You are here

Article Text

Article menu

Download PDFPDF

Research reports
Educational inequalities in mortality in northern, mid and southern Italy and the contribution of smoking
  1. Bruno Federico1,2,
  2. Johan P Mackenbach2,
  3. Terje A Eikemo2,
  4. Gabriella Sebastiani3,
  5. Chiara Marinacci4,
  6. Giuseppe Costa5,
  7. Anton E Kunst6
  1. 1Department of Human Sciences, Society and Health, University of Cassino and Southern Lazio, Cassino, Italy
  2. 2Department of Public Health, Erasmus Medical Center, Rotterdam, The Netherlands
  3. 3National Institute of Statistics, Rome, Italy
  4. 4Local Health Unit TO3, Turin, Italy
  5. 5Department of Clinical and Biological Sciences, University of Turin, Turin, Italy
  6. 6Department of Public Health, Academic Medical Center, Amsterdam, The Netherlands
  1. Correspondence to Dr Bruno Federico, Department of Human Sciences, Society and Health, University of Cassino and Southern Lazio, via Sant'Angelo snc, Cassino 03043, Italy; b.federico{at}unicas.it

Abstract

Background Previous studies have shown that mortality inequalities are smaller in Italy than in most European countries. This may be due to the weak association between socioeconomic status and smoking in Italy. However, most published studies were based on data from a single city in northern Italy (Turin). In this study, we aimed to assess the size of mortality inequalities in Italy as a whole, their geographical pattern of variation within Italy, and the contribution of smoking to these inequalities.

Methods Participants in the National Health Interview Survey 1999–2000 were followed up for mortality until 31 December 2007. Using Cox regression, we computed the age-adjusted relative index of inequality (RII) for all-cause mortality with and without controlling for smoking status and intensity. Education was used as an indicator of socioeconomic status.

Results Among 72 762 individuals aged 30–74 years at baseline, 4092 died during the follow-up. The age-adjusted RII of mortality was 1.69 (95% CI 1.44 to 2.00) among men and 1.43 (95% CI 1.13 to 1.82) among women. Among men, inequalities were larger in both northern and southern regions than in the middle of the country, whereas among women they were larger in the south. After controlling for smoking RII decreased to 1.63 (95% CI 1.38 to 1.92) among men and increased to 1.54 (95% CI 1.21 to 1.96) among women. The geographical variation in mortality inequalities was not affected by smoking adjustment.

Conclusions Mortality inequalities in Italy are smaller than in most European countries. This is due, among other factors, to the weak socioeconomic pattern of smoking over the past decades in Italy.

Keywords

https://doi.org/10.1136/jech-2012-201716

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Keywords

    Introduction

    Despite the large reductions in mortality observed over the past decades in industrialised countries,1 persistent inequalities are found within these populations, with lower mortality rates among individuals who are more advantaged in terms of education, occupation and income.2–4

    The size of mortality inequalities vary with time, place and the specific cause of death considered. A recent European comparative study showed that inequalities were particularly large in the east, while they were relatively small in the south.4 On the other hand, inequalities in lung cancer mortality and ischaemic heart disease showed a north-south gradient within Europe, being largest in northern Europe.5 ,6 The socioeconomic pattern of lung cancer mortality was even reversed in Madrid among the elderly.6

    Among those factors that can ‘explain’ inequalities in mortality, an important role is played by health behaviours, although their contribution varies widely across time and place.7–11 Smoking alone contributes to a significant part of the socioeconomic difference in mortality, particularly among men.12 ,13 However, recent studies in France and Spain showed that the contribution of smoking to mortality inequalities in these countries was limited.14 ,15 The delay of the smoking epidemic in southern Europe16 may explain why mortality inequalities in these countries are smaller than in northern European countries.

    With few exceptions,17 previous comparative studies on mortality inequalities were limited by the fact that the only available data for Italy were based on residents of Turin, a large city in the north west. In this urban population, a follow-up study linking mortality records with census data has been carried out since the early 1970s.18 Given that inequalities in mortality are larger in urban than in rural areas,19 mortality inequalities reported for the city of Turin may be larger than those for the whole country.

