Correspondence  |   December 1997
An Algorithm for Quantifying Blood Pressure Lability 
Author Notes
  • Department of Anesthesiology, University of Oklahoma Health Sciences Center, Post Office Box 53188, Oklahoma City, Oklahoma 73152,
Article Information
Correspondence   |   December 1997
An Algorithm for Quantifying Blood Pressure Lability 
Anesthesiology 12 1997, Vol.87, 1593-1594. doi:
Anesthesiology 12 1997, Vol.87, 1593-1594. doi:
To the Editor:-Reich et al. [1 ] have developed and preliminarily validated an algorithm for quantifying blood pressure lability. The problem is clinically significant, and the authors' use of receiver-operating characteristic curves to finetune their system for optimal results seems meritorious.
However, the authors' system does not appear to be an expert system by the conventional use of the term. Consequently, the keyword classification of the article appears incorrect. Further, the first sentence of the conclusion from the abstract, which reads, “One potential application of expert systems to anesthesia practice is a smart alarm to detect blood pressure lability,” is a nonsequitur because of this.
An expert system is a computer-based system, typically having certain characteristics:(1) the system performs a difficult task typically done by humans, where experts are provably better than amateurs, (2) the style of programming emphasizes symbolic rather than numeric computation, (3) the style of programming attempts to separate actual medical knowledge from program elements determining how that knowledge is used, in part to make system modification and explanation of system behavior easier, and (4) the systems are usually large and complex. Medical expert systems from Mycin [2 ] to Attending [3 ] have shared most, if not all, of these defining characteristics.
Running down this list of characteristics, none of these criteria seems applicable to the system developed by Reich et al. First, it is not clear that finding blood pressure lability is difficult or that experts are provably better than amateurs. It could easily be argued that anesthesiologists are better than others at managing or even preventing blood pressure lability, but the author cites no study to support the notion that anesthesiologists are better than nurses or even laymen at defining this lability. Second, the system developed is a numeric algorithm. Third, the system is explicitly referred to as an algorithm, and that approach is antithetical to one separating actual medical knowledge from “control knowledge” or information describing the way the program will execute. Finally, the system developed is neither large nor complex.
The system described by Reich et al. seems totally unlike any expert system in existence, and it seems inappropriate to classify this system as a smart alarm.
Aaron I. Cohn, M.D., M.A.
Department of Anesthesiology; University of Oklahoma Health Sciences Center; Post Office Box 53188; Oklahoma City, Oklahoma 73152
(Accepted for publication August 13, 1997.)
Reich DL, Osinski TK, Bodian C, Krol M, Sarier K, Roth R, Konstadt SN: An algorithm for assessing intraoperative mean arterial pressure lability. Anesthesiology 1997; 87:156-61.
Shortliffe EH, Davis R, Axline SG, Buchanan BG, Green CC, Cohen SN: Computer-based consultations in clinical therapeutics: Explanation and rule acquisition capabilities of the MYCIN system. Comput Biomed Res 1975; 8(4):303-20.
Miller PL: Critiquing anesthetic management: The “ATTENDING” computer system. Anesthesiology 1983; 58(4):362-9.