Injuries, accidental and intentional, result in more years of human life lost in the United States than any other disease. The purpose of the American College of Surgeon's Advanced Trauma Life Support (ATLS) course is to enable physicians to provide immediate expert response in the initial evaluation, resuscitation, and stabilization of severely injured patients, prior to their initial definitive management. The goal of our system, TraumAID, is to optimize this next phase. TraumAID is a surgical consultation system for the initial definitive care of injuries. It currently deals with penetrating injuries to the abdomen and chest.
Our specific aim i s to improve its performance through the design, implementation and validation of a module that can take a set of diagnostic and therapeutic procedures and integrate them into a sensible and coherent plan. In doing this, our goals are two-fold. We hope both to increase TraumAID's clinical accuracy and practical utility - i.e., the range of cases on which it can produce patient-specific protocols equivalent to those of trauma experts - and to develop a general approach to the problem of formulating plans that integrate both diagnostic and therapeutic procedures. The planner we propose to design and build complements the logical reasoning already done by the system. Logical reasoning produces recommendations for diagnostic and therapeutic procedures that address the system's suspicions and conclusions. The purpose of the planner is to weave these recommendations into a clinically valid plan. Since a plan that contains diagnostic procedures cannot, by definition, have its final therapeutic goals known initially, the system consisting of both reasoner and planner is designed to operate incrementally: as the physician follows the plan, findings and results that emerge from diagnostic and therapeutic procedures that address them. These the planner will use in forming a revised plan. In forming its plans, the planner's decisions will be based on standard practices of trauma care, logistics, and a rich knowledge of procedures. It will use these to create an coarse ordering of the recommended procedures, resolve conflicts between them, and then optimize the resulting plan. While the resulting plan may be revised again, in response to new information, an entire plan is presented so that the physician may at all times be aware of what may still be pending and why. As part of this proposal, an initial validation will be done of the combined system (reasoner plus planner), just as such validations have been done of its abdomen and chest modules.

Agency
National Institute of Health (NIH)
Institute
National Library of Medicine (NLM)
Type
Research Project (R01)
Project #
5R01LM005217-03
Application #
2237720
Study Section
Biomedical Library and Informatics Review Committee (BLR)
Project Start
1991-06-01
Project End
1994-11-30
Budget Start
1993-06-01
Budget End
1994-11-30
Support Year
3
Fiscal Year
1993
Total Cost
Indirect Cost
Name
University of Pennsylvania
Department
Biostatistics & Other Math Sci
Type
Schools of Engineering
DUNS #
042250712
City
Philadelphia
State
PA
Country
United States
Zip Code
19104
Gertner, A S; Webber, B L; Clarke, J R et al. (1997) On-line quality [corrected] assurance in the initial definitive management of multiple trauma: evaluating system potential. Artif Intell Med 9:261-82
Clarke, J R; Webber, B L; Gertner, A et al. (1994) On-line decision support for emergency trauma management. Proc Annu Symp Comput Appl Med Care :1028