    Other studies were carried out in parts of northern and central Italy.20–22 Only recently, data on a follow-up study on a large and representative sample of the entire Italian population became available.23 Using these data, Marinacci et al24 showed that Italian individuals in the lowest educational category had a mortality rate for any cause about twice that of individuals in the highest educational category.

    However, it is not known whether the size of inequalities varies within Italy, as suggested by De Vogli et al.25 A significant geographical variability in overall mortality is present, with higher rates in the south, which is the poorest and least developed area of the country.26 In 2008, life expectancy at birth ranged from 77.2 years in Campania (south) to 79.8 in Marche (centre) among men and from 82.6 in Campania to 85.1 in Trentino-Alto Adige (north) among women. This may imply higher inequalities in the south than in the north-centre.

    With regards to mortality inequalities, regional differences within Italy may result from the socioeconomic patterning of smoking among women over the past decades. Between 1980 and 2000, inequalities in smoking among women were larger in the north and smaller in the south.27 This pattern may result in smaller inequalities in mortality among women in southern Italy than in the rest of the country. To our knowledge, no previous study compared the size of mortality inequalities between northern, mid and southern Italy. Such a comparison would show whether the longitudinal study of Turin, from which most published studies on mortality inequalities were derived, provides representative estimates of the size of inequalities for Italy as a whole.

    In this study, we first aimed at estimating both the relative and the absolute magnitude of educational inequalities in overall mortality in Italy. Second, we aimed at assessing differences in the size of mortality inequalities between northern, central and southern regions. A third specific objective was to assess the contribution of smoking to educational inequalities in overall mortality in Italy as a whole and separately by geographical area.

    Methods

    Data sources

    We used data from the Italian Longitudinal Study. This is a follow-up study linking participants of the 1999–2000 National Health Interview Survey (NHIS) with mortality and hospital records.23 Sociodemographic data and information about smoking behaviour were collected in the NHIS, a cross-sectional survey that was carried out by the National Institute of Statistics in 1999–2000. This survey was based on a representative sample of the Italian non-institutionalised population (n=140 011), using both interviews and a self-compiled questionnaire. Non-response was 13%.

    Information about the outcome was obtained by means of deterministic data linkage with the national archive of mortality records. Using unique individual codes, subjects who participated in the 1999–2000 NHIS were linked to the mortality record archive for the period 1999–2007.

    Study variables

    Data on smoking status (never, current and former smoker), number of cigarettes smoked per day, and age at smoking cessation were collected. These data were combined to classify subjects in either one of the following categories: never smoker, ‘heavy’ current smoker (those who smoked ≥20 cigarettes/day), ‘light’ current smoker (those who smoked <20 cigarettes/day), recent quitter (those who quit <10 years before the survey), long-term quitter (those who quit ≥10 years before the survey). Occasional smokers (n=6004), as well as pipe and cigar smokers (n=532) were classified as never smokers. A dichotomous variable (ever/never smoker) was used for descriptive purposes, whereas the smoking variable with five categories was used in regression modelling (see below).

    Education was used as an indicator of socioeconomic position, and it was categorised in four classes (primary, lower secondary, upper secondary and tertiary) on the basis of the highest level achieved.

    Subjects were classified into three macro-regions (north, centre and south), according to their region of residence. This classification corresponds to the following first-level codes of the nomenclature of territorial units for statistics: ITC and ITD (north), ITE (centre), ITF and ITG (south).

    Study population

    Included in the analyses were individuals between 30 and 74 years of age at the time of the interview in 1999–2000. Individuals from the provinces of Trento and Bolzano were excluded because of incomplete personal identification data. Approximately 8% of the remaining 78 000 individuals were additionally excluded because they had missing or incorrect identification codes, leading to a final sample of 72 762 subjects.

