Background: There is little evidence about the effects of environmental interventions on population levels of physical activity. Major transport projects may promote or discourage physical activity in the form of walking and cycling, but researching the health effects of such "natural experiments" in transport policy or infrastructure is challenging.
Methods: Case study of attempts in 2004-2005 to evaluate the effects of two major transport projects in Scotland: an urban congestion charging scheme in Edinburgh, and a new urban motorway (freeway) in Glasgow.
Results: These interventions are typical of many major transport projects. They are unique to their context. They cannot easily be separated from the other components of the wider policies within which they occur. When, where, and how they are implemented are political decisions over which researchers have no control. Baseline data collection required for longitudinal studies may need to be planned before the intervention is certain to take place. There is no simple way of defining a population or area exposed to the intervention or of defining control groups. Changes in quantitative measures of health-related behavior may be difficult to detect.
Conclusions: Major transport projects have clear potential to influence population health, but it is difficult to define the interventions, categorize exposure, or measure outcomes in ways that are likely to be seen as credible in the field of public health intervention research. A final study design is proposed in which multiple methods and spatial levels of analysis are combined in a longitudinal quasi-experimental study.