Abstract

Background: Emergency admissions related to chronic liver failure (CLF) are common, expensive and associated with frequent readmission. There are no multicenter randomized controlled trials (RCT) investigating models to reduce liver-related emergency admissions (LREA). The primary aim of the Australian Liver Failure (ALFIE) trial was, therefore, to assess the efficacy of a nurse coordination model of care to reduce LREAs.

Methods: The study was a multicenter RCT occurring between 2018 and 2022. Patients were recruited following an inpatient admission with a CLF complication. The intervention at each site was a multifaceted care coordination model involving a nurse. Key components included intensive post-discharge monitoring (weekly phone calls for a minimum of 3 months), rapid access to care pathway, enhanced patient and carer education and self-management support. The intervention was applied continuously for the duration of the trial. Secondary aims were to assess the effects of this model on other measures of hospital usage, mortality, patient-reported outcomes and quality of care.

Results: 146 patients (75 Intervention group, 71 Control group) were recruited. The combined cohort had the following characteristics: mean age 54.9 years, 68% male, median MELD score 19.0 and median Child-Pugh score 9.0. The main causes of CLF were alcohol (68%), MAFLD (16%) and HCV (11%). The median (IQR) follow-up time for individuals in the Intervention and Control groups was 2.0 years. For the primary endpoint, LREA, there was a non-significant 11% reduction in LREA for the Intervention group vs. Control group (IRR 0.89, 95% CI 0.53-1.50, P=0.666). Improvement trends were also seen for the Intervention group for ICU admissions (IRR=0.62 P=0.491), 7-day readmissions (IRR=0.72, P=0.62), and length of stay (IRR=0.86, P=0.56). The leading causes of LREAs were ascites (43%), encephalopathy (22%) and variceal bleeding (11%). There was an increased risk of LREA due to encephalopathy in the Control vs. Intervention group (Hazard ratio=1.87, 95% CI=1.18-2.96, P=0.007); see Figure. There were no significant differences observed between groups for actuarial survival, or quality-of-life measures (CLDQ, EQ5D-5L utility, EQ-5D-VAS, QALY gains). All quality-of-care measures were improved in the Intervention group with significant improvement for HCC surveillance adherence (P=0.05), performance of bone density (P=<0.001) and vitamin D testing (P=<0.001).

Conclusion: This care coordination model showed benefits for CLF patients, particularly for reductions in LREA due to encephalopathy and improved quality of care. Further studies are needed to define this intervention model's optimal components, patient groups and settings. Further studies examining model cost-effectiveness and qualitative experiences of patients and care providers are in progress.