    Statistical analyses

    We estimated the size of educational inequalities in mortality in both relative and absolute terms. As a measure of relative inequality, we computed the relative index of inequality (RII).28 This is a regression-based measure of inequality that takes into account the distribution of the population according to the socioeconomic indicator of choice. It expresses the ratio of mortality for an individual at the bottom of the social ladder, compared to an individual at the top of the ladder. In order to compute the RII, we assigned a new index of education to each subject, which was equal to the proportion of the total population with an educational level higher than that of the subject. This index, which ranges between 0 (for a person at the top of the social scale) and 1 (for a person at the bottom), was then entered in a Cox regression model. The HR for this continuous variable estimates the RII. This methodology was previously used by Marmot et al.29

    In each Cox regression model, age at baseline was entered as a continuous variable. Subjects alive at the end of follow-up (31 December 2007) were treated as censored. The proportional hazard assumption was tested using the global test based on Schoenfeld residuals.30 The likelihood ratio test was used to assess whether there was significant heterogeneity among regions (north, centre, south and main islands) in the RII, comparing a model with to a model without the interaction term between education and region.

    The contribution of smoking to mortality inequalities was assessed by comparing a model without smoking with one that included smoking. Smoking was entered into the model using four dummy variables, with ‘never smoker’ as the reference category. The percentage change in the coefficient for education between the two models (with and without smoking) was computed. Bootstrap analysis was performed in order to obtain 95% CI for the percentage change. One thousand replications were performed.

    In addition to the RII, we computed the slope index of inequality (SII), which is an absolute measure of inequality. This index was calculated as follows: SII=2 × m × (RII − 1)/(RII+1), where m is the age-standardised mortality rate (per 1000 person-years).4

    All analyses were stratified by sex. The statistical software Stata V.11 was used.

    Results

    The study sample had an almost equal representation of men and women (table 1). During the 570 461 person-years of follow-up 4092 subjects died, with approximately three-quarters of deaths occurring in the age group 60–74 years.

    View this table:
    Table 1

    Descriptive statistics of the sample (n=72 762)

    Subjects from southern regions had on average a lower educational level than those from the rest of the country (table 2). Among men, smoking patterns were similar between regions, with more than half of male subjects being either a smoker or a former smoker. Ever smoking was negatively associated with educational level. Rates of ever smoking were lower among women than among men, and they were higher in the north-centre than in the south. In all regions, the association between ever smoking and education was positive among women. Among both sexes, the proportion of heavy current smokers was larger among the low-educated than among the high-educated (data not shown).

    View this table:
    Table 2

    Distribution (%) of respondents and age-standardised rates (%) of ever smokers by sex and educational level in north, central and south Italy

    Mortality rates are shown in table 3, separately for men and women, together with the SII. A clear educational gradient in mortality was found among men, with mortality rates decreasing as educational status increased. Absolute inequalities were larger in both northern and southern regions than in central Italy. Contrary to men, educational differences in mortality did not show a regular gradient among women. The highest mortality rates were observed among the lowest educated subjects, as expected, but the lowest rates were observed among subjects with lower and upper-secondary education. Women in the south had the highest absolute inequalities.

    View this table:
    Table 3

    Age-standardised mortality rates (per 1000 person-years) for any cause by sex and educational group in north, central and south Italy, and SII

    RII in mortality are shown in table 4, together with 95% CI, with and without adjustment for smoking. In Italy, the age-adjusted RII of mortality was 1.69 (95% CI 1.44 to 2.00) among men and 1.43 (95% CI 1.13 to 1.82) among women. Among men, RII were larger in both northern and southern regions than in central regions (p value of the likelihood ratio test 0.097), whereas among women they were larger in the south than in the rest of the country (p=0.015).

    View this table:
    Table 4

    Relative index of inequality with 95% CI for overall mortality, with and without adjustment for smoking

    Smoking contributed little to relative inequalities in mortality, because both the size and the geographical pattern of RII were not substantially altered by adjustment for smoking. However, the effect of smoking adjustment differed by sex: RII decreased to 1.63 (95% CI 1.38 to 1.92) among men, which corresponds to an 8% reduction on the log scale, whereas it increased to 1.54 (95% CI 1.21 to 1.96) among women, which corresponds to a 21% increase on the same scale. Interaction terms between education and the smoking dummies were then entered into the models in order to assess whether the effect of education differed by smoking category, but none of these terms reached conventional statistical significance (results not shown).

