Partners and close family members of long-term cancer survivors: health status, psychosocial well-being and unmet supportive care needs

Authors


School of Health and Social Care, Oxford Brookes University, Jack Straw's Lane, Oxford, OX3 0FL, UK. E-mail: ewatson@brookes.ac.uk

Abstract

Background

A cancer diagnosis can have a profound impact on partners and close family members of patients. Little is currently known about the long-term impact.

Objectives

The objective of this study is to describe health status, levels of anxiety and depression, unmet supportive care needs and positive outcomes in the partners/family members of breast, prostate and colorectal cancer survivors 5–16 years post-diagnosis.

Methods

Patients in a linked study were asked to invite a partner or other close family member to complete a self-administered postal questionnaire. Data were analysed by cancer site and time since diagnosis. Matched comparisons were made between cancer patients in the linked study and their partners.

Results

An expression of interest was received from 330 partners/family members, and 257 questionnaires (77.9%) were returned. Health status and levels of anxiety and depression were comparable with population norms. Respondents reported an average of 2.7 unmet needs from 34 possible options. Hospital parking, information about familial risk, help managing fear of recurrence and coordination of care were the most cited unmet needs. There was little variation in health status, psychological morbidity and unmet needs by cancer site or time since diagnosis. Concordance between patients and partners was low for anxiety but higher with respect to positive outcomes and some unmet needs.

Conclusions

Most partners/family members of long-term cancer survivors report few ongoing issues. However, a small proportion (<10%) have high levels of anxiety and/or moderate or strong unmet needs. Strategies for identifying this group and addressing their needs are required, while allowing the majority to resume normal life. Copyright © 2011 John Wiley & Sons, Ltd.

Introduction

Improvements in early diagnosis, screening and treatments have seen substantial increases in 10-year survival rates for many cancers. Seventy-three per cent of breast cancer patients, 54% of prostate cancer patients and 45% of colorectal cancer patients now survive 10 years or more [1]. Overall, there are currently over two million cancer survivors in the UK, of whom 1.2 million were diagnosed 5 or more years ago [2, 3]. However, survivorship research has identified that cancer impacts the physical, psychological and social well-being of patients both in the short term and longer term. Even when the cancer is successfully treated and the patient discharged, side effects from the treatment may remain [4], late effects may emerge [5] and worries about recurrence may continue for many years [6].

Cancer can also have a significant impact on partners and other family members of cancer patients [7-9]. A diagnosis of cancer can provoke stress and anxiety for the partner [10], with some studies suggesting that distress [11, 12] and fear of recurrence [13] may be higher in partners than in the patient, even after treatment has ended. In addition to coping with their own stress, partners or close family members are often the primary provider of support to the patient, with little support for or recognition of their own needs, and may experience a lower quality of life [13] and substantial unmet needs [14-16]. Research also suggests that most partners and family members of cancer patients nevertheless come to terms with their experience and may even experience positive effects [17, 18].

In drawing comparisons between their experiences, a number of studies have reported a strong correlation between partners and cancer patients in their psychological adjustment and in their level of psychological distress [19]. However, partners' cancer experiences have also been found to differ from that of patients in terms of depression [10] and anxiety [20], and in relation to post-traumatic growth [17].

In a study of older individuals who had survived cancer for an average of 10 years and their family members (109 dyads), Bowman and colleagues found that family members appraised the cancer experience as more stressful than did cancer survivors themselves. Family members appeared to be less able to let go of the diagnosis and treatment phase of the disease and thus it continued to have consequences for their appraisal of the overall cancer experience [12]. A recent study by Hodgkinson and colleagues in Australia also looked at longer-term cancer survivors and their partners (154 dyads, 1–11 years post-diagnosis, mean 4.2 years, 50% gynaecological cancers) [20]. This study found that partners reported high levels of anxiety and supportive care needs and that within partnerships, couples reported both shared and unique needs, although agreement on shared needs was low [20].

However, most of the existing literature on partners and close family members focuses on the period around diagnosis and treatment or end-of-life care and relatively little is known about their cancer experience in the extended survivorship period.

