FRAGMENT 
How does Euro NCAP results correlate to real life injury risks  - a paired comparison study OF CAR-TO-CAR CRASHES
Anders Lie - Swedish National Road Administration and Karolinska Institutet, Sweden
Claes Tingvall - Monash University Accident Research Centre, Australia

 Montpellier 20 September 2000
DISCUSSION

 Euro NCAP is an initiative to drive vehicle safety beyond current regulation, by offering the market more extensive information about adoption of best practice. Euro NCAP has never been intended to predict real life outcome on a car by car basis. It should not even theoretically be able to do that in the current form with a star ranking system with no reflection on representative weights on different aspects. However, it is still important to evaluate whether the general aim of promoting vehicles with higher safety standards is fulfilled. The most natural way to do this is to compare real life outcomes with the ratings or scores in Euro NCAP. While this study shows a very good overall relationship, it should not be seen as proof that there is a predictive value in Euro NCAP, especially not for individual car model scores. There might be several reasons for this general relationship, the most likely being that car manufacturers developing cars with high safety standards also do well in Euro NCAP. That does not mean that a vehicle that was designed entirely for good Euro NCAP results will perform well in real life crashes. It is, however, clear that a car that performs well in real life also can be highly scored in Euro NCAP, something that is becoming increasingly important. Similar results have been seen in other parts of the world (Newstead and Cameron 1998, Kahane et al 1994, O`Neill et al 1994).

     While Euro NCAP is a process that should lead to car manufacturers aiming for best practice, and does develop and drive further development of best practice, it is of importance that the real life outcome is constantly monitored. Monitoring is of importance also in finding any indications of sub-optimisation or negative consequences of designing cars to create good NCAP scoring. A co-operation around real life follow-up, could help pinpoint car models that are built only for good results in crash test.

     There has been concern that the test speeds in Euro NCAP might lead to sub-optimisation, possibly leading to better vehicle performance in high severity crashes and worse in low severity impacts. The analysis of risk functions in this study does not suggest that this is the case at this moment. However, it is still important to monitor this issue. It would also be beneficial to study any sign of reduced compatibility. Compatibility has been raised as a possible concern, but earlier studies have shown that this is not a necessary consequence (Lie et al 1996).

     Even if all crashes over several years were included, relatively few crashes are available for this kind of study. When looking at severe and fatal crashes in a country like Sweden, this is evident. It is though important to know that the crashes used in this study, is a sample from a much larger number of crashes. A more detailed study, using the same methodology, could be performed with a larger data set. If all European countries could merge their data, more reliable and precise results could be achieved. Risk levels for every size group could then be calculated.

     In the Swedish police material no information is given about the specific injury or the point of impact. If data were available some indication of the front and side protection levels could be added and individual types of injuries could be studied.

     Further research could look for relations between risk levels within the different size groups used by Euro NCAP. Also the risk levels for individual car models can be studied if large data sets are available. This could help in the comparison between size groups. It would also be interesting to test some old cars with known real life performance. By doing this further verification could be achieved.

     No major differences could be seen concerning minor injuries. This confirms earlier research, and suggests that the focus in Euro NCAP and more generally in car development, towards reduction of more severe injuries, can be seen also in real life crashes. There is, however, a concern that the results did not show any benefit to minor injuries. While these might be seen as less important to reduce, they still contain some injuries that generate long term health losses. This is especially relevant to neck injuries in rear end and frontal collisions, that traditionally are defined as minor injuries. However, they contribute to the total loss of health (Krafft 1998). It is important that the Euro NCAP process in the future focuses on these kinds of injuries as well.

     It is clear that the reductions achieved in serious and fatal injuries are substantial. It is to some extent surprising that it is possible to discriminate between cars built at the same time, and with differences that are at a level where they can influence safety. The magnitudes of the safety differences are at a level that they become one of the major instruments for the future of traffic safety. While there was no difference between cars that were ranked with 2 stars and older vehicles, four star cars seems to reduce the risk of serious and fatal injuries by more than 30%.

     When applying risk functions to the statistical data, it was shown that in order generate the same risk for serious and fatal injuries for the low and high rated vehicles, the change of velocity would have to be increased by approximately 12% for the 3 and 4 star group. This could be seen as an indication as to the level, which cars have been designed to match as a result of the higher test speed than has been used before. The risk functions did not show that higher performing vehicles produced more injuries in lower impact severity segments, although the method used did not allow for looking at very low severity impacts.

     The importance of weight should not be underestimated, and while this factor is not taken into account in crash tests into fixed barriers, 100kg more weight in a car to car impact will generate a 7% lower risk of injury. In single vehicle crashes, which account for a high proportion of crashes, the mass should not have any significant influence on safety.

Pobierrz  caly materiał
 how does euro ncap results correlate to real life injury risks.doc