The Multi-Trellis Software Architecture and the Intelligent Cardiovascular Monitor

Abstract:

A real-time, intelligent cardiovascular monitor is complex. It must process multiple waveforms, recognize artifacts, extract pertinent parameters, recognize a patient’s clinical state, analyze the problem and formulate a response. This paper presents the multi-trellis (a collection of process trellises), a software architecture for building such a monitor. A process trellis is a uniform hierarchical framework for heterogeneous program modules. The multi-trellis extension allows one to compile several process trellis programs with widely varying run-time requirements into a single executable program that it is efficient, predictable and usable. Our prototype consists of two process trellises. The lower trellis contains processes to analyze three different analog signals: the blood pressure from a non-invasive monitor and an arterial catheter, and the ECG. The upper trellis contains processes to help detect evolving hemodynamic trends, identify abnormalities, and present a succinct summary to the clinician. Our prototype shows that the multi-trellis is a demonstrably useful software architecture for building these real-time, intelligent monitors.

Key-Words

Real-time Monitors - Software Architecture - Artifact Detection - Signal Processing

• REFERENCES

• 1 Factor M, Gelernter DH, Kolb C, Miller PL, Sittig DF. Real-time Performance, Parallelism and Program Visualization in Medical Monitoring. Yale University Research Report YALEU/DCS/RR-808. June 1990.
• 2 Factor M. The process trellis software architecture for real-time monitors. In: Proceedings Second ACM SIGPLAN Symposium on Principles & Practice of Parallel Programming. Seattle WA. (ed) March 1990: 147-55 (also in SIGPLAN Notices, 25 [3]).
• 3 Factor M, Sittig DF, Cohn AI, Gelernter D, Miller PL, Rosenbaum SH. A parallel software architecture for building intelligent medical monitors. Int J Clin Mon Comp 1990; 07: 117-28.
• 4 Wesseling KH, Ty Smith N. Availability of intra-arterial pressure waveforms from catheter-manometer systems during surgery. J Clin Monit 1985; 01: 11-6.
• 5 The American Society of Anesthesiologists Newsletter. 1986; 50: 12.
• 6 Kestin IG, Miller BR, Lockhart CH. Auditory alarms during anesthesia monitoring. Anesthesiology 1988; 69: 106-9.
• 7 O’Carrol TM. Survey of alarms in an intensive therapy unit. Anesthesia 1986; 41: 742-4.
• 8 Poler SM, Dunford-Shore B, Lappas DG. Architecture of an intelligent anesthesia monitoring system: Concepts for distributed real-time multiprocessing. In: Proceedings of the 1988 Rochester FORTH Conference 1988; 110-2.
• 9 Fagan LM. Representing time-dependent relations in a medical setting (Ph. D. Diss.). Stanford University; 1980
• 10 Sittig DE, Pace NL, Gardner RM, Beck E, Morris AH. Implementation of a computerized patient advice system using the HELP clinical information system. Comp Biomed Res 1989; 22: 474-87.
• 11 Dawant B, Uckum S. A Framework for Intelligent Multi-channel Biological Signal Interpretation. Vanderbild University Center for Intelligent Systems Technical Report #CIS-90-08. 1990
• 12 Rutledge G, Thomsen G, Beinlich I, Farr B, Sheiner L, Fagan LM. Combining qualitative and quantitative computation in a ventilator therapy planner. In: Proceedings of the 13th Annual Symposium on Computer Applications in Medical Care. Kingsland LC (ed) 1989; 315-9.
• 13 Berger MP, Gelfand RA, Miller PL. Combining statistical, rule-based, and physiological model-based methods to assist in the management of diabetes mellitus. Comp Biomed Res 1990; 23: 346-57.
