Article Text

Download PDFPDF

Supporting family caregivers of hospitalised palliative care patients: a psychoeducational group intervention
  1. Peter L Hudson1,2,
  2. Thomas Trauer3,
  3. Elizabeth Lobb4,
  4. Rachel Zordan1,
  5. Anne Williams5,
  6. Karen Quinn6,
  7. Michael Summers6 and
  8. Kristina Thomas7
  1. 1Centre for Palliative Care, The University of Melbourne and St Vincent's Hospital, Melbourne, Australia
  2. 2Queen's University, Belfast, UK
  3. 3Department of Psychiatry, University of Melbourne, Melbourne, Australia
  4. 4Department of Medicine, University of Notre Dame, Melbourne, Australia
  5. 5Clinical Nursing and Midwifery Research Centre, Edith Cowan University, Jundalup, Australia
  6. 6Multiple Sclerosis Society of Australia, Melbourne, Australia
  7. 7Centre for Palliative Care, St Vincent's Hospital, Melbourne, Australia
  1. Correspondence to Professor Peter Hudson, Centre for Palliative Care, The University of Melbourne and St Vincent's Hospital, Melbourne 3065, Australia


Background Many family caregivers of palliative care patients experience poor health and have other unmet needs, requiring health professionals' support. However, there are few evidence-based supportive interventions to address these issues.

Purpose The purpose of this project was to undertake preliminary testing of a psychoeducational group education programme, delivered in an in-patient setting, designed to prepare family caregivers for the role of supporting a relative currently receiving hospital-based palliative care.

Methods A pilot phase was conducted to develop the intervention and explore its utility. Thereafter the single session intervention was delivered in five palliative care units in three states of Australia and its effectiveness was examined using a pre–post design. Outcome variables included caregiver preparedness, competence and unmet needs. Psychological wellbeing was measured in order to determine if there were any deleterious psychological outcomes.

Results One hundred and twenty-six participants completed Time 1 data and 107 (84.9%) completed Time 2 data (post-intervention). There were statistically significant improvements in caregivers' sense of preparedness (p=<.001; effect size (ES) 0.43) and a significant reduction in unmet caregiver needs (p=014; ES 0.22). There was no significant effect on psychological wellbeing and the improvement on competence fell short of statistical significance.

Conclusions This study reinforces the notion that psychoeducational interventions for this population can potentially be applicable, acceptable and effective. However, the number of participants who were recruited and attended each session was fewer than anticipated, resulting in methodological implications. It is recommended that the intervention undergo further empirical inquiry, such as via a controlled trial.

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.


There are multiple reasons why family caregivers of patients receiving palliative care should receive more support from healthcare professionals and be better prepared for their role.1 Many family caregivers are prone to physical and psychological morbidity2,,6, are responsible for numerous tasks such as symptom management3, are financially disadvantaged3 4 7, become socially isolated8, report unmet needs (typically aligned with lack of information about the caregiver role)9, have levels of need equal to and/or greater than the needs of patients10, and have very limited firsthand exposure to death and dying.3 9 In addition, family caregivers have the potential (with suitable support) to gain positive outcomes from the role.4 9 11 They are pivotal to achieving ‘successful’ home care (where most people prefer to die)12,,14 and make a remarkable economic contribution to healthcare.15 Finally, when well-supported, family caregivers may significantly enhance the well-being of patients.12 16

Support for family caregivers who are assisting a relative/friend requiring palliative care may incorporate practical, educational, psychological, spiritual, financial or social strategies (based on unmet needs and a desire by the caregiver(s) for assistance). Healthcare professionals should aim to enhance the caregiver's capacity to undertake their role, respond to its challenges and maintain caregiver health, during and after their relative/friends' death.17

Even though there is a requirement for healthcare professionals to support family caregivers, the reality is that in many instances this support is less than optimal.1 A review of interventions to reduce caregiver strain and burden in those caring for cancer patients failed to identify a suitable intervention to recommend for practice.18 One author concluded that current evidence barely supports the claim that palliative care services provide effective support for family caregivers (19, p 162). A recent systematic review of psychosocial interventions for family caregivers revealed that empirical inquiry regarding effective ways to provide support to family caregivers is still in its infancy.19 An international study to determine the research priorities related to family caregivers reinforced the need to develop and evaluate interventions focused on preparing family caregivers for their role.20 Several psychoeducational studies have focused on preparing family caregivers who are supporting a relative at home, yet many patients are cared for in hospital settings and hence interventions to prepare and support their family caregivers are warranted.21 22

Given this background we sought to undertake testing of the effectiveness of a psychoeducational group programme, delivered in the in-patient setting, designed to prepare family caregivers for the role of supporting a relative currently receiving hospital based palliative care. We previously undertook a pilot study to:(1) develop the specific education session content and delivery protocol;(2) pilot the intervention in three clinical sites; and(3) evaluate its potential accessibility and acceptability.23 For the current study we hypothesised that family caregivers who attended a single education session would show improvements in their sense of preparedness and competence related to their care-giving role and also report reduced unmet needs.


