Addi van Bergen

Chapter 7 146 Finally, in chapter 6, we explored possible applications of the SEI-HS. We tested the SE, measured with the SEI-HS, against four traditional social stratifiers (low education, low income, low labour market position and non-Western migration background) in terms of their ability to identify high-risk/high need population segments. We compared the relative risks (RR) and (hypothetical) population attributable fractions (PAF) for cardiovascular risk, cancer, low self-rated health, anxiety and depression symptoms, and low personal control and studied their overlap and their combined effect. Data from the G4 PHM 2016 were used for this study (N=33,285). The analyses showed significant associations of SE with all health indicators and personal control, with particularly strong RRs for anxiety and depression symptoms (7.95) and low personal control (6.36). The corresponding PAFs were 42% and 35%, respectively. The SEI-HS was significantly better at identifying population segments with anxiety and depression symptoms and low personal control than the four traditional stratifiers were and performed equally well in identifying other health problems. The combination of SE and low labour market position proved to be most impactful: this population segment accounted for 67% of all adults exhibiting anxiety and depression symptoms and 60% of all those exhibiting low personal control, as well as substantial proportions of the other health indicators, while making up only 19.5% of the adult population in the G4. These findings have implications for health care practice, public health and social interventions in large cities. GENERAL DISCUSSION To date, there has not been a generally accepted measure of SE in national and international public health research and monitoring. This dissertation presents the construction and validation of a short index for SE suitable for embedding in the Dutch PHM. The SEI-HS makes use of items that are already present in the PHM, i.e., on loneliness, social capital, financial situation and housing. The number of additional items to be included is thus limited, and by preventing overlap, respondent acceptance is not jeopardised. The psychometric measurement properties of the SEI-HS were found to be satisfactory to good, although there is still some room for improvement in one of the four dimension scales, i.e., the Normative Integration scale (Chapter 3). In this dissertation, we show evidence for the association between high SE/low SI and adverse mental health outcomes (Chapter 2). Where other researchers became stuck in a forest of divergent ideas, a lack of definitions and different ways of measuring SE, by delineating the concept of SE, we were able to synthesise the existing evidence on the relation between SE and health. The findings from our own research confirm the association between SE and adverse mental health and strengthen the plausibility of the association between SE and adverse physical health, i.e., with severe functional limitations (Chapters 3 and 4) and diabetes, high blood pressure, obesity and cancer (Chapter 6).