Critical approaches to road injury trends, forecasts and scenarios: two urban cases in Norway

Urban regions experience high population growth rates and possess distinct characteristics regarding mobility, mode of transport, road environments and road safety plans. Urban areas are prioritized in the Norwegian Transport Plan’s strategies for transport and by implementation of General Urban Environmental Agreements with casualty reduction targets. The study examines possibilities and limitations in developing and analysing urban trends, forecasts and scenarios for killed and seriously injured road users in Norway. The Oslo and Trondheim regions are case studies for the period 1998–2025. Police reports of killed and seriously injured casualties 1998–2012 in the regions (N Oslo = 1801, N Trondheim = 790) formed the basis of the analyses. The mean injury rate for Oslo region was 10.8 per 100,000 pop. per year for the period 2008–2012, whereas the corresponding figure for Trondheim was 15.7, the highest among all large urban areas. The road authorities’ accident targets towards 2025, initiated by the National Plan of Action for Road Safety are decomposed into regional units. On the basis of the two regions’ statistics relating to road casualties, the applied principle of equal proportional share indicates a 40% injury reduction towards 2025. Exponential models, ARIMA models, negative binomial regression, and scenario approaches were employed to estimate possible trends and changes in casualties in the studied urban regions. These models formed a framework to reveal critical assumptions, sensitivity, uncertainties, and realism in the road authorities’ strategies for casualty reduction targets and for trend extrapolations and predictions. The analyses consider sources of critical conditions to rate the magnitude of the following conditions: (1) the trend towards the prevailing road safety status (casualty numbers, casualty rate levels, and stability) problematized by small numbers and random variation; (2) the exposure effects of urban road network composition and urban–rural mix; (3) the exposure effect of population growth towards 2025, change in population, and risk-exposed road users; and (4) the exposure effect of shifts in mode of transport, with uncertainties regarding non-linear relationships between casualty numbers and exposure measures. Some effects of exogenous road safety measures of all types are considered, such as diminishing return and reduced sum effects of measures. One conclusion is that the various model approaches for the two urban reg