%0 Report %A Filtness, Ashleigh %A Papadimitriou, Eleonora %A Leskovsek, B. %A Focant, N. %A Martensen, Heike %A Sgarra, V. %A Usami, D. Shingo %A Soteropoulos, A. %A Stadlbauer, S. %A Theofilatos, Athanasios %A Yannis, George %A Ziakopoulos, A. %A Diamandouros, K. %A Durso, C. %A Goldenbeld, Charles %A Loenis, B. %A Schermers, G. %A van Petegem, J.-H. %A Elvik, R. %A Hesjevoll, I.S. %A Quigley, Claire %A Papazikou, Evita %D 2017 %T Identification of infrastructure related risk factors, Deliverable 5.1 of the H2020 project SafetyCube %U https://repository.lboro.ac.uk/articles/report/Identification_of_infrastructure_related_risk_factors_Deliverable_5_1_of_the_H2020_project_SafetyCube/9354641 %2 https://repository.lboro.ac.uk/ndownloader/files/16964441 %K untagged %K Design Practice and Management not elsewhere classified %X The present Deliverable (D5.1) describes the identification and evaluation of infrastructure related risk factors. It outlines the results of Task 5.1 of WP5 of SafetyCube, which aimed to identify and evaluate infrastructure related risk factors and related road safety problems by (i) presenting a taxonomy of infrastructure related risks, (ii) identifying “hot topics” of concern for relevant stakeholders and (iii) evaluating the relative importance for road safety outcomes (crash risk, crash frequency and severity etc.) within the scientific literature for each identified risk factor. To help achieve this, Task 5.1 has initially exploited current knowledge (e.g. existing studies) and, where possible, existing accident data (macroscopic and in-depth) in order to identify and rank risk factors related to the road infrastructure. This information will help further on in WP5 to identify countermeasures for addressing these risk factors and finally to undertake an assessment of the effects of these countermeasures. In order to develop a comprehensive taxonomy of road infrastructure-related risks, an overview of infrastructure safety across Europe was undertaken to identify the main types of road infrastructure-related risks, using key resources and publications such as the European Road Safety Observatory (ERSO), The Handbook of Road Safety Measures (Elvik et al., 2009), the iRAP toolkit and the SWOV factsheets, to name a few. The taxonomy developed contained 59 specific risk factors within 16 general risk factors, all within 10 infrastructure elements. In addition to this, stakeholder consultations in the form of a series of workshops were undertaken to prioritise risk factors (‘hot topics’) based on the feedback from the stakeholders on which risk factors they considered to be the most important or most relevant in terms of road infrastructure safety. The stakeholders who attended the workshops had a wide range of backgrounds (e.g. government, industry, research, relevant consumer organisations etc.) and a wide range of interests and knowledge. The identified ‘hot topics’ were ranked in terms of importance (i.e. which would have the greatest effect on road safety). SafetyCube analysis will put the greatest emphasis on these topics (e.g. pedestrian/cyclist safety, crossings, visibility, removing obstacles). To evaluate the scientific literature, a methodology was developed in Work Package 3 of the SafetyCube project. WP5 has applied this methodology to road infrastructure risk factors. This uniformed approach facilitated systematic searching of the scientific literature and consistent evaluation of the evidence for each risk factor. The method included a literature search strategy, a ‘coding template’ to record key data and metadata from individual studies, and guidelines for summarising the findings (Martensen et al, 2016b). The main databases used in the WP5 literature search were Scopus and TRID, with some risk factors utilising additional database searches (e.g. Google Scholar, Science Direct). Studies using crash data were considered highest priority. Where a high number of studies were found, further selection criteria were applied to ensure the best quality studies were included in the analysis (e.g. key meta-analyses, recent studies, country origin, importance). Once the most relevant studies were identified for a risk factor, each study was coded within a template developed in WP3. Information coded for each study included road system element, basic study information, road user group information, study design, measures of exposure, measures of outcomes and types of effects. The information in the coded templates will be included in the relational database developed to serve as the main source (‘back end’) of the Decision Support System (DSS) being developed for SafetyCube. Each risk factor was assigned a secondary coding partner who would carry out the control procedure and would discuss with the primary coding partner any coding issues they had found. Once all studies were coded for a risk factor, a synopsis was created, synthesising the coded studies and outlining the main findings in the form of meta-analyses (where possible) or another type of comprehensive synthesis (e.g. vote-count analysis). Each synopsis consists of three sections: a 2 page summary (including abstract, overview of effects and analysis methods); a scientific overview (short literature synthesis, overview of studies, analysis methods and analysis of the effects) and finally supporting documents (e.g. details of literature search and comparison of available studies in detail, if relevant). To enrich the background information in the synopses, in-depth accident investigation data from a number of sources across Europe (i.e. GIDAS, CARE/CADaS) was sourced. Not all risk factors could be enhanced with this data, but where it was possible, the aim was to provide further information on the type of crash scenarios typically found in collisions where specific infrastructure-related risk factors are present. If present, this data was included in the synopsis for the specific risk factor. After undertaking the literature search and coding of the studies, it was found that for some risk factors, not enough detailed studies could be found to allow a synopsis to be written. Therefore, the revised number of specific risk factors that did have a synopsis written was 37, within 7 infrastructure elements. Nevertheless, the coded studies on the remaining risk factors will be included in the database to be accessible by the interested DSS users. At the start of each synopsis, the risk factor is assigned a colour code, which indicates how important this risk factor is in terms of the amount of evidence demonstrating its impact on road safety in terms of increasing crash risk or severity. The code can either be Red (very clear increased risk), Yellow (probably risky), Grey (unclear results) or Green (probably not risky). In total, eight risk factors were given a Red code (e.g. traffic volume, traffic composition, road surface deficiencies, shoulder deficiencies, workzone length, low curve radius), twenty were given a Yellow code (e.g. secondary crashes, risks associated with road type, narrow lane or median, roadside deficiencies, type of junction, design and visibility at junctions) seven were given a Grey code (e.g. congestion, frost and snow, densely spaced junctions etc.). The specific risk factors given the red code were found to be distributed across a range of infrastructure elements, demonstrating that the greatest risk is spread across several aspects of infrastructure design and traffic control. However, four ‘hot topics’ were rated as being risky, which were ‘small work-zone length’, ‘low curve radius’, ‘absence of shoulder’ and ‘narrow shoulder’. Some limitations were identified. Firstly, because of the method used to attribute colour code, it is in theory possible for a risk factor with a Yellow colour code to have a greater overall magnitude of impact on road safety than a risk factor coded Red. This would occur if studies reported a large impact of a risk factor but without sufficient consistency to allocate a red colour code. Road safety benefits should be expected from implementing measures to mitigate Yellow as well as Red coded infrastructure risks. Secondly, findings may have been limited by both the implemented literature search strategy and the quality of the studies identified, but this was to ensure the studies included were of sufficiently high quality to inform understanding of the risk factor. Finally, due to difficulties of finding relevant studies, it was not possible to evaluate the effects on road safety of all topics listed in the taxonomy. The next task of WP5 is to begin identifying measures that will counter the identified risk factors. Priority will be placed on investigating measures aimed to mitigate the risk factors identified as Red. The priority of risk factors in the Yellow category will depend on why they were assigned to this category and whether or not they are a hot topic. %I Loughborough University