This study therefore aimed to

  1. describe health status, rates of possible or probable anxiety, possible or probable depression, unmet supportive care needs and positive effects experienced by the partners/family members of long-term cancer survivors;
  2. examine differences in the aforementioned measures by cancer site (breast, colorectal and prostate) and by time since diagnosis (5–16 years);
  3. and assess concordance in possible or probably anxiety, possible or probable depression, unmet needs and positive outcomes in matched pairs of patients and partners.

With increasing numbers of cancer survivors, a good understanding of long-term needs is essential to inform national initiatives seeking to improve care provision and maximise the number of patients, partners and families that can resume a normal life after cancer.

Methods

Respondents and procedure

Partners and close family members were recruited through a linked study that included a sample of long-term survivors of cancer from two cancer registries in the UK, covering the Thames Valley and Yorkshire/North East England regions [21]. Two thousand four hundred adults with a diagnosis of breast, colorectal or prostate cancer 5–7, 9–11 or 14–16 years prior to September 2009 were sampled to collect experiences of patients at three different periods since diagnosis. The sample was first checked to exclude recent deaths and patients who had moved out of the area or had metastatic disease. The GP of remaining patients were then contacted and asked to exclude any patients that they considered unfit to take part, for example because of current serious ill health or recent bereavement.

A total of 1275 patients were invited through their GP to take part in the study and were sent a self-administered questionnaire. Of the patients, 659 (51.7%) returned completed and useable questionnaires. Patients were also asked to invite the person whom they felt had supported them most during their cancer experience to take part in the linked study. In the 330 cases where the patient nominated a partner or family member, a self-administered questionnaire was sent, with a reminder to non-respondents at 1 month. Two hundred fifty-seven (77.9%) questionnaires were returned. Close family members received a slightly modified version of the questionnaire sent to partners and spouses, using more appropriate wording and removing questions about impact on sexual relationships. The questionnaire data were double-entered into spss (SPSS Inc., Chicago, IL, USA) for analysis.

The study was granted ethics approval by the Oxfordshire B Research Ethics Committee (08/H0605/18).

Measures

Where possible we used the same measures in this study as those used for the linked study with cancer survivors [21] to allow comparisons within couples.

Health status

EuroQol-5 Dimensions and Visual Analogue Scale

The EuroQol-5 Dimensions (EQ-5D) is a self-report measure of health status using five questions which give an indication of any problems in the following dimensions: mobility, health care, usual activities, pain/discomfort and anxiety/depression [22]. Each dimension is scored from 1 (no problem) to 3 (extreme problem) which can be used to calculate a single index score (Range = −0.594–1.000) [21]. The EQ-5D also generates a self-rating of health-related quality of life scored from 0 to 100 on a Visual Analogue Scale (VAS) akin to a thermometer. The EQ-5D has previously been used in research on cancer patients [23, 24].

Anxiety and depression

Hospital Anxiety and Depression Scale

The Hospital Anxiety and Depression Scale (HADS) is a validated, widely used screening instrument consisting of 14 items used to identify levels of anxiety and depression [25]. Individuals with a score of 8–10 are classified as possible cases and those with a score of ≥11 as probable cases. The HADS has also been used in studies of both cancer patients and partners [26].

Unmet supportive care needs

Cancer Survivors' Partners Unmet Needs measure

The Cancer Survivors' Partners Unmet Needs (CaSPUN) is a self-report measure of needs specifically designed to capture likely issues for partners of long-term cancer survivors [27]. It was derived from the measure of unmet needs developed for cancer patients, which was used in the linked study [21]. For each of the 34 listed potential needs, respondents are asked to indicate whether in the last month that ‘need’ was not applicable, was being met or was unmet. For unmet needs, respondents are asked to indicate how important they perceive the unmet need to be (weak, moderate or strong). For this study, we adapted the wording of some questions to make them suitable in the British context and added one question relating to information about the familial risk of cancer. We also excluded an additional section on positive growth.

Positive outcomes

Five questions were asked relating to possible areas of positive outcomes. These questions were derived from each of the five factors of the post-traumatic growth inventory, a much longer and more detailed measure than we could incorporate within our questionnaire [28].

Socio-demographic characteristics

Standard questions were asked regarding the respondent's socio-demographic characteristics and the cancer patient's cancer site and time since diagnosis.

