This first ever randomised trial of effectiveness has shown that volunteers may help at the end of life. But its methodological problems may say even more about the challenges facing volunteer services.
The results showed a trend in favour of the intervention and it was concluded that:
“Doctors and other clinicians can confidently refer people in their last year of life to volunteer services for support. They can expect that these services may slow a person’s decline in quality of life.
It was also recommended that Policymakers should continue to promote the involvement of volunteers in end-of-life care.
The study limitations were said to include “study power, blinding, missing data, attrition and intervention fidelity”.
But I think the limitations of the study actually give important clues about volunteer support services for people nearing the end of life.
1. The researchers acknowledge that the “primary focus on biomedical and physical aspects of end-of-life care ignores the social context within which dying takes place.”
But the study itself continues to look at professional-defined outcomes, when it focuses on WHO and other standardised measures — rather than whatever the individual may have defined as important.
2 and 3.The exclusion criteria for the study included language criteria, and having an anticipated prognosis of <4 weeks
The language criterion is surely one reason the study found that “volunteer services … did not facilitate access for those who were from black and minority ethnic communities”.
Volunteer services like Edencare, which focuses especially on BME people, take care to do language-matching — as this takes account of the “social context” of the dying person.
Despite the prognosis exclusion criterion, 20% of the enrolled participants died during the study. Anecdotally, I think that substantial numbers of people may only want external / volunteers input when they near the end of life — when it all gets too much for them. Are we not going to offer or evaluate help when it is most needed?
4. Statistical power: Recruiting partipants. The researchers intended to recruit 350 or more participants per arm (that means 700). This would allow adequate ‘statistical power’ for the outcomes that were selected. However, only 329 eligible people were approached to see if they wanted to take part, and only 196 consented.
For me, these numbers show how unrealistic the services and investigators were – and also how many people don’t want what the services are offering.
Which surely means we need to think again about exactly what services are offered?!
Some of the people didn’t want to participate because they didn’t want to receive the service (n = 46), they died or move away (n = 38 – this seems two strange groups to lump together!), or didn’t want to take part in research (n = 30).
Few carers (n = 33) entered / wanted to enter the study “and hence their data are not given here.”
The recruitment period was constrained by “governance and contractual delays and time limited funding”. This isn’t a surprise – but why, o why does it just keep on being a problem?
4. There was an even bigger issue with volunteers availability and matching. Around 40% of those in the ‘immediate’ arm received no volunteer intervention before week 4 assessment.
So its not just clients that don’t want to take part, but volunteers too.
Although there were no differences in the selected outcomes, there were ‘trends’.
These included a significantly reduced deterioration in physical Quality of Life for those in the ‘immediate’ group
I was also very interested that the researchers had to “[control] for site” — which implies that there were important differences between the sites – one of which matched only 3 participants-volunteers.
The trial compared participants randomly allocated to receive a volunteer support intervention either immediately or after a 4 week wait
Trained volunteers were matched to individuals by a volunteer co-ordinator. They provided care tailored to the needs of the individual but offered from a suite of options including befriending, practical support and signposting. Input could differ at each visit. Volunteer support was typically provided face-to-face, one-to-one, in the home, but telephone contact and meeting outside the home were possible. Most contacts were befriending visits in the home, typically for 2–3 hours per week
196 individuals were included in the study. They were all adults. 61% (n = 109) were female. Their mean age was 72 years
The results showed a small reduction in the rate of decline, rather than any actual improvement in Quality of Life or Loneliness. When the number of hours of volunteer input were taken into account, there was a significant reduction in the physical quality of life for the intervention group.
It was recommended that future trials should focus on exploring dose issues such as hours and frequency of contact as well as the type of support offered.
Read the whole paper at: https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-016-0746-8