MMI 406 - Decision Support Systems and Health Care

Course Description:

Provided an introduction to clinical decision support systems (CDSS) in health information technology (HIT). Instruction is given in formal decision analysis techniques as they apply to decisions in the medical domain. Clinical decision support systems are introduced and issues relating to their design and implementation discussed. The mathematical foundations upon which they are based was examined. Evidence-based guidelines and performance measurement techniques were covered. A framework for designing and implementing CDSS was introduced. Principles learned from this framework were applied in writing a final paper that described a prototype decision support system, including justification for its use and a description of steps followed in its design, implementation and performance measurement.
Professor: Gerasimos Petratos, MD, MS
Term: Winter 2012
Grade: A-
Text: 1) Hunink, H. & Glasziou, P. (2009). Decision making in health and medicine: Integrating evidence and values (7th printing). Cambridge, England: Cambridge University Press. 2) Berner, E. S. (ed.). (2007). Clinical decision support systems: Theory and practice (2nd ed.). New York, New York: Springer, Health Informatics Series. 3) Osteroff, J. A., Pifer, E. A., Teich, J. M., Sittig, D. F., & Jenders, R. A. (2005). Improving outcomes with clinical decision support: an implementer's guide. Chicago, IL: HIMSS.

Learning Objectives:

  • Analyze medical decisions for the application of pertinent clinical decision support technology.
  • Evaluate IT resources needed for the implementation of decision support systems.
  • Apply medical decision support interventions at appropriate points in the diagnosis and treatment continuum.
  • Gain stakeholder support in leading the development of decision support interventions.
  • Develop tools for the improvement of decision support technology.
  • Formulate a comprehensive plan for the development, implementation, and evaluation of a decision support system.
  • Promote effective use of decision support technology in a health care organization.

What I Learned:

This course was one of the most challenging classes in the program. It leveraged mastery and application of course content from 402 ("Intro to Clinical Thinking"), 403 ("Intro to Medical Informatics") and 409 ("Introduction to Biostatistics"). However, it was also one of my favorite classes. With a professional background in Software Engineering, I have always had a keen interest in improving systems to become more intelligent. There is a whole discipline dedicated to Artificial Intelligence (AI), and Clinical Decision Support Systems (CDSS) benefits from some of the advance research originating from AI.

The class was heavy with readings both from our assigned texts and from professional journal articles sourced from different topics all pertaining to the emerging science of Medical Informatics. Some of the readings focused on the historical development of CDSS. A prominent leader in CDSS is Intermountain Healthcare (IHC) based in Salt Lake City, Utah, and upon discovering this, my interest deepened further, as I had at one time considered pursuing a career change to work for them. IHC is world renowned for their forward thinking and patient-centered care model. The readings, classroom discussions, and lectures also enlightened me with a perspective of how CDSS has evolved and is being used today to improve patient outcomes across disparate delivery modalities.

Class evaluation consisted of a number individual assignments that drew upon our learnings from other courses. There were seven assignments in all, and a comprehensive final where our recollection of course principles was tested. One exercise which was extremely valuable and not learned elsewhere was the application of the theory of decision tree analysis. I found this assignment to be extremely pertinent in my own life when considering a decision with multiple possible outcomes. The tool helps one consider the “utility” (or one’s own preference) towards an outcome, and then traces probabilities from the result to help the individual (or patient) make an informed choice based on quality of life indicators ascribed to the utility. In doing the exercise, I used a software package known as TreeAge . In essence, CDSS successfully helps patients navigate a labyrinth of choices toward a targeted outcome that makes the most sense for them based on their preferred outcomes. I often envisioned how tools like this could be integrated into an Electronic Medical Records (EMRs) or Computer Physician Order Entry (CPOEs) applications to help patients visualize their options more efficiently.

My passion for CDSS had heightened at this point, and out of self-interest I started to explore open source projects that existed in the field. Motivated by a strong desire to learn more,. It was then that I discovered OpenCDS, which was a project sponsored through the University of Utah. The project's founder is Dr. Kenosha Kawamoto, MD, PhD , who also is accredited as the inventor of SEBASTIAN technology that originated from a research project at Duke University. SEBASTIAN is short for System for Evidence-Based Advice through Simultaneous Transaction with an Intelligent Agent across a Network. It is a "Web service-based framework for (i) encoding medical knowledge into a machine-executable format, and (ii) integrating this knowledge into various clinical applications to enable clinical decision support." (Open Clinical, 2006)  Through my personal endeavors, I managed to get in contact with Dr. Kawamoto and the lead developer for the project, and was granted access to the project's source code. I downloaded the code, compiled the project, and began to get hands on experience with a CDSS. I also leveraged my learnings by including them in my personal assignment papers..

For the final course project, I teamed with other students with a goal of applying my newly-learned concepts to a research paper. My one stipulation was that the team was to choose a CDS topic that could be leveraged in a real world setting. Eric Abbott , Nancy Casazza > , and myself wrote a paper entitled "A Clinical Decision System to Improve PN Core Measure Reporting". The project focused on a conceptual product called "To The Core" (2TC) which solves the problem of collecting, aggregating, and reporting of core measures from disparate data locations. 2TC incorporated my OpenCDS knowledge and the various elements of the CDS that are edited by clinicians. This fulfilled the role of knowledge informaticists, who map core measure rules from CMS-based rules in order to comply with core measure reporting – a requirement for P4P and PQRI as defined by MU under the 2009 ARRA HITECH Act. One benefit of CDSS is that knowledge domain experts can create rules using domain specific language (or DSL). DSL makes rules creation intuitive and easy to maintain over time. Thus, the knowledge authored by informaticists becomes a tangible asset that can be kept, updated, and shared across the HCO or with other HCOs on the same open source platform – one of the fundamental goals of HIT semantic interoperability that is repeatedly espoused throughout the MMI program.

In conclusion, I learned a lot during this course, and my pursuit with OpenCDS will continue. As I prepare to wrap up the program, I intend to become a collaborator and contributor to the OpenCDS project. While I chose a different Capstone Project, I expect to pursue a personal interest project in CDS by helping an organization in Zambia, Africa implement OpenCDS using a portable integrated Electronic Health Record (EHR) system known as SmartCare

WP-Backgrounds by InoPlugs Web Design and Juwelier Schönmann