Investigation of tobacco and alcohol co‐consumption in Thailand: A joint estimation approach

<jats:title>Abstract</jats:title><jats:sec><jats:title>Introduction and Aims</jats:title><jats:p>Most studies on tobacco and alcohol consumption empirically model the two behaviours separately. The primary aim of this study was to investigate the patterns of tobacco and alcohol co‐consumption in Thailand, modelling the two behaviours jointly and identifying whether they complemented or substituted each other.</jats:p></jats:sec><jats:sec><jats:title>Design and Methods</jats:title><jats:p>Data were extracted from the 2007, 2011, 2014 and 2017 waves of the nationally representative Tobacco and Alcohol Consumption Survey in Thailand. A bivariate ordered probit model, under which the frequencies of tobacco and alcohol consumption were jointly estimated, was performed on the repeated cross‐sectional sample as well as each of the cross‐sectional samples.</jats:p></jats:sec><jats:sec><jats:title>Results</jats:title><jats:p>Based on the repeated cross‐sectional sample (<jats:italic>n</jats:italic> = 411 981), the analyses suggested that tobacco and alcohol consumption were complementary behaviours. The predicted conditional probabilities revealed that the more frequently one drank, the higher the likelihood that he/she also smoked at a higher frequency, and vice versa.</jats:p></jats:sec><jats:sec><jats:title>Discussion and Conclusions</jats:title><jats:p>Given the complementary nature of tobacco and alcohol consumption in Thailand, tobacco‐ and alcohol‐prevention strategies should be realigned and better co‐ordinated for greater efficiency. The net impacts of a given tobacco or alcohol policy should be incorporated into the optimal policy design to avoid duplication of efforts. Also, future studies of tobacco and alcohol consumption should consider modelling the two behaviours together using a joint estimation approach.</jats:p></jats:sec>