Related Experiment Videos
Conditioned life tables from registries with unidentified random losses
G M Tallis1, P Leppard, T J O'Neill
1Department of Statistics, University of Adelaide, South Australia.
Statistics in Medicine
|April 30, 1993
Abstract:
We identify two types of random loss that may afflict registries which use 'passive follow-up'. The effects of these errors on estimates related to survival are examined mathematically and numerically. A procedure for correcting the attendant biases is suggested, and its strengths and limitations are explored.