Analysis

Scores from the EQ-5D were converted into a weighted index and compared with UK population norms [29]. HADS scores were calculated, and respondents were categorised into possible, probable or non-cases of anxiety and depression and compared with population norms using 95% CI [30]. Although the CaSPUN measure captures total needs as well as unmet needs, in this study we focussed only on unmet needs. CaSPUN data were analysed as the mean total number of unmet needs across the whole sample, as the proportion of respondents rating each individual item as unmet and as the proportion indicating a moderate or strong unmet need for each item.

Statistical analysis was conducted using spss v17 [31]. Descriptive statistics for the sample was produced and compared using z-tests for proportions between the partner and patient samples. T-tests were performed to test for significant differences between means between patient and partner samples and of unmet needs in groups based on time since diagnosis and site of cancer.

Regression analysis was performed to test for significant associations between reported numbers of unmet needs with caseness for anxiety, health status (EQ-5D) and age. Poisson regression was chosen because of the measure of unmet need comprising count data and was tested with a Wald chi-squared test. Too few respondents reported possible or probable depression to test for possible association between unmet needs and depression.

Data from the linked patient study were matched with partner data for couples where both partner and patient had provided completed questionnaires. Matched analysis focussed only on couples, that is husbands, wives and partners, to ensure consistent relationship status and allow comparison with other studies. The relationship between anxiety in the partner and anxiety in the patient was assessed in terms of the relative risk of possible or probable anxiety in the partner where the patient met the criteria for possible or probable anxiety. Concordance with respect to unmet needs and positive outcomes was expressed as the proportion of patients who agreed with their partner. Twenty-three questions were common to both the patient (CASUN) and partner (CaSPUN) versions of the measure of unmet need, and both patients and partners completed the same positive outcome questions. Concordance on reported unmet need was measured only for those couples where the partner reported an unmet need. The same approach was used for assessing concordance on positive outcomes. In both cases, the measure used was the proportion of patients who agreed with their partner in stating an unmet need or positive outcome for a given item.

Results

Key sample characteristics are presented in Table 1. Of the total patient sample, 63.6% were living with a spouse or partner, and 44% were still attending hospital follow-up [21]. Eighty-two per cent of our partner/family member sample was partners or spouses of cancer patients, and most of them were in stable long-term relationships with 79.8% indicating they had been with their partner more than 10 years. Respondents were predominantly White, and the largest proportion was retired.

Table 1. Sample characteristics
 PartnersPatients [20]
  1. SD, standard deviation.

 n%n%
Age (years)    
<50249.3192.9
50–7011444.423135.1
>7010540.940962.1
Not specified145.4n/an/a
Mean (SD)65.5(12.1)71.6(9.9)
Gender    
Male11845.930245.8
Female13954.135754.2
Site of patient's cancer    
Breast10641.225839.2
Colorectal8131.520531.1
Prostate7027.219629.7
Time since patient's diagnosis    
5–7 years15560.338858.9
9–11 years5822.614221.5
14–16 years4417.112919.6
Employment status    
Retired13251.449276.3
Employed7629.610516.5
Other4919.0487.3
Ethnicity    
White25297.963197.5
Other52.1162.4
Location    
Northern8432.726640.4
Thames Valley17367.339359.6
Relationship to index patient    
Partner/spouse (total)21282.5  
Husband10942.4  
Wife9637.4  
Partner72.7  
Other family member (total)3815.4  
Daughter/son2810.9  
Sibling41.5  
Other/non-specified72.7  
Total sample257659  

The linked patient study had a particular interest in experiences of discharge and so oversampled patients in the 5–7 year since diagnosis group (2:1:1), accounting for the higher response from this group of patients, also reflected in the partner sample.

Health status, anxiety and depression

Table 2 presents summary health status scores and possible or probable cases of anxiety and depression. The mean EQ-5D index and VAS scores in the partner and family member sample were similar to those reported for the general population [29] and also did not vary from the population norms when analysed by age bands.

Table 2. Health status scores, and levels of anxiety and depression
 NPartners/family members (95% CI)UK norms (95% CI)
  1. EQ-5D, EuroQol-5 Dimensions; EQ-VAS, EuroQol-Visual Analogue Scale; HADS, Hospital Anxiety and Depression Scale.

