If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to: Catherine Walshe, PhD, The International Observatory on End of Life Care, Division of Health Research, Lancaster University, Bailrigg, Lancaster LA1 4AY, UK.
Trends in symptoms and functional ability are known toward the end of life, but less is understood about quality of life, particularly prospectively following service referral.
Objectives
This study compares quality of life trajectories of people with and without cancer, referred to volunteer-provided palliative care services.
Methods
A secondary analysis of the ELSA trial (n = 85 people with cancer and n = 72 without cancer) was performed. Quality of life data (WHOQOL-BREF) were collected at baseline (referral), four weeks, eight weeks, and 12 weeks. Sociodemographic data were collected at baseline. We specified a series of joint models to estimate differences on quality of life trajectories between groups adjusting for participants who die earlier in the study.
Results
People with cancer had a significantly better quality of life at referral to the volunteer-provided palliative care services than those with nonmalignant disease despite similar demographic characteristics (Cohen d's = 0.37 to 0.45). More people with cancer died during the period of the study. We observed significant differences in quality of life physical and environmental domain trajectories between groups (b = −2.35, CI −4.49, −0.21, and b = −4.11, CI −6.45, −1.76). People with cancer experienced a greater decline in quality of life than those with nonmalignant disease.
Conclusion
Referral triggers for those with and without cancer may be different. People with cancer can be expected to have a more rapid decline in quality of life from the point of service referral. This may indicate greater support needs, including from volunteer-provided palliative care services.
Experientially, quality of life for those with life-limiting illness is critically important, as people may prioritize quality of life over treatment that extends its quantity.
Palliative care services have a focus on quality of life, but access can be limited or late. Recent data indicate that people with cancer remain likely to receive aggressive end-of-life interventions including chemotherapy, repeated hospitalizations, ICU admission, and late hospice or palliative care enrollment.
Does involving volunteers in the provision of palliative care make a difference to patient and family wellbeing? A systematic review of quantitative and qualitative evidence.
While the intervention effect was small, the longitudinal quality of life data of those referred are a novel addition to knowledge, helping understanding of referral timing and change over time. Studies of quality of life trends for those in the palliative phase of illness are scarce, despite being perceived as an important outcome of trials of palliative care interventions.
Effect of specialist palliative care services on quality of life in adults with advanced incurable illness in hospital, hospice, or community settings: systematic review and meta-analysis.
Understanding trends in quality of life is important both clinically and to improve research so that interventions can be carefully timed, contextualized, and evaluated.
Response shift is a particular concern, where people reappraise illness and their quality of life and accommodate its challenges and their perceptions, leading to problems interpreting standard measures over time.
Recommendations for managing missing data, attrition and response shift in palliative and end-of-life care research: Part of the MORECare research method guidance on statistical issues.
Trajectory data for people toward the end of life tend to focus on functional change, augmented with understandings of social, psychological, and physical change, rather than quality of life.
Different experiences and goals in different advanced diseases: comparing serial interviews with patients with cancer, organ failure, or frailty and their family and professional carers.
Physical, social, psychological and existential trajectories of loss and adaptation towards the end of life for older people living with frailty: a serial interview study.
Is there a need for early palliative care in patients with life-limiting illnesses? Interview study with patients about experienced care needs from diagnosis onward.
Potential triggers for the holistic assessment of people with severe chronic obstructive pulmonary disease: analysis of multiperspective, serial qualitative interviews.
Typically, these studies only focus on one diagnosis, track symptoms or functional change rather than quality of life, or use limited data, often from those accessing in-patient clinical services. Some explain the trajectory retrospectively from point of death, but this is less useful clinically due to the inherent problems of prognostication accuracy.
Identifying patients with advanced chronic conditions for a progressive palliative care approach: a cross-sectional study of prognostic indicators related to end-of-life trajectories.
This study adds to knowledge by reporting longitudinal, prospectively collected, quality of life data, of those expected to be in their last year of life who were referred to volunteer-provided palliative care services.
