The future number of older people will be determined by future levels of mortality. These numbers directly determine resources required in some areas such as pensions, and have a major influence on areas such as the need for informal care. Overall mortality change will inevitably affect population characteristics in ways relevant to this proposal: for example, decreasing levels of mortality will lead to more older people being married, and hence affect the volume and types of care in years to come.

Differential mortality change will also affect the overall population structure in all areas of the project. Examples include changing sex differentials in mortality and between social groups: at present, men in Social Class I in Britain can expect to live over 8 years longer than those in Social Class V, which had major implications for equity in use of services and availability of resources. There are substantial differentials in mortality by the key factors addressed in this project, living arrangements, income levels, and health status, although the extent of these among older people is not well-established or recognised in some cases, in part due to lack of data sets of appropriate size and/or quality.

Aims

The aim of this Work Package is to produce a range of alternative forecasts of mortality using a variety of approaches, both overall and specific by variables such as cause of death and marital status. These results will be set into an international context.

Methods

The project is using a range of statistical methods developed to elucidate mortality trends, including models based on period and cohort effects, and use of overall and cause-specific mortality, specifically:

To model overall patterns of mortality, including both cohort and period data and to assess the importance of such effects and their implications for future mortality forecasts. In addition to British official statistics available from 1841, Sweden, Japan and France from the MPIDR/Berkeley Human Mortality Database (HMD) are also being used.
Since living arrangement are an important component of the project, we are modelling marital-status differences using cross-national analyses of mortality trends using a range of recently-developed statistical methods such as numerical derivatives based on super-smoothers and penalised splines which overcome some of the problems such as the ‘identification problem’ in Age-Period-Cohort (APC) models with a data-base for nine countries with over one billion person-years of experience constructed for the EU-funded FELICIE programme.
The sensitivity of overall mortality forecasts to assumptions about future prevalence and incidence of a key set of diseases, and in particular, how they may influence sex differentials in mortality, are being undertaken by comparing results obtained for m cause-specific information using data from Britain and France. In addition we are incorporating where relevant, information on a small number of major life-threatening diseases defined through Work Package 2.
We are investigating mortality differentials by socio-demographic variables and assessing how they may change in decades to come, linked to Work Packages 3 and 5.

Linkages

All other parts of the project depend on the projected number of older people, and are therefore using these results. While the links between mortality and morbidity are by no means clear-cut, in some cases it is possible to model these in an integrated way (one example is diabetes, itself linked to obesity – a factor which can lead to major variations in both the health and mortality status of older people in years to come).

Outputs

In addition to inputs to other models, this Work Package is producing outputs of interest in their own right. Analysing the role of cause-specific mortality in a more integrated way than before, can elucidate processes by which cohort effects may work through particular diseases. The work on socio-economic differentials provides an understanding of whether socio-economic differentials continue up to the highest ages, and, in particular, whether these differentials are changing over time, as we have found in the case of marital status.

Contact

Professor Mike Murphy,
Department of Social Policy,
London School of Economics,
Houghton Street,
London WC2A 2AE

Tel: 020-7955-7661 (direct line)
Email: m.murphy@lse.ac.uk

^ Back to top