Integrating health effects within an agent-based land use and transport model

Abstract

Transport is a determinant of urban health, and the past two decades have seen a rise in transport health impact assessment (HIA) tools. However, a common limitation of HIAs is the use of aggregation which ignores variations over space, time, and/or demographics. Furthermore, HIAs generally treat travel behaviour exogenously, meaning they can assess the effects of hypothetical behaviour shifts but not the policies that cause behaviour to change. This article presents an open-source agent-based integrated land use, transport, and health impacts model that assesses health exposures and impacts related to physical activity, air pollution, and traffic injury risk. The agent-based framework allows exposures and impacts to be assessed with high spatiotemporal and sociodemographic resolution. To achieve a tight integration, the model adapts (traditionally car-oriented) transport modelling methods to capture health-relevant behaviours such as active travel more realistically. In addition, exposure modelling methods have been adapted to make use of the high spatial and temporal detail available in agent-based transport models. To demonstrate this integrated model, this article presents an example simulation which evaluates the health impacts of a work from home policy in Munich. Overall, there is on average a small net health detriment for individuals required to work from home and a smaller net health benefit for everybody else. Agent-based modelling enables HIAs to jointly assess transport, environmental, and population health outcomes with high granularity, helping policymakers identify who benefits and suffers from proposed policies. This work demonstrates the feasibility of combining agent-based transport and health modelling methods in a seamless framework.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 6 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.