Team for Research in
Ubiquitous Secure Technology

Citation
Janos Laszlo Mathe. "A Model-Integrated, Guideline-Driven, Clinical Decision-Support System". Talk or presentation, 11, November, 2010.

Abstract
In our earlier paper we have reported the use of a formal model-based development method for a guideline-driven patient management system, called Sepsis Treatment Enhanced through Electronic Protocolization (STEEP). During the last 8 months we have been engaged in the preparation and execution of a clinical trial at two Intensive Care Units of the Vanderbilt University Medical Center partially funded by NIH. The process required passing of the institutional HIPAA (privacy), security and quality review process and the integration of STEEP into Vanderbilt's clinical information systems including the EMR, Ordering, Medical Administration and Alert Management systems. The experience gave us a number of interesting conclusions that led us to enrich our science agenda on trustworthy health information systems. In this presentation we will summarize the experience gained with the clinical trial and will elaborate on the following two specific challenges: (1) Development of the Clinical Process Modeling Language (CPML) has been a significant effort. CPML integrates three kinds of knowledge: medical (sepsis), execution platform (STEEP engine) and privacy rules. Our primary conclusion is that for future problem domains (such as the cancer management we are currently analyzing) instead of applying CPML or developing new domain specific modeling languages from skretch we need to decompose the knowledge CPML into reusable sub-languages. These sub-languages will be representing essential aspects of the problem space: Medical Ontology, Execution Platform Semantics and Privacy Rules. Reusable model libraries built for these separate aspects will then be used to generate the integrated domain specific models. (2) Integration of STEEP into Vanderbilt's Health IT infrastructure has been a significant challenge. Reusability of the effort can be ensured by formally defining abstraction layers for each major components—including the relevant privacy rules—and integrate these layers into the system architecture as models.

Electronic downloads

• 14%20-%20Mathe%20%28Vanderbilt%29.pptx · application/vnd.openxmlformats-officedocument.presentationml.pre · 3445 kbytes

• HTML

  Janos Laszlo Mathe. <a
  href="http://www.truststc.org/pubs/768.html"
  ><i>A Model-Integrated, Guideline-Driven, Clinical
  Decision-Support System</i></a>, Talk or
  presentation,  11, November, 2010.

• Plain text

  Janos Laszlo Mathe. "A Model-Integrated,
  Guideline-Driven, Clinical Decision-Support System".
  Talk or presentation,  11, November, 2010.

• BibTeX

  @presentation{Mathe10_ModelIntegratedGuidelineDrivenClinicalDecisionSupport,
      author = {Janos Laszlo Mathe},
      title = {A Model-Integrated, Guideline-Driven, Clinical
                Decision-Support System},
      day = {11},
      month = {November},
      year = {2010},
      abstract = {In our earlier paper we have reported the use of a
                formal model-based development method for a
                guideline-driven patient management system, called
                Sepsis Treatment Enhanced through Electronic
                Protocolization (STEEP). During the last 8 months
                we have been engaged in the preparation and
                execution of a clinical trial at two Intensive
                Care Units of the Vanderbilt University Medical
                Center partially funded by NIH. The process
                required passing of the institutional HIPAA
                (privacy), security and quality review process and
                the integration of STEEP into Vanderbilt's
                clinical information systems including the EMR,
                Ordering, Medical Administration and Alert
                Management systems. The experience gave us a
                number of interesting conclusions that led us to
                enrich our science agenda on trustworthy health
                information systems. In this presentation we will
                summarize the experience gained with the clinical
                trial and will elaborate on the following two
                specific challenges: (1) Development of the
                Clinical Process Modeling Language (CPML) has been
                a significant effort. CPML integrates three kinds
                of knowledge: medical (sepsis), execution platform
                (STEEP engine) and privacy rules. Our primary
                conclusion is that for future problem domains
                (such as the cancer management we are currently
                analyzing) instead of applying CPML or developing
                new domain specific modeling languages from
                skretch we need to decompose the knowledge CPML
                into reusable sub-languages. These sub-languages
                will be representing essential aspects of the
                problem space: Medical Ontology, Execution
                Platform Semantics and Privacy Rules. Reusable
                model libraries built for these separate aspects
                will then be used to generate the integrated
                domain specific models. (2) Integration of STEEP
                into Vanderbilt's Health IT infrastructure has
                been a significant challenge. Reusability of the
                effort can be ensured by formally defining
                abstraction layers for each major
                components—including the relevant privacy
                rules—and integrate these layers into the system
                architecture as models.},
      URL = {http://www.truststc.org/pubs/768.html}
  }
  

Posted by Larry Rohrbough on 7 Dec 2010.
Groups: trust
For additional information, see the Publications FAQ or contact webmaster at www truststc org.