebel Mobile CRM for Pharmaceutical Sales Mobile Device Management Mobile Applications (Android OS) . Android Mobile App & UI. 10 Offerings Detail Conceptual .
The Clinical Program is administered by the Clinical Training Committee (CTC) under the leadership of the Director of Clinical Training (DCT) and the Associate Director of Clinical Training (ADCT). The program consists of three APA defined Major Areas of Study: Clinical Psychology (CP), Clinical Child Psychology (CCP), Clinical Neuropsychology .
The Tivoli Decision Support 2.1 User’s Guide describes the features and concepts behind the Tivoli Decision Support 2.1 product. Step-by-step procedures for using the Tivoli Decision Support Discovery Interface are also provided. Hereafter, the Tivoli Decision Support Discover Administrator is called the
intelligent vehicles in recent years. Intelligent vehicles tech-nologies are based on the information of the ego vehicle and its surroundings, such as the lanes, roads, and other vehicles, using the sensors of intelligent vehicles [4], [5]. The sensors in intelligent vehicles can be divided into internal and external sensors.
Ceco Building Carlisle Gulf States Mesco Building Metal Sales Inc. Morin Corporation M.B.C.I. Nucor Building Star Building U.S.A. Building Varco Pruden Wedgcore Inc. Building A&S Building System Inland Building Steelox Building Summit Building Stran Buildings Pascoe Building Steelite Buil
Decision Support System (DSS) Decision Making Every decision and action that humans make is inherently related to decision making. But humans have weaknesses (subjective, bias, forgetful, imprecise, and slow) in many cases of decision making. Therefore humans need to be assisted by computer-based tools called DSS to improve
1. What is decision theory?.5 1.1 The decision disciplines 5 1.2 Decision processes 7 1.3 Decision matrices 11 1.4 Classification of decision theories 13 1.4.1 Normative and descriptive theories 14 1.4.2 Individual and collective decision-making 15 1.4.3 Degrees of knowledge 16 2.
Oct 18, 2014 · A decision problem is characterized by decision alternatives, states of nature, and resulting payoffs. The decision alternatives are the different possible strategies the decision maker can employ. The states of nature refer to future events, not under the control of the decision maker, which
Decision theory and Decision analysis Decision Analysis De nition (B. Roy):\consists in trying to provide answers to questions raised by actors involved in a decision process using a model" Answers:\Optimal solution" or \Good decision" is absent Models:formalized or not Brice Mayag (LAMSADE) Introduction to Decision Modeling Chapter 0 18 / 36
