One-to-one volunteer support programs for people with cancer: A review of the literature

3.5.1.1. Methodologies for one group descriptive papers 

Nine of the 10 papers in this category (see Table 3) were descriptive studies. Although one paper employed a randomized, waitlist controlled trial design [49], baseline and post-intervention data for the waitlist group were included with data from the intervention group. Five papers (50%) used quantitative methods, three (30%) qualitative and two (20%) used a combination of both. Six papers were retrospective, involving one data collection point; and four were prospective, involving at least two data collection points. While program purpose varied, the research generally reported program benefits, satisfaction or development. One paper was an exception, reporting volunteers’ knowledge and attitudes [42].

Table 3. Research papers—papers with one group descriptive data

	Paper reviewed	Design	Qualitative/quantitative	Unit of analysis	Measures	Number and timing of data points & timing of intervention	Data reported	Key findings	Research rating (out of 13)	Cochrane criteria
	Brown-Hunter and Price [42]	Descriptive-assessing impact of volunteer training	Quantitative	68 volunteers	Their own scale	Pre-test survey before training. Post-test after training. Unclear if support commenced before post-test.	Volunteer knowledge of & attitudes towards cancer	Pre- to post-test improvement in volunteer knowledge of (significant) and attitudes towards cancer	3	No
	Cameron et al. [43]	Descriptive-assessing perceived objectives and benefits of program	Qualitative	14 people with cancer	Interviews with people with cancer and volunteers	Retrospective interviews a few months to several years after program	Experiences with program according to volunteers and people with cancer	Benefits: provide information, emotion and practical support, understanding, empathy, hope, encouragement and reassurance. Service provides understanding, sensitive listener who is non-judgement. Can talk to someone who ‘has been there’. Would prefer closer to diagnosis	4	No
	Dunn et al. [44]	Descriptive-assessing perceived benefits of the program	Quantitative	223 people with cancer	Focus groups held to formulate their own scale and a quality of life scale	1 data point at variable time after service participation (but within 4 months of diagnosis)	Views of people with cancer on helpfulness and unhelpfulness of the program	Overall the program was considered helpful and individual less anxious after visit.	7	No
								Helpful=visited at the time when most needed, volunteer explain about cancer and reassured normality, person felt similar to volunteers/similar experience of cancer.
								Unhelpful=differences in belief, person left with uncertainty about future, difficult to contact.
	Fusco-Karmann and Tamburini [45]	Descriptive-nurses and volunteers rate role of volunteers in hospitals and at home	Quantitative	246 people with cancer	Their own scales	One data point after service usage. Timing not indicated	Demographics, person's opinions of importance of volunteers, volunteer and nurse views on activities of volunteers	Home visit volunteers rated higher than hospital volunteers by people with cancer. Volunteers rate their activity levels higher than nurses rate volunteer activity levels.	4	No
				119 volunteers
				273 health professionals
	Giese-Davis et al. [46]	Descriptive, impact of peer support provision on support recipients and providers	Quantitative	25 support providers	CES-D, FACT-B, Post traumatic stress Disorder Checklist-civilian version, CECS, WAI, Brief Cancer Behaviour Inventory, their own scales	Baseline before intervention, 3, 6, and 12 month follow-up.	Baseline data for volunteers and patients on all scales. Change in measures over time for volunteers and patients	At baseline, patients had higher depression and post-traumatic stress and lower emotional well-being than volunteers. Patients’ post-traumatic stress decreased and well-being increased.	8	No
						Contact reports after each contact.		No change in volunteers’ well-being. Volunteers became less satisfied with their medical care.
				29 women with breast cancer
	Kleiman et al. [47]	Descriptive-development of visitor program	Quantitative and qualitative	Volunteer and person with cancer data (number in analyses unclear)	Not indicated	Not indicated	Process of program development, service usage, volunteer experiences with program	Main finding: volunteers used the service to work out their own issues	0	No
	May et al. [48]	Descriptive-satisfaction with program	Quantitative and qualitative	36 people with cancer	Their own scales and interviews	Volunteers complete forms after visit and interviews with people with cancer and family 1–6 months after visit	Demographics, content of visits, person with cancer and relative satisfaction with program	Majority responded favourably to program and said it improved their outlook	3	No
				2 volunteers
				13 relatives
	Schover et al. [49]	Pre-post test design assessing impact of counselling program	Quantitative	48 women with breast cancer	Brief Symptom Inventory, FACIT-Sp, Female Sexual Functioning Inventory, Menopause Symptom Checklist, Dyadic Adjustment Scale, own scales	Baseline (before intervention), post waitlist (for wait list controls), post-counselling, 3-month follow-up.	Demographics, baseline data on scales compared with normative data, change in mean scores on measures across baseline, post counselling and 3-month follow-ups	Depression and menopause symptoms decreased between baseline and post counselling, sexual functioning increased between post counselling and 3-month follow-up. Sexual knowledge increased between baseline and post counselling	9	No
	Smith et al. [50]	Descriptive-development of and experiences with training program and service	Qualitative	Person with cancer and volunteer data (number in analyses unclear)	Attending Behaviour Scale, Empathy Scale, their own surveys	Volunteers assessed before and after training. Volunteers and people with cancer complete forms after each contact. People with cancer complete evaluation after each contact. No baseline survey for people with cancer	Experiences with program according to volunteers and people with cancer	Volunteers found training acceptable, people with cancer reported benefiting from experience	4	No
	Sutton and Erlen [51]	Descriptive-changes in interpersonal relationships and quality of life associated with peer support	Qualitative	31 patient-volunteer support provider dyads	Interpersonal Relationship Inventory, own open ended questions assessing interpersonal relations and quality of life (MOJ) and contact logs	Baseline demographics completed at study entry. Patients and volunteers complete MOJ during first and last week of intervention. Volunteers and patients complete contact logs.	Demographics, perceptions of interpersonal relationships and quality of life before and after intervention. Perception of support process.	Most support provided over the phone. Patients expected to receive information and support and volunteers expected to provide support. Expectations met for volunteers and patients. Most dyadic relationships supportive and themes discussed changed little over time.	7	No

