Authors: Desmond O'Connor, National Roads Authority; Clodagh Staunton, PMCE Consulting Engineers; and Louise Doherty, Roadplan Consulting

A report was commissioned by the National Roads Authority (NRA) in 2012 to examine how collision data is collected in Ireland. This report, entitled ‘Contributory factors analysis for road traffic collisions’, aimed to identify gaps in the data collected about contributory factors for road traffic collisions.

Since formal collision records began in Ireland in 1968, contributory factors, amongst other details, have been collected by An Garda Síochána through the ‘PC16’ form. Within this form are five contributory factor fields that are attributed to the cause of a collision: driver error, pedestrian error, road factors, environmental factors and vehicle factors.

The investigating gardaí use these five fields to describe, in their opinion, the circumstances that led to the collision. Similar to Ireland, the police are the primary collectors of collision data in many other countries. There is often high demand on police resources in Ireland and as a result of these time constraints, many contributory factors relating to the collision go unrecorded (see Figure 1).

Collision causation factors recorded by the police are vitally important to help determine the most appropriate follow-up action to pursue in terms of engineering, education or enforcement measures. Despite the move by European roads authorities towards a safe system and forgiving roadside approach, historic collision data still plays a critical role in the implementation of remedial measures, and should be carried out alongside proactive initiatives.

The report examined collision data collected by An Garda Síochána for the years 2007 to 2010. In this time, some 6,934 injury collisions occurred on national roads. Of these collisions, some 31% were not assigned a contributory factor. For the collisions that did include contributory factors, driver action was flagged as the most common, being attributed as the only cause in over 59% of collisions.

Pedestrian error was the second most frequent contributory factor, which accounted for just over 4% of collisions, and a total of 3.4% of collisions listed the road infrastructure as being one of the contributory factors.

[caption id="attachment_14743" align="alignright" width="926"] Figure 1: Venn diagram showing Garda opinion of factors that contributed to a ‘large extent’ or ‘some extent’ in collisions on national roads from 2007-2010 (click to enlarge)[/caption]

Figure 1 shows that contributory factors were not recorded for a large percentage of collisions, leaving a significant gap in the data. For comparison, the report looked to a US study on collision contributory factors by Treat et al (1979) and the Highways Agency Safety Risk Model from the UK (2007-2010). The results are presented in Table 1 below.

The table shows the Irish collision data with comparable results for US and UK studies. Similar trends can be identified between the countries. Where driver/human error, the environment or road factors are listed as the single contributory cause to a collision, there is a positive correlation between the countries.

[caption id="attachment_14747" align="aligncenter" width="11182"] Table 1: Comparison of Garda opinion of contributory factors with that reported by Treat at al for the USA and comparable data from the UK Highways AgencyTable 1: Comparison of Garda opinion of contributory factors with that reported by Treat at al for the USA and comparable data from the UK Highways Agency[/caption]

There are also some diverse results within the table, for example, where road infrastructure and driver/human error are said to have contributed to a collision. The US attributed this combination of contributory factors to 26% of collisions, whilst the same factors were assigned to 14.5% of collisions in the UK.

In Ireland, the combined contributory factors of driver error and road environment accounts for less than 2% of collisions. Based on this comparison, it can be assumed that some of the 2,168 collisions studied for this report with no contributory factor assigned would likely be attributed to driver and road contributory factors.

The collision data review has highlighted and quantified the extent of the missing component of contributory factors.

Collision contributory factors are a complex issue and it is unlikely for even a single-vehicle collision to have only one causation factor involved. In addition, it is important to note the difference between collision causation factors and factors that influence the severity of collisions. For example, if a vehicle were to lose control and collide with a tree, the severity of the injury would be a consequence of the impact with the tree. The tree would not have caused the collision, but it would have played significant factor in the determining the extent of any injuries to the vehicle occupants.

Although the report has highlighted a significant gap in some of the data collected, it does rank Ireland as a leading country in the breadth of information collected at the scene of a collision. Perhaps, then, Ireland should look more towards utilising available technologies to help better gather these important nuggets of data, rather than rely on the individual for all the data.

For example, what role will vehicle Event Data Recorders (EDR) play in gathering data leading up to the collision in the future? Although not a mandatory requirement for car manufacturers to install EDR systems, the National Highway Traffic Safety Administration estimate that already 85% of vehicles in the US have them onboard. The European Commission has begun a phased introduction of standardised EDR systems, initially in rented vehicles and subsequently in commercial and private vehicles.

Whether the data comes from EDRs or continues to be collected by the police, it is the collective knowledge and understanding of the contributory factors that will forge better and more effective countermeasures to help reduce the number and severity of collisions on our roads.  Research like this challenges all stakeholders involved to think and rethink about current and future practice for data collection.

ROAD COLLISION DATA COLLECTION IN IRELAND

Collision data has been a key component of Irish road safety research for decades. The NRA Safety section use collision data to prioritise extensive programmes of safety treatments across the national road network. It is crucial that collision data is suitable for the purpose of managing the safety of the network. Generally speaking, the quality of Irish collision data is very good, but regular users are aware of issues within the dataset, such as blank or missing records.

