Article Sidebar
Published:
Jan 15, 2009Article Details
Copyright of published material remains in the Authors’ name. This allows authors to use their work for their own non-commercial purposes without seeking permission from the Publisher, subject to properly acknowledging the Journal as the original place of publication.
Authors are free to copy, print and distribute their articles, in full or in part, for teaching activities, and to deposit or include their work in their own personal or institutional database or on-line website. Authors are requested to inform the Journal/Publishers of their desire/intention to include their work in a thesis or dissertation or to republish their work in any derivative form (but not for commercial use).
Material submitted for publication in the SAJCC is accepted provided it has not been published or submitted for publication elsewhere. Please inform the editorial team if the main findings of your paper have been presented at a conference and published in abstract form, to avoid copyright infringement.
Main Article Content
Validating the use of the APACHE II score in a tertiary South African ICU
Abstract
Background. In order to evaluate both outcome of intensive care unit (ICU) patients and ICU care, the riskadjusted mortality can be calculated using the APACHE II equation. Our aim was to: (i) describe the case mix of admissions to our ICU; (ii) investigate the impact of such variation on outcome; and (iii) validate the use of the APACHE II risk prediction model in a developing country. Methods. Prospective data collection of consecutive adult admissions over 13 months in a tertiary, predominantly medical, ICU. Survivors and non-survivors were compared for age, sex and diagnoses. ICU mortality was calculated for diagnostic categories and for the whole group. Risk of death was calculated
according to the APACHE II method. The goodness of fit of the APACHE II equation was assessed with a calibration curve. The discrimination of the model was assessed with a receiver operator characteristic (ROC) curve. Results. There were 304 admissions with an average APACHE II score of 17.4 and 37% ICU mortality. Diagnostic groups with high ICU mortalities included medical patients (42%), severe sepsis (59.4%), community-acquired
pneumonia (CAP) (53%), pulmonary tuberculosis (45%) and immunocompromised patients (62%). A calibration curve for the APACHE II equation, applied to our data, shows that the predicted ICU mortality was within the 95% confidence interval (CI) of the actual mortality. The only exception was the group with a 70% predicted risk of ICU death. The area under the ROC curve was 0.83 (95% CI: 0.78 - 0.88). The standardised mortality ratio was
0.98 (95% CI: 0.79 - 1.17). Conclusions. This study validates the use of the APACHE II model to accurately describe the risk of ICU death of the patient population in a tertiary ICU in a developing country. Patients with severe sepsis and/or CAP had a significantly higher mortality. The main reason for this appeared to be a high risk of death at ICU admission. The
principles of appropriate management of early sepsis should be taught to all doctors through continuing medical education.
SAJCC Vol. 21 (1) 2005: pp. 46-54
