The goal of this proposal is to perform a comprehensive evaluation of the APACHE III Clinical Decision and ICU Management Support System (System). The System is an automated, computerized, clinical and management program that provides previously unavailable predictive and evaluative information to intensive care (ICU) clinicians and hospital administrators. This turnkey system provides objective risk estimates for key clinical outcomes (i.e., hospital death, length of stay, resource utilization) in real time for critically ill adults treated in intensive care units. The system was developed from the APACHE III (Acute Physiology, Age, and Chronic Health Evaluation) Prognostic Scoring System. The long-term objective of this proposal is to determine whether and how the timely provision of a new form of computerized, patient specific, prognostic information will influence existing relations between computation and language and between formal logic and current clinical practices. We will evaluate the impact of the APACHE III system on clinical decision making, patient outcomes, administrative and institutional decision making, user acceptance, system design, and its integration with existing laboratory and ICU monitoring systems.
These aims will be addressed through a combination of direct data processing from the APACHE III system and through a coordinated series of interviews with ICU clinicians. The APACHE III system's capabilities speak directly to two fundamental issues of computerized information systems - The interaction of formal (e.g., objective) logic and current clinical decision making and the effect of computation on language and communication. Among the important issues that this evaluation of APACHE III will address are: How does the numeric and color-coded representations of patient status influence physician knowledge, attitudes, and communication? Are the barriers to acceptance of computerized information systems primarily technical or perceptual? What are the major positive and negative influences of a computerized decision support system and how can future evaluative efforts best detect them? The results of this evaluation will be of central importance to medical informatics and will be generalizable to a wide variety of future, clinical alert, decision support, and artificial intelligence systems.