A pre–post design was employed to provide a preliminary test for the effectiveness of the intervention. This approach is in keeping with the Medical Research Council (UK) recommendations for developing and evaluating complex interventions24 whereby interventions are developed and tested based on pilot work prior to conducting a randomised controlled trial(s). Ethics approval was obtained at the participating clinical sites.

The intervention

The process for developing the intervention, including its theoretical framework, is comprehensively outlined in our pilot work.23 The intervention components were unchanged based on the favourable results from the pilot phase. The intervention consisted of five specific topics: (1) what is palliative care?, (2) the typical role of family caregivers, (3) support services available to assist caregivers, (4) preparing for the future and (5) self-care strategies for caregivers. Each education session was of 1.5 h duration (with approximately 15 min either side for data collection instructions). The main written resource for the education programme was a guidebook developed with input from current and bereaved family caregivers and its utility was tested within the randomised controlled trial described in the pilot study.21

In addition to receiving information and communicating experiences with other family caregivers, participants were given the option to share their contact details so that they could liaise with other caregivers in their own time, if desired. At the end of the programme, family caregivers were invited to arrange a mutually agreeable time to meet individually with relevant members of the multidisciplinary team to discuss their needs. This also provided an opportunity to make formal referrals and to schedule family meetings. Health professionals experienced in palliative care and with demonstrated skills in facilitating groups and responding to common family caregiver psychosocial issues, conducted the programmes. To help ensure a degree of uniformity, facilitators also received a resource kit, including the intervention protocol, an outline and script for the information session and a ‘PowerPoint’ presentation. At each of the three participating clinical sites, an education facilitator and a research assistant were employed. The education facilitators, who were experienced palliative care nurses and the research assistants, all attended a 1-day training programme incorporating communication skills, their roles and responsibilities, the education session objectives, content and format and the research protocols (eg, recruitment procedures, data collection).

Sample, setting and recruitment

Participants were recruited from five in-patient palliative care units in three Australian states: New South Wales, Victoria and Western Australia. These palliative care units provided multidisciplinary specialist palliative care for patients with intractable symptoms or for whom death was imminent and dying at home was not feasible. Patients with advanced cancer admitted (for the first time) to in-patient palliative care units and their primary family caregivers, were approached to participate. Patients were not required to complete any data collection tools; however, their informed consent was required to confirm access to the primary family caregiver and to their medical record. Patients were screened by clinical staff for their eligibility within 24 h of admission. Those eligible were given an information sheet and, if they agreed, written consent was obtained. Their family caregiver was contacted thereafter by letter (including information and consent form and reply-paid envelope). If willing to participate, the caregiver was phoned and scheduled to attend the next education programme. Prior to the session medical files were checked to determine if the patient had died in the interim.


To be eligible the patient must (A) have had a cancer diagnosis, (B) been at least 18 years of age, (C) be able to read and speak English (exclusion due to limited study resources for funding interpreter), (D) be able to identify a relative or friend as their primary caregiver and (E) be willing for that person to be invited to participate. Exclusion criteria consisted of (A) exhibiting signs of significant emotional distress which is likely to impact upon their capacity to provide informed consent and (B) presence of a cognitive disturbance (ie, delirium or dementia). The caregiver must have been at least 18 years of age, able to read and speak English, able to provide valid, informed consent, and been willing to be recognised as the patient's primary family caregiver. Exclusion criterion comprised exhibiting signs of significant emotional distress which was likely to impact upon their capacity to provide informed consent, complete data or participate in the intervention.

Data collection

At the commencement of the education programme (Time 1) participants completed a set of self-report questionnaires including the following: sociodemographic details, instruments to measure intervention outcome variables25,,27 as outlined in table 1. The intervention did not specifically target psychological wellbeing; however, the General Health Questionnaire (GHQ; 12-item version)28 29 was included in order to evaluate whether the intervention caused any detrimental psychological effects, in order to discern if it was ‘safe’ to administer. The three selected outcome variables were in keeping with the aims and theoretical framework underpinning the intervention.30 The psychometric properties of the instruments for measuring the dependent variables have been tested in Australian family caregiver palliative care populations and were deemed suitable.25 31

Table 1

Study variables and instruments used to test the effectiveness of the intervention

At completion of the education programme family caregivers were given a sealed envelope containing the same set of instruments administered at Time 1; they were asked to complete these in 3 days' time (Time 2), returning them in the enclosed reply-paid envelope. A reminder to complete was made by phone, where required. Three days were chosen in order to limit attrition due to patients' death (given the relatively short duration of the admission). In addition, we routinely collected specific patient data, such as the interval between the date the caregiver commenced participation in the education programme, and the date of the patient's death. The same recruitment and data collection approaches for administration of the instruments used for the pilot study were incorporated for the current study.

