Data and diagnostics

This study is based on a cohort of the first 10,000 participants with validated baseline data. The spatial variable was defined as the place of residence (zip-code) of the participants. Included in the study were persons who have lived in their current place of residence for at least ten years to assume a residential effect on NCDs. The city of Hamburg is structured in 104 districts and seven municipalities (administrative structure). To achieve sufficient sample sizes per district, some districts were summarised into clusters with at least 3,000 inhabitants by Erhart et al. (2013) [22], resulting in 68 analysed district clusters in total (Fig 1).

We considered those NCDs with the largest impact on morbidity and mortality, as defined by the WHO [5]. We focused on the six most common diagnoses within the HCHS study population: COPD, CHD, diabetes mellitus (here we combined Type 1 and Type 2 diabetes), heart failure and hypertension. In addition, we chose depression as representative for mental disorders. Hypertension takes on a special role here, as it can be seen as both a disease and a risk factor. Due to the high prevalence, we included hypertension as a NCD. Cancer was however not considered due to the low number of cases at the district cluster level.

Several personal risk factors were selected for further analysis. The selection was based on literature [3, 24] and on the model for classifying risk factors for NCDs [25]. We included socio-demographic variables (age, gender, education) and risk behaviour variables (smoking, alcohol consumption). Educational status was defined according to the International Standard Classification of Education (ISCED) [26] (Table 1). For the regional social conditions in those district clusters, we applied a score developed by Erhart et al. (2013) [22]. The score is based on a principal component analysis and takes into account 26 variables (e.g. education, income). The factor scores were grouped into a three-level form: good, moderate, and poor social conditions (Fig 1).