Investigating MPM-III, sofa and delta sofa scoring systems as predictors of mortality at a tertiary hospital in Nairobi, Kenya: a retrospective cohort study
Date of Award
Master of Medicine (MMed)
Dr. Wangari Waweru
Prof. Rodney Adam
Dr. Peter Kussin
Anaesthesiology (East Africa)
Background: Despite major advances, management of critically ill patients worldwide is associated with high mortality rates. Assessment of disease severity is crucial in the analysis of ICU mortality. Although a variety of severity of illness scoring systems have been developed for this assessment and validated across Europe and the USA, few studies exist to show support of their utilization in African countries. There is conflicting evidence regarding the most reliable severity of illness scoring system suitable for resource limited countries as their mortality predictive values differ with different populations. SOFA is already in use at the AKUHN ICU and has easily attainable variables, while MPM-III is a simple, non-laboratory test dependent scoring system.
Objectives: The primary objective of this study was to compare SOFA and MPM-III severity of illness scoring systems as predictors of mortality in critically ill adult patients admitted to the AKUHN ICU. Secondary objectives were to describe the clinical and demographic characteristics of patients admitted to the AKUHN ICU, to identify factors associated with increased risk of mortality in patients admitted to the AKUHN ICU, to compare MPM-III and delta SOFA scoring systems as predictors of mortality in critically ill adult patients admitted to the AKUHN ICU and to compare SOFA and delta SOFA scoring systems as predictors of mortality in critically ill adult patients admitted to the AKUHN ICU.
Primary outcome: ICU mortality
Study setting: AKUHN ICU.
Study design: A retrospective cohort study.
Sample size: Using a formula for determining statistical difference between two areas under receiver operating characteristic curves, a sample size of 543 was used.
Study population: Critically ill adult patients admitted to the AKUHN ICU between January 2015 and September 2017.
Data collection and entry: The ICU admission register was used to obtain the names and medical records of patients admitted during the study period. A comprehensive review of patients’ files and the AKUHN Electronic Medical Records System (CARE®) was then conducted to extract data as per the data collection tool. MPM-III scores were calculated using an online MPM-III scoring calculator. Data was entered into an MS Excel spreadsheet to form a database. STATA software version 14.2 was used for data analysis.
Results: Male patients accounted for 60% of total admissions. The most common reason for admission was respiratory failure at 33%. Medical and surgical admissions constituted 61% and 39% of the total admissions respectively. Mortality rate was 32.4%. Median LOS was 4 days for both survivors and non-survivors. Factors associated with increased odds of mortality were LOS < 5 days, ICU admissions from HDU and Casualty. SOFA and MPM-III scoring systems had fair discrimination (AUROCs: SOFA 0.771, MPM-III 0.740) while delta-SOFA had poor discrimination with an AUROC of 0.687. The Hosmer-Lemeshow chi-square statistics for SOFA, MPM-III and delta SOFA were p = 0.099, 0.225 and 0.054 respectively.
Conclusion: In this study, MPM-III and SOFA scoring systems had good discrimination and were well calibrated while delta-SOFA had poor discrimination and borderline calibration. Therefore, SOFA and MPM-III demonstrated acceptable accuracy in predicting ICU mortality. In settings where it may not be possible to obtain the basic laboratory tests required as part of SOFA variables, MPM-III scoring system is best suited.
Lukoko, L. (2018). Investigating MPM-III, sofa and delta sofa scoring systems as predictors of mortality at a tertiary hospital in Nairobi, Kenya: a retrospective cohort study (Unpublished master's dissertation). Aga Khan University, East Africa.