Patients and Methods
Design
This prospective, longitudinal, multicenter study of people understood to be in their last year of life assessed quality of life in the 12 weeks following referral to a volunteer-provided befriending service. Data were collected in the context of a pragmatic, randomized, prospective waitlist trial; the study protocol and reports are available, and the trial was prospectively registered.
Protocol for the End-of-Life Social Action Study (ELSA): a randomised wait-list controlled trial and embedded qualitative case study evaluation assessing the causal impact of social action befriending services on end of life experience.
‘Being with’ or ‘doing for’? How the role of an end-of-life volunteer befriender can impact patient wellbeing: interviews from a multiple qualitative case study (ELSA).
The aim of this exploratory analysis was to prospectively compare the quality of life of people (with and without cancer) referred to these services.
Participants and Setting
Eligible participants were adults (≥18 years) referred to the volunteer-provided palliative care services where the answer to the surprise question “Would you be surprised if the patient dies within a year?”, assessed by the referring health care professional, was “no.”
They could have any diagnosis. They had to understand or speak a language in which our main outcome measure (the WHOQOL-BREF) was available and have an anticipated prognosis of more than four weeks. The volunteer-provided palliative care services were provided in 11 community settings across England (nine hospices, one alcohol and substance use charity, one NHS Trust). Participants continued to receive all usual care during the study. NHS Research Ethics Committee approval was received, and governance approvals from each participating site. All participants gave written consent.
Data Collection
Data were collected at study entry, and four and eight weeks following that point. Those in the “wait” arm of the trial also provided data at 12 weeks. Quality of life was assessed using the World Health Organisation Quality of Life (WHOQOL-BREF) Scale, a relatively short, but broad (26-item) non–disease specific, validated self-reported measure of quality of life and well-being.
The World Health Organization's WHOQOL-BREF quality of life assessment: psychometric properties and results of the international field trial. A report from the WHOQOL Group.
Data are reported across physical, psychological, environmental, and social relationship domains. Loneliness was assessed with the De Jong Gierveld six-item Loneliness Scale, a short, well-used, reliable, and valid measurement instrument.
Social support was assessed using the eight-item modified Medical Outcomes Study Social Support Survey, a short validated scale covering two domains of emotional and social support.
Moser AF, Stuck AE FAU - Silliman R, Silliman RA FAU - Ganz P, Ganz PA FAU - Clough-Gorr K, Clough-Gorr KM. The eight-item modified Medical Outcomes Study Social Support Survey: psychometric evaluation showed excellent performance. (1878-5921 [Electronic))]. J Clin Epidemiol 2012;65:1107–1116.
Additional data included self-reported contact with health and social care services and other networks over the previous two weeks, and baseline sociodemographic data (age, gender, disease diagnosis, education, marital status, living status, spirituality, and ethnicity). Study instruments were self-completed by participants, baseline questionnaires were explained to participants when written consent was taken at a home visit, and subsequent questionnaires were posted to participants' home address, self-completed, and returned by post to the study team.
Statistical Analysis
Participants were characterized in terms of reported primary diagnosis (cancer vs. noncancer), accounting for no detected difference in primary and secondary outcomes between intervention and control groups.
Baseline characteristics between diagnostic groups were compared using t-test or chi-square test, irrespective of original random treatment allocation. To test for diagnostic status (cancer vs. noncancer) effect on quality of life trajectories, we specified a series of joint models.
These joint models simultaneously model the longitudinal outcome (quality of life) of interest and risk of death, by adjusting for participants who die earlier in the study. In end-of-life studies and those involving older people, a significant proportion of participants may die, with survivors contributing disproportionately larger amounts of data than decedents. The tendency for healthier persons to live longer and contribute more data may lead to a “healthy survivor” effect in estimates obtained from a longitudinal analysis,
so it is important this is accounted for in modeling these data. As part of the joint model, we specified a linear mixed model with intercept and random slopes. We tested the main effect of time and the interaction of time with diagnostic status to evaluate for potential differences in quality of life trajectories between groups. In the Cox model, we added diagnostic group as a time-independent covariate. We used a (pseudo) adaptive Gauss-Hermite optimization algorithm.
