Improving Perioperative Mortality Rate Data Capture in Hawassa, Ethiopia: A Mixed‐Methods Study

BACKGROUND: In 2015, Ethiopia created a national strategic plan aimed at increasing surgical capacity. This encompassed a set of surgical indicators collected at each hospital, which facilitates assessment of the current state of surgical practices and enables monitoring of progress towards improving outcomes. This project aimed to investigate one of these key surgical indicators, the perioperative mortality rate (POMR), in the largest referral hospital in the Sidama region. METHODS: Perioperative mortality was defined as death after major surgery prior to hospital discharge. The number of surgical deaths and surgical volume for the most recent reporting year (July 2022-May 2023) were extracted from registries and discharge information. POMR was compared between the following data sources: (1) paper registries from the operating rooms, surgical wards, ICU, (2) aggregate discharge information from the Liaison Office, (3) monthly indicator reports to the health management information system (HMIS), and (4) reports on the national data collection system (DHIS2). Additionally, qualitative interviews with healthcare professionals and data officers were conducted to evaluate current practices and challenges in collecting POMR. RESULTS: The aggregate reported 1-year POMR was 0.9% (56/6438) for registries, 0.9% (57/6336) for discharge information, 0.5% (33/6437) for HMIS reports, and 0.6% (35/5935) for DHIS2 reports. Qualitative interviews (n = 17) reported regular tracking of perioperative deaths within surgical departments and surgical data quality checks by quality officers. However, many interviewees identified the lack of an electronic system as a significant barrier to accurate, timely data collection. Other proposed solutions include implementing closer monitoring of data quality, providing additional equipment and training for staff on data entry, and promoting timely completion of monthly reports. CONCLUSION: Explanations behind the discrepancies in POMR between data sources in this study include inconsistencies in data entry, lack of consolidation of postoperative deaths from different service points (wards, ICU), and reports generated without a mechanism to include updated registry or discharge counts. These findings, along with the diverse perspectives from our interviews, point toward the need for a well-trained workforce and the development of a digital record system to improve POMR data capture.