Patterns of Readmission and Death after Admission with AKI: A Data-Linkage Study

Background: Despite episode-level data, and health-service level studies, patient-level incidence of AKI across Australia or its major jurisdictions has not been characterized to date. Important outcomes including discharge to residential aged care, hospital readmission, and mortality remain undefined. Data linkage allows clarification of patient-level outcomes across a jurisdiction to permit system-level policy and planning and potentially refine resource allocation. Methods: Data linkage analysis of jurisdiction-wide administrative data in Victoria, Australia. The study linked the Victorian Admitted Episodes Dataset (VAED), Victorian Death Index (VDI), Victorian Integrated Non-admitted Health (VINAH), and Victorian Cancer Registry (VCR) in July 2016 to June 2017. Index admissions with AKI were defined with follow up to 36 months. Results: 2,817,000 admissions were complicated by AKI in 198,966 (7.1%) episodes of care in Victorian hospitals. 38,033 individual adult patients accounted for the 96,549 AKI episodes identified in adults admitted overnight to acute care. AKI incidence was greatest amongst elderly (age ≥75 years; 57.6%), overseas-born, and comorbid patients, but paradoxically less amongst socioeconomically disadvantaged patients. Inpatient mortality complicated 1,984 (5.2%) of the index admissions. Most AKI survivors were readmitted [23,757 (65.9%)] at least once (median 3, range 0 – 33), and a majority [20,244 (85.2%)] experienced recurrent AKI within 12 months. At 12-months, overall mortality was 27.2%, and 22.9% amongst survivors of the index episode with 28.0% of these deaths within 30 days. At 3 years, overall mortality was 46.2%. Cancer was the leading cause of death followed by cardiovascular disease, sepsis, dementia, and kidney failure. Conclusion: Despite likely underreporting of AKI in Victoria, AKI was a commonly reported and relapsing condition of predominantly elderly patients. An index episode of AKI was associated with recurrent admission and recurrent AKI in most patients. Discharge to a residential aged care facility was common. AKI was associated with high mortality, most commonly attributable to cancer rather than cardiovascular disease. Education, interventional studies, and policy can better reflect the epidemiology of AKI. Further analysis should clarify previously unexplored interactions between AKI, cancer, residential aged care, avoidable readmission, and mortality.