Benchmarking hospital safety and identifying determinants of hospital-acquired complication: the case of Queensland cardiac linkage longitudinal cohort

BACKGROUND: Hospital-acquired complications (HACs) are costly and associated with adverse health outcomes, although they can be avoided. Administrative linkage health data have become more accessible and can be used to monitor and reduce HAC. AIMS: This study aims to use linkage administrative data to benchmark the safety performance of hospitals and estimate the feasible magnitude that HAC can be reduced. We also identify risk factors associated with HACs, and estimate the effects of HACs on adverse health outcomes and hospital costs. METHODS: This is a retrospective linkage cohort study. The cohort includes 371,040 inpatient multiple-day admissions of 83,025 cardiovascular disease patients admitted to public hospitals in 2010 with follow-ups until 2015. Data envelopment analysis was applied to benchmark the patient safety performance of hospitals. Logistic regression was used to examine the odds of HAC and its effects on in-hospital mortality and 30-day readmission. Generalised linear models were used to identify the impacts of HACs on hospital costs and the length of hospital stay. FINDINGS: On average, 9.3% of multiple-day hospital admissions were associated with HACs. The average HAC rate can be reduced by two percentage points if all hospitals achieve the safety record of best-practice hospitals. Old age and multiple comorbidities were major driving factors of HACs. CONCLUSIONS: Cardiovascular disease patients with HAC have a higher risk of death, stay longer in hospitals and incur higher health care costs. The average HAC rates can be reduced by two percentage points by learning from best-practice hospitals operating in the same region.