Introduction Refractory ascites (RA) confers a median survival of 6 months. Repeated large volume paracentesis (LVP) is commonly required. This typically occurs following unplanned admission, or via day case (DC) services. This study aims to assess the impact of DC LVP services on economic and clinical outcomes in the last year of life (LYOL), alongside other potential clinical and demographic predictors

Method The ONS mortality database was searched for deaths occuring secondary to cirrhosis in England between 2012–15. Identifiers were linked to the hospital episode statistics database. Patients who died from cirrhosis and underwent LVP in the LYOL were included. 2014 HRG tariffs were applied to each hospital episode in the LYOL and total patient costs calculated. Outcomes were: i) Total cost in LYOL ii) Bed days in LYOL iii) Place of death (POD). Linear multiple regression analysis was conducted to examine predictors for cost and bed days. Use of DC LVP services within LYOL (≥1 DC LVP=planned care (PC) group vs emergency care (EC)), cause of death, index of multiple deprivation, sex, ethnicity, age, days from index presentation to death and total episodes, were entered. Logistic regression was used to analyse POD outcomes. Secondary analysis investigated outcomes in the PC group as proportion of planned care increased.

Results 1 27 699 episodes from 13 841 patients were analysed (EC=11,198, PC=2,643). In combination, variables within the linear model accounted for 53.4% of variation in total cost (mean cost in LYOL=£21,137, adj R²=0.53), and 25.6% of variation in bed days (mean=35.3, adj R²=0.26). DC LVP use was significantly and independently predictive of reduced total cost and bed days in LYOL (coeff: -£6432,–21.1 days p<0.001). Lower deprivation and hepatocellular carcinoma (HCC) were also independently predictive of reduced cost and bed days (p<0.001). Logistic regression for POD outcomes demonstrated odds ratios of 0.53 (95%CI:0.46–0.59) for death in hospital, and 1.85 (1.64–2.1) for death at home or in a hospice in the PC group. HCC was the strongest predictor of death outside hospital. Secondary analysis demonstrated significant correlations between increasing proportion of planned care with reduced cost, bed days and probability of dying outside hospital.

Conclusion Strong associations between DC LVP use and improved economic and end of life outcomes are demonstrated. Increasing access to DC services is likely to be cost effective, and may improve quality of care. Implications for community hepatology and the specialist nurse workforce are discussed

Disclosure of Interest None Declared