Social support following diagnosis and treatment for colorectal cancer and associations with health‐related quality of life: Results from the UK ColoREctal Wellbeing (CREW) cohort study

Abstract Objective Social support is acknowledged as important in cancer survivorship, but little is known about change in support after cancer diagnosis and factors associated with this, particularly in colorectal cancer. The CREW cohort study investigated social support up to 2 years following curative intent surgery for colorectal cancer. Methods A total of 871 adults recruited pre‐treatment from 29 UK centres 2010 to 2012 consented to follow‐up. Questionnaires at baseline, 3, 9, 15, and 24 months post‐surgery included assessments of social support (Medical Outcomes Study‐Social Support Survey, MOS‐SSS) and health‐related quality of life (HRQoL). Socio‐demographic, clinical and treatment details were collected. Longitudinal analyses assessed social support over follow‐up, associations with participant characteristics, and HRQoL. Results Around 20% were living alone and 30% without a partner. Perceived social support declined in around 29% of participants, with 8% of these reporting very low levels overall from baseline to 2 years (mean MOS‐SSS overall score < 40 on a scale from 0 to 100). Older age, female gender, greater neighbourhood deprivation, presence of co‐morbidities, and rectal cancer site were significantly associated with reductions in perceived support. Poorer HRQoL outcomes (generic health/QoL, reduced wellbeing, anxiety, and depression) were significantly associated with lower levels of social support. Conclusions Levels of social support decline following colorectal cancer diagnosis and treatment in nearly a third of patients and are an important risk factor for recovery of HRQoL. Assessment of support early on and throughout follow‐up would enable targeted interventions to improve recovery, particularly in the more vulnerable patient groups at risk of poorer social support.


| BACKGROUND
Social support is widely regarded as beneficial to people living with and beyond cancer, particularly for psychological wellbeing, 1 and has been proposed as an important aspect of recovery, 2 as well as in planning survivorship care. 3 Higher levels of support have been shown to be associated with better health outcomes and healthrelated quality of life (HRQoL) following cancer treatment, with most studies in breast cancer. 4,5 However, few studies have investigated the role of social support in colorectal cancer, despite being the most common cancer type that affects both genders. Results from small cross-sectional studies in colorectal cancer have shown associations between lower social support and poorer psychological wellbeing and HRQoL. [6][7][8] Longitudinal studies showing similar associations are limited by short or incomplete follow-up (≤ 1 year) or no pre-treatment baseline data. [9][10][11][12] Furthermore, there is very little published evidence on changing levels of social support following a cancer diagnosis. Studies in breast cancer suggest that support levels decline following diagnosis and treatment, [13][14][15][16] although 1 study reported levels remaining moderately high over 3 years' follow-up. 5 Limited evidence from other cancer sites also varies; support levels were reported to be stable up to 1 year follow-up in a prostate cancer study (N = 134, 17 ), but declined pre-treatment to post-treatment in head and neck cancer (N = 32 18 ). Longitudinal data on social support in colorectal cancer are extremely sparse; a study in China (N = 227) reported declining support up to 1 year following surgery, particularly in women and those with a lower family income. 19 Alternative supportive self-management models for cancer aftercare are being implemented in the UK, such as patient-triggered follow-up whereby patients initiate contact if they experience symptoms or have concerns. Compared with the routine follow-up approach, patients are likely to have less contact with health care professionals, and therefore it is more important than ever to determine patterns of social support following treatment. Identification at diagnosis of individuals who might be at greater risk of poor and declining levels of support as well as the potential effects of this on recovery of HRQoL will help tailor management packages for these patients.
The UK ColoREctal Wellbeing study (CREW) is a large-scale cohort study investigating factors associated with recovery of health and wellbeing following colorectal cancer. The domains of assessment were informed by a conceptual framework of recovery following cancer diagnosis and treatment, 2 which hypothesised that a number of factors including social support would affect recovery. The main aim of this paper was to describe patient-reported social support pre-surgery up to 2 years' follow-up, in particular to investigate any change in support over time and which individuals were more likely to report poorer levels of support. A secondary aim was to investigate associations between social support over follow-up and HRQoL outcomes.

