Associations between community contextual factors and depressive symptoms: postcode-based data linkage of POKAL survey data with publicly available German routine data.
Depression prevalence varies across and within geographical regions, with contextual factors potentially playing a relevant role. Therefore, we assessed how informative a postcode-based data linkage approach is to study the association between contextual factors and depressive symptoms in Germany.
In this cross-sectional study, we linked individual-level survey data from a multi-setting clinical sample (n = 2,323) with population-level routine data from two different German datasets to assess depression severity. Contextual data included area-level deprivation, civic participation, population structure and the built environment.
Including contextual factors in the models increased their model fit (AICc Model 2 = 13,640.80) compared to the demographic factors-only model (AICc Model 1 = 13,718.24; Δ AICc = 77.18). Assessing each contextual factor alone revealed that area-level deprivation yielded the most parsimonious fit.
The data linkage approach is a technically viable and resource-friendly approach within the German mental health research infrastructure. In this multi-setting clinical sample, incorporating objective contextual variables improved the models. However, choosing the appropriate spatial resolution is essential, as coarse aggregation may obscure meaningful contextual associations with depression severity. To ensure the linkage is meaningful, administrative indicators must be carefully chosen based on existing literature to guarantee they are proxies of environmental factors that are associated with mental health.
In this cross-sectional study, we linked individual-level survey data from a multi-setting clinical sample (n = 2,323) with population-level routine data from two different German datasets to assess depression severity. Contextual data included area-level deprivation, civic participation, population structure and the built environment.
Including contextual factors in the models increased their model fit (AICc Model 2 = 13,640.80) compared to the demographic factors-only model (AICc Model 1 = 13,718.24; Δ AICc = 77.18). Assessing each contextual factor alone revealed that area-level deprivation yielded the most parsimonious fit.
The data linkage approach is a technically viable and resource-friendly approach within the German mental health research infrastructure. In this multi-setting clinical sample, incorporating objective contextual variables improved the models. However, choosing the appropriate spatial resolution is essential, as coarse aggregation may obscure meaningful contextual associations with depression severity. To ensure the linkage is meaningful, administrative indicators must be carefully chosen based on existing literature to guarantee they are proxies of environmental factors that are associated with mental health.
Authors
Schoenweger Schoenweger, Sachse Sachse, Brisnik Brisnik, Bühner Bühner, Ditzen-Janotta Ditzen-Janotta, Dreischulte Dreischulte, Eder Eder, Falkai Falkai, Gensichen Gensichen, Gökce Gökce, Haas Haas, Henningsen Henningsen, Junker Junker, Krcmar Krcmar, Lukaschek Lukaschek, Pfeiffer Pfeiffer, Pitschel-Walz Pitschel-Walz, Schneider Schneider, Schillok Schillok, Teusen Teusen, von Schrottenberg von Schrottenberg, Jung-Sievers Jung-Sievers
View on Pubmed