Article ID: | iaor20021868 |
Country: | United Kingdom |
Volume: | 33 |
Issue: | 3 |
Start Page Number: | 387 |
End Page Number: | 392 |
Publication Date: | May 2001 |
Journal: | Accident Analysis and Prevention |
Authors: | Guria Jagadish, Mara Kelly |
Keywords: | accidents |
A control chart system was developed in an earlier paper to identify the occurrence of actual risk changes or deviation from the expected levels of road crash fatalities during the course of a year. This paper discusses the development of a prediction method for estimating number of fatalities during a year. The method then provides a mechanism for estimating the likelihood of meeting a pre-set target or any other outcome. At any point of time in a year, the attainability of a target for the year depends not only on safety outcomes during the remainder of the year but also on outcomes up to that point in the year. High week to week variability would indicate that it can be difficult to achieve the intended maximum goal. The control charts developed earlier indicate if the current trend differs from the expected trend or that required to achieve the target. This paper determines the probability of achieving the target given the past outcomes of the year. It also determines the level of outcome for the year based on the past trend and in the absence of any special programmes. The paper discusses the New Zealand application as an example.