    Discussion

    Summary of the main results

    In this study, we found significant educational inequalities in mortality for both sexes. Compared to the rest of the country, mortality inequalities were larger in both northern and southern regions among men and in southern regions among women. After smoking adjustment, mortality inequalities decreased among men, whereas they increased among women.

    Evaluation of data problems

    We analysed a large and representative sample of the entire national population. Approximately 129 000 subjects were followed up, which corresponds to 92% of the sample of the 1999–2000 NHIS. The remaining subjects could not be followed over time because they had missing or incomplete codes. Completeness of follow-up was not related to education.24 Furthermore, the observed mortality experience of the sample was shown to be very similar to the mortality of the general Italian population over the same time period (Marinacci et al, submitted for publication).

    A limitation of this study is the use of a single indicator of socioeconomic status. Although education is a widely applied indicator of socioeconomic position, other mortality studies used occupation, either as a single indicator3 or in combination with other indicators.14 Occupation reflects social standing and prestige and it is also correlated with income. However, it does not easily allow classifying subjects who are retired, unemployed or housewives, whereas education permits the classification of all individuals in the population. Importantly, the choice of education allowed us to replicate the methodology adopted in a recent paper,4 and to make comparisons with the size of mortality inequalities existing in other European countries (see below).

    Another limitation of this study is the use of self-reported information on smoking, which may be inaccurate.31 Non-differential misclassification may have resulted in an underestimation of the effect of smoking on educational inequalities in mortality. A Finnish study found no evidence that misclassification of smoking is related to education.32 If differential misclassification occurred in the Italian NHIS, then our estimates of the contribution of smoking to mortality inequalities would be biased.

    Comparison with other studies and explanation of the results

    In most European comparative studies, estimates of mortality inequalities for Italy were derived from the city of Turin. These studies showed that inequalities in mortality were smaller in Turin than in most European countries.2 ,4 ,33 Data for Italy as a whole were reported in an earlier study, which found that around 1980, mortality inequalities in Italy were larger than elsewhere.17 However, this latter study is affected by severe loss in the follow-up.34 In the present study, we applied the same methodology used in a recent European comparative study.4 In figure 1, the RII estimated in our study are placed next to the RII estimates reported in the above-mentioned European comparative study. The figures clearly support the previous suggestion that inequalities in mortality were smaller in Italy than in northern and eastern Europe. The size of mortality inequalities in Italy are similar to those reported in other Mediterranean populations with longitudinal studies among men. The Turin longitudinal study provides estimates of inequalities that are very close to those found for the entire country, with the exception of women, for whom inequalities may be underestimated, compared to those for Italy as a whole.

    Figure 1
    Figure 1

    Comparison between the age-adjusted relative index of inequality (RII) of overall mortality estimated in this study (right) and those reported by Mackenbach et al4 (left) for men and women in 16 European countries.

    Smoking alone contributes significantly to inequalities in mortality,13 but this contribution varies widely over time and place. Adjustment for smoking reduced mortality inequalities by 32% in England, but only by 4% in France.14 In Spain, adjustment for smoking slightly decreased mortality inequalities in men while it increased mortality inequalities among women,15 a finding remarkably similar to that reported in this study. In Italy, Spain and other southern European countries, the association between a low socioeconomic status and smoking was much weaker than in northern European countries among men, and it was even reversed among women.35 Therefore, weak or even reversed socioeconomic inequalities in smoking in Italy contributed to its smaller inequalities in mortality.

    A nationwide study in Italy reported a north–south gradient in inequalities in smoking among women, with smaller inequalities in the south than in the north.27 Despite their smaller inequalities in smoking, women in southern regions showed the largest inequalities in mortality. This suggests that factors other than smoking shaped the geographical pattern of inequalities in mortality within Italy.