| Study design and participants
CREW is a multicentre, prospective cohort study of adults with non-metastatic colorectal cancer. Details of eligibility criteria, recruitment strategy, and sample size are provided elsewhere. 20 In brief, eligible individuals were approached before primary surgery from 29 UK cancer centres between November 2010 and March 2012. Written consent was obtained, and baseline questionnaires completed prior to surgery whenever possible. Follow-up questionnaires were completed at 3, 9, 15, and 24 months post-surgery (longer-term follow-up is ongoing). Socio-demographic information was also collected at consent. Participants reported whether they lived alone, and self-reported co-morbidities were recorded from 3 months onwards; clinical and treatment details were taken from medical notes. Ethical approval was granted by the UK NHS NRES Committee South Central-Oxford B (REC ref: 10/H0605/31).

| Measures
Full details of the questionnaire measures used in the CREW study are provided elsewhere. 20 The measures presented in this paper are described in brief: The Index of Multiple Deprivation 21 is the official measure of relative deprivation for small areas in England and uses postcodes to calculate an overall deprivation score based on 7 weighted domains of deprivation including income, employment, education, health, crime, barriers to housing and services and living environment.
Availability of social support was assessed in various ways: participants reported whether or not they lived alone, and number of close friends and family. The MOS Social Support Survey (MOS-SSS 22 ) yields an overall measure of social support (mean of all 19 items), subscales representing emotional/informational support, tangible support, affectionate support and positive social interaction (higher scores represent greater support, range 0-100), and an individual item relating to the extent to which participants feel they have "someone to do things with to help you get your mind off things" (Appendix S1).
The EQ-5D 23 measures generic health status/QoL, comprising 5 domains (mobility, self-care, usual activities, pain/discomfort, anxiety/ depression), each scored as none/some/severe problems, which can be summarised overall as presence/absence of problems on ≥1 of the domains.
The Personal Wellbeing Index-Adult (PWI-A 24 ) contains 8 items of satisfaction corresponding to standard of living, health, achieving in life, relationships, safety, community-connectedness, future security, and spirituality/religion. A higher overall score of wellbeing denotes better wellbeing (range 0-100; < 70 represents reduced wellbeing).

| Statistical methods
Published guidance for missing items in subscales were applied where available; otherwise, if ≥75% of items had been completed, mean scores were imputed from completed items. MOS-SSS scores were calculated according to published guidelines; binary variables were also created indicating whether a participant had responded "none" or "a little of the time" to all items within a subscale (versus "some/most/ all of the time"). The Index of Multiple Deprivation was categorised into quintiles. Descriptive analyses summarised measures of social support at each time-point. A sensitivity analysis was performed including only those questionnaires completed within specific timeframes around the expected date; as this made very little difference to the results, all data were included for the analyses presented.

