How immigrants adapt their smoking behaviour: comparative analysis among Turkish immigrants in Germany and the Netherlands

Abstract

Background: Smoking behaviour among immigrants is assumed to converge to that of the host country’s majority population with increasing duration of stay. We compared smoking prevalence among Turkish immigrants residing in two different countries (Germany (DE)/the Netherlands (NL)) between and within countries by time spent in Turkey and DE/NL. Methods: The German 2009 micro-census and the Dutch POLS database (national survey, 1997–2004) were analysed. An interaction variable with dichotomised length of stay (LOS) in Turkey (age: 0-17; 18+) and categorised LOS in the host country (immigration year: 1979 and earlier, 1980-1999, 2000-2009; the latter only for Germany) was generated. Age standardised smoking prevalences and sex-specific logistic regression models were calculated. Results: 6,517 Turkish participants were identified in Germany, 2,106 in the Netherlands. Age-standardised smoking prevalences were higher among Turkish immigrants in the Netherlands compared to those in Germany: 62.3% vs. 53.1% (men/lower education); 30.6% vs. 23.0% (women/lower education). A similar trend was observed for the majority population of both countries. The chance of being a smoker was lower among Turkish men with short LOS in Turkey and middle LOS in Germany/the Netherlands compared to those with short LOS in Turkey and long LOS in Germany/the Netherlands (NL: OR = 0.57[95% CI = 0.36-0.89]; DE: OR = 0.73[95% CI = 0.56-0.95]). Contrary to that, the chance of being a smoker was higher among Turkish men with long LOS in Turkey and middle LOS in Germany/the Netherlands compared to those with long LOS in Turkey and long LOS in Germany/the Netherlands (NL: OR = 1.35[95% CI = 0.79-2.33]; DE: OR = 1.44[95% CI = 1.03-2.02]). The effects for Turkish women were similar, but smaller and often non-significant. Conclusion: Turkish immigrants adapt their smoking behaviour towards that of the Dutch/German majority population with increasing duration of stay. This was particularly obvious among those who left Turkey before the age of 18 years - a group that needs tailored interventions to prevent further increases in smoking. Those who left Turkey as adults and spent a short time in the host countries show 'imported' smoking patterns. A limitation of this study is the use of cross-sectional data: a cohort effect cannot be ruled out. Our findings have to be confirmed with longitudinal data.