3.5.1.2. Findings for one group descriptive papers 

The data presented in these papers suggest that support recipients and volunteers value these programs. Brown-Hunter and Price [42] found a significant improvement in volunteers’ knowledge of cancer after training and service provision. Dunn et al. [44] found that support recipients were less anxious after volunteer visits and that volunteers provided shared experiences. Unhelpful experiences were when volunteers and support recipients had differences in beliefs or when volunteers were difficult to contact [44]. Papers found that support recipients received the support program favorably [48], generally felt supported [51], and benefited from the program [43], [46], [50]. In one study, people with cancer rated volunteers that visited homes higher on psychosocial support than those at hospitals [45]. In another study, counseling by a peer volunteer improved knowledge of sexuality issues, menopause symptoms and emotional well-being [49].

One paper did not report good program outcomes with regard to volunteers [47]. This program used people with current cancers as volunteer visitors to dying cancer patients; the authors reported that the volunteers used the service to resolve their own issues. The authors focused on problems encountered with this population as volunteers [47].

One paper reported the effect of peer support on volunteers and recipients [46]. This quantitative study found positive benefits for support recipients and no adverse outcomes for volunteers. Two studies found that people with cancer perceived the contact with peers as more effective than did the peer volunteers [46], [51].

3.5.1.3. Rating of research quality for one group descriptive papers 

The overall research quality rating (see Table 1 for rating checklist) for these papers was poor to fair, ranging from 0 to 8 out of 13 (mean=4.8). In general, these papers provide little evidence for the effectiveness of these programs.

Not surprisingly, because of their assigned category, none of the papers presented comparison group data; although one involved randomization, no control data were presented [49]. Four papers (40%) presented aims and response rates, six (60%) described measures used and seven (70%) provided a clear indication of data collection points in relation to program commencement. Four studies [42], [46], [49], [51] had a pre-program (baseline) measure and three presented some data from baseline measures [46], [49], [51]. In one study [42] data related to volunteer knowledge and attitudes, rather than outcomes for support recipients. Seven papers (70%) presented some post-intervention scores. One paper presented results from slopes analyses to examine change over time [46]. Seven papers (70%) provided a clear indication of the number of participants in analyses.

3.5.2.1. Methodologies for comparative papers 

All six papers in this category (see Table 4) reported quantitative data and involved comparison between people with cancer, participating in volunteer programs and not participating. Only one paper [56] reported on a non-peer-support program. All papers but one [56] were retrospective, with time since program participation varying across studies. Of the five retrospective papers, three surveyed or interviewed a specific population, ascertained program users, and compared these to non-users in the sample [31], [53], [55]. One study recruited two groups of people with cancer: program participants and a comparison sample of non-program participants recruited through hospital records [52]. Another study compared people recruited from hospitals running the support program and hospitals where the program was unavailable [54].