| Characteristics of the sample
A total of 857 participants consented to follow-up excluding 15 who withdrew at baseline. Response rates were 89% at baseline, 84% at 3 months, 82% at 9 months, 80% at 15 months, and 74% at 24 months.
Baseline questionnaires were completed prior to primary surgery in 70% of participants, and within 3 months after surgery by a further 26%; reasons for post-surgical baseline questionnaires included admission for emergency surgery. Participants with and without a 24-month questionnaire were broadly similar in terms of demographic and clinical characteristics, with similar levels of social support at baseline. The mean age of participants at study entry was 68 years, with 60% male.
The sample comprised 65% colon and 35% rectal cancer, disease stage was 14% Duke's A, 53% Duke's B, and 32% Duke's C. By 2 years, 79 participants had experienced a recurrence, 65 had died, and 105 had withdrawn for reasons such as a deterioration in health, co-morbidities, significant life events, or that the participant felt the questions were no longer relevant as they had recovered from their cancer. Full details of participants are described elsewhere. 32 3.2 | Levels of social support from baseline to 2 years after surgery At baseline, 70.6% of participants were married or living with a partner.
Around 20% of participants reported that they lived alone (21.3% at 3 months and 23.5% at 24 months). The median number of close friends and family reported at baseline was 6 and 7, respectively, and remained stable over follow-up. At baseline, 7.8% of participants reported that they had "someone to help them get their mind off things" none/a little of the time, increasing to 17.3% at 2 years (Table 1). Proportions reporting none/a little of the time to all items within the MOS-SSS subscales at baseline and 2 years were 2.5% and 10.1% for emotional/informational support, 5.1% and 12.4% for tangible support, 4.5% and 12.9% for affectionate support, and 6.5%

| Associations between baseline participant characteristics and levels of social support over follow-up
Participants estimated to be in Group 3 (mid/declining support) were significantly older (52% aged >70) and with a greater proportion of women (47%) compared with Group 1 (very high/constant support; 39% aged >70 and 36% women, P = 0.046 for both) ( Table 2). For Group 4 (low/declining support), participants were significantly more likely to have higher neighbourhood deprivation (52% in fourth or fifth quintiles versus 37% for Group 1; P = 0.049), co-morbidities (84% versus 69%, P = 0.021), and to have had rectal rather than colon cancer

| Associations between levels of social support and HRQoL outcomes
Poorer HRQoL outcomes within 2 years were significantly associated with lower levels of social support throughout follow-up, adjusting for baseline participant characteristics (Table 3). Poorer generic health/HRQoL (problems on ≥1 EQ-5D domains) was significantly associated with lower overall support (P < 0.001), lack of affectionate support (P = 0.024), lack of positive social interaction (P < 0.001), and MOS overall social support index  not having "someone to take their mind off things" (P = 0.008).
Reduced personal wellbeing and high levels of depression were consistently significantly associated with lower levels of support domain scores were lower than reported in a breast cancer cohort of similar age to the CREW sample. 33 An overall lack of perceived social support was also reported in a cross-sectional study of mixed cancer types in Korea, where levels of support were lower for colorectal cancer than for other cancers. 8 In addition to considering the overall cohort, we also found distinct subgroups of participants, some of which had relatively stable levels of support over time (an estimated 71%), and others with poorer and declining support (21% mid/declining support and 8% low/declining support).

| Associations between participant characteristics and levels of social support
Our findings suggested that participants with lower and declining levels of social support were more likely to be older, female, with in seeking and accepting support from others, and interventions to support carers. Our findings highlight the importance of asking about perceived levels of support; it cannot be assumed that for individuals who have a wide circle of friends and family, there is naturally a perception of high levels of support available to them. Health care professionals could ask patients to identify who they feel they can turn to for help or support when then need it. In addition, in view of our findings, this conversation should not just take place at diagnosis but be re-visited throughout follow-up.

| CONCLUSIONS
A substantial proportion of people report poor and reduced social support following diagnosis and treatment for colorectal cancer. Social support is an important risk factor for recovery and HRQoL outcomes in colorectal cancer, particularly mental health. Vulnerable patients such as older people, those living in more deprived areas, and those with co-morbidities have less social support and, therefore, need more targeted professional support throughout their care, with social support implemented from the point of diagnosis as part of a holistic approach to health care provision.

CONFLICT OF INTEREST
None.

AUTHOR CONTRIBUTIONS
JH and SS wrote the paper, with contributions and comments from LC, JC, AD, DF, CG, AR, PWS, JW, and CF. JH performed the statistical analyses.

ETHICS STATEMENT
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.

ORIGINAL PUBLICATION STATEMENT
The manuscript contains original unpublished work and is not being submitted for publication elsewhere at the same time.