Understanding the alcohol harm paradox: an analysis of sex- and condition-specific hospital admissions by socio-economic group for alcohol-associated conditions in England

Background and Aims: In many countries, conflicting gradients in alcohol consumption and alcohol‐associated mortality have been observed. To understand this ‘alcohol harm paradox’ we analysed the socio‐economic gradient in alcohol‐associated hospital admissions to test whether it was greater in conditions which were: (1) chronic (associated with long‐term drinking) and partially alcohol‐attributable, (2) chronic and wholly alcohol‐attributable, (3) acute (associated with intoxication) and partially alcohol‐attributable and (4) acute and wholly alcohol‐attributable. Our aim was to clarify how (1) drinking patterns (e.g. intoxication linked to acute admissions or dependence linked to chronic conditions) and (2) non‐alcohol causes (e.g. smoking and poor diet which are risks for partially alcohol‐attributable conditions) contribute to the paradox. Design: Regression analysis testing the modifying effects of condition‐group (1–4 above) and sex on the relationship between area‐based deprivation and admissions. Setting: England, April 2010–March 2013. Participants: A total of 9 239 629 English hospital admissions where a primary or secondary cause was one of 36 alcohol‐associated conditions. Measurements: Admissions by condition and deciles of Index of Multiple Deprivation (IMD). Socio‐economic gradient measured as the relative index of inequality (RII, the slope of a linear regression of IMD on admissions adjusted for overall admission rate). Conditions were categorized by ICD‐10 code. Findings: A socio‐economic gradient in hospitalizations was seen for all conditions, except partially attributable chronic conditions. The gradient was significantly steeper for conditions which were wholly attributable to alcohol and for acute conditions than for conditions partially alcohol‐attributable and for chronic conditions. Gradients were steeper for men than for women in cases of wholly alcohol attributable conditions. Conclusions: There is a socio‐economic gradient in English hospital admission for most alcohol‐associated conditions. The greatest inequalities are in conditions associated with alcohol dependence, such as liver disease and mental and behavioural conditions, and in acute conditions, such as alcohol poisoning and assault. Socio‐economic differences in harmful drinking patterns (dependence and intoxication) may contribute to the ‘alcohol harm paradox’