We have designed and implemented a medically-based model for generating human casualty simulations. In determining the state of a casualty, the model has been designed to consider the type of injuries sustained and the medical interventions performed. Many of the appropriate physiological, physical, and behavioral effects are portrayed visually on a graphical human body; additional graphical features will be added in the next two quarters. The simulation also has an interactive interface that allows a medic trainee to further probe a casualty using various queries. The system responds with a written or visual response.
The casualty simulation begins with the specification of medical conditions resulting from injuries sustained by the victim. Currently, the modeled medical conditions include tension pneumothorax, hemothorax, pericardial tamponade, and abdominal bleeding. The simulation displays plots of the victim's vital signs over time. The vital signs are plotted in terms of levels on the Trauma Scale [6] endorsed by the American Trauma Society for estimating injury severity. The trauma scale reflects basic assessments of the respiratory, circulatory, and central nervous systems. Fig. 5 shows the Trauma Scale rating system.
A Jack human figure portrays the physical and behavioral manifestations of its injuries and responds appropriately to commands representing various diagnostic and therapeutic procedures. A text response is written to the screen for queries regarding respiratory rate, respiratory expansion, blood pressure, capillary refill, and verbal response. Queries assessing eye opening and motor response result in changes in the displayed Jack figure. These commands include:
The model is built using Parallel Transition Networks (PaT-Nets), a package for creating and running parallel state-machines [2,4]. Four types of networks currently compose the casualty model:
A casualty may receive multiple conditions as result of injury; thus, the controller network is designed to handle combinations of injury and treatments in determining the effects on a casualty. For instance, a penetrating abdominal wound that extends into the chest may result in hemothorax, pericardial tamponade, and abdominal bleeding. Currently, when multiple conditions occur, the lowest value of each trauma scale value is used as the overall value. The current networks were built based on estimates from one of the project P.I.s, Dr. John Clarke, who is an experienced trauma surgeon.
As an illustration of our model, Fig. 6 shows the tension pneumothorax and needle aspiration PaT-Nets. The machine starts in the Base Node. A clock starts at the time of injury. After specific time intervals (1, 3, 5, 8, 12, and 15 minutes), the PaT-Net transitions to a new state where the appropriate vital signs are decremented. Thereafter, the PaT-Net immediately returns to the Base Node. If the patient is given a needle aspiration, the tension pneumothorax PaT-Net exits and spawns a needle aspiration PaT-Net. The needle aspiration PaT-Net starts in its Base Node. At one and two minute intervals, it transitions to a new state and increments the appropriate vital sign values. Fig. 7 shows the plot of the vital signs reflecting the Respiratory System when the patient develops a tension pneumothorax and receives no treatment. Fig. 8 shows the Respiratory System plot when the patient receives a needle aspiration 9 minutes after the time of injury. Note the marked improvement in the patient's state in this case.
Figure 6: Tension Pneumothorax and Needle Aspiration PaT-Nets
Figure 7: Tension Pneumothorax Plot -- No Treatment
Figure 8: Tension Pneumothorax Plot -- Needle Aspiration at 9 minutes
