A retrospective cross-sectional study was conducted at a single referral hospital. All cases within the general surgery, pediatric surgery, neurosurgery, and cardiothoracic and vascular surgery units were analyzed. The study included all patients who were operated on and those who died within 30 days of surgery in the time course between May of 2021 and April of 2022.

The study was conducted at Tikur Anbessa Specialized Hospital, the largest teaching hospital in Ethiopia. It is located at the center of Addis Ababa, the capital city of the country. The hospital has over 700 beds and provides services to over 500,000 patients annually at a tertiary-level designation.

All surgical patients within the 4 surgical units selected for the study were included initially to determine the rate of POMR. Following this, perioperative mortalities (POMs) occurring during the course of the study were included and analyzed.

All surgical patients who underwent surgical intervention with open or minimally invasive techniques within the study period. All deaths following surgical intervention within 1 month after surgery regardless of the cause of death were included in the study and evaluated.

All patients with surgical disease who were treated non-surgically were excluded. In addition, all patients who died at the time of arrival at the hospital or before surgical intervention were excluded. Obstetric and gynecology, urology, and orthopedic patients were excluded. Obstetrics and gynecology cases were excluded because they are out of the jurisdiction of the Department of Surgery, and ethical clearance for these cases could not be acquired. Urology and orthopedics were excluded because there were no mortality cases during the course of the study. Finally, 17 cases with poor or incomplete documentation deemed difficult for analyses were also excluded.

The independent variables for this study were gender, age, American Society of Anesthesiology (ASA) score, comorbidity, type of admission, indications for surgery, and surgical procedures performed.

The dependent variables were the rate of postoperative death, the cause of death, the postoperative day of death, and the length of hospital stay.

The data source was from medical records, and both electronic and paper-based retrieved after the medical record numbers were acquired from the operation logbooks.

After the data was collected, it was cleaned, coded, and entered into IBM Corp. Released 2015. IBM SPSS Statistics for Windows, version 23.0. Armonk, NY: IBM Corp. Both descriptive and inferential statistics were utilized for the interpretation of the data.

For all categorical variables, measures of central tendency with mean and standard deviation were used in addition to frequency distribution. Inferential statistics with univariable and multivariable logistic regression was then performed for the risk identification regarding the time of death and associated variables.

The ethical approval for this study was acquired from Addis Ababa University, College of Health Sciences, ethical review board. The study was conducted in accordance with the Helsinki Declarations, National and Institutional Guidelines, while keeping all the information retrieved for the study confidential.