• 14 Henderson SE, Crapo RO, East TD, Morris AH, Wallace CJ, Gardner RM. Computerized clinical care protocols in an intensive care unit: How well are they followed?. In: Proceedings of the 14th Annual Symposium on Computer Applications in Medical Care. Miller RA. (ed) 1990: 284-8.
• 15 Sittig DF, Gardner RM, Morris AH, Wallace CJ. Clinical evaluation of computer-based respiratory care algorithms. Intern J Clin Mon Comp 1990; 07: 177-85.
• 16 Merri M, Fardin DC, Mottley JG, Titlebaum EL. Sampling frequency of the electrocardiogram for spectral analysis of the heart rate variability. IEEE Trans Biom Engin 1990; 37: 99-105.
• 17 Beneken JEW, Gravenstein JS. Sophisticated alarms in patient monitoring: A methodology based on systems engineering concepts. In: The Automated Anesthesia Record and Alarm Systems. Gravenstein JS, Newbower RS, Ream AK, Smith NT. (eds) Boston: Butterworth; 1987: 211-28.
• 18 Carriero N, Gelernter DH. How to Write Parallel Programs. MIT Press. 1990
• 19 Factor M, Gelernter DH. Experience with the process Trellis architecture. Yale University, Department of Computer Science Research Report YALEU/DCS/RR-818. 1990
• 20 Factor M, Gelernter DH, Kolb C, Miller PL, Sittig DF. Real-time data fusion in the ICU. IEEE Computer 1991; 24: 45-54.
• 21 Factor M, Gelernter DH. The process Trellis: A software architecture for intelligent monitors. In: Proceedings IEEE International Workshop on Tools for Artificial Intelligence: Architectures, Languages, and Tools. IEEE Computer Society 1989; 174-81.
• 22 Factor M. The Process Trellis; Software Architecture for Parallel, Real-Time Monitors (Ph. D. Diss.). Yale University, Department of Computer Science. 1990
• 23 Erman LD, Hayes-Roth F, Lesser VR, Raj Reddy D. The Hearsay-II Speech-Understanding System: Integrating Knowledge to Resolve Uncertainty. Comp Surveys 1980; 12: 213-53.
• 24 Cohn AI, Rosenbaum SH, Factor M, Miller PL. DYNASCENE: An approach to computer-based intelligent cardiovascular monitoring using sequential clinical “scenes”. Meth Inform Med 1990; 29: 122-31.
• 25 Sittig DF, Factor M. Physiologic trend detection and artifact rejection: A parallel implementation of a multi-state Kalman filtering algorithm. Comp Meth Progr Biomed 1990; 31: 1-10.
• 26 Cohn AI, Rosenbaum SH, Miller PL. An alternative parallel-computing approach to intelligent hemodynamic monitoring. Anesthesiology 1990; 73: A541.
• 27 Boehmer RD. Continuous, real-time noninvasive monitor of blood pressure: Penaz method applied to the Finger. J Clin Monit 1987; 03: 282-7.
• 28 Pan J, Tompkins WJ. A real-time QRS detection algorithm. IEEE Trans Biomed Engin 1985; 32: 230-6.
• 29 Kinias P, Fozzard HA, Norusis MJ. A realtime pressure algorithm. Comp Biol Med 1981; 11: 211-20.
• 30 Sittig DF, Clyman JI, Cheung KH, Miller PL. Receiver operating characteristic (ROC) curve helps optimize blood pressure detection algorithms. Anesthesiology 1990; 73: A456.
• 31 Rampil IJ. Intelligent detection of artifact. In: The Automated Anesthesia Record and Alarm Systems. Gravenstein JS, Newbower RS, Ream AK, Smith NT. (eds). Boston, MA: Butterworths; 1987
• 32 Sittig DF. Characteristic Fourier transform spectrum helps identify arterial blood pressure artifact. J Clin Mon 1990; 07: 130-2.
• 33 Lipman BS, Massie E, Kleiger RE. Clinical Scalar Electrocardiography. (6th ed.).. Chicago: Year Book Medical Publ; 1979
• 34 Sittig DF, Cheung KH. A parallel implementation of a multi-state Kalman filtering algorithm to detect ECG arrhythmias. Int J Clin Mon Comp (in press).