We report parameter estimates, standard errors, and 95% CIs.
Results
Of those referred to the volunteer-provided palliative care services (n = 329), 196 consented to take part in the study, and 157 provided evaluable data for this analysis. At each time point, missing data were noted, but participants continued to be enrolled in the study unless advised otherwise, as data sets could be and often were returned at subsequent time points. Twenty percent of enrolled participants (39 of 196) died during the study. The overall flow through the study is presented in Figure 1. Data were collected during 2015–2016.
Baseline demographic, quality of life, loneliness, and social support data for the 157 participants who provided diagnostic information, enabling these analyses, are presented in Table 1. There were no significant differences on demographic characteristics between those with cancer and with other nonmalignant conditions. Of those without cancer, the study included those with respiratory disease (n = 26), neurological disease (n = 21), heart failure (n = 10), liver disease (n = 7), and other forms of life-limiting illness (n = 8).
Table 1Baseline Demographic, Quality of Life, Loneliness, and Social Support Data for Those Who Provided a Diagnosis
mMOS-SS = modified Medical Outcomes Study Social Support Survey.
an = 143, bn = 154, cn = 152, dn = 145, en = 140, fN = 82 & N = 70, gN = 83 & N = 72, hN = 82 & N = 71, iN = 83 & N = 71, jN = 76 & N = 69, kN = 76 & N = 67, lN = 76 & N = 65, mN = 79 & N = 66, n N = 78 & N = 65, oN = 81 & N = 68, p N = 80 & N = 66.
The WHOQOL BREF comprises four individually scored domains. Domain scores are calculated by computing the mean scores within the domain, noting that negatively phrased questions are reverse scored. Domain scores are transformed to a 0–100 scale according to the formula in the WHOQOL Manual. Lower scores indicate a worse quality of life.
There were significant differences between baseline scores on areas of quality of life (d's = 0.37 to 0.45), social loneliness (d = 0.37), and emotional social support (d = 0.44) between those with and without cancer. Those with nonmalignant life-limiting disease typically had worse quality of life, were lonelier, and had less social support on referral to the volunteer-provided palliative care services. All differences reflect small effect sizes.
We evaluated the distribution of quality of life at different time points by group. Despite the impact of attrition on our sample, we observed a normal distribution, with good coverage of spread of scores (Table 2, Fig. 2).
Table 2Distribution of Quality of Life Scores at Different Time Points
By the end of the study, 31 people in this sample had died, 27 of those with cancer and four of those with nonmalignant disease. Data on change in quality of life are reported in Table 3 for the physical, psychological, and environmental domains of the WHOQOL-BREF.
While people with cancer have a generally higher quality of life at referral to the volunteer-provided palliative care service (baseline), their quality of life deteriorates significantly more rapidly over the (relatively short) data collection period to the end of the study compared to those without cancer (Figs. 3 and 4).
Fig. 3Change in quality of life (physical domain) over study data collection time points.
These are novel longitudinal quality of life data from those anticipated to be in their last year of life who were referred to volunteer-provided palliative care services. At referral to the service (baseline), demographic characteristics of those with and without cancer are similar, but those with nonmalignant life-limiting illnesses had a worse quality of life, were lonelier, and had less social support. This may indicate they had more need for palliative care services at the time of referral. During the relatively short (12-week) period of data collection, however, those with a cancer diagnosis had a more rapidly deteriorating quality of life. More people with cancer died during the study period. These are different trajectories of quality of life for people with cancer and noncancer diagnoses who were nevertheless identified as requiring similar volunteer-provided palliative care services.
The baseline quality of life scores of those referred to the volunteer-provided palliative care services can be compared to reference data. This demonstrates that quality of life of both those with and without cancer in this sample is worse than people who are healthy or in the general population,
This may indicate that the triggers for referral, at least in terms of need relating to quality of life, are similar both where referral to volunteer-provided palliative care services and a specialist palliative care service are considered. Minimal clinically important differences are not reported for the WHOQOL-BREF, but it has been reported to be sensitive to change in health status.
a magnitude similar to the estimated change observed in our sample at the end of the study.