    Other risk factors for disease may also contribute to the smaller size of mortality inequalities reported in Italy. Contrary to northern Europe, dietary habits do not show large differences by educational level.36 On the other hand, inequalities in overweight/obesity, physical inactivity and heavy alcohol consumption are larger in Italy than in most European countries.37 ,38 Inequalities in alcohol consumption are reflected by inequalities in mortality from cancer of the upper aerodigestive tract and the liver, which, on the basis of data from the Turin study, were among the highest in Europe.39 Studies carried out in Brianza (northern Italy) showed a strong negative association between education and obesity among women,40 as well as a more favourable pattern of alcohol use and high-density lipoprotein cholesterol among the low-educated men,41 but it is unknown whether these socioeconomic patterns also apply to other regions.

    More upstream determinants of mortality inequalities include welfare policies, which are less generous in Italy than in Nordic countries. Despite the fact that healthcare is universally guaranteed and free, inequalities in access to adequate healthcare were reported.42 ,43 On the other hand, family support and informal social networks are widespread in this country, and they may mitigate inequalities in access to state-organised welfare services.44

    Another explanation for the smaller inequalities in mortality observed in Italy compared to other European countries is the fact that educational attainment might be less important as a social stratifier than in central and northern Europe. The proportion of the adult population with a low education is larger in Italy than in most other European countries.45 Therefore, lower educated individuals may include those with widely different levels of material wealth. However, it was reported that education is still a strong predictor of social class in Italy.46

    Among women, large inequalities in mortality were observed in the south, with values similar to those reported for eastern European countries. On the other hand, there were no sizable inequalities in mortality for women in northern Italy. Material living circumstances and the quality of social and healthcare services may be important contributors to the geographical heterogeneity in mortality inequalities in Italy. Deprived living conditions largely explain why poor self-reported health is more frequently reported in southern regions than in the rest of Italy.47 Economic circumstances, including employment opportunities, are markedly less favourable in the south than in the centre-north. Health and social systems of northern regions constantly top the national ranking, according to residents’ opinions. In addition, patients’ mobility patterns suggest that the quality of hospital care is generally higher in northern and central regions than in the south.48 It is possible that lower educated individuals of southern regions may not afford to travel or they may not have the resources to follow the appropriate medical path. In particular, this may apply to women, who have lower economic and cultural resources than men.49

    Conclusions

    This study provided further evidence that mortality inequalities are smaller in Italy than in most European countries. This is probably due, among other factors, to the weak socioeconomic pattern of smoking over the past decades in Italy. On the other hand, the larger inequalities in mortality observed in Italy in southern regions, especially among women, call for other explanations, such as other health behaviours, inequalities in working and living conditions, and differential access to social and healthcare services.

    What is already known on this subject?

    • Previous comparative studies have shown that mortality inequalities are smaller in Italy than in most European countries. However, most of these studies were based on data from a single city in northern Italy (Turin) and not on data from the whole country.

    • The association between socioeconomic status and smoking has been weaker in Italy than in other European countries for the past decades, which may contribute to explain the smaller inequalities in mortality.

    What this study adds?

    • Previous estimates of inequalities in mortality for Italy, based on the Turin longitudinal study, approximate the size of inequalities reported in the present study.

    • We showed that inequalities in mortality were significantly larger in southern regions than in the rest of the country among women and in both southern and northern regions among men.

    • Smoking contributed little to mortality inequalities among men and even attenuated mortality inequalities among women.

    Acknowledgments

    The authors would like to thank Gaetano Roscillo for editorial assistance in the preparation of the manuscript and Giovanni Capelli for performing bootstrap analyses.

