In order to facilitate a more rapid and systematic transfer of new medical knowledge and capabilities into mainstream clinical practice, a new Medical Information Network Decision Support (MINDS) system was developed. This system is intended to be a platform for storing and fusing medical data in a standardized way, and for providing probabilistic diagnostic and treatment decisions to assist doctors and medical researchers in understanding and treating disease.

A key part of this project is the identification of existing standards and protocols that are currently utilized by the medical industry. These standards will be evaluated for their usefulness in the current MINDS project as a first step in making the MINDS system compatible with the existing medical network infrastructure so that it will more easily interface with existing equipment, tools, and information networks. Another key part of the project is an evaluation of prior and existing clinical decision support systems. An analysis will be done to determine which of the existing techniques are useful within the proposed MINDS system architecture. Finally, the project will identify key areas of improvement that will be pursued in future funding cycles in order to develop improved decision support algorithms, data formats and modernized data structures, and general system architecture.

Six core areas have been identified where clinical information systems have provided some level of computerized or automated support to the medical community. Each of these core areas is further regulated by an underlying set of quality control rules, ethical practices, and government regulation. The key contributions that the MINDS system can provide to the broader CIS are in the medical/clinical decision support area, and in the research and training support systems. These data analysis processes can also be mapped to medical processes that can be supported by a data analysis process.

The Computer-based Physician Order Entry (CPOE) and electronic prescribing tools have been developed successfully and are not of primary interest in the MINDS system. However, the automation of data analysis to provide doctors with support of diagnosis and disease propagation, treatment planning for desired outcomes, and preventative care are key interest areas in the MINDS system.

The general decision architecture is modeled after the JDL Fusion model, which has been adapted to fit the appropriate medical terminology. The JDL model helps to associate a data level with the processes involved in analyzing that data. These data analysis processes can also be mapped to medical processes that can be supported by a data analysis process. The modularity of the system is further enhanced by the definition of different software layers. This will improve the overall maintainability of the system.

This work was done by H. K. Armenian of TechFinity, Inc. for the U.S. Army Medical Research and Materiel Command. ARL-0066