Sample size and analysis

The sample size estimation was based on a conventional level of statistical significance (α=0.05) and a minimum power of 0.8. Based on our earlier work in this field, we believed that the effect size (ES) was likely to be moderate (∼0.5). Applying Cohen's32 power calculation methods to a paired t test design indicated that paired data from 70 participants would detect a small to medium effect. No formal power analysis was undertaken because: (a) we had no prior information about expected ES and (b) our approach was going to be to recruit as many participants as possible given the funding and time factors.

Descriptive statistical approaches (average, percentages and distributional plots) were conducted to describe and profile the sample of caregivers who attended the programmes and the palliative care patients. Repeated measures analysis of variance including research site as a factor was conducted to examine differences between pre-intervention and 3-day follow-up, using the self-report measures. Multiple regression analyses were undertaken to predict change in outcome score from the pre-test variables. In order to assess whether those who do not return Time 2 questionnaires were consistently different from those who do, we performed post hoc comparisons (unpaired t tests) of the baseline data of those who did and those who did not return questionnaires. All significance tests were two tailed.


In light of the results of the pilot phase, the components of the intervention and data collection remained unchanged for the effectiveness component of the research and consequently the data from the 12 participants (pilot phase) were incorporated in the analysis. A total of 52 educational sessions were conducted, with between one to three participants in attendance. Thirteen of these were conducted with only one caregiver because on these occasions other caregivers who were expected to attend did not do so. One hundred and twenty-six caregivers completed the intervention. Figure 1 provides a summary of the number of family caregivers approached, final number of those who agreed to participate and other key information such as reasons for declining participation. Of the 126 participants who completed Time 1 data, 107 (84.9%) also completed Time 2 data.

Figure 1

Recruitment and data collection overview

There were 19 caregivers who did not produce any post-intervention outcomes data (n=19). Comparisons showed that those who did not produce post-intervention outcomes data were: (A) younger (58.7 vs 47.5 years, t=3.1 df=120, p=.002); (B) less likely to have ever been married (90.6% vs 73.7%, χ2=4.43, df=1, p=.035; (C) more likely to be full time employed (14.7% vs 42.9%, χ2=6.44, df=1, p=.011); and (D) more likely to have no religious affiliation (81.3% vs 52.6%, χ2=7.49, df=1, p=.006). The potential significance of these findings is unclear.

The sociodemographic profile of the participants and selected details regarding the patient they were caring for is outlined in table 2. As outlined, approximately two-thirds of caregivers were women. Approximately one quarter of family caregivers either ceased employment or reduced working hours in order to support their relative/friend.

Table 2

Sociodemographic profile: family caregiver and patient

There were three main outcome measures Family Inventory of Needs (FINB)—how well need was met), a caregiver preparedness scale (PREP) and a caregiver competence scale (COMP) and the GHQ. In order to determine whether it was appropriate to use the total scores for these measures we examined their internal consistencies by Cronbach's α, at both pre-test and post-test. All the internal consistencies of these measures were above 0.87.

Differences between participants' pre-test and post-test scores on all measures are outlined in table 3. With regard to main outcome results there was a small (ES=0.22) but statistically significant rise in FINB, suggesting an improvement in how well needs have been met. There was a moderate and significant rise (two points, ES=0.43) in PREP, showing an improvement in preparedness. Although there was a rise in COMP, suggesting an improvement in perceived competence in the caregiver role, this result was not significant.

Table 3

Changes in pre–post intervention scores

There were no significant changes in GHQ scores, indicating that the intervention had neither a beneficial nor a harmful psychological impact.

Despite the significant improvements in needs being met and preparedness at the group level, the mean changes are small and some participants' scores deteriorated. Therefore we examined whether any pre-test demographic variables predict change in score, which could suggest which kinds of participants are most responsive to this kind of intervention, and potentially who are harmed by it.

There were some marked changes in FINB (+27 to −12), PREP (+16 to −14) and COMP (+5 to −11). For all three measures, no pre-test variables were significantly associated with change, which is not to say that certain other information, such as closeness to death, may not have some predictive value.

Multiple regression revealed only one variable, having some religious belief, as relating to change in GHQ score. Specifically, the GHQ scores of the 84 participants with religious belief went down by 0.42, while the scores of 19 without belief went up by 2.05. This difference is statistically significant (t(101)=2.32, p(2-tail)=0.022).