EQ-5D index (mean score)2560.83 (0.80–0.86)0.86 (0.85–0.87) [30]
EQ-VAS (mean)24978.8 (76.9–80.6)82.5 (81.9–83.0) [30]
Anxiety HADS247  
Possible case 8–10 (n)3714.4% (10.1–18.8)20.6% (18.7–22.5) [31]
Probable case 11+ (n)249.3% (5.8–13.0)12.6% (11.0–14.1) [31]
Depression HADS251  
Possible case 8–10 (n)103.9% (1.5–6.3)7.8% (6.6–9.0) [31]
Probable case 11+ (n)83.1% (1.0–5.3)3.6% (2.7–4.5) [31]

Cases of both possible and probable anxiety and depression, as measured by the HADS, were consistent with those reported for the UK general population [30]. Only a small number of respondents recorded scores that would suggest possible (n = 10) or probable (n = 8) caseness for depression. There was no significant difference in the proportions of possible or probable cases of anxiety by cancer site, or time since diagnosis.

Unmet supportive care needs

One hundred twenty-seven (49.4%) respondents reported at least one unmet supportive care need, with an average of 2.7 unmet needs per respondent (standard deviation (SD) = 5.2). Respondents that expressed at least one unmet need listed an average of 5.5 unmet needs (SD = 6.2). Table 3 presents unmet needs by cancer site and time since diagnosis, showing no significant differences between these groups. The 10 most commonly listed unmet needs, and how often these unmet needs were classified as moderate or strong, are presented in Table 4. The number of unmet needs was associated with anxiety (Wald-χ2 = 15.6, p < 0.0001), but not with health status (Wald-χ2 = 1.3, p = 0.25) or age (Wald-χ2 = 0.76, p = 0.38).

Table 3. Unmet supportive care needs, by type of cancer and time since diagnosis
 NMean unmet needs (95% CI)At least one unmet need (%)At least one moderate or strong unmet need (%)
All2572.7 (2.1–3.3)127 (49.4)68 (26.5)
Type of cancer
Breast1063.2 (2.1–4.3)59 (55.7)33 (31.1)
Colorectal812.2 (1.3–3.1)35 (43.2)21 (26.0)
Prostate702.5 (1.3–3.7)33 (47.1)14 (20.0)
Years since diagnosis
5–7 years1552.8 (2.0–3.6)80 (50.6)48 (31.0)
9–11 years582.3 (1.2–3.4)25 (43.1)8 (13.8)
14–16 years443.1 (1.2–5.0)22 (50.0)12 (27.3)
Table 4. Ten most common unmet supportive care needs and strength of needs
Total n = 257Unmet needModerate or strong need
n%n%
I need more accessible hospital parking5621.84517.5
I need information about the chance of one of the members of my family developing cancer4517.52710.5
I need help to manage my concerns about the cancer coming back4015.6176.6
I need to know that my spouse's/partner's doctors all talk to each other to coordinate their care3714.4238.9
Due to my spouse's/partner's cancer, I/we need help getting life and/or travel insurance3312.8228.6
I need to feel like I am involved in my spouse's/partner's health together with the medical team3011.7176.6
I/We need help to find out about financial support and/or government benefits to which I/we are entitled2911.3145.4
I need my spouse/partner to have a case manager (single contact) to whom I can go to find out about services whenever they are needed2710.5176.6
I need local health care services that are available when my spouse/partner requires them249.3155.8
I need any complaints regarding my spouse's/partner's care to be properly addressed238.9166.2

Positive outcomes

Positive outcome findings are reported in Table 5. For all but one of the five positive outcome questions, the largest proportion of respondents reported ‘no change’. Two-thirds of respondents reported having a greater appreciation of life, almost half felt they had become a stronger person, and a quarter felt the way in which they related to other people had improved.

