The benefits of improved car secondary safety

https://doi.org/10.1016/S0001-4575(02)00030-1Get rights and content

Abstract

The term ‘secondary safety’ refers to the protection that a vehicle provides its occupants when involved in an accident. This paper studies information from the British database of road accident reports between 1980 and 1998, to estimate the reduction in the number of occupant casualties over these years which may be attributed to improvements to secondary safety in cars.

The paper shows that the proportion of driver casualties who are killed or seriously injured (KSI) is lower for modern cars than for older cars. The reduction of this proportion is used to assess the improvement in secondary safety. Statistical models are developed to represent the proportion with ‘year of first registration’ as one of the independent variables, although only an incomplete assessment of the benefits of improved secondary safety can be made with the available data. The assessment compares the number of casualties that would have been expected if secondary safety had remained at the level found in cars first registered in 1980 with the actual casualty numbers. It is estimated that improved secondary safety reduced the number of drivers KSI by at least 19.7% in 1998, in comparison with what might have occurred if all cars had had that lower level of secondary safety. This figure relates to all cars on the road in 1998, and rises to 33%, when confined to the most modern cars (those which were first registered in 1998).

Introduction

Various regulations have been introduced over at least 20 years with the aim of improving car secondary safety (also known as crashworthiness). Secondary safety refers to the protection that a car provides its occupants when involved in an accident, whereas primary safety refers to the features such as braking systems that should help the driver to avoid becoming involved in an accident. This paper examines accident data from 1980 to 1998 to assess the extent to which car occupant casualties have been reduced by the improved safety features of modern models. The data come from the British database of police reports (known as STATS19) of road accidents that involve personal injury or death.

This research was originally carried out for the UK Department for Transport, Local Government and the Regions (formerly, Department of the Environment, Transport and the Regions), as part of a project to investigate the numerical context for setting national casualty reduction targets for 2010 (Broughton et al., 2000). The original analyses have been revised and updated using accident data from 1997 and 1998.

In this paper, secondary safety is investigated via the proportion of injured car occupants who were killed or seriously injured (KSI) (the ‘severity proportion’) in any particular year. If secondary safety has improved over the years, then more modern cars should protect their occupants better than older cars in the accidents occurring in any particular year, so this proportion should be lower for newer cars. Other road safety measures may well have affected the proportions over the years, such as improvements to the road system, but these should affect all cars equally—irrespective of when they were manufactured. Thus, the benefits of improved secondary safety will be identified by separating the component of the severity proportion that relates to a car’s newness from the component that relates to the year when the accident occurred. Newness will be represented by the year when the car was first registered, which is known in most cases from the STATS19 accident reports.

An accident-involved car may have been first registered 5 years or more after that model was first introduced, depending on the manufacturer’s marketing strategy. The analyses will compare the secondary safety of the new cars sold in a year, rather than the new car models introduced in that year. Thus, the effects of any development in the ‘state of the secondary safety art’ will only be visible with a lag of some years.

One key question which cannot be answered by studying the severity proportion is whether improved secondary safety might have affected the total number of car occupant casualties, not simply the proportion of fatal or serious casualties within a known casualty total. For example, a measure could reduce casualties of all severities to the same degree: the severity proportion would be unaffected, although the measure was effective. The STATS19 reports are only for accidents involving personal injury, and in UK there is no reliable source of national data for damage-only accidents. It does seem plausible that car design might be improved sufficiently for occupants to escape injury in accidents, where otherwise they would have been slightly injured, thereby reducing the total number. Indeed, the increase in seat belt wearing by roughly 50% in 1983, following implementation of the 1981 Transport Act, was equivalent to a sudden improvement in the secondary safety of half the car fleet. Broughton (1990) found that the increased wearing rate had reduced the number of casualties of each severity—so the casualty total had fallen. Rutherford et al. (1985) reached the same conclusion in an extensive hospital-based study.

Thus, the benefits of improved car secondary safety cannot be quantified fully by studying the severity proportion. Nonetheless, the equations for estimating the casualty benefits of improved secondary safety will be developed in a way that allows for the possibility that the car occupant casualty total may have been reduced, so that the sensitivity of the results to uncertainty over the effect on the casualty total can be tested.

Analyses will focus upon the driver casualties, principally because the number of passengers per car can vary and these variations might bias the results if passenger casualties were included. This is discussed more fully by Broughton (1996a).

Several methods have been developed in recent years to assess the secondary safety of cars by statistical analysis of road accident data. Broughton (1996a) described two of them, but showed that “the indices measure the relative secondary safety of current car models, but cannot evaluate the general progress in improving secondary safety. They compare models with the average for a particular year, but other measures are needed to determine whether this average changes over the years.” This conclusion applies equally to the other assessment methods of which the author is aware. In contrast to these existing methods, this paper attempts to evaluate the ‘general progress’ by analysing data from accidents that occurred over almost two decades.

Section snippets

Exploratory data analysis

The source of the data used in this study is the STATS19 database of police reports of road accidents involving personal injury. The STATS19 reporting system has operated throughout UK since 1949; car driver casualty data for accidents that occurred in 1980–1998 have been analysed.

During this period, the British vehicle registration system incorporates a prefix or suffix to denote the year of first registration, and since 1979 this was reported by the police for accident-involved vehicles. More

Statistical modelling

The previous section showed that there has been a general reduction in the severity of car drivers’ injuries, as well as a reduction related to the year of first registration of the car. A statistical model is required to disentangle the two effects. The appropriate way to investigate the data in more detail is to develop logistic regression models for two dependent variables:

  • P1: proportion of injured drivers who were killed,

  • P2: proportion of injured drivers who were KSI.

The GLIM program (

Estimated casualty benefits

The results of the GLIM modelling will now be used to reassess the car accidents from 1981–1998, and to estimate the number of extra casualties that would have been expected if secondary safety had not improved, but had remained at the level achieved by 1980–1981 year cars. It is assumed that the same accidents would still have occurred, but that more drivers would have been killed and seriously injured because of the lower level of secondary safety. The calculation is done for each year of

Discussion

This paper has demonstrated that the proportion of injured car drivers who are KSI in modern cars is clearly less than in older cars. The only remaining question is whether it is reasonable to attribute the associated casualty reductions to the efforts of regulators and manufacturers to produce safer vehicles, or whether independent factors may have contributed to the reductions.

The only plausible candidate is vehicle mass. This is well known to influence secondary safety, since the laws of

Conclusions

This paper has used car accident data from UK between 1980–1998 to estimate the effect on the number of driver casualties of the gradual improvement in the secondary safety features of the national car fleet. Simple analyses of data from individual years showed that the proportion of injured car drivers KSI is lower in more modern cars. Statistical models were developed to represent the proportion of car driver casualties who were KSI, with the ‘year of first registration’ of a car and ‘year of

Acknowledgements

The work described in this paper was carried out in the Safety Group of Transport Research Laboratory (TRL) under a contract from the UK Department for Transport, Local Government and the Regions (formerly Department of the Environment, Transport and the Regions).

References (6)

There are more references available in the full text version of this article.

Cited by (0)

View full text