The impact of migration in all-cause mortality: The Turin Longitudinal Study, 1971e2005 Domenica Rasulo a, * , Teresa Spadea a , Roberta Onorati a , Giuseppe Costa b a Regional Epidemiology Unit, ASL TO3 Piedmont Region, Via Sabaudia 164, 10095 Grugliasco, Turin, Italy b Department of Clinical and Biological Sciences, University of Turin, Italy article info Article history: Available online 17 January 2012 Keywords: Italy Turin Mortality risk Internal migrants Age at arrival Length of stay Birth cohorts Migration abstract North-western Italy has a long history of domestic influx, however little is known on how migrant mortality compares to mortality at the local level. While geographic mortality gradients may play a role, conceptualizations developed for international migration may also be relevant. Using this theoretical framework, the study investigated immigrant-native differentials in the north-western city of Turin through a 34-year follow-up that was facilitated by the Turin Longitudinal Study. The study population comprised inhabitants of age 30e74 years at the 1971 census. Survival trajectories were modelled through the Gompertz distribution and were examined for nativity status, birthplace, length of stay and age at arrival (the last two variables were combined). All estimates were adjusted for socio-economic factors. Overall, the risk of dying for internal migrants was lower, compared to locals, and consistent with geographic gradients. However, this pattern hid signifi- cant differences mediated by both age at arrival and length of stay. The advantage appeared to be exclusive to young and adult migrants, despite differentials narrowing over time. Immigrants who arrived after age 44 suffered instead a progressively greater excess risk, compared to natives, as residence increased. The dissipation of the health advantage found in internal migrants, along with poor health outcomes amongst people older at arrival, raises concern about immigrants from developing countries who need to endure a more demanding journey and adjustment to the new environment. The study indicated, through lengthy longitudinal data, that immigrant-native differentials were best explained by the stratified variable ‘length of stay by age at arrival’ and this should inform future studies. Ó 2012 Elsevier Ltd. All rights reserved. Introduction A number of studies have investigated differentials in mortality between international migrants and natives (Deboosere & Gadeyne, 2005; Marmot, Adelstein, & Bulusu, 1984; Singh & Miller, 2004; Singh & Siahpush, 2001; Stirbu et al., 2006; Uitenbroek & Verhoeff, 2002). Within a single country, differ- ences are, of course, less pronounced, as risk factors for mortality are usually more uniform. However, Italy stands out in this respect due to a well-established geographical gradient, with the highest rates in the North for men and in the South for women (Istat, 2010; Mortara, 1960). Similarly, there exists a gradient between the rural and the industrial areas of the North, with higher mortality in the latter (Caselli & Egidi, 1979; Parodi et al., 2004). These disparities can be captured by immigrant-native differentials in north-western Italy (Vigotti, 1992, 1996), whose industrial sector attracted workers from the South and the rural North from the mid-1940s to the 1970s (Bonifazi & Heinz, 2000; Rosori, 1978). Besides birthplace gradients, the little research that exists on immigrant-native differentials in Italy is inconclusive, owing to lack of adequate data. In fact, prior studies mainly had access to cross- sectional data, such as death certificates by areas of residence and of birth, which narrowed the research objectives. However, theoretical work on international migrants can reasonably inform a study on internal migrants since both streams are in response to similar forces (e.g. labour movement) and face the same challenges (Lu, 2008). We thus use theories about inter- national migration to pinpoint research questions on domestic influx in Italy. It is well documented that migrants have lower mortality rates compared to natives (Akresh & Frank, 2008; Bos, Kunst, Keij- Deerenberg, Garssen, & Mackenbach, 2004; Deboosere & Gadeyne, 2005; Marmot et al., 1984; Singh & Siahpush, 2001; * Corresponding author. E-mail address: domenica.rasulo@ons.gsi.gov.uk (D. Rasulo). Contents lists available at SciVerse ScienceDirect Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed 0277-9536/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.socscimed.2011.10.045 Social Science & Medicine 74 (2012) 897e906