F. BYIRINGIRO1, J. BYIRINGIRO1 1University Of Rwanda,Department Of Surgery,KIGALI, , Rwanda
Introduction: A Hospital mortality Surveillance is a tool used for reporting hospital deaths. The trends, the causes and predictive risk factors are well understood with reliable tools that accurately report hospital mortality data, and stimulate regular discussions behind them.
Methods: Through a retrospective descriptive study, we collected 404 medical records of patients who died at CHUK during an 8-months period (January 1 to August 31, 2014). Descriptive statistics and baseline characteristics are used for analysis; Students t-tests, chi-square tests, non-parametric tests and time to event with Kaplan Meier and regression where appropriate. We used Epidata and STATA 12 for data management.
Results: We found only 404 medical records of deceased patients while admitted in the hospital. We report 39.6% females and 60.15% males; though 19.36% were traumas, 11.39% Cerebral vascula accidents, 6.94 cardiovascular diseases, 4.73 intestinal obstructions, 3.82% shocks, 3.47% multiorgan failures, and 2.48% comas of unknown causes. The Documented cause of Death were 2.25% RTA, 1.5% stroke, 1.2% Hematemesis, 1.2% severe TBI, 16.3% with different causes of Death, 18.7% unclear causes of death, and 50.87% unspecified diseases. 61.9% were unstable and critically ill patients, 26.46% stable but critically ill, 10.5% stable but not dischargeable, and 1.17% stable and dischargeable. Critically ill patients were 88.3%. Death occurred within the first 6h, 12h, 24h, or 5days respectively at 19%, 6%, 2.8%, and 45.7%. The association of critical status and imminent death was at first 6h (OR: 3.16; p-0.001), at first 12h (OR: 1.78; p-0.283), at first 24h (OR: 1.66; p-0.460), and at first 5days (OR: 0.68; p-0.147). There were 77.98% incoherencies between the working diagnosis and causes of death, and 36.4% missing data of the early warning score. The hospital structure and environment explain the correlation and associations detected between the factors related directly or indirectly to hospital deaths. The imminent death within the first 6h is significantly related to the critical presentation of the deceased patients (P<0.001). In fact, the shorter the hospital stay, the higher the patient related factors to death; and the longer the hospital stay, the higher the hospital-structure and working environment related factors to death.
Conclusion: Hospitals have different ways and tools of reporting mortality data; and the most reliable tools are hospital mortality registries. In hospitals where these tools don't exist, the report of mortality remains a challenge regarded the constant search of accurate data related to deaths and the sensibility the data carry. The survey states clearly the challenges behind the inaccuracy of hospital mortality data, and the study proposes long-term solutions to understand and decrease the current hospital mortality rate.