Combining a name algorithm with a capture-recapture method to retrieve cases of Turkish descent from a German population-based cancer registry

An increasing proportion of the 2 million Turkish residents in Germany is r eaching the age in which cancer becomes a common health problem. However, d ata on cancer incidence and survival among Turkish residents are lacking du e to incomplete reporting of nationality in German cancer registries. In th e population-based cancer registry of the Saarland, retrieval by reported n ationality yielded only 38% (95% confidence interval (CI): 31-45%) of the e stimated number of Turkish cases in the registry; furthermore, nationality information was found to be inaccurate, and completeness dependent on the v ital status of cases. A newly developed algorithm based on family names ret rieved 85% (95% CI: 79-90%) of Turkish cases. Combining the two sources in a capture recapture approach yielded 91% (95% CI: 86-94%) of the estimated total number of Turkish cases. Hence, the name-based algorithm provides a n ew and attractive tool for valid registry-based cancer research among Turks in Germany. (C) 2000 Elsevier Science Ltd. All rights reserved.