The impact of alcohol care teams on emergency secondary care use following a diagnosis of alcoholic liver disease - a national cohort study

Claire Currie, Alisha Davies, Cono Ariti, Martin Bardsley

The increasing mortality rates from alcohol-related liver disease (ARLD) are a public health concern. To address this, alcohol care teams (ACT) case-find and lead management of alcohol issues for these patients. Local assessments of ACTs have shown reductions in emergency admissions and emergency department attendances. We examine the impact of ACTs on emergency hospital activity following a diagnosis of ARLD. Administrative Hospital Episode Statistics (HES) data were extracted. Information on ACT provision at English NHS hospital trusts and sites in 2009/10 was taken from a survey by Public Health England. We undertook a difference-in-difference analysis to compare emergency hospital activity for a cohort of individuals diagnosed with ARLD who presented to hospitals either with or without an ACT in the one year before and after a first ARLD diagnosis during 2009/10. Over the study period, 9,165 individuals eligible for inclusion in our study had a first diagnosis of ARLD. 4,768 presented to one of 41 hospital trusts with an ACT (59 sites) and 4,397 presented to one of 50 non-ACT hospital trusts (65 sites). Whilst age and sex demographics were similar between the two cohorts, the ACT hospital cohort had a higher proportion of individuals in the most deprived quintile (41.6 % v 28.5 % p < .0001). In the difference-in-difference analysis, the presence of an ACT at a hospital trust was not associated with a change in all-cause emergency admissions (0.020 (95 % CI -0.070, 0.111), p = 0.656), alcohol-related emergency admissions (-0.025 (95 % CI -0.104, 0.054), p = 0.536) or all-cause emergency department attendances (0.042 (95 % CI -0.087, 0.171), p = 0.521). Sensitivity analyses by sex and hospital site did not affect the study findings. In this study, the presence of an ACT at the NHS hospital trust where individuals have their first recorded diagnosis of ARLD does not appear to be associated with subsequent emergency hospital activity within these populations. Further analysis focussing on the components and specific effects of ACT interventions on individuals and systems both pre- and post-diagnosis of ARLD may reveal important avenues to improve care.