Interobserver variability in the use of APACHE II scores

Abstract

2and the simplified acute physiology score (SAPS) are widely used to assess outcome and quality of care in intensive care units. In spite of the widespread use and general acceptance of these scoring systems, there is little information on their reliability and on inter-observer variability in their use. We assessed interobserver variation during application of the most frequently used scoring system, APACHE II, in our intensive care unit. Two groups of doctors were studied: residents (n=9) with limited experience of intensive care (average: 4 months), and intensivists (7), who should be experts in the use of scoring systems. Over 6 weeks, all doctors were given the charts of ten chosen patients and asked to assess APACHE II scores. We obtained 16 APACHE II scores of each individual patient. Analysis was with Student’s unpaired t test. There was wide variability between scores (mean 14, SD 6·0; table). There were no significant differences in score variations between intensivists and residents. The largest differences arose in the interpretation of data acquired in RESEARCH LETTERS the operating room (some took these data into account; others did not). Another cause of confusion arose in the interpretation of data which were inconsistent with the general trend: for example, tachycardia which was found only once during a 24-hour period was erroneously disregarded by some doctors. The accordance of chronic health points (2 or 5) was also a frequent source of problems. We conclude that assessment of APACHE II scores in individual patients varies widely; this applies both to less experienced doctors and to experts. Use of the APACHE II scoring system requires regular training, adherence to strict guidelines, and an understanding of which data should be used and which disregarded. Assessments of quality of medical care based on these scoring systems should be viewed with some care. 1 Kollef MH, Rainey TG. The role of outcomes research in the