Follow us
Publications Commentary Research People Events News Resources and Videos About IFS
Home Publications An overview of the ELSA 'end of life' data

An overview of the ELSA 'end of life' data

Rowena Crawford and Pierfrancesco Mei
Report

The ageing of the population is one of the biggest changes facing society today. People are living longer, remaining healthier at older ages, and working and engaging in society for longer. This is an achievement that should be celebrated, both in its own right and for the opportunities it presents, with older generations able to provide assistance to younger generations and pass down their accumulated experience and wisdom. However, such changes are, of course, not without challenges. The needs of an older population are different, and this has profound implications for the planning of economic, health and social policy by government, and for the decisions made by the population as a whole. There are also marked inequalities in the experiences of those with different circumstances, which presents an additional set of challenges when adapting policy to address issues such as income security, social isolation, and health and social care provision.

For the last decade, the English Longitudinal Study of Ageing (ELSA) has been an invaluable source of information on the economic, social and health circumstances of the older population. ELSA is a representative survey of the private household population of England aged 50 and over. It began in 2002 and has interviewed the same households every two years since. The original sample of over 11,000 individuals has been followed for 14 years so far, and individuals have also been added to refresh the younger ages to ensure that the study remains representative of people aged 50 and older. ELSA collects data on household composition, employment and pensions, housing, income and wealth, self-reported doctor-diagnosed diseases and symptoms, cognitive performance and walking speed, health behaviours, care giving and receipt, social contacts and activities, quality of life and expectations. ELSA data have been instrumental in furthering understanding of issues such as: labour market participation at older ages; the accumulation and spend-down of wealth in retirement; the effect of state pension policy reforms; the genetic, biological and social pathways that lead to frailty; the lifestyles of older cohorts; forecasting trends in disability and life expectancy; how well-being changes as people age; and the importance of loneliness and social isolation in determining health and well-being.

Many respondents have died since ELSA started in 2002. Figure 1.1 illustrates how the proportion of the original ELSA sample known to have died has evolved over time. A nice feature of ELSA is that respondents are asked for permission to link their records to official mortality data (the National Health Service Central Register), which provide date of death even if an individual had stopped responding to the ELSA survey some time previously. As of the end of 2012 (the last time linked mortality data were made available), 24% of the original ELSA sample had died (2,763 deaths). A further 123 deaths have also been recorded amongst individuals who joined the ELSA sample after 2002 as ‘refreshment’ individuals.

While an individual can clearly no longer respond to ELSA once they have died, a follow-up interview is attempted with a proxy respondent – a relative or friend of the deceased respondent – to elicit information on the circumstances of the deceased individual. In particular, these ‘End of Life’ (EoL) interviews collect data on the individual’s health in the two years preceding their death, expenses around and at the end of life, and the size and allocation of the individual’s assets after their death. These EoL interviews are a feature that is unique among UK household surveys, but in common with some other international ageing studies such as the Health and Retirement Study in the US.

EoL interviews were conducted alongside the second, third, fourth and sixth waves of ELSA. Over the four waves, EoL interviews were completed for a total of 988 individuals. Of these, 977 were in respect of core ELSA respondents who had died before the end of 2012 – representing a third of the sample of ELSA respondents known to have died by that time.

In this report, we provide an introduction to the ELSA EoL data. Chapter 2 is devoted to discussing the quality of the data. The two important issues considered are: (i) the extent to which the subsample of deceased ELSA respondents for whom EoL data are available are similar, in terms of their observed characteristics, to the full sample of deceased respondents; and (ii) how confident we may be in the responses given to the EoL questionnaire by the proxy respondent. In other words, are the EoL data ‘any good’ in terms of how representative they are, and how accurate are they? Chapter 3 provides more detail on the contents of the EoL questionnaire and summarises some of the data that are collected. This covers four main areas: circumstances around death, health in the last years of life, end-of-life expenses, and assets and their distribution after death. Chapter 4 draws some conclusions.