Table 4. Research papers—non-randomized comparative studies

	Paper reviewed	Design	Qualitative/quantitative	Unit of analysis	Measures	Number and timing of data points & timing of intervention	Data reported	Key findings	Research rating (out of 13)	Cochrane criteria
	Ashbury et al. [52]	Descriptive-compared people who did and did not participate in service	Quantitative	367 people with cancer (175 in program)	FLIC, FACT, Duke Social Support, their own scale	One data point at unknown time after service participation	Demographics, function social support, satisfaction	76% of people with cancer satisfied with program, women having surgery & chemo less satisfied than those having surgery and radio. Higher function status and relationship with Dr for people in service	8	No
	Edgar et al. [53]	Descriptive-compared people who did and did not participate in service	Quantitative	177 people with cancer (138 in program)	POMS, FACT, LOT, Hope and Cope questionnaire	Measures completed at 4 and 12 months post diagnosis. Unclear where program participation falls	Demographics, mood, impact of cancer, optimism, usefulness of service	Service users had significantly higher distress and fatigue and significantly poorer health and well-being. Users physical wellbeing improved significantly from Time 1 to Time 2	8	No
	Geiger et al. [54]	Descriptive-compared people at program and non-program sites	Quantitative	307 people with cancer (82 in program)	Their own scale	Once four months post diagnosis. After service participation	Demographic, satisfaction, and views on program	Those in program sites had significantly higher satisfaction with and ratings of most hospital services and support	8	No
	Rankin et al. [55]	Descriptive-population based survey compared service users and non-users	Quantitative	544 people with cancer (196 in program)	Their own scale	One data point between 6 and 12 months post diagnosis. After service participation	Demographics, sources of support, experiences with program	Majority of non-users said they received support elsewhere. Significantly more likely to use program if had mastectomy. Few demographic differences between users and non-users. 53% of users benefited from sharing experiences with peer	9	No
	Rogers et al. [31]	Descriptive-compared program users and people receiving other forms of peer support	Quantitative	652 people with cancer (476 in program)	Interviews	Once 6–16 months post surgery. Unclear how far after program participation	Opinions of benefits and criticisms of program. Effect of different types of peer support on outcomes	Most people not involved in program report not having the opportunity to be involved. 58% of participants said the program was helpful. Of benefit to the greatest number participants (75%) was seeing the recovered survivor. No significant difference in outcomes across groups.	7	No
	Ventafridda et al. [56]	Controlled trial – compared outcomes of people receiving and not receiving service	Quantitative	52 people with cancer (35 in program)	Integrated Pain Score, Karnofsky, QLI, Spitzer, LASA	Three data points- second is 2 weeks after first, and then third is 4 weeks after that. Unclear if there is a true baseline.	Pain, performance status, anxiety, side effects, quality of life.	Significantly less pain and higher quality of life in intervention group.	6	No
				Family members (number in analyses unclear)

In the non-retrospective study, people with cancer living near the program were allocated non-randomly to the intervention while people in rural areas were controls [56]. This was possibly a controlled before-and-after study [56]. The authors reported data from weeks 0, 2 and 6, suggesting there was a baseline measure. Yet timing of program commencement is not clear and there is no evidence of a data collection point before volunteer support [56]. Edgar et al. included two assessment points but both times were after program participation [53].

Sample sizes for four studies were over 177 people with cancer, with at least 138 in the volunteer program [31], [52], [53], [55]. Although one study involved 307 participants, only 82 were in the intervention group [54]. One paper involved only 52 patients, with 35 in the intervention group [56].

Three papers [52], [53], [56] used published instruments and reported differences in outcomes such as function, mood, support, optimism and quality of life for program and non-program users. The other three papers reported satisfaction, experiences with the program and perceived program benefits and shortcomings.

3.5.2.2. Findings for comparative papers 

Few negative findings were reported in these papers. Ashbury et al. reported that 76% of support recipients were satisfied with the program [52], Rankin et al. indicated that 58% of women reported benefiting from sharing experiences with a peer [55]. While 58% of participants in Rogers et al. found the program helpful, there were no significant differences in outcomes between program participants and people receiving other peer support [30]. In another study, people at program sites reported significantly higher levels of satisfaction with most hospital services compared to people at non-program sites [54]. It is possible, however, that other differences existed between the two sites.

