With yearly rates ranking clearly above world average in Europe, suicide constitutes a substantial public health problem. Because of that, prevention has become a major concern for German mental health institutions. A requirement for successful prevention strategies is to address all key factors that contribute to suicidality. It is highly relevant in this respect that suicidal behaviour itself exhibits a social gradient: drawing on the relevant literature, low socioeconomic status (SES) and a high extent of social isolation (SI) are related to increased suicide risks (Lorant et al. 2005; Li et al. 2011; Qin et al. 2003; Agerbo et al. 2007). The purpose of this study was therefore to add to these findings and to further investigate associations of SES and SI with suicide in order to define starting points for public health interventions. It was consequently hypothesized that lower individual levels of SES and higher individual levels of SI are correlated with increased suicide rates. SI potentially compromises the perception of social support in stressful live events associated with low SES (Cohen et al. 2006; Kumari et al. 2010). Since such life events correlate with suicidal behavior (Beautrais et al. 1997; Cohen et al. 2019), the effects of low SES were further hypothesized to be aggravated in individuals with high SI levels (SES x SI interaction).
In order to test the hypotheses, all 149.033 suicide deaths between 1997 and 2010 (T = 14 years) were extracted from the official German death record as coded by ICD categories E950 - E959 for 1997 and X60 - X84 for the years from 1998 onwards, respectively. Information on SES and SI was gained by merging the dataset with Germany’s main household survey, i.e. the Microcensus. In accordance with the existing literature, established indexes on occupational status (ISEI, Ganzeboom & Treiman 1996) and educational achievements (CASMIN, König et al. 1988) were applied as well as items on income, minor employment, unemployment, the number of received public transfers and the reception of social bene fits due to unemployment (ALG I/II) in order to capture SES. SI was proxied with variables measuring single marital status, living in a one-person-household and relocations throughout the year before the survey was conducted.
Due to German data protection regulations that do not permit the analysis of death record data based on individual level information, suicide deaths were examined as aggregated rates at the level of N = 390 administrative districts. In order to deal with two problems associated with this kind of statistical analysis, Prentice and Sheppard’s model for aggregate data (1995) was applied accounting for potential estimation biases due to differences in baseline suicide rates between districts and between time periods. The model specification further corrected for spatial effect correlations. An important limitation to this procedure is that the estimates represent a blend of effects at the individual and district levels. However, an adequate solution is only available through the application of individual level data.
The statistical analysis turned out the following results: The positive effect on suicide rates of unemployment and the negative effect of income as two out of seven SES proxies and the positive effect of living in a one-person-household as one out of three SI proxies validate the proposed hypotheses on the relations of SES and SI with suicide rates. Confirming the hypothesis on SI mediating SES effects, the model revealed positive effects on suicide rates of income decreases in single individuals. Likewise, we observed positive effects on district suicide rates for decreasing levels of CASMIN in district population shares who had relocated throughout the past year. In
contradiction to the theoretical claims, however, increases in CASMIN scores were found to result in positive effects on suicide rates just as a history of relocation prior to suicide was related to decreasing suicide rates. Furthermore, decreases in income were found to result in negative effects on suicide rates in the district population of persons living in a one-person-household.
The results indicating associations of SES and SI with increases in district suicide rates represent appropriate starting points for the definition of suicide prevention strategies. Thus, particularly the unemployed, individuals with low incomes, persons living in one-person-households and relocated individuals with lower educational levels should be targeted by public health interventions. Moreover, the observations of the present study clearly demonstrate the significance of longitudinal individual level data for public health policies. Respective research incorporating such data would permit a better understanding of the causal mechanisms resulting in suicidality and help to further investigate the robustness of the shown results. By this means, prevention
strategies could be better adapted to the specfic needs of the individuals under concern. Regarding the findings contradicting the theoretical claims, it needs to be mentioned that associations of low SES and high SI levels with increases in suicide risks can not be ruled out at the individual level. Rather, the observed inconsistent effects might be attributable to differences in district compositions than to differences in characteristics of the respective subjects. Also a statistical separation of compositional effects from effects of individual traits would be made possible by including individual level data in future work.:Abbrevations II
Tables II
1 Introduction 1
1.1 Suicide - A Global Health Burden 1
1.2 Risk Factors and Etiology of Suicide 1
1.3 Suicide Prevention 2
1.4 Social Disparities in Suicide 2
1.4.1 Socioeconomic Status 2
1.4.2 Social Isolation 3
1.4.3 Health Inequalities and Health Inequities 4
1.4.4 Causation and Selection 5
1.4.5 Individual Life Courses 7
1.5 Stress and Diathesis 8
1.5.1 Critical Life Events 9
1.6 Neurobiological Correlates of Suicidality 9
1.6.1 Neurobiological Correlates of SES and SI 10
1.7 SES, SI and Social Support 11
1.8 Aims of the Thesis 11
1.9 Methods 12
2 Original Publication 14
Summary 23
References 26
Supplementary Materials - Further Statistical Tests & Models 41
Structural Breaks in Suicide Numbers 41
Age- and Gender-Adjustment of District Suicide Rates 42
Alternate Model Specifications
Anlagen i
Erklärung über die eigenständige Abfassung der Arbeit i
Spezifizierung des eigenen wissenschaftlichen Beitrags iii
Danksagung iii
| Identifer | oai:union.ndltd.org:DRESDEN/oai:qucosa:de:qucosa:72693 |
| Date | 04 November 2020 |
| Creators | Näher, Anatol-Fiete |
| Contributors | Universität Leipzig |
| Source Sets | Hochschulschriftenserver (HSSS) der SLUB Dresden |
| Language | English |
| Detected Language | English |
| Type | info:eu-repo/semantics/acceptedVersion, doc-type:doctoralThesis, info:eu-repo/semantics/doctoralThesis, doc-type:Text |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | https://doi.org/10.3389/fpsyt.2019.00898 |
Page generated in 0.0028 seconds