Table 5. Positive outcomes
 Yes % (n)No change % (n)No %(n)
Do you feel that you have a greater appreciation of life?66.6 (168)32.5 (82)0.8 (2)
Do you feel that you are a stronger person?43.0 (108)49.8 (125)7.2 (18)
Do you feel that the way in which you relate to other people has improved?25.9 (65)72.5 (182)1.6 (4)
Do you feel that your religious faith or sense of spirituality has strengthened?16.7 (42)78.5 (197)4.8 (12)
Do you feel that new opportunities have become available which wouldn't have otherwise?7.9 (20)86.9 (219)5.2 (13)

Comparisons between patient and partner couples

Matched analysis was possible on complete data from 186 patients and their partners. There was no significant difference between the characteristics of the matched couple population and the main sample.

Patients and partners independently reported identical rates of possible or probable cases of anxiety (20.9%, 39/186). However, relative risk analysis indicated that there was no greater risk of anxiety in the partners of patients whose HADS score indicated possible or probably anxiety (RR = 0.68, 95% CI 0.37–1.23).

Table 6 presents within-couple concordance on unmet needs for the nine most common unmet needs reported by partners and where a similar question was asked of the patients. (One of the top 10 needs (see Table 4) was asked only in the partner questionnaire). Agreement between couples was highest (≥50%) on the need for help to manage fears of recurrence, the need for coordinated care and for complaints to be properly addressed. Within-couple concordance on positive outcomes was highest for a strengthened feeling of religious faith or spirituality (60%, 21/35), having a greater appreciation of life (43.4%, 79/182) and improvement in relating to other people (42.2%, 19/45).

Table 6. Within-couple concordance for the most common unmet needs, for couples where partner has reported the need as unmet
 Number of partners reporting the need as unmetNumber of matched patients reporting the need as unmet% Concordant couples
I need more accessible hospital parking371745.9
I need help to manage my concerns about the cancer coming back271659.2
I need to know that all my/my partner's doctors talk to each other to co-ordinate care241458.3
Due to the cancer, I/we need help getting life and/or travel insurance25832.0
I/we need to feel like I am managing my/my partner's health together with the medical team23834.7
I/we need help to find out about financial support and/or government benefits to which I/we are entitled19526.3
I/we need an ongoing case manager (single contact) to whom I/we can go to to find out about services whenever they are needed19736.8
I need local health care services that are available when I/my partner need them14642.8
I need any complaints regarding my/my partner's care to be properly addressed16850.0

Discussion

This is the first study conducted in the UK that addresses the health and supportive care needs of the partners or family members of long-term cancer survivors. Overall, the study found that most partners or family members of long-term survivors of breast, colorectal and prostate cancer, with no signs of recurrence at the time of the questionnaire, have similar health status and levels of anxiety and depression to that reported by the general UK population. Most also reported low levels of unmet need relating to their cancer experience, although almost half reported at least one unmet need. Hospital parking, information about familial risk, help managing fear of recurrence and coordination of care were the most cited unmet needs. A positive relationship was found between levels of moderate or strong unmet need and anxiety. Respondents also indicated they had been able to derive some positive gain from their cancer experience, with two-thirds reporting they now had a greater appreciation of life and over 40% feeling they had become a stronger person. Matched analysis of data from couples showed no evidence of increased risk of anxiety amongst partners of patients with possible or probably anxiety as measured by the HADS. For couples where the partner had reported the need as unmet, moderate levels of concordance were found on some, but not all of the nine most common unmet needs; moderate levels of concordance were also found on positive outcomes for couples where the partner reported a positive outcome.

Hodgkinson et al. previously reported on the psychological morbidity and unmet needs in a study of Australian cancer survivors [20]. We report here similar levels of anxiety and depression in our population and a slightly lower mean number of unmet needs. Both studies report a positive relationship between unmet need and anxiety. However, the Australian sample was younger (mean = 60.4 years, SD = 10.3), included patients closer to diagnosis (mean = 4.2 years post-diagnosis) and also included gynaecological cancers. That study found longer-term survivors (over 3 years) had higher levels of anxiety and distress than those 1–3 years from diagnosis. Our study, which found no difference between those diagnosed 5–7, 9–11 and 14–16 years previously, suggests that a majority of partners may have come to terms with their cancer experience by 5 years and remain this way even 10–15 years after diagnosis—although this would need confirmation in a prospective longitudinal study.

Although a meta-analysis of studies reported strong correlations between partners and cancer patients in relation to psychological adjustment and distress [19], we did not find a relationship between patients and their partners in relation to possible or probable anxiety. This could be due in part to differences in the way anxiety and psychological morbidity are measured in other studies and in part to the longer time since diagnosis in this sample.