People with life-limiting illness who do not have cancer are known to be referred to services at a point in time where their functional status is typically worse than those with cancer. For example, those referred to specialist palliative care services with primary diagnoses other than cancer have been found to be less functional at time of referral (odds ratio: 1.6; 95% CI: 1.1–2.3). This was felt to be because of the slower and more varied trajectory of noncancer serious illness, typified by greater disability.
Our data are important as they identify that this differentiation in baseline status (whether in quality of life or functional status) at point of referral to services is also true for referral to a very different volunteer-provided palliative care services as for referral to specialist palliative care. Specialist palliative care is typically triggered later in the disease course for those without cancer,
but this may not necessarily have been expected for those referred to a volunteer-provided palliative care service. People with noncancer life-limiting disease appear to have been deteriorating for a longer period, with more impact on quality of life, before need is recognized and referral to services made. Reasons for this may include having more time to adapt to a lower functional status, less appreciation of the life-limiting nature of the illness, less routine assessment of need, or discrimination against some disease, for example, COPD where lifestyle behaviors such as smoking contribute to risk.
Even over the relatively short period, data were collected for this study; people with cancer demonstrated a more rapid deterioration in quality of life. Comparisons with other studies are challenging, as many only include people with a single diagnosis, typically cancer, but not with a comparison to other disease trajectories.
Changes in health-related quality of life and quality of care among terminally ill cancer patients and survival prediction: multicenter prospective cohort study.
which is broadly characterized as a gradual decline, accelerating in the last months of life. There are studies however which do not demonstrate such changes, with no changes in quality of life found in those referred to a community palliative care service,
Where studies do compare people with and without cancer, typically steeper declines are found in functional status or quality of life for those with cancer,
Our study adds to this scarce comparative literature, strengthening the evidence base on the different trajectories of those with and without cancer at the end of life. In particular, our study adds data on quality of life rather than the more typical functional status, prospectively gathered, rather than judged retrospectively from death.
The strengths of this study are in the relatively large sample, with different life-limiting illnesses, providing data prospectively over a number of time points. The follow-up time points were carefully and deliberately chosen to be short, given that the study was at the end of life, and care effects need to be rapid to be worthwhile, but it is possible that a longer term follow-up may reveal different trends. A potential limitation is that these data were provided in the context of an interventional trial, and this may affect people's responses to outcome measures in unanticipated ways. Diagnosis data were unavailable for 20 participants, and they had to be removed from this data set; we did not have direct access to clinical data to address this issue. This lack of access to clinical data meant that we do not know if participants were receiving any potentially disease-modifying treatments that could affect quality of life. It is known that the predictive value of the surprise question in identifying those who may die is not perfect, with worse performance in noncancer illness.
It may be that some of those referred are not in their last year of life, with differences in prediction between those with and without cancer. This may be why fewer of those who died were in the noncancer group. However, the baseline data from both groups, and comparison to population norms, nevertheless indicate that these are groups eligible for palliative care, and we carefully controlled for survival differences in our joint model. Our data reflect “real-life” referral patterns to a novel intervention, rather than the sample being representative of these diagnostic populations, and there is strength in these data because of this diversity. We aggregated quality of life data for those with nonmalignant conditions, as it is known that those with chronic conditions do have similar quality of life trends,
but it must be noted that their patterns of quality of life may not be the same. Our study adds to knowledge methodologically by using joint models that take into account the effect of participants who die earlier in the study. Information on the deaths of those in the study was provided contemporaneously, but exact dates of death were not known for some, and it is possible that some of those for whom we had missing data may have died.
People with life-limiting illness appear to be referred to volunteer-provided palliative care services both close to death (for those with cancer) and with a poor quality of life (for those with nonmalignant disease). The rapid decline in quality of life experienced by those with cancer may indicate that their support needs are greater in this phase of life, potentially necessitating additional support from volunteers in addition to clinical services.
