Write-up: seminar 2

Seminar 2 – Public Health and Transport Modelling
Seminar two involved workshops, report-backs, lead-offs, summaries and debates. Much of this was audio recorded and can be found here. The following represent just some of the points made in group discussions.
• In health, modelling often starts from supposing a behaviour change – e.g. more active travel – then look at what are the benefits. But transport tries to look at what predicts behaviour. Health is often marginal to transport schemes – it’s added on and doesn’t get to shape plans.
• Modelling’s status as evidence differs in the two disciplines. In health modelling doesn’t guide decisions so much – there is less weight put on models as sources of evidence, more emphasis on trials. Health modellers could learn from how modelling has been made credible in transport.
• There is a lack of communication between the two groups – public health practitioners/modellers and transport modellers don’t understand each other. If people just bring their own bit into the modelling process, interdisciplinary thinking gets limited.
• The disconnect is partly because of different purposes, different scales and metrics. Often the data needed and used is different. For health benefits, you need to know who is taking up walking or cycling, how active they are already, because health benefits depend on it. Transport modelling’s focus on shortening travel times seems at odds with a health approach, which seeks to encourage more travel by active modes.
• There are developments that make collaboration easier – network-based transport modelling can feed into health assessments via better measures of pollution exposure, but this doesn’t happen.
• There is a distinction between transport modelling, and transport appraisal, and decisions should be made on the basis of the latter. If health outcomes are monetised they can be fed back into the transport appraisal, but often they’re not (e.g. noise).
• The cutting edge of transport modelling is emergent properties in terms of how junction change affects systems. But we need to understand people too; for example, not just junction layout but how you do cycling in this city and where. How can we model social norms (e.g. perceptions of walking) and how they change? Transport models don’t take into account social networks and social influence – could models incorporate this and differential influence between groups?
• What gets left outside the model, and gets brought in later via case studies? And how does that effect the status of different modes? What levels of evidence do you need for different things? What kinds of short cuts do we use – for example, people may act on perceptions of risk or journey times more than realities; what difference would that make to modelling if we could incorporate it?
• Do we have a data problem for active travel? Is it inherently harder to model active modes, and to get information about them (for example, because they are more intangible)? Is data really as big a barrier as we think (we don’t necessarily get things right where we do have the data).
• We should count the health disbenefits of pro-car policies rather than just the health benefits of walking or cycling scheme. This affects how we see risk – walking and cycling are counted as risky, but shouldn’t this risk be attached to driving, as walking in cycling in themselves are very safe?
• What should we expect from models? Should we see models as helping us understand relationships between assumptions and data and outputs, rather than always making predictions? How much accuracy should we expect from models? Should we focus on point estimates or understanding where the key uncertainties lie and how those factors affect decision-making?
• Much transport modelling focuses on things like small-scale changes to junctions, and impact on traffic routing and travel times. But the data collected for these kinds of exercises doesn’t help us learn about health impacts, nor how we might transform walking and cycling participation. At the macro end of the scale the DfT National Transport Model doesn’t help us think about different futures either.

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>