With regard to unmet supportive care needs, we found evidence of shared needs on some items but not others. Hodgkinson and colleagues reported low agreement between partner and cancer survivor ratings [20]. In a qualitative study, Sharpe also found that agreement on unmet need within couples was low [16].

Our findings provide further support for previous research [17-19], which found that, in the long term, the majority of partners put their cancer experience behind them. It is also encouraging that many of our participants reported positive outcomes following their experiences. This may be due to the process of psychological adjustment known as response shift [32, 33], which occurs when, following their encounter with cancer, patients and their family members shift their priorities to areas of life which they experience as positive and recalibrate their expectations in line with their changed circumstances.

Despite the overall encouraging findings from our study, there is a minority of partners with significant levels of anxiety and unmet needs. Apart from a need for more accessible hospital parking, the most commonly reported unmet needs in our study were information about familial risk and help managing fears of recurrence, mentioned by approximately one in six respondents. Familial risk could be addressed by the provision of written and verbal information with a referral to genetic counselling services if required. Screening strategies are also currently being developed [34] to identify those with high levels of fear of recurrence so that targeted interventions can be delivered. The other most commonly identified unmet needs relate to the provision of comprehensive and coordinated cancer care. The planned introduction of survivorship care plans, currently being trialled in the UK, which summarise information about the patient's diagnosis, treatment and ongoing management requirements and incorporate psychosocial issues as well as clinical care [35], should facilitate enhanced communication between secondarycare, primary care, patient and family members in the future. This could offer partners the opportunity to raise any ongoing concerns or needs and encourage them to consult at a future point should any new concerns arise.

Methodological issues

A key strength of this study is the recruitment of a large number of respondents through a registry database, across different periods of the survivorship trajectory. However, the design was such that partners were recruited by invitation from patients, and so it is very unlikely that the study will have captured partners of patients where couples separated after the cancer diagnosis, whether or not cancer was a significant factor in the break-up of the relationship. The ethnic homogeneity of the sample also limits the applicability of the findings to minority ethnic groups who may live in inner city areas with stretched health services or have language or cultural barriers that restrict their ability to navigate support and information services.

Although there were no significant differences between respondents and non-respondents in relation to cancer site or time since diagnosis, it cannot be assumed that levels of health status, anxiety and unmet need were also similar between the two groups. Despite emphasising in the study invitation letter that the researchers were just as interested to hear from those who did not feel they had any current problems or issues, it is possible that those who are dissatisfied or have continuing unmet needs were more likely to reply.

There may also be some limitations to the measure of unmet needs we used. Although the questionnaire emphasised that the study focussed on present issues, it is possible that respondents found it difficult to focus on current unmet needs and instead may have reflected back to the time of initial treatment and follow-up. For example, we were surprised that hospital parking emerged as the most common unmet need, particularly as 52.1% no longer attended hospital follow-up appointments. Alternatively, responses may have been influenced by unmet needs arising from current health problems unrelated to cancer. Also, the CaSPUN measure was developed in Australia and there may be cultural differences in understanding and expression of need between the two countries. It is noteworthy that the question we added to the measure on the need for information on familial risk was one of the most frequently reported unmet needs, and others choosing to use the CaSPUN measure may also wish to add this item.

This study was cross-sectional in nature and although we explored variation in outcomes according to time since diagnosis, more costly prospective studies would be required to definitively address changes in outcome variables over time.

Conclusion

This study indicates that the spouses, partners and close family members of long-term survivors of breast, colorectal and prostate cancer diagnosed between 5 and 16 years ago and with no signs of recurrence, generally have similar health status and levels of anxiety and depression to that seen in the general population. Current unmet needs related to their cancer experience are also generally low. However, a minority (less than 10%) does have higher levels of anxiety and moderate or strong unmet needs and could potentially benefit from greater information and support. Health care services need to establish ways of identifying and supporting these partners and family members, whilst leaving the majority to put their cancer experience behind them and move on with their lives.

Acknowledgement

We would like to thank Dr Lesley Smith for providing assistance with statistical analysis.

Ancillary