This report is an other special report of the series planned by the operational committee of IRTAD. IRTAD is an International Road Traffic Accident Database; an initiative of the Road Transport Research division of the OECD. The aim of the database is to provide easy accessible road safety data for comparative road safety research. Detailed international comparisons is a valuable source for policy making. It may guide the selection of priorities and can be used to investigate the effectiveness of national policies. IRTAD contains the totals for categories of road fatalities in most OECD-countries generally from 1970 onwards and categorised by variables such as mode, age, road type. It contains also the relevant national variables that may indicate exposure to road risks. There are plans to extended the IRTAD database to data on severe road injuries as well. It is generally assumed that national road fatalities are completely recorded. However, for injuries selective national underreporting of accidents, as well as differences in definitions for the accident variables (this also holds for many variables on road fatalities), is a serious problem in comparative studies. This special report deals with the characteristics of the underreporting phenomenon from an international perspective. As such it is a valuable contribution, since unwarranted comparisons may be the result otherwise. The international use of data can he limited if we do not know the: completeness; reliability; if incomplete the representativeness or selectivity; and reporting invariance over time. One of the aims of IRTAD, towards much effort for road fatality data is devoted, is the enhancement of the reliability, the recording invariance over time and the correction of aggregated subtotals for nationally different definitions. It is felt, and more or less justified by this report, that IRTAD is an international database for road fatalities satisfying the requirements of almost completeness, high reliability and stability over time. As this report show, the national injury data are far from complete and generally also not representative. Its selectivity is nationally different, depending on mode, age, road type and severity. A further step for improvement could be to study the feasibility of national correct ion factors for (sub)totals on the basis of select ion fractions by mode, age, road type and severity in order to improve the international comparability of injury data. An additional problem is the possibility of (nationally different) changing selectivity over time of the underreporting. As this report may indicate by its review of underreporting studies from identical countries on successive periods, this lack of stability in the selectivity of the reporting is at least present in same North-West European countries. The feasibility of an approach for improvement of injury data by national correction factors, however, may be a dead alley. An other direction for improvement may he the obligation of hospitals to report all in-patients from road accidents to the police, who then could establish a nearly complete reporting of all accidents with at least one in-patient injury as outcome (that is one person with an injury severity that ask for at least a full day treatment in a hospital). The results of this report indicate that one at least ought to question the validity of any international comparisons on general road safety, if it is not based on fatalities only. (Author/publisher)
Samenvatting