Social Support is Associated with Survival in Patients Diagnosed with Gastrointestinal Cancer

Abstract

Purpose

The aim of this study was to examine the link between psychological, behavioral, and social factors and survival in patients diagnosed with gastrointestinal cancer.

Methods

A cohort of gastrointestinal cancer patients were administered a battery of questionnaires that assessed trauma, depression, social support, sleep, diet, exercise, quality of life, tobacco and alcohol use, pain, and fatigue. Analyses included Pearson’s correlations, analyses of variance, Kaplan Meier survival, and Cox regression analyses.

Results

Of the 568 patients, the majority were male (57.9%) and Caucasian (91.9%), with a mean age of 61 (S.D. = 10.7). The level of perceived social support was comparable to patients with other medical conditions. Sociodemographic predictors of social support included the number of years of education (r = 0.109, p = 0.05), marital status (F(6,387) = 5.465, p ≤ 0.001), and whether the patients’ income met the family’s basic needs (F(1,377) = 25.531, p < 0.001). Univariate analyses revealed that older age (p < 0.001), male gender (p = 0.007), being black (p = 0.005), diagnosis of hepatocellular carcinoma (p = 0.046), higher body mass index (p = 0.022), larger tumor size (p = 0.032), initial treatment including chemotherapy rather than surgery (p < 0.001), and lower level of perceived social support (p = 0.037) were associated with poorer survival. Using multivariate Cox regression and adjusting for all factors found to be significant in univariate survival analyses, older age (p = 0.024) and lower perceived social support (HR = 0.441, 95% CI = 0.233, 0.833; p = 0.012) were the factors that remained significantly associated with poorer survival.

Conclusion

There are several biological and psychosocial factors that predict cancer mortality. Social support appears to be a robust factor affecting mortality in gastrointestinal cancer patients.

Background

Social support has been shown to foster psychological adjustment to chronic medical illnesses [1, 2] and to be a protective factor against depression, physical morbidity, and mortality [3, 4]. The Institute of Medicine has recently highlighted critical areas for improvement in cancer care by recommending 10 distinct goals, one of which is improving the psychosocial support of patients diagnosed with cancer [5].

Prior research has shown social support to be positively associated with cancer survival [6]. One study demonstrated that women with breast cancer who had close relatives, friends, or living children had a decreased risk of recurrence of breast cancer and of all-cause mortality [7]. Social support has also been associated with an improved overall survival in patients with acute myeloid leukemia [8]. A study in Japan showed similar findings, with low social support being associated with a higher risk of both mortality and incidence of colorectal cancer in men [9].

Although the positive association of social support and survival has been reported in patients diagnosed with leukemia, lymphoma, lung, colorectal, breast, and mixed cancer types, this association has been found to vary based on cancer site [6]. This is the first study to our knowledge to investigate the role of social support in survival in patients with gastrointestinal tumors [6, 10–12]. Previous studies of social support in cancer patients had smaller sample sizes and controlled for little or no other psychological or behavioral factors that have been associated with survival [6, 12–14]. Furthermore, many of these studies did not factor in demographic or disease-specific factors in their analyses [11, 15].

Our study is unique in that it has incorporated demographic, disease-specific, behavioral, and psychological factors that have been associated with survival in an attempt to control for potential confounders. It also includes a large sample size and the power needed to examine these factors with multivariable survival analyses. Gastrointestinal cancer patients often have a poor prognosis so an understanding of factors that improve patient quality and duration of life is critical [16]. The aims of the study included (1) comparing the perceived social support of patients with GI cancer with other chronic illnesses, (2) examining potential demographic and disease- and treatment-specific predictors of social support, (3) investigating the association between social support and stress and depressive symptoms, and (4) testing the link between social support and survival after adjusting for demographic, disease-specific, behavioral, and psychological factors associated with survival in those diagnosed with cancer.

Methods

Design and Participants

Participants

The patients were enrolled at the University of Pittsburgh Medical Center’s Division of Hepatobiliary and Pancreatic Surgery. Patients were enrolled in the study between January 2008 and June 2012 and November 2012 and October 2013. Inclusion criteria were (1) biopsy or radiographically-proven diagnosis of gastrointestinal cancer, (2) age 21 or older, and (3) no evidence of hallucinations, delusions, or thought disorder.

