Letter to the Editor - BMJ The Role of Social Protection in Mitigating the Effects of Unemployment on Suicides Before and During The Great Recession in Italy Roberto De Vogli12, PhD, MPH School of Medicine, Department of Public Health Sciences, University of California Davis Department of Epidemiology and Public Health, Division of Population Health, University College London, London, UK Manuscript (words): 753 Figures: References: 10 Correspondence to: Roberto De Vogli, School of Medicine, Department of Public Health Sciences, University of California Davis One Shields Ave Med Sci 1-C Build Davis, CA 95616 (email: r.devogli@ucl.ac.uk) In a recent article published in the BMJ, Arie asks: “Has austerity brought Europe to the brink of a health disaster?” The author also writes about the plight of suicides due to the crisis and budget cuts in Italy making specific references to the suicide of a couple of Italian pensioners hit by austerity measures and struggling with financial difficulties [1] In a recent short report we showed a statistically significant association of the economic crisis with suicides and suicide attempts due to economic reasons in Italy [2] So far, however, no study investigated the impact of austerity on suicides Here we present novel data on the role of social protection in moderating the association between unemployment and suicides across Italian regions In order to address this question, I first retrieved annual data on suicides per 100,000 and unemployment for all Italian regions from the Italian Institute of National Statistics (ISTAT) [3] Data on unemployment and suicides were available for all regions from 2000 to 2010 Linear regression models were used to assess the association between change in unemployment and change in suicides stratified by regions Models corrected for fixed effects, [4] [5] including social conditions and differences in surveillance systems and data quality, that may predispose some regions to higher or lower unemployment and suicide rates I also employed robust standard errors clustered by region to adjust for the lack of independence of the sample [6] All models were analyzed using STATA version 12.0 Figure shows that, similarly to previous analyses conducted in Europe, [7] unemployment and suicides were falling in Italy before the economic crisis, but reversed upon the onset of the financial downturn in 2007 Across Italian regions, levels of unemployment correlated strongly with levels of suicides per 100,000 in the period studied (r =0.7359; p=0.01) [Figure about here] There was a statistically significant association between within-region change in unemployment and change in suicide Every 1% increase in unemployment between 2000 and 2010 was associated with an increase of 0.1 suicides per 100,000 (95% CI; 0.0001 to 0.2, p=0.05) When using a 1-year or 2-year lagged measure of unemployment as independent variable there was no significant difference However, there was substantial heterogeneity in terms of change in unemployment and suicide rates across Italian regions The strongest positive correlation between unemployment and suicides was in Sicily (r=0.759) In regions such as Trentino Alto-Adige, however, the association was negative (r=-0.44) I thus investigated the role of social protection identified by previous authors as potential effect modifier for the association between unemployment and health outcomes [8] Data on social protection (per capita expenditure in Euro for social services and benefits delivered by single and associated municipalities in 2008) and social cohesion (percentage of population aged 14 or older who participated to free-of-charge activity for volunteer groups or associations in 2010) came from the SocialCohesion.Stat database, a warehouse of statistics produced by the National Social Security Institute (INPS), the Ministry of Labor and Social Policy and ISTAT [9] As shown in figure 2, the association between unemployment and suicides from 2000 to 2010 was negatively correlated to levels of social protection (r=-0.49; p=0.02) across Italian regions I then added a timeinvariant interaction term between unemployment and social protection and one for unemployment and social cohesion (separately) in the basic model between unemployment and suicides In regions investing more than 135 Euro per capita on social services and benefits, each 1% rise in unemployment was associated with a reduction of 0.41 suicides per 100,000 (95% CI; -0.07 to -0.76, p=0.02) [Figure about here] Results have important implications for research and policy Although various Italian media outlets and even ISTAT claim that suicides have not increased in Italy, our study is in line with international evidence on the adverse effects of the Great Recession on suicides in Europe [7] Findings are also consistent with previous research indicating that strong social protection can mitigate the consequences of economic crises on health outcomes [8] The case of Trentino Alto-Adige, which has the second highest per capita expenditure on social services and benefits and the highest volunteerism rate in Italy, is in line with evidence from Sweden, [7] which showed that investment in social protection can decouple unemployment from suicides As Italy is still struggling with the sixth consecutive year of economic contraction, new policies (alternative to austerity) buffering the victims of the financial crisis from the adverse effects of unemployment, foreclosure and bankruptcy are more urgent than ever Figure Trends in suicides per 100,000 and unemployment rate in 20 Italian regions (2000-2010) Data source: Italian National Institute of Statistics (ISTAT), 2000-2010 Figure Correlation between unemployment and suicides (2000-2010) by per capita expenditure on social services and benefits (2008) in 20 Italian regions Data sources: Unemployment and suicide rates come from the Italian National Institute of Statistics (ISTAT), 2000-2010 Data on per capita expenditure (Euro) for social services and benefits delivered by single and associated municipalities come from socialcohesion.stat database (INSP) Contributors: RDV drafted the paper, analyzed the data and revised the paper Conflicts of Interest Statement: I declare that I have no conflicts of interest Acknowledgments: RDV is supported by a grant from the Economic and Social Research Council (ESRC) (RES-070-27-0034) References: ... I then added a timeinvariant interaction term between unemployment and social protection and one for unemployment and social cohesion (separately) in the basic model between unemployment and suicides. .. statistics produced by the National Social Security Institute (INPS), the Ministry of Labor and Social Policy and ISTAT [9] As shown in figure 2, the association between unemployment and suicides from... austerity) buffering the victims of the financial crisis from the adverse effects of unemployment, foreclosure and bankruptcy are more urgent than ever Figure Trends in suicides per 100,000 and unemployment