Intuition might suggest that people are more likely to die the older they become, but in fact the death rate stabilizes above a certain age leading to a “mortality plateau”. Now, Jonathan Coe and Yong Mao at the University of Cambridge and Mike Cates at the University of Edinburgh have developed a computer model that reveals this levelling-off in the rate of death with increasing age. This result, which previous models have been unable to achieve, may help to explain mortality rates better than ever before. (J.B. Coe et al. 2002 Phys.Rev.Lett. 89 28).
In 1995, T.J. Penna developed a computer-simulated model of ageing, which goes a certain way in explaining mortality. In the model, each individual is given a string of binary numbers that stays fixed during his or her lifetime. Each number, a ‘0’ or ‘1’, represents the change in state of health of the individual at each age. A ‘0’ indicates no decline in health, while a ‘1’ indicates the onset of a genetically inherited disease. As time progresses, the model counts the number of ‘1’s, which represents the number of diseases that each individual has. When this reaches a certain threshold, the individual dies. The main shortcoming of the Penna model is that it predicts that all members of a genetically identical population must die at exactly the same, pre-determined, age.
Researchers have tried to overcome this problem in subsequent modelling work but have not been able to reproduce a definite mortality plateau, only a gently decelerating rate of death with increasing age.
What Coe and co-workers have done is to find an exact solution to the Penna problem. The Edinburgh-Cambridge team produced a more general formula in which they deduce that individuals have an arbitrary “survival function”. The researchers adjusted Penna’s model so that genetically identical individuals can die at different ages rather than always at the same age. The researchers believe that the factors causing this variable death rate may be essential in understanding the mortality plateau observed in many species.
In the long term, the result may provide important information for insurance companies and pension funds.