Procedure

Both studies were approved by the University of Pittsburgh’s Institutional Review Board prior to the commencement of any study activities. Patients were referred to the study team by their medical team. If the patient agreed to speak to a member of the study team, the risks and benefits of the study were explained to the individual, and written informed consent was obtained from the patient prior to completing the questionnaires.

Instruments

Demographic, Disease, and Treatment-Specific Factors

A questionnaire was used to collect sociodemographic data, which included patients’ age, gender, race, ethnicity, educational level, occupation, income, and health insurance status. Disease-specific and treatment-related information was gathered from the patients’ electronic medical record including diagnosis, body mass index, presence or absence of cirrhosis, maximum tumor size, number of lesions, vascularity of lesions, and vascular invasion. Survival was measured from the time of diagnosis of cancer until death. Death was determined by records in the electronic medical record or the Social Security Death Index.

Social Support

Social support was measured by using a 12-item version of the Interpersonal Support Evaluation List (ISEL) [17]. It measures the perceived availability of material aid, of someone to talk to about one’s problems, and of people one can do things with. The subscales have internal consistency and test–retest reliabilities that range from 0.70 to 0.80, with moderate intercorrelation.

Depressive Symptoms

The Center for Epidemiologic Studies-Depression (CES-D) is a valid and reliable 20-item self-report questionnaire designed to assess the weekly frequency of depressive symptoms [18, 19]. A score of 16 or greater represents depressive symptoms in the clinical range [17].

Fatigue

The Functional Assessment of Cancer Therapy-Fatigue (FACT-F) subscale is a valid and reliable 13-item questionnaire [20]. The FACT-F has been used extensively in a range of cancer populations. Scores can range between 0 and 52, with higher scores indicating greater fatigue [20].

Sleep

The Pittsburgh Sleep Quality Index (PSQI) is an 18-item self-rated questionnaire which assesses sleep quality and disturbances over a 1-month time interval [21]. Adequate levels of internal consistency, test–retest reliability, and validity have been reported for the PSQI in cancer patients [21].

Physical Activity

Physical activity data were obtained via the International Physical Activity Questionnaire (IPAQ), a reliable and valid measure that assesses facets of a previous week’s physical activity [22]. The overall score was based on metabolic equivalents (METS).

Tobacco Use

Current smoking was assessed with regard to the average number of cigarettes per day as documented in patient medical records. The duration the patients had smoked and the amount per week were also assessed. For the purposes of this study, we were interested in current tobacco use and the change in tobacco use over an 18-month period. Tobacco use was analyzed as a dichotomous variable with regard to patient’s response (yes or no) to current use of tobacco at each time point.

Quality of Life

The Functional Assessment of Cancer Therapy-Hepatobiliary (FACT-Hep) [23], a valid and reliable instrument, was used to assess changes in symptoms and side effects of treatment. The FACT-Hepatobiliary includes both the FACT-General [24] (a 27-item instrument that measures four dimensions of quality of life: physical, social, emotional, and functional well-being) and a module with 18 items specific to hepatobiliary disease symptoms as well as side effects of the treatment [23, 25].

Data Analysis

Data were entered, verified, and analyzed using SPSS version 23 (IBM Corp, Armonk, NY). Descriptive statistics were performed to obtain measures of central tendency, distribution, and proportions for each variable. Pearson’s correlation and analyses of variance (ANOVA) were performed to test associations between social support and demographic and psychological factors. Cox regression analyses were employed to test the associations between demographic, disease-specific, psychological, and behavioral factors with survival. A multivariable analysis was conducted with all factors that were significantly associated with survival in univariate analyses. Marital status was not included due to the significant association (multicollinearity) with social support.

Results

Sociodemographic and Disease-Specific Factors

A total of 568 participants with gastrointestinal cancer were included in the study. Sociodemographic characteristics can be found in Table 1. The majority of the patients were male (64.8%) and Caucasian (89.6%). The mean age was 62 years (SD = 11). The majority of patients were diagnosed with primary liver cancer or cancer of the bile ducts (44%) followed by other primary cancers with liver metastases (34.9%); neuroendocrine (11.8%); and cancers of the gallbladder, stomach, and pancreas (8.8%).