This short, single session psychoeducational group intervention appeared to have a moderate but significant impact on reducing the unmet needs of caregivers and improving their perceived sense of preparedness. Competence scores did improve but not to the point of statistical significance. These results further support the notion that interventions for caregivers (if designed in a targeted way) can have multiple positive outcomes.33 The findings also add to the accumulating body of work on the benefits of preparing family caregivers for their role.22 34

Psychological distress was not lessened by the intervention, nor did it increase, suggesting that the intervention did not have a deleterious effect. However, the intervention did not specifically target distress; hence an improvement in this domain was not expected. This finding is in agreement with the findings of others who have concluded that it is difficult to reduce caregiver anxiety and/or depression and, perhaps, this can only be lessened if interventions specifically focus on these factors and are of a sufficient duration.33 Psychological distress scores did increase for those family carers who did not report a religious affiliation. Ideally, we would like to see a replication of this finding before making any conclusive interpretation, but it does suggest that religious belief may have some protective effect.

Many participants had moderate to high levels of psychological distress prior to undertaking the intervention (refer GHQ scores) yet this very brief intervention was able to create a significant change in their sense of preparedness and level of unmet needs. Also, family caregivers attended in spite of their feelings of psychological distress, indicating a need for this type of supportive intervention. These positive findings led to the subsequent development of a web-based resource designed to promote the implementation of programmes in other in patient palliative care services (see

There are several limitations of this study which can help guide future research. With regard to the utility of the programme, it seemed applicable, appropriate and acceptable to most of the participants, in several instances the programme was delivered to only one or two caregivers (due to the fact that just prior to the session some caregivers were unable to attend as a consequence of unforeseen events). In these instances there was no opportunity for mutual support. Furthermore, offering the intervention to only one or two caregivers has resource, cost and feasibility implications. Nonetheless, we contend that it is potentially inappropriate not to offer the intervention to those who did attend just because some caregivers were unable to be there.

The measures used to test the effectiveness of the intervention were research instruments. Thus what constitutes a clinically meaningful change requires further empirical investigation.

An additional limitation of this study was a less than desirable recruitment rate. A large proportion of eligible caregivers declined to participate in the study. Although this is not unexpected with this population it does raise the possibility of selection bias and should be taken into account for future studies in order to minimise external validity issues.35 We contend that more caregivers might attend interventions of this ilk if they are not linked to a research study. In other words, the process of recruiting participants and completing relevant evaluation forms may have been a deterrent for some.

Uptake to these types of sessions might also be improved if the programmes were run routinely and frequently (for example, fortnightly) in a room close to the in-patient unit. It would be advantageous if future studies incorporated an economic evaluation.

Although there were several statistically significant results, the ES/change was moderate. Nonetheless, this was a very brief intervention that took place in the context of many other factors encountered by these family carers, so a dramatic impact was not expected. While the ES was modest, the intervention may be valuable in terms of it being seemingly easy to administer.

The reason for testing the effectiveness of the intervention after several days of completion was pragmatic; the majority of the caregivers' relatives were likely to die within days and we therefore wanted to reduce attrition. However, future empirical testing of the intervention should include an additional data collection point in order to consider the potential longer term impact, particularly with regard to bereavement outcomes.36 Moreover, future testing should incorporate an experimental design (ideally via a controlled trial), given the limitations of pre–post approaches,37 and also examine the long term outcomes of attending this type of programme.

The study described herewith did not examine whether or not the intervention led to referrals to members of the multidisciplinary palliative care team. This would be worth exploring, for if the intervention acts as a referral source this may be a resource effective strategy. It would also be worth investigating whether or not the intervention has any impact on assisting caregivers to agree to trial care at home as an alternative to hospital care.


Although family caregivers are relied upon to provide much of the care to palliative care patients they receive inadequate preparation, information and support.33 However, the findings of the project described herein and other recent research19 22 33 38 demonstrate that it is possible to prepare family caregivers for their role and reduce their unmet needs, seemingly without a large impact on resources. Nonetheless, research is required to further test these approaches in other settings, by other means (eg, web based), by other methods (eg, controlled trials) and with other caregiver populations, to ensure interventions of this kind are feasible to deliver.


Clinical sites/partners: Calvary Health Care Bethlehem, Victoria, Australia; St John of God, Murdoch Community Hospice, Western Australia and Camden Hospital and Braeside Hospital, New South Wales, Australia. Research assistants and other clinicians who supported the project: Ms Naomi McGowan, Ms Jane Hunt, Ms Nichole Petrie, Ms Meera Agar, Ms Bev Richards, Ms Jacqueline Mazzer, Ms Lisa Willenberg. Associate investigators/collaborators: Professor Linda Kristjanson, Dr Lyn Oldham, Dr Kirsten Auret, Associate Professor Kelli Stajduhar, Ms Jackie Kearney (Department of Health, Victoria, Australia).



  • Funding This study was supported by the National Health and Medical Research Council.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.