Journal of Clinical Monitoring and Computing

, Volume 14, Issue 4, pp 283–294

A Cross-Validated Multifactorial Index of Perioperative Risks in Adults Undergoing Anaesthesia for Non-Cardiac Surgery

Authors

  • Bernhard Schwilk
    • Department of AnaesthesiologyUniversity of Ulm
  • Rainer Muche
    • Department of Clinical Documentation and StatisticsUniversity of Ulm
  • Hans Treiber
    • Department of AnaesthesiologyUniversity of Ulm
  • Alexander Brinkmann
    • Department of AnaesthesiologyUniversity of Ulm
  • Michael Georgieff
    • Department of AnaesthesiologyUniversity of Ulm
  • Ulrich Bothner
    • Department of Medical InformaticsUniversity of Utah
Article

DOI: 10.1023/A:1009916822005

Cite this article as:
Schwilk, B., Muche, R., Treiber, H. et al. J Clin Monit Comput (1998) 14: 283. doi:10.1023/A:1009916822005

Abstract

Objective. To develop a severity index of anaesthetic risk that predicts relevant perioperative adverse events in adults. Design. Prospective cross-sectional study. Setting. Department of anaesthesiology at one university hospital. Patients. 26907 consecutive anaesthetic procedures in patients over 15 years of age and a complete preoperative evaluation. Patients undergoing cardiac and obstetric surgery were excluded. Measurements and main results. Demographic data, preoperative health status, type of anaesthesia, operative procedures, and perioperative incidents (standardised on a national basis) were acquired by means of a computerised anaesthetic record system. Occurrence of at least one perioperative event with impact on postanaesthetic care was computed by a multivariate logistic regression model against 17 variables with different characteristics representing possible risk factors. Fourteen variables proved to be independent risk factors. The weighting of the variables was expressed in scores which added up to form a simple index for each patient. Patients without major risk factors (0–10 points) had a 0.3% risk of suffering from a relevant incident. Patients with more than 60 points had a 28.6% risk. The results were well demonstrated by cross-validation. Conclusions. The index seems to reflect the risk of relevant perioperative incidents. It can be used for audit purposes. In daily routine, the index could focus our attention on patients with increased perioperative risk. However, it is limited in detecting particular constellations of factors which interact on each other with regard to perioperative risk.

Anaesthesiaepidemiologyperioperative eventsrisk factorscomplications
Download to read the full article text

Copyright information

© Kluwer Academic Publishers 1998