TY  - CONF
A1  - Grömping, U
A1  - Weimann, U.
A1  - Menzler, S.
T1  - Split-register study : a new method for estimating the impact of rare exposure on population accident risk based on accident register data
N2  - The data situation for quantifying the proportion of accidents avoided by the introduction of active safety systems is incomplete, since there is generally no data available on the accidents avoided by the technology in question. In this paper, a split-register approach is suggested and compared with the classical case-control approach known from epidemiologic applications. Provided a set of assumptions hold, which can reasonably be made in such data situations, the split register approach allows inferences on the population accident risk. For both approaches the benefits of basing the analysis on the results of a logistic regression to adjust for confounding factors are outlined. The biasing effects of violating key assumptions are discussed and the split-register approach is demonstrated using the example of the active safety system ESP with data from the German in-depth accident study GIDAS.
KW  - Aktives Sicherheitssystem
KW  - Bevölkerung
KW  - Datenbank
KW  - Deutschland
KW  - Elektronik
KW  - Fahrstabilität
KW  - Inventar
KW  - Konferenz
KW  - Risiko
KW  - Unfall
KW  - Accident
KW  - Active safety system
KW  - Conference
KW  - Data bank
KW  - Electronics
KW  - Germany
KW  - Inventory
KW  - Population
KW  - Risk
KW  - Vehicle handling
Y1  - 2005
UR  - https://bast.opus.hbz-nrw.de/frontdoor/index/index/docId/443
UR  - https://nbn-resolving.org/urn:nbn:de:hbz:opus-bast-4434
N1  - Weitere beteiligte Körperschaft: Institut für Biometrie, Epidemiologie und Informationsverarbeitung <Hannover>
ER  -