Clinical Decision Support Systems in Biomedical Informatics and their LimitationsOverviewClinical DecisionsRequirements for excellent decision-makingGoalDefinition of Decision Support SystemsHistorySlide 8TypesRequirements of a CDSSRequirements: Patient Data AcquisitionRequirements: Patient Data ValidationRequirements: Medical Knowledge ModelingRequirements: Medical Knowledge ElicitationRequirements: Medical Knowledge RepresentationRequirements: Medical Knowledge ReasoningRequirements: System PerformanceRequirements: Integration to the WorkflowExisting SystemsPathfinder (1992)Pathfinder’s Deductive Reasoning ModelIliad (1988)DiagnosisPro (1993)DiagnosisPro’s User InterfaceHeart Disease Program (HDP) (1980’s – 90’s)Heart Disease Program’s Differential SummaryClinical Knowledge Summaries (CKS) (2007)Clinical Knowledge Summaries’ User InterfaceDxplain (1987)VisualDx (2006)VisualDx’s User InterfaceINTERNIST-1 / QMR Project (1974 - 80’s)EON System (1996)EON System ArchitectureSnapshot of our clinical decision support systemsLimitationsLimitations: Patient’s RoleSlide 38Limitations: UsabilitySlide 40Limitations: System PerformanceSlide 42Limitations: Knowledge SharingSlide 44Limitations: Knowledge MaintenanceSlide 46Limitations: SecuritySlide 48Concluding RemarksCDSS-1CSE5095Clinical Decision Support Systems in Clinical Decision Support Systems in Biomedical Informatics Biomedical Informatics and their Limitationsand their LimitationsAlberto De la Rosa AlgarínComputer Science & EngineeringUniversity of Connecticut, Storrsalberto.delarosa.algarin@engr.uconn.eduCDSS-2CSE5095Overview Overview Clinical Decisions Clinical Decisions What types of clinical decisions exist?Requirements for excellent decision-makingDefinition of Decision Support SystemsHistoryHistoryFirst possibility of a CDSSFirst prototype and the shortcomingsBetter CDSS (MYCIN, HELP, Leeds System)Existing SystemsExisting SystemsPathfinder, Iliad, DiagnosisPro, CKS, HDP, etc.LimitationsLimitationsPatient’s Role, Usability (and performance), Knowledge sharing and maintenance and SecurityCDSS-3CSE5095Clinical DecisionsClinical DecisionsTwo types of clinical decisions:Two types of clinical decisions:Diagnosis decisionsDiagnosis processDiagnosis decisionsDiagnosis decisionsDone analyzing to determine the cause of sicknessDiagnosis processDiagnosis processUsed to determine which questions to ask in order to make better diagnosis decisionsCDSS-4CSE5095Requirements for excellent decision-makingRequirements for excellent decision-makingAccurate data:Accurate data:Bad data is useless obviouslyGood data is equally useless if there is no knowledge on how to apply it.Pertinent knowledgePertinent knowledgeThe overload of information affects the process of decision making in a negative way.Overload of information can be seen in the ICUAppropriate problem-solving skillsAppropriate problem-solving skillsThe glue between the correct use of accurate and pertinent knowledge.CDSS-5CSE5095GoalGoalThe goal of clinical decision support systems (CDSS) The goal of clinical decision support systems (CDSS) is to emulate the clinician’s thought process during is to emulate the clinician’s thought process during diagnosis.diagnosis.CDSS-6CSE5095Definition of Decision Support SystemsDefinition of Decision Support SystemsA decision support system is a system in which one or A decision support system is a system in which one or more computers and computer programs assist in more computers and computer programs assist in decision making by providing information.decision making by providing information.They can exist as hardware-software solutions or stand They can exist as hardware-software solutions or stand alone software applications.alone software applications.CDSS-7CSE5095HistoryHistoryThe possibility first appeared in 1959 [Ledley & The possibility first appeared in 1959 [Ledley & Lusted]Lusted]With the use of symbolic logic, probability theory and value theory, the foundations of medical diagnosis could be understood.The first prototype appeared in 1964 [Walker et al.]The first prototype appeared in 1964 [Walker et al.]Issues with logistics, scientific shortcomings related to medical diagnosis, and the lack of integration to the workflow made the widespread use and adoption virtually impossible.CDSS-8CSE5095HistoryHistoryAfter this, several CDSS appeared that tackled the After this, several CDSS appeared that tackled the previous pitfalls (MYCIN, Leeds System and HELP)previous pitfalls (MYCIN, Leeds System and HELP)MYCIN [Shortliffe, 1976]MYCIN [Shortliffe, 1976]A consultation system for patients with infectionsLeeds Abdominal Pain System [De Dombal et al., Leeds Abdominal Pain System [De Dombal et al., 1972]1972]A system for the diagnosis of acute abdominal painHELP [Warner, 1979]HELP [Warner, 1979]A system to alert clinicians in case of abnormalities in patient recordsCDSS-9CSE5095TypesTypesInformation Management SystemsInformation Management SystemsProvide an environment for the storage and retrieval of information. Decision is left to the clinician.Focusing Attention SystemsFocusing Attention SystemsAlert clinicians when a conflict arises. Follow simple logic.Patient-specific Recommendation SystemsPatient-specific Recommendation SystemsOffer advice to a single patient using the patient’s medical history.Can use simple logic, decision theory, cost-benefit analysis, etc.CDSS-10CSE5095Requirements of a CDSSRequirements of a CDSSClinical decision support systems must satisfy the Clinical decision support systems must satisfy the following requirements in order to be widely accepted following requirements in order to be widely accepted and used:and used:Patient Data Acquisition and ValidationMedical Knowledge Modeling, Elicitation, Representation and ReasoningSystem PerformanceIntegration to the WorkflowCDSS-11CSE5095Requirements: Patient Data AcquisitionRequirements: Patient Data AcquisitionThere is no standard way to acquire data.There is no standard way to acquire data.Current methods range from keyboard to natural Current methods range from keyboard to natural language processing.language processing.Some health care professionals even use Some health care professionals even use intermediaries like nurses or secretaries.intermediaries like nurses