Ashbury et al. found that, compared to non-users, program participants had significantly higher functioning status and better relationships with their doctors [52]. Edgar et al. reported that program users had significantly higher distress and fatigue and significantly poorer well-being than non-users at time 1, but by time 2 well-being improved significantly for the user group, with no significant difference between groups for mood [53]. The retrospective non-randomized design of these studies makes it difficult to determine if the program or other factors influenced these outcomes.

Ventafridda et al. also found differences between program users and non-users at follow-up [56]. Support recipients in their program had significantly less pain and higher quality of life than people not in the program. While promising, because group assignment was based on residency it is possible that differences in urban and rural living influenced findings.

3.5.2.3. Rating of the research quality for comparative papers 

Although higher than the previous category of papers, the average research quality rating for these papers was fair, ranging from 6 to 9 out of 13 (mean=7.7). None were eligible for Cochrane review.

Five papers presented aims and response rates, five described measures and one provided a clear indication of when data were collected in relation to program commencement. No study had a clear pre-program measure, although, as noted previously, it is possible that one study used a baseline [56]. All papers presented post-program scores and comparisons of scores across groups.

3.5.3.1. Methodologies for RCTs 

Four of the 20 research papers were RCTs (see Table 5). All reported quantitative data, with one also presenting qualitative data [58]. Two papers involved comparison of several conditions [32], [33], [57], two compared the volunteer program to a control or waitlist group [58], [59]. Only one reported on a non-peer-support program [32].

Table 5. Research papers—randomized controlled trials

	Paper reviewed	Design	Qualitative/quantitative	Unit of analysis	Measures	Number and timing of data points & timing of intervention	Data reported	Key findings	Research rating (out of 13)	Cochrane criteria
	Mantovani et al. [32]	Prospective, open, randomized, parallel group design-compared people receiving (A) psychopharmacological treatment, (B) A plus volunteer support, and (C) B plus psychotherapy	Quantitative	72 people with cancer (number in each group unclear)	Karnofsky, Scott-Huskisson, BDI, STAI, QLI, FLIC, SCI	Before cancer treatment, 2.5 months later and 2.5 months after that. Unclear when volunteer support is introduced in relation to surveys	Demographics, performance status, pain, depression, anxiety, functioning, impact of side effects.	Significant improvement in performance status, pain, depression, and state and trait anxiety for people in conditions B and C but not A. No significant differences for quality of life or functional living	7	Yes
	McArdle et al. [57]	RCT-compared 4 conditions—(1) routine ward staff, (2) 1 plus breast care nurse, (3) 1 plus volunteer, and (4) 2 plus 3	Quantitative	202 people with cancer (67 in group 1), (70 in group 2), (67 in group 3), (69 in group 4)	GHQ, HADS	After 1st post-operative clinic visit and at 3-, 6-, and 12-month post surgery. Unclear where intervention fell but possibly before first data point	Demographics, general health, anxiety and depression	Outcomes improved significantly in group receiving breast care nurse support but not group with volunteer or group with usual care	9	Yes
	Rudy et al. [58]	RCT-compared people in peer support program with wait listed group	Quantitative and qualitative	59 people with cancer (29 in program)	Social Support Survey, evaluation surveys completed by people with cancer and volunteers	One data points, within a week of 2nd treatment. After scheduled intervention contacts	Social support, experiences with program	Those in intervention have significantly more sources of social support than control group. People with cancer benefited from information and emotional support.	10	Yes
				29 volunteers
	Weber et al. [59]	RCT-compared people in peer support and usual care	Quantitative	30 people with cancer (15 in program)	Modified Inventory of Socially Supportive Behaviors, Stanford Inventory of Cancer Patient Adjustment, Geriatric Depression Scale, UCLA Prostate Cancer Index, Charlson Index	Three data points- baseline (pre-intervention), then 4 weeks later and 4 weeks after that	Demographic, depression, self-efficacy, social support, urinary and sexual function/bother, PSA	Significant difference between groups for depression at 4 weeks. No significant difference in social support and depression at 8 weeks. No significant difference for urinary bother or function or for sexual function. Significant difference between groups at week 8 for sexual bother.	13	Yes

