Doing things differently: Using data to commission high quality end of life care

by Dr Michael Cooke
Head of Analytics


Michael Cooke



There is currently a rare alignment of what terminally ill people want and what the health social care system needs.  The challenge for commissioners and those working in end of life care is to seize this opportunity to redesign and rebalance it.


We believe that this rebalancing should be rooted in the better use of available data so that we improve equality of access to services and outcomes for terminally people and their families and address variation in experience across geographical areas, and across the characteristics of terminally ill people, including their age and their diagnoses.

The National Bereavement Survey (VOICES) is a postal survey of bereaved relatives which is commissioned by NHS England and delivered by the Office for National Statistics (ONS).   The 2012 survey achieved 22,292 responses (a response rate of 45.7%).  It records bereaved relatives’ perceptions of the care their loved one received in the last months of their life.   Key measures include levels of dignity and respect afforded by different professionals and in different settings, pain management and the degree to which services were coordinated.

ONS has recently combined the 2012 and 2013 waves of VOICES survey data and reconfigured the data so it is available at both a Local Area Team and Clinical Commissioning Group level. Doing this has enhanced the potential for comparison of bereaved relatives’ perceptions of end of life care across different parts of England and provided us with a much greater understanding of local differences in experience.

VOICES enhances what we already know about end of life care, including where people die and how this varies byage, gender, cause of death and level of deprivation.

Having CCG level data which is based on the views of families with direct experience of end of life care means commissioners can base local service planning on local data and real experiences.

As part of our on-going work to ensure commissioners have access to the most accurate and useful data possible, we have produced a mini Atlas to present the new CCG level VOICES data. This enables commissioners to see what bereaved relatives in their area think of the care their loved one received from local services in the last months of their lives and to compare areas both locally and nationally.

Marie Curie End of Life Care Mini-Atlas

The latest CCG level VOICES data is now available via the Marie Curie End of Life Care Mini-Atlas

At Marie Curie we are always looking to find different ways to display national and local level data and enhance our understanding of end of life care.  Learning from the experience of others, we have started to use funnel plots, which enable us to take into account the fact that we would expect to see greater variation and therefore more extreme experience ratings in areas which received a smaller number of responses to the VOICES survey.  This helps to identify unexpected outliers in a more rigorous way than looking at the top and bottom performers without the context of sample size.

It is worth noting that even though two years of data were pooled for the VOICES CCG level analysis, some of the sample sizes are still relatively small.  43 CCGs received less than 100 responses, compared with 23 CCGs that received more than 300 responses. To help make sense of these responses, we have produced our own funnel plot using the AHPO tool on the Public Health England website.

This chart shows that there are three areas with positive ratings of overall experience of care in the last three months of life that are very unlikely to be due to random variation: Dorset, South Devon and Torbay, and North Hampshire CCGs. The positive rating in South Devon and Torbay CCG is of particular interest as its health and social care is integrated and provided through one team, with part of their work focusing on providing coordinated end of life care.  It would be interesting to further explore whether integrated care delivery leads to better experiences and perceptions of end of life care.

Overall quality of care in the last three months of life by CCG

There are also a number of areas where experience is rated poorly to an extent that is very unlikely to be due to random variation. Five of these six areas are in London and the South East. It is interesting to note that when we look at the statistics around death in the usual place of residence between 2010-11 and 2013-14, these six CCGs consistently recorded lower than average figures for people dying in their usual place of residence. By comparison, the three CCGs discussed in the previous paragraph, consistently scored higher than the average figures for this measure.  Again, it would be good to explore the relationship between where people die in an area and the experience of end of life care. We know from the original VOICES analysis that, at a national level, relatives of people who die at home are much more likely to rate the overall quality of care positively.

By undertaking such analysis, organisations like Marie Curie, local commissioners and those responsible for end of life care at a national level, can begin to unpick why particular areas score higher than others. They can do so by exploring the relative roles and impact of demographic characteristics, disease profiles, levels of hospice provision, the impact of community based nursing and the degree to which local services are integrated. Good data and sound analysis must be the foundation for giving terminally ill people and their families more of what they want and need.