In the lead up to the current Road Safety Strategy, the NRA commissioned a report to review collision data capture in Ireland and to benchmark this data against that of eight leading European countries in road safety.

Benchmarking road-safety data and research is common practice at an international level. Organisations such as EuroRap (European Road Assessment Programme) and programmes including IRTAD (International Traffic Safety Data and Analysis Group) and CARE (Community Road Accident Database) are examples of collaborations in benchmarking safety levels using data from a number of key performance indicators.

The report examined the collision data collected by An Garda Síochána through the PC16 form. The PC16 data is entered into the Garda database and the Road Safety Authority (RSA) road collision facts are derived from this source. The study also included collision data gathered by local authorities using the LA16 form. This form was introduced by the NRA in 2007 for fatal collisions in response to the delay in receiving the official collision data.

The data fields from the PC16 and LA16 forms were compared with similar fields from collision databases of eight European countries. From this, a ‘generic list’ of 224 data fields was produced. The generic list was considered to be the maximum number of data fields used to record collision details. The 224 fields were classified into five categories: Incident Reference Information; Circumstances; Vehicles; People; and Contributory Factors.

In general, the benchmarking report shows that Ireland is performing very well in comparison to its European counterparts in terms of the extent of collision data collected by An Garda Síochána. The table shows that the collision data recorded through the Garda PC16 form is the most comprehensive in Europe, with 111 of the 224 generic list fields accounted for. These results are solely based on the number of fields used to collect information on the vehicle, people, location, contributory factors and circumstances related to a collision.

The table provides an interesting overview of the range of collision data collected across Europe, although it is difficult to ascertain the quality of the data in each country from this brief comparison. The LA16 form, for instance, appears to perform poorly in relation to the generic list, recording only 36 of the 224 data fields. In reality however, the LA16 form provides the NRA with essential information about fatal collision locations. The LA16 process also ensures that a local authority engineer has met with a member of the Garda investigation team at the site of the collision and remedial actions, if needed, can take place without undue delay.

CONTRIBUTORY FACTOR INFORMATION

A closer look at the results indicate that there is scope for collision data to be collected more efficiently in Ireland, particularly in relation to collision contributory factors. Contributory factor information can determine if an engineering countermeasure is required at a location. The low number of contributory factor fields in the Irish database, coupled with instances of blank or missing records, means that it can be very difficult to identify the cause of a collision from PC16 records alone.

The chart shows the data fields from Ireland, the United Kingdom and the generic list. From the generic list, we can deduce that very little information is collected on contributory factors across the board. Currently, some 20% of UK data fields describe the factors that may have led to a collision occurring. In Ireland, only 11% of data fields relate to contributory factors. Perhaps improving the contributory factor fields would make the data more beneficial in terms of road safety research.

Considering the results, it may be beneficial to investigate how collision data is actually used in Ireland. A significant amount of resources are exerted in collision data collection and it is essential that the resulting database offers a return for the time and money invested in it. Understanding the applications of collision data may allow for Gardaí to collect key information with a smaller amount of fields rather than vast amounts of data that might never be used.

In doing this, authorities could potentially remove fields that can later be derived from other fields, thus making the collision data more meaningful. Further investigation into the data fields used in other countries may help to identify possible deficiencies in Irish collision data.

A review of collision data collection in Ireland is currently being implemented under Action 118 of the Road Safety Strategy. This work is a collaborative effort between the RSA, An Garda Síochána and NRA Safety. This review is a step in the right direction but the need for investment in the collection of collision data, its analysis and dissemination of the results is a critical part of road safety and must not be neglected.

Similarly, it is important to consider obtaining collision information from additional sources, such as hospital databases, to gain the most comprehensive information about road traffic collisions ad their consequence. Such progress is critical for the effective implementation of future road safety programmes.

A copy of the report is available at www.nra.ie/safety. For additional information about the research, please see: www.nra.ie/safety/research/irish-collision-data-revi/

Desmond O’Connor has 20 years’ experience working with spatial data. He began his career with ESB International before moving to Dublin City Council; then he joined the National Roads Authority in November 2001. O’Connor has a master’s degree in Spatial Information Management from Dublin Institute of Technology. Currently, he is involved on a number of CEDR research projects as well as his work with the NRA on network safety ranking and road safety inspections. Clodagh Staunton joined PMCE in 2013 and is currently working on secondment with the NRA as a spatial analyst. She received a BA in Geography and Economics from NUI Galway in 2009 and recently graduated with an MA in Economic and Environmental Modelling. Staunton is currently involved with both network safety ranking and road safety inspections. Louise Doherty is a geographic information systems (GIS) spatial analyst. She has been employed with Roadplan Consulting since 2013.  She is currently working on secondment with the NRA Safety section, where she is involved in network safety ranking and provides GIS support for road safety inspection. Doherty graduated with a BA in Geography from Dublin City University in 2006 and she was awarded an MSc in Environmental Resource Management from University College Dublin in 2012.