All studies were prospective and all but one [58] involved three or four data collection points. Only one paper provided a clear indication of a baseline measure [59]. Mantovani et al. [32] and McArdle et al. [57] did not clearly indicate when questionnaires were administered in relation to program implementation. In Mantovani et al.'s study, people in groups A, B and C commenced psychopharmacotherapy 10 days prior to chemotherapy [32]. Questionnaire one was completed immediately before the start of chemotherapy. Volunteer support (group B) was provided for the duration of cancer treatment but commencement time is unspecified. Psychotherapy (group C, which also included a volunteer support component) was provided on a weekly basis for the duration of chemotherapy but commencement time is unspecified. Presumably, volunteer support and psychotherapy commenced after questionnaire one, but this is unclear. It is clear, however, that there was no baseline for psychopharmacotherapy and this was the only condition for group A.

McArdle et al. are also unclear regarding timing of the intervention in relation to baseline questionnaire administration [57]. This study consisted of four conditions: (1) routine care (RC), (2) RC plus volunteer, (3) RC plus breast care nurse, and (4) RC plus volunteer and nurse. Questionnaire one was completed at the first post-operative clinic visit. Volunteer support (conditions 2 and 4) commenced after discharge but it is unclear whether before or after questionnaire one. Breast care nurse support (conditions 3 and 4) commenced peri-operatively, as did routine ward care, suggesting these interventions preceded questionnaire one.

The sample sizes for these papers were relatively small. McArdle et al.'s study involved the largest total sample (N=202) but, divided across four conditions, group sizes were considerably smaller (67–70) [57]. Mantovani et al.'s study involved 72 people with cancer, distributed across three groups [32]. One paper had 29 program participants and 30 waitlist controls [58]. The paper with the highest research quality rating [59] had the smallest sample (15 in the program and 15 controls).

All papers presented data from published, validated instruments, although one paper [58] mainly presented process evaluation data. While this data provided useful feedback regarding the program's benefits, very little constituted RCT data. All other papers presented outcome data, including performance status, pain, mood, functioning, bother, impact of side effects, general health and social support.

3.5.3.2. Findings for RCTs 

Mantovani et al. reported some significant improvements in performance status, pain, depression and anxiety for people in conditions B and C (i.e. psychopharmacotherapy plus volunteer support or volunteer support and psychotherapy) but not in group A (psychopharmacotherapy alone) [32]. Although the combination of interventions means it is not possible to ascertain where the effect lies, both groups showing improvement involved the volunteer program. The lack of improvement for group A is perhaps not surprising as their intervention commenced prior to questionnaire one. McArdle et al. found significant improvements for people supported by the breast care nurse but not for those receiving volunteer support or routine care [57]. However, as the conditions commenced at different times, and there was no apparent baseline for the nurse condition, it is difficult to draw definite conclusions about the impact of these interventions.

Rudy et al. mainly presented post-intervention data and reported that people in the intervention had significantly more sources of social support post-intervention than people in the control group [58]. While this may reflect the social support provided in the intervention, as analyses did not control for baseline social support, this is unclear. Support recipients reported benefiting from the support provided by volunteers [58].

Despite its small sample size, it may be easier to draw more valid conclusions about the intervention described by Weber et al. [59] as it lacks some of the methodological confusion of the RCTs described above. This study found a significant difference between intervention and control groups for depression midway through the program (4 weeks), but not at its end (8 weeks) [59]. No significant differences were found for social support, urinary function, or sexual function at 4 or 8 weeks, but a significant difference was observed between groups on sexual bother at 8 weeks. While needing confirmation in a larger RCT, these findings suggest that the peer-support model used may have some immediate impact on depression and a more gradual impact on sexual bother.

3.5.3.3. Rating of the research quality for RCTs 

Research quality ratings for the RCTs ranged from 7 to 13 (mean=9.8). Only one paper scored 13 [59]. All four papers met the requirements for inclusion in a Cochrane review [15]. The limited RCT data presented in one paper, however, made inclusion questionable [58].

As these papers were RCTs they scored higher than the other research papers from the outset. In addition, all presented aims and described the measures used. Three papers described response rates, and two provided a clear indication of when data were collected in relation to the program [58], [59]. Only one paper provided pre-program data [59], including comparisons across groups. All papers provided data post-program, including comparisons across groups. The number of participants in analyses was clear for all except one paper [32].