    References

    1. ↵
      1. Oeppen J,
      2. Vaupel JW
      . Demography. Broken limits to life expectancy. Science 2002;296:1029–31.
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Mackenbach JP,
      2. Kunst AE,
      3. Cavelaars AE,
      4. et al
      . Socioeconomic inequalities in morbidity and mortality in western Europe. The EU Working Group on Socioeconomic Inequalities in Health. Lancet 1997;349:1655–9.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Kunst AE,
      2. Groenhof F,
      3. Mackenbach JP
      . Mortality by occupational class among men 30–64 years in 11 European countries. EU Working Group on Socioeconomic Inequalities in Health. Soc Sci Med 1998;46:1459–76.
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Mackenbach JP,
      2. Stirbu I,
      3. Roskam AJ,
      4. et al
      . Socioeconomic inequalities in health in 22 European countries. N Engl J Med 2008;358:2468–81.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Avendano M,
      2. Kunst AE,
      3. Huisman M,
      4. et al
      . Socioeconomic status and ischaemic heart disease mortality in 10 western European populations during the 1990s. Heart 2006;92:461–7.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Mackenbach JP,
      2. Huisman M,
      3. Andersen O,
      4. et al
      . Inequalities in lung cancer mortality by the educational level in 10 European populations. Eur J Cancer 2004;40:126–35.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Ernstsen L,
      2. Bjerkeset O,
      3. Krokstad S
      . Educational inequalities in ischaemic heart disease mortality in 44,000 Norwegian women and men: the influence of psychosocial and behavioural factors. The HUNT Study. Scand J Public Health 2010;38:678–85.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Laaksonen M,
      2. Talala K,
      3. Martelin T,
      4. et al
      . Health behaviours as explanations for educational level differences in cardiovascular and all-cause mortality: a follow-up of 60 000 men and women over 23 years. Eur J Public Health 2008;18:38–43.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Stringhini S,
      2. Sabia S,
      3. Shipley M,
      4. et al
      . Association of socioeconomic position with health behaviors and mortality. JAMA 2010;303:1159–66.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Kilander L,
      2. Berglund L,
      3. Boberg M,
      4. et al
      . Education, lifestyle factors and mortality from cardiovascular disease and cancer. A 25-year follow-up of Swedish 50-year-old men. Int J Epidemiol 2001;30:1119–26.
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Pekkanen J,
      2. Tuomilehto J,
      3. Uutela A,
      4. et al
      . Social class, health behaviour, and mortality among men and women in eastern Finland. BMJ 1995;311:589–93.
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Popham F
      . To what extent can deprivation inequalities in mortality and heart disease incidence amongst the working aged in Scotland be explained by smoking? Relative and absolute approaches. Health Place 2011;17:1132–6.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Jha P,
      2. Peto R,
      3. Zatonski W,
      4. et al
      . Social inequalities in male mortality, and in male mortality from smoking: indirect estimation from national death rates in England and Wales, Poland, and North America. Lancet 2006;368:367–70.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Stringhini S,
      2. Dugravot A,
      3. Shipley M,
      4. et al
      . Health behaviours, socioeconomic status, and mortality: further analyses of the British Whitehall II and the French GAZEL prospective cohorts. PLoS Med 2011;8:e1000419.
      OpenUrlCrossRefPubMed
    15. ↵
      1. Regidor E,
      2. Kunst AE,
      3. Rodriguez-Artalejo F,
      4. et al
      . Small socio-economic differences in mortality in Spanish older people. Eur J Public Health 2012;22:80–5.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Cavelaars AE,
      2. Kunst AE,
      3. Geurts JJ,
      4. et al
      . Educational differences in smoking: international comparison. BMJ 2000;320:1102–7.
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Kunst AE,
      2. Mackenbach JP
