Related Experiment Video
Updated: Nov 29, 2025

Measurement of Lifespan in Drosophila melanogaster
Published on: January 7, 2013
Are estimates of 'potential gains in life expectancy' potentially unreliable?
1Columbia University, United States; National Bureau of Economic Research, United States; CESifo, Germany.
Objectives:
The objective of this study was to investigate whether estimates of 'potential gains in life expectancy' (PGLE) are potentially unreliable.
Study Design:
We compare the sum of the PGLEs from actual 1999-2017 reductions in U.S. mortality from all causes to the actual change in life expectancy of the entire population.
Methods:
Multiple-decrement life table techniques are used to calculate the PGLEs from the actual reductions in age-specific rates of mortality from 113 causes. Then, the sum of the estimated PGLEs is compared to the actual change in life expectancy (LE) of the entire population.
Results:
The sum of the estimated gains in LE at birth is 40% larger than the actual increase in LE at birth: 2.7 years vs 1.9 years. The sum of the estimated gains in LE at age 65 years is 40% smaller than the actual increase in LE at age 65 years: 1.0 years vs 1.7 years.
Conclusions:
The fact that the sum of the estimated gains in LE from all diseases differs substantially (in both directions) from the actual increase in LE suggests that estimates of PGLE are potentially unreliable.
Related Concept Videos
Actuarial Approach
Consider the example of a high-risk surgical procedure with significant early-stage mortality. A two-year clinical study is conducted,...
Kaplan-Meier Approach
Life Tables
Regression Toward the Mean
Applications of Life Tables
Prediction Intervals
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.

