Identifikasi Faktor Risiko Kematian di ICU RSUP Dr. Sardjito

Abstract

Background: ICU is one of patients care unit that is costly and have limited capacity. Thus, time spending, man resources and tools should be allocated wisely based on the needs of the patients that would be admitted to ICU. One of the means that can be used to increase care quality of the ICU is creating a risk prediction system to assess and analyze risk factors that associated with mortality in ICU. This system makes comparative assessment and intensive care evaluation can be done. For the first step, we need to identify risk factors that affect mortality in ICU, that can be used as a new scoring models.

Methods: This study based on 356 patients that admitted to ICU RSUP dr. Sardjito from 1 January to 31 December 2019. In this cohort retrospective study, variables will be tested with regression logistics test, with univariable and multivariable approach. From this test, we achieved variables that identified as risk factors of mortality in ICU. Those identified variable undergo assessment of strength as a predictor with Area Under The Curve (AUC) method. If we find that, the discrimination strength is quite strong, we continue the test with calibration test using Hosmer-Lemeshow to achieve compoarative value within observed and expected mortality. The whole statistic process use SPSS application version 26.0.

Results: Factors that identified as risk factors for mortality in the ICU Dr.Sardjito Hospital were intraoperative use of vasopressor/inotropic drugs drugs, respiratory failure, GCS (Glasgow Coma Scale) decrease, kidney failure and intraoperative PRC transfusion with p < 0.005 in both univariable and multivariable tests. The AUC for mortality prediction in this study was 0.896 (95% CI; 85,3-94%). This value is classified as strong as a predictor factors, so it is continued to the calibration test with Hosmer-Lemeshow test and showed a p value of 0.53 (p > 0.05) which means that this risk prediction factors has a good fit between the observed and the expected.

Conclusion: The intraoperative use of vasopressor/inotropic drugs, respiratory failure, GCS decrease, kidney failure and intraoperative PRC transfusion are predictive factors for mortality in the ICU Dr.Sardjito Hospital. The discrimination ability of these factors is strong and also have a good fit in predicting the incidence of mortality in the ICU Dr.Sardjito Hospital.