      . The size of mortality differences associated with educational level in nine industrialized countries. Am J Public Health 1994;84:932–7.
      OpenUrlPubMedWeb of Science
    18. ↵
      1. Costa G,
      2. Cardano M,
      3. Demaria M
      . Storie di salute in una grande città, Torino. Torino: Osservatorio Socio-economico Torinese. Ufficio Statistico Città di Torino, 1998.
    19. ↵
      1. Senior M,
      2. Williams H,
      3. Higgs G
      . Urban-rural mortality differentials: controlling for material deprivation. Soc Sci Med 2000;51:289–305.
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Michelozzi P,
      2. Perucci CA,
      3. Forastiere F,
      4. et al
      . Inequality in health: socioeconomic differentials in mortality in Rome, 1990–95. J Epidemiol Community Health 1999;53:687–93.
      OpenUrlAbstract
    21. ↵
      1. Merler E,
      2. Benvenuti A,
      3. Baldi P,
      4. et al
      . Socioeconomic inequalities in health in the Tuscany Longitudinal Study (SLTO): persistence and changes over time in overall mortality and selected causes (lung cancer, liver cirrhosis, AIDS and overdose). Epidemiol Prev 1999;23:207–14.
      OpenUrlPubMed
    22. ↵
      1. Candela S,
      2. Cavuto S,
      3. Luberto F
      . Condizioni socioeconomiche e mortalità nella popolazione di Reggio Emilia. Azienda Unità Sanitaria Locale di Reggio Emilia. 2005. http://www.epicentro.iss.it/territorio/CondizioniocioeconomicheMortalit%C3%A0_ReggioEmilia.pdf (accessed 14 Jun 2012).
    23. ↵
      1. Marinacci M,
      2. Sebastiani G,
      3. Demaria M,
      4. Caranci N,
      5. Proietti G,
      6. Pappagallo M,
      7. et al
      . Disuguaglianze sociali nella mortalità di un campione della popolazione italiana. 2010. http://www.dors.it/alleg/0200/venerdi_6a_marinacci.pdf (accessed 14 Jun 2012).
    24. ↵
      1. Marinacci C,
      2. Sebastiani G,
      3. Ferracin E,
      4. et al
      . Distance among Italian people through geographic and socio-economic inequalities. Epidemiol Prev 2011;35(Suppl. 2):S68–75.
      OpenUrl
    25. ↵
      1. De Vogli R,
      2. Gimeno D,
      3. Kivimaki M
      . Socioeconomic inequalities in health in 22 European countries. N Engl J Med 2008;359:1290–1.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      ISTAT. Mortality: social and demographic characteristics. 2011. http://www.istat.it/it/archivio/24843 (accessed 5 Sep 2011).
    27. ↵
      1. Federico B,
      2. Kunst AE,
      3. Vannoni F,
      4. et al
      . Trends in educational inequalities in smoking in northern, mid and southern Italy, 1980–2000. Prev Med 2004;39:919–26.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Mackenbach JP,
      2. Kunst AE
      . Measuring the magnitude of socio-economic inequalities in health: an overview of available measures illustrated with two examples from Europe. Soc Sci Med 1997;44:757–71.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Marmot MG,
      2. Shipley MJ,
      3. Hemingway H,
      4. et al
      . Biological and behavioural explanations of social inequalities in coronary heart disease: the Whitehall II study. Diabetologia 2008;51:1980–8.
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Therneau TM,
      2. Grambsch PM,
      3. Fleming TR
      . Martingale-based residuals for survival models. Biometrika 1990;77:147–60.
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Patrick DL,
      2. Cheadle A,
      3. Thompson DC,
      4. et al
      . The validity of self-reported smoking: a review and meta-analysis. Am J Public Health 1994;84:1086–93.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Vartiainen E,
      2. Seppala T,
      3. Lillsunde P,
      4. et al
      . Validation of self reported smoking by serum cotinine measurement in a community-based study. J Epidemiol Community Health 2002;56:167–70.
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Huisman M,
      2. Kunst AE,
      3. Andersen O,
      4. et al
      . Socioeconomic inequalities in mortality among elderly people in 11 European populations. J Epidemiol Community Health 2004;58:468–75.
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Meneghel G
      . La mortalité en Italie selon des facteurs socio-économiques. Espace Populations Sociétés 1991;9:173–80.
      OpenUrlCrossRef
    35. ↵
      1. Giskes K,
      2. Kunst AE,
      3. Benach J,
