Evaluating the Validity of Apache Ii as a Predictor of Icu Mortality for the Critically Ill Patients at Knh’scritical Care Units

Abstract

Introduction:Though most patients survive their critical illness after intensive care, another proportion are readmitted or die in hospital. While some of the risk factors for poor outcome are known, few are modifiable. It is now common to use tools that predict short term and long-term survival to guide care and cost management. The performance of these tools is quite variable especially when applied to different cohorts of patients. Acute Physiology and Chronic Health Evaluation (APACHE) II tool is most commonly used for audit and clinical purposes and to provide general measure of severity of disease. Study Objectives: The objective of the study was to evaluate the validity of APACHE II scoring system as a predictor of mortality in KNH ICU. Study Methods: This was a quantitative retrospective cross-sectional study. Stratified random sampling was used to select 180 files of patients admitted in ICU for the period January to December2017. The selected fileswere evaluated after approval from the Ethics and Review Committee. An APACHE II score was calculated by summing up the various diagnostic category weightings. A binary logistics regression was done to examine whether APACHE II score correctly predicts mortality. Tests of Calibration to assess the correspondence between the expected probability of mortality and actual observed mortality was done using a Hosmer-Lemeshow goodness of fit test. The area under the resultant curve (AUC) of the receiver operating characteristic (ROC) curve was calculated to assess for discrimination. Results:The mortality rate reported in this study was high at 31.1% with a mean APACHE II score of 20.58, the expected mortality rate also stood at 31.1%. Moreover, the mean APACHE II score was significantly higher in non-survivors when compared to survivors (p < 0.001). The study revealed that APACHE II score is an excellent predictor for mortality as ROC curve was found to be 0.889 with an optimum cut-off value of 20.5 with sensitivity 87.5% and specificity 79.0%. Recommendations: The researcher recommends that APACHE II be used in stratifying patients according to their degree of severity of illness at admission to ICU.