Description: Presented at Living to 100 and Beyond symposium 2002. Discussion of two papers, 'Approaches and Experiences in Projecting Mortality Patterns for the Oldest Old' by Buettner and 'Using Dynamic Reliability in Estimating Mortality at Advanced Ages' by Lin. The first paper discusses the UN methodology for projecting mortality above age 80 and presents another approach for addressing the problem of table closing. The second paper suggests a new statistical model for fitting and estimation of mortality rates at advanced ages.

Description: Presented at Living to 100 symposium, January 2011. Using data from Social Security's Death Master File, this study shows that mortality deceleration in humans observed at advanced ages may be an artifact caused by age exaggeration, data heterogeneity or use of improper estimates of hazard rate. The phenomenon of mortality deceleration at older ages vanishes for datasets of higher data quality.