Intervening in the trip to work. A system dynamics approach to commuting and public health

Abstract

Background: Car use is the dominant mode of transport in short, habitual journeys to work in many cities, including Auckland. It allows access to a range of employment and training while enabling families to manage other responsibilities. However, car dependent commuting has significant negative effects for commuters, the wider community, and local and global ecosystems. Existing evidence about harms is sufficient to seek a commuting mode shift for environmental, health and equity benefits. Although some of these benefits are already considered in transport planning to a varying extent, there are considerable challenges of complexity and implicit trade-offs, among competing interests and between outcomes. Cross-disciplinary discourses have identified principles for effective policy decisions in complex systems such as this, including a systems approach, transdisciplinarity, community participation and a social justice focus. These principles formed the basis for the thesis. Aims: 1. To develop a comprehensive conceptual model of the trip to work and public health that synthesises knowledge from epidemiology, communities and policy makers 2. To develop a commuting and public health simulation model that could quantify a range of outcomes for some particular policy options 3. To use this model to identify effective policy levers for intervening in commuting patterns to improve public health outcomes 4. To test the utility of the modelling methodology for integrating public health outcomes in a more participatory approach to transport policy making Methods A review of the complex links between commuting and health was undertaken using two ecosystem health frameworks. I then investigated how these links have been incorporated into New Zealand's transport policies. Participatory system dynamics (SD) modelling was used to combine policy, community and academic knowledge. SD modelling enables the simulation of complex systems characterised by feedback, delay, non-linear relationships and tipping points. Interviews and workshops with community, health and policy stakeholders led to the development of causal loop diagrams (CLDs) connecting influences on mode share with broad wellbeing outcomes. One element of the CLDs was developed into a simulation model incorporating best evidence for Auckland. Policy scenarios were simulated to consider their relative effectiveness for meeting identified policy targets and public health outcomes. The model was shared with regional and national decision makers to influence policy. Results The review and policy analysis identified a complex range of connections between commuting and public health and found these have been poorly incorporated into transport planning in New Zealand to date. Nine commuting and public health themes emerged from the qualitative modelling process: pedestrian and cyclist safety; relative attractiveness of public transport; neighbourhood security; time pressure and employment accessibility; workplace support for different modes; participation and leadership in planning; environmental and cultural wellbeing; and "car culture". These were developed into CLDs. Causal loops relating to cycling were developed into a simulation model, incorporating the following scenarios to 2051: business-as-usual; the planned regional cycle network; physically separated cycle lanes on all arterial roads; and region-wide “self-explaining” local roads. Combining the last two policies was most likely to meet existing policy targets for perception of safety, climate change and cycling mode share. Other benefits accrue from improved physical activity and air quality. The combined policy changes the shape of the cycling injury curve to significantly reduce the injury cost to increasing commuter cycling. Health benefits outweigh the costs by about 22:1 in this scenario. The model was behaviourally robust, but sensitive to assumptions about safety in numbers, the effectiveness of policies on perception of safety and the link between survey perceptions and revealed mode share. Conclusions This is the first time SD modelling has been used to integrate the public health outcomes of transport policy. The SD modelling process was successful in achieving agreement about the important dynamics in the commuting and public health system by a transdisciplinary group of stakeholders. I identified some prerequisites for the successful use of participatory SD modelling in urban settings. These include a stable regional governance structure; combining regional level, spatially homogenous modelling with place-based narratives that allow stakeholders to identify with modelled relationships; the involvement of powerful policy stakeholders accompanied by the incorporation of the method into the policy process. The cycle commuting simulation model represents the first integrated assessment of public health outcomes for specific active transport policies. It was able to demonstrate the comparative costs and benefits of policies to increase commuter cycling, identifying trade-offs between public health outcomes. Creating safe cycling infrastructure will be crucial for increasing commuter cycling. A universal approach that progressively transforms Auckland’s arterial and local roads over the next 40 years would be needed if cycling is to assist with achieving the region’s quantified sustainable transport targets. This area-wide change would be cost-effective, returning in the order of $20.00 in quantified public health benefits for every dollar spent.