In recent years, substantial improvements in mortality from certain causes have been observed in many developed populations. The most notable, at least in some populations, has been mortality from ischaemic heart disease. A question frequently posed in respect of such improvements is the quality of the extended lifetime. Analysis of the available data requires a relatively sophisticated model of morbidity and mortality and their interaction. The purpose of this paper is to describe methods which allow investigators to partition differentials in life expectancies in healthy and unhealthy states (and trends in such differentials) according to their source: differentials in disease onset rates, recovery rates and mortality rates (and their trends). The methods allow an investigator to explain, for example, the contributions of different diseases to changes in life expectancy, and for a particular disease, the contribution of disease prevention, successful disease treatment (leading to recovery) and mortality prevention. The techniques have their origins in the work of Kitagawa (1955) and more recent investigations by Pollard (1982, 1986, 1988).