Cirrhosis with ascites in the last year of life: a nationwide analysis of factors shaping costs, health-care use, and place of death in England

Abstract

Liver disease mortality increased by 400% in the UK between 1970 and 2010, resulting in rising pressures on acute hospital services, and an increasing need for end-of-life care. We aimed to assess the effect of demographic, clinical, and health-care factors on costs, patterns of health-care use, and place of death in a national cohort of patients with cirrhosis and ascites in their last year of life. We did a retrospective, nationwide analysis of all patients who died from cirrhosis in England between 2013 and 2015, who required large-volume paracentesis in their last year of life. The outcomes measured were health-care costs accrued in the last year of life, number of inpatient days in last year of life, 30-day readmission rate, and occurrence of unplanned hospital death (probability of dying in hospital after unplanned admission). Using generalised linear and logistic regression models, we examined the effect of 12 independent variables on each outcome: sex, ethnicity, age at death, index of multiple deprivation quintile, year of death, liver disease causing death, place of death, time from index presentation in last year of life to death, whether enrolled in a day-case paracentesis service (care group), paracentesis ratio (number of day-case large-volume paracentesis procedures as a proportion of the total number of procedures in the last year of life), number of hospital episodes in the last year of life (not involving large-volume paracentesis), and number of large-volume paracentesis procedures in the last year of life. Between Jan 1, 2013, and Dec 31, 2015, 13 818 people in England died from liver disease and had large-volume paracentesis within their last year of life. For all patients, mean cost of the last year of life was £21 113 (SD 16 881), 17 888 (52·5%) of 34 068 readmissions occurred within 30 days of discharge, and 10 341 (74·8%) of 13 818 deaths occurred in hospital, of which 10 045 (97·1%) followed an emergency hospital admission. Patients who attended a day-case large-volume paracentesis service within their last year of life had significant reductions in cost (-£4240, 95% CI -4829 to -3651; p<0·0001), number of inpatient bed days (-16·98 days, -18·45 to -15·51; p<0·0001), probability of early readmission (odds ratio [OR] 0·35, 95% CI 0·31 to 0·40; p<0·0001), and probability of dying in hospital after unplanned admission (0·31, 0·27 to 0·34; p<0·0001), compared with patients who had unplanned care. For patients enrolled in day-case services, improvements in outcomes correlated with the proportion of large-volume paracentesis procedures done in a day-case (vs unplanned) setting. The use of day-case large-volume paracentesis services in the last year of life was associated with lower costs, reduced pressure on acute hospital services, and a lower probability of dying in hospital, compared with patients who received exclusively unplanned care in their last year of life. Wider adoption of day-case models of care could reduce costs and improve outcomes in the last year of life. David Telling Charitable Trust.