Abstract
ObjectivesWe aimed to evaluate hospital mortality rates, readmission rates and length of hospital stay (LOS) among adult medical patients admitted to a teaching hospital in Ethiopia.DesignWe performed a retrospective study using routinely collected electronic data.SettingData were collected from Yekatit 12 Hospital Medical College between January 2021 and July 2023.ParticipantsThe analysis included 3499 (4111 admissions) adult medical patients with complete data.Outcome measuresWe used mortality rates, readmission rates and LOS to measure the quality of the outcomes for the top 15 admission diagnoses. A multivariable Cox proportional hazard model was used to identify the statistically significant predictors of mortality with p values<0.05 and a 95% CI. The Kaplan-Meier curve was used to estimate the failure rate (mortality) of the admitted patients.ResultsThe median age of patients was 50 years and men accounted for 1827 (52.3%) of all admitted cases. Non-communicable diseases accounted for 2537 (72.5%) admissions. In descending order, stroke, 644 (18.29%); heart failure, 640 (18.41%); and severe pneumonia, 422 (12.06%) were the three most common causes of admission. The readmission rate was 25.67% (1056/411), and 61.9% of them were readmitted within 30 days of index discharge. The overall median LOS was 8 days. The median LOSs in the index admission (11 vs 8 days, p value=0.001) of readmitted patients was significantly higher than not readmitted. The in-hospital mortality rate was 438 (12.5%), with the highest number of deaths occurred between days 30 and 50 of admission. The mortality rate is significantly higher among patients with communicable diseases (adjusted HR, 1.64, 95% CI: 1.34, 2.10) and elderly patients (≥65 years) (adjusted HR, 1.79, 95% CI: 1.44, 2.22). Septicemia, chronic liver diseases with complications and HIV with complications were the three common causes of death with a proportional mortality rate of 55.2%, 27.93% and 22.46%, respectively.ConclusionsMortality, median LOSs and readmission rate were comparable to other national and international studies. Multicentre compressive research using these three quality patient outcomes is required to establish national standards and evaluate institutional performance.