A Stochastic Control Program to Predict Outcome and to Support Therapeutic Decisions: A Preliminery Report
Background and Objectives. Early noninvasive hemodynamic monitoring with an outcome predictor and a therapeutic decision support system may be useful to identify and correct hemodynamic deficiencies in emergency patients. The first aim was to apply a stochastic (probability) search and display model to predict outcome as early as possible. The second aim was to explore the usefulness of a therapeutic decision support system to evaluate the relative effectiveness of various therapies. Methods. A stochastic control and display program based on noninvasive hemodynamic monitoring was applied in 100 consecutive critically ill patients admitted to the emergency department of an inner city public hospital. The program continuously displayed the noninvasive hemodynamic data and the patient's predicted survival probability (SP) that was based on the patient's diagnosis, covariates, and hemodynamic data. The accuracy of the SP at the initial resuscitation on admission to the emergency department (ED) was evaluated by the actual outcome at hospital discharge. The therapeutic decision support program evaluated the relative effectiveness of various therapies on based on their hemodynamic and SP responses and outcome of patients with similar clinical-hemodynamic states. Results. The cardiac index, mean arterial pressure, arterial saturation, transcutaneous oxygen and carbon dioxide tensions were appreciably higher in survivors than in nonsurvivors in the initial resuscitation. Heart rate was higher in the nonsurvivors. The calculated Survival Probability (SP) of survivors averaged 81 ± 1.4% in the first 24-hour observation period. It was 58 ± 2.2% for nonsurvivors during this period. Misclassifications were 10/100 or 10%.
Key WordsStochastic analysis and control program outcome prediction therapeutic decision support system noninvasive hemodynamic monitoring thoracic bioimpedance estimation of cardiac output pulse oximetry transcutaneous oxygen and carbon dioxide tensions tissue perfusion
Unable to display preview. Download preview PDF.
- 1.Conners AF, Jr, Speroff T, Dawson NV, Thomas C, et al. The effectiveness of right heart catheterizaion in the initial care of critically ill patients. JAMA 1996; 276: 899–897.Google Scholar
- 5.Artegas A, Bernard GR, Brigham RL, Carlet J, Falke K, et al. The American-European Consensus Conference on ARDS. Parts 1 and 2. Am J Respir Crit Care Med 1994; 149: 818–824 and 1998; 157: 1332–1347.Google Scholar
- 12.Berlauk JF, Abrams JH, Gilmour IJ, et al. Pre-operative optimization of cardiovascular hemodynamics improves outcome in peripheral vascular surgery. Ann Surg 1991; 214: 189.Google Scholar
- 16.Bayard DS, Botnen A, Shoemaker WC, Jelliffe RW. Stochastic analysis of therapeutic modalities using a database of patient responses. IEEE Symposium on Computer Based Medical Systems (CBMS), 2001; 11: 439–444.Google Scholar
- 17.Shoemaker WC, Bayard DS, Botnan A, Wo CCJ, et al. Mathematical program for outcome prediction snf thersprutiv support for trauma beginning within one hour of admission. Crit Care Med, in press.Google Scholar
- 19.Wang X, Van De Water JM, Sun H, et al. Hemodynamic monitoring by impedance cardiography with an improved signal processing technique. Proc IEEE Eng Med and Biol 1993; 15: 699–700.Google Scholar
- 21.Tremper KK, Shoemaker WC. Transcutaneous oxygen monitoring of critically ill adults with and without low flow shock. Crit Care Med l981; 9: 706–709.Google Scholar
- 23.Tremper KK, Huxtable RF. Dermal heat transport analysis for transcutaneous O2 measurements. Acta Anesth Scand 1978; supp1: 68: 4–9.Google Scholar
- 24.Tremper KK, Waxman K, Bowman R, Shoemaker WC. Continuous transcutaneous oxygen monitoring during respiratory failure, cardiac decompensation, cardiac arrest, and CPR. Crit Care Med 1980: 8: 337.Google Scholar
- 25.Sutton RS, Barto AG. Reinforced Learning. MIT Press, Cambridge, 1998.Google Scholar
- 26.Bertsekas DP, Tsitsiklis JN. Neuro-dynamic Programming. Athena Scientific, Belmont MA, 1996.Google Scholar