Smoking and heavy drinking patterns in rural, urban and rural-to-urban migrants: the PERU MIGRANT Study

BACKGROUND Previous studies have found mixed results about cigarette and alcohol consumption patterns among rural-to-urban migrants. Moreover, there are limited longitudinal data about consumption patterns in this population. As such, this study aimed to compare the smoking and heavy drinking prevalence among rural, urban, and rural-to-urban migrants in Peru, as well as the smoking and heavy drinking incidence in a 5-year follow-up. METHODS We analyzed the PERU MIGRANT Study data from rural, urban, and rural-to-urban migrant populations in Peru. The baseline study was carried out in 2006-2007 and follow-up was performed five years later. For the baseline data analysis, the prevalence of lifetime smoking, current smokers, and heavy drinking was compared by population group using prevalence ratios (PR) and 95% confidence intervals (95% CI). For the longitudinal analysis, the incidence of smoking and heavy drinking was compared by population group with risk ratios (RR) and 95% CI. Poisson regression with robust variance was used to calculate both PRs and RRs. RESULTS We analyzed data from 988 participants: 200 rural dwellers, 589 migrants, and 199 urban dwellers. Compared with migrants, lifetime smoking prevalence was higher in the urban group (PR = 2.29, 95% CI = 1.64-3.20), but lower in the rural group (PR = 0.55, 95% CI = 0.31-0.99). Compared with migrants, the urban group had a higher current smoking prevalence (PR = 2.29, 95% CI = 1.26-4.16), and a higher smoking incidence (RR = 2.75, 95% CI = 1.03-7.34). Current smoking prevalence and smoking incidence showed no significant difference between rural and migrant groups. The prevalence and incidence of heavy drinking was similar across the three population groups. CONCLUSIONS Our results show a trend in lifetime smoking prevalence (urban > migrant > rural), while smoking incidence was similar between migrant and rural groups, but higher in the urban group. In addition, our results suggest that different definitions of smoking status could lead to different smoking rates and potentially different measures of association. The prevalence and incidence of heavy drinking were similar between the three population groups.