      4. et al
      . Trends in smoking behaviour between 1985 and 2000 in nine European countries by education. J Epidemiol Community Health 2005;59:395–401.
      OpenUrlAbstract/FREE Full Text
    36. ↵
      1. Prattala R,
      2. Hakala S,
      3. Roskam AJ,
      4. et al
      . Association between educational level and vegetable use in nine European countries. Public Health Nutr 2009;12:2174–82.
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. Roskam AJ,
      2. Kunst AE,
      3. Van Oyen H,
      4. et al
      . Comparative appraisal of educational inequalities in overweight and obesity among adults in 19 European countries. Int J Epidemiol 2010;39:392–404.
      OpenUrlAbstract/FREE Full Text
    38. ↵
      1. Bloomfield K,
      2. Grittner U,
      3. Kramer S,
      4. et al
      . Social inequalities in alcohol consumption and alcohol-related problems in the study countries of the EU concerted action ‘Gender, Culture and Alcohol Problems: a Multi-national Study’. Alcohol Alcohol 2006;41(Suppl. 1):i26–36.
      OpenUrlAbstract/FREE Full Text
    39. ↵
      1. Menvielle G,
      2. Kunst AE,
      3. Stirbu I,
      4. et al
      . Socioeconomic inequalities in alcohol related cancer mortality among men: to what extent do they differ between Western European populations? Int J Cancer 2007;121:649–55.
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      1. Ferrario M,
      2. Sega R,
      3. Chatenoud L,
      4. et al
      . Time trends of major coronary risk factors in a northern Italian population (1986–1994). How remarkable are socioeconomic differences in an industrialized low CHD incidence country? Int J Epidemiol 2001;30:285–97.
      OpenUrlAbstract/FREE Full Text
    41. ↵
      1. Veronesi G,
      2. Ferrario MM,
      3. Chambless LE,
      4. et al
      . Gender differences in the association between education and the incidence of cardiovascular events in Northern Italy. Eur J Public Health 2011;21:762–67.
      OpenUrlAbstract/FREE Full Text
    42. ↵
      1. Materia E,
      2. Spadea T,
      3. Rossi L,
      4. et al
      . Health care inequalities: hospitalization and socioeconomic position in Rome. Epidemiol Prev 1999;23:197–206.
      OpenUrlPubMed
    43. ↵
      1. Ancona C,
      2. Agabiti N,
      3. Forastiere F,
      4. et al
      . Coronary artery bypass graft surgery: socioeconomic inequalities in access and in 30 day mortality. A population-based study in Rome, Italy. J Epidemiol Community Health 2000;54:930–5.
      OpenUrlAbstract/FREE Full Text
    44. ↵
      1. Lindstrom M
      . Social capital and health-related behaviors. In: Kawachi I, Subramanian SV, Kim D, eds. Social capital and health. Boston, MA: Springer, 2008:215–38.
    45. ↵
      European Commission. Database: Adult Education Survey 2007. http://epp.eurostat.ec.europa.eu/portal/page/portal/education/data/database (accessed 14 Jun 2012).
    46. ↵
      1. Schizzerotto A,
      2. Barone C
      . Sociologia dell'istruzione. Bologna: Il Mulino, 2006.
    47. ↵
      1. Franzini L,
      2. Giannoni M
      . Determinants of health disparities between Italian regions. BMC Public Health 2010;10:296.
      OpenUrlCrossRefPubMed
    48. ↵
      1. Lo Scalzo A,
      2. Donatini A,
      3. Orzella L,
      4. et al
      . Italy. Health Systems in Transition. European Observatory on Health Systems and Policies; 2009.http://www.euro.who.int/__data/assets/pdf_file/0006/87225/E93666.pdf (accessed 14 Jun 2012).
    49. ↵
      1. Costa G,
      2. Marinacci C,
      3. Caiazzo A,
      4. et al
      . Individual and contextual determinants of inequalities in health: the Italian case. Int J Health Serv 2003;33:635–67.
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Contributors BF, JPM, TAE and AEK had the original idea and they devoloped the study protocol. BF performed data analysis and drafted the manuscript. CM, GS and GC contributed to data collection. All authors contributed to the preparation of the manuscript and read and approved the final version.

    • Competing interests None.

    • Ethics approval This study is part of the National Statistical Programme 2011–2013, which was approved by the Italian Data Protection Authority.

    • Provenance and peer review Not commissioned; externally peer reviewed.