O-30 ALCOHOL-HARM PARADOX IN LATIN AMERICA: HOW TO STUDY IT DESPITE DATA LIMITATIONS? THE CHILEAN EXPERIENCE

Abstract

Introduction and Objectives: Research on the “Alcohol-Harm Paradox” (AHP) investigates why low-income individuals have more alcohol-related harm despite lower alcohol consumption (AC). Possible explanations have been evaluated in Europe and the US, but data constraints make it difficult in Latin America (LATAM). This study aimed to design a strategy to study the AHP in LATAM's restricted-data context, recognize its strengths and limitations, and identify possible explanations in the Chilean experience. The AHP is expected to be explained by the unequal distribution of comorbidities, risk behaviors, consumption patterns, rurality, education, access to health, social capital, and mental health. Materials and Methods: We first evaluated our hypothesis at the individual level with data from the 2016-17 National Health Survey. We conducted logistic regression models to assess whether the hypothesized explanatory factors mediated the effect of AC on liver disease. Second, we aggregated at the municipal level registry data on deaths from alcohol-related liver disease (Ministry of Health Statistics) and survey data on AC and the hypothesized explanatory factors (National Drug Survey and National Survey of Socioeconomic Characterization) to test our hypothesis using mortality as the outcome of negative binomial regression models. Results: The first analysis suggests that the AHP exists among Chilean men and it is explained by the unequal distribution of metabolic syndrome, diabetes, obesity, smoking, heavy episodic drinking, rurality, education, social support, and depression. The second analysis reinforces these findings and highlights the explanatory potential of healthcare-access inequality (Figure 1). Conclusions: The proposed analyzes support our hypothesis in Chile. They can be replicated in other LATAM countries as an effective restricted-data strategy to start investigating the AHP. However, cross-sectional survey analyzes are limited by reverse causation and aggregate data analyses by ecological fallacy. Better access to administrative data with patient identifier is needed to generate accurate longitudinal evidence on the explanatory mechanisms.