Investigation into a Principal Components Methodology for Predicting Disease in Case-Based Decision Support, 10-R9757

Printer Friendly Version

Principal Investigators
Micah A. Spears
David A. Tong, Ph.D.
Gwen Y. Ross
Nathan A. Price

Inclusive Dates:  10/01/07 – Current

Background - Medical diagnosis is the process where a medical provider forms a differential diagnosis using the information garnered from a conversation with the patient, physical examination and laboratory test results. Because of inherent time constraints and the emerging availability of Electronic Medical Record (EMR) data, health-care providers are readily employing clinical decision support systems (DSS) to assist them in the diagnostic process. Research has shown leading diagnostic systems to be sufficient medical reference systems (i.e., electronic manuals), yet, in practice these systems fail to provide an informed differential diagnosis in a timely manner. Modern differential diagnostic systems are ineffective at accurately diagnosing disease because these systems cannot efficiently process "noisy" data. During the diagnostic process, the information shared between the diagnostic indicators represents a source of statistical "noise" that can significantly reduce the accuracy of the resulting differential diagnosis.

Approach - This project is developing a novel medical software algorithm capable of automating the interpretation, simplification, and noise reduction of diagnostic data for the clinician, thus requiring minimal human intervention to produce an accurate differential diagnosis. An ongoing software prototype, Tυχη("tie-kee"), is being developed to provide a testing and demonstration vehicle for the project. A clinical knowledge base has been constructed using the Centers for Disease Control's (CDC) National Ambulatory Medical Care Survey (NAMCS) dataset and relational database technology.

The specific objectives for the software prototype are:

  • Model a relational knowledge base to store more than one million diagnostic cases.
  • Develop a principal components algorithm to analyze the diagnostic cases.
  • Develop a multiple variable regression algorithm to compute the diagnosis odds.
  • Develop an explanatory model to communicate the rationale for a diagnosis recommendation.
  • Evaluate the diagnostic efficacy of the new algorithm when compared to two baseline algorithms.

Accomplishments - The project has met all of its objectives to date. Additional focus has been placed on the explanatory model to better communicate how the signs and symptoms influence the diagnostic selection model. The additional focus will better position the software prototype for integration into an existing clinical environment and help promote buy-in from experienced medical providers.

The prototype is a rich, visual DSS developed using the Java® programming language, R statistical package, MySQL® database management system, and the Swing® graphical toolkit. The prototype assists a medical provider in rapidly generating a differential diagnosis, referring a patient to a medical specialist, and qualifying a patient for hospital admission.

As a direct result of this project, SwRI will have the capability to demonstrate medical diagnosis technology and pursue external projects in clinical decision support, telehealth/telemedicine, medical training, and biosurveillance.


Diagnostic Process
 

Differential Diagnosis

2008 Program Home