Disclosures and Acknowledgments
The authors acknowledge the contribution of the people who took part in this research, often at a time of great challenge in their lives, thank you. This study would not have been possible without the support of the participating sites, who were responsible for identifying participants, taking consent, and managing study participants and site-specific documentation as well as training volunteers and managing the befriending services. With thanks to Dr. Evangelia (Evie) Papavasiliou, Research Associate on the project to 15.9.15 who was involved in trial initiation procedures and initial data collection, and Paul Sharples, Research Intern on the project 11.1.16–15.4.16 who assisted with data entry and initial data analysis.
This research was funded by the UK Cabinet Office, who also provided grants to sites to cover the cost of the intervention. The views expressed are those of the authors, not necessarily those of the Cabinet Office. The funders had no role in collection, analysis, or interpretation of data or in the writing of the report. Reports are seen by them before submission, but they have no contribution to the writing or amendment of the report.
Ethical approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The trial was prospectively registered. ISRCTN 12929812 http://www.isrctn.com/ISRCTN12929812 Health Research Authority research ethics approval was granted 12.3.15 by NRES Committee Yorkshire & The Humber—South Yorkshire. REC reference 15/YH/0090. IRAS project ID 173058. Site specific approvals were granted by NRES Committee Yorkshire and the Humber—South Yorkshire. Informed consent was obtained from all individual participants included in the study.
Availability of data and materials: Patient-level data are stored in the ELSA database developed by the study authors on a secure server maintained by Lancaster University. Presented data are fully anonymized. The corresponding author may be contacted to forward requests for data sharing.
The authors declare they have no competing interests.
References
Gawande A.
Quantity and quality of life: duties of care in life-limiting illness.
Does involving volunteers in the provision of palliative care make a difference to patient and family wellbeing? A systematic review of quantitative and qualitative evidence.
Effect of specialist palliative care services on quality of life in adults with advanced incurable illness in hospital, hospice, or community settings: systematic review and meta-analysis.
Recommendations for managing missing data, attrition and response shift in palliative and end-of-life care research: Part of the MORECare research method guidance on statistical issues.
Different experiences and goals in different advanced diseases: comparing serial interviews with patients with cancer, organ failure, or frailty and their family and professional carers.
Physical, social, psychological and existential trajectories of loss and adaptation towards the end of life for older people living with frailty: a serial interview study.
Is there a need for early palliative care in patients with life-limiting illnesses? Interview study with patients about experienced care needs from diagnosis onward.
Potential triggers for the holistic assessment of people with severe chronic obstructive pulmonary disease: analysis of multiperspective, serial qualitative interviews.
Identifying patients with advanced chronic conditions for a progressive palliative care approach: a cross-sectional study of prognostic indicators related to end-of-life trajectories.
Protocol for the End-of-Life Social Action Study (ELSA): a randomised wait-list controlled trial and embedded qualitative case study evaluation assessing the causal impact of social action befriending services on end of life experience.
‘Being with’ or ‘doing for’? How the role of an end-of-life volunteer befriender can impact patient wellbeing: interviews from a multiple qualitative case study (ELSA).
The World Health Organization's WHOQOL-BREF quality of life assessment: psychometric properties and results of the international field trial. A report from the WHOQOL Group.
Moser AF, Stuck AE FAU - Silliman R, Silliman RA FAU - Ganz P, Ganz PA FAU - Clough-Gorr K, Clough-Gorr KM. The eight-item modified Medical Outcomes Study Social Support Survey: psychometric evaluation showed excellent performance. (1878-5921 [Electronic))]. J Clin Epidemiol 2012;65:1107–1116.
Changes in health-related quality of life and quality of care among terminally ill cancer patients and survival prediction: multicenter prospective cohort study.
30394-4&id=gr1.jpg){kind=link}
30394-4&id=gr2.jpg){kind=link}
30394-4&id=gr3.jpg){kind=link}
30394-4&id=gr4.jpg){kind=link}