Article Text

Download PDFPDF

Free papers 22–24 – Research methods
Planning the first national end of life care survey
  1. Katherine Hunt1,
  2. Natalie Shlomo2 and
  3. Julia Addington-Hall3
  1. 1Department of Health Sciences, University of Southampton, UK
  2. 2Southampton Statistical Sciences Research Institute, University of Southampton, UK
  3. 3Department of Health Sciences, University of Southampton, UK


Background The End of Life Care Strategy highlighted a need to evaluate care experiences by accessing the views of those who use end of life care services. The Strategy identified the Views of Informal Carers – Evaluation of Services (VOICES) questionnaire, which is completed by bereaved relatives, as a potential method of evaluating these experiences. The DH commissioned this study to explore the feasibility of a national VOICES survey.

Aims To develop the most appropriate methods for a national end of life care survey by considering recruitment, sampling, online methods, ethics, increasing participation and support for participants.

Methods VOICES was re-designed following user/professional discussion groups and analysis of existing VOICES datasets. 1422 deaths registered in two PCTs were identified by the Office of National Statistics using stratified sampling methods. Coroner-registered deaths were excluded. Deaths were assigned to one of two trial groups to determine the most appropriate recruitment approach (opt-in vs opt-out). Online completion was offered to all informants. Local organisations representing minority ethnic groups collaborated in publicising the survey, interpreting services were provided and advertising posters were translated into five languages. A series of support structures was initiated.

Results Response rate was 33% and response was significantly higher in the ‘opt-out’ trial group (40%, p<0.001). There were no complaints in either group: only two informants called the support line. 19% of responders completed online. Minority group responses were low. As planned, data weighting was used to account for non-response bias. Excluded coroner-registered deaths were significantly different to included deaths on a series of parameters.

Conclusions The opt-out method is the recommended recruitment approach. Experiences of minority groups should be gathered using alternative methods. Coroner-registered deaths should be included and data should be weighted. VOICES detected differences between PCTs. It will be used in 2011/2012 together with these methods in the first national end of life care survey.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.