Table 1.

Demographic and disease-specifc characteristics

Demographic and disease-specific characteristic
Age (mean, SD)	61 (10.7)
Male (n, %)	329 (57.9)
Caucasian (n, %)	522 (91.9)
Education less than high school (n, %)	23 (4.1)
Diagnosis (n, %)	36 (6.3)
Pancreatic, gallbladder, stomach	204 (35.8)
Hepatocellular or cholangiocarcinoma	241 (42.4)
Other primary cancers with metastasis	87 (15.1)
Neuroendocrine
Maximum tumor size in centimeters (mean, SD)	2.3 (2.9)
Number of lesions (mean, SD)	1.9 (1.4)
Vascular invasion (n, %)	67 (11.8)
Treatment (n, %)	197 (34.7)
Chemotherapy	23 (4.1)
Radiation	260 (45.7)
Surgery

Norms and Sociodemographic Predictors of Social Support

The means and standard deviations for the Interpersonal Social Evaluation List (ISEL) are comparable to patients with other medical conditions (Table 2). Amongst sociodemographic factors, years of education (r = 0.109, p = 0.05), marital status (F(6, 387) = 5.465, p < 0.001), and whether one’s income met the family’s basic needs (F(1, 377) = 25.531, p < 0.001) were significantly associated with social support. Currently living with a partner (mean = 29.8, S.D. = 4.3) and being married (mean = 29.7, S.D. = 5.9) had the highest levels of reported perceived social support (Fig. 1). Being able to meet the basic needs of the family was associated with higher social support (mean = 29.4, S.D. = 5.9).

Table 2.

Means and standard deviations of the Interpersonal Support Evaluation List total score for this cohort and other medical conditions

Cohort	Mean	Std. Deviation
Gastrointestinal cancer patients (present study)	28.7	6.3
Respiratory illness	29.2	5.4
Osteoarthritis	27.6	5.8
Breast cancer	30.3	5.6

Fig. 1.

Interpersonal Support Evaluation List (ISEL) total scores based on marital status

Social Support and Psychological Correlates

An inverse relationship was found between depression and each of the subscales and total score of the ISEL: appraisal (r = − 0.361, p < 0.01), belonging (r = − 0.406, p < 0.01), tangible (r = − 0.342, p < 0.01), and total social support (r = − 0.414, p < 0.01). A similar pattern was observed between stress and social support with the subscales and total score of the ISEL: appraisal (r = − 0.328, p < 0.01), belonging (r = − 0.371, p < 0.01), tangible (r = − 0.287, p < 0.01), and total social support (r = − 0.374, p < 0.01).

Social Support and Survival

The multivariate model was tested using all significant demographic, disease-specific, behavioral, and psychological factors that have been found in the past literature to be significantly associated with survival. Age was found to be a significant predictor of survival. As a person’s age at diagnosis increased, their odds of death increased by about 4% per year (HR = 1.039, 95% CI = 1.006, 1.074). Those who were not currently receiving treatment (stable disease) had 80% lower odds of death than those who were receiving TACE (HR = 0.195, 95% CI = 0.046, 0.834) or ⁹⁰Y (HR = 0.23, 95% CI = 0.043, 1.224). Higher perceived social support had a 60% lower odds of death (HR = 0.441, 95% CI = 0.233, 0.833). See Table 3.

Table 3.

Multivariate Coxregression of predictors of survival

Factor	Beta	p-value	HR	95% CI
Age	0.038259	0.024	1.039	1.005	1.074
Gender	− 0.45256	0.232	0.636	0.303	1.335
Race (Caucasian)		0.522
Black	0.833344	0.254	2.301	0.549	9.643
Other	0.124869	0.885	1.133	0.208	6.185
Body Mass Index	− 0.00904	0.729	0.991	0.943	1.042
Tumor size	0.015873	0.763	1.016	0.914	1.13
Vascular invasion	0.192272	0.673	1.212	0.497	2.956
Diagnosis (HCC/CCC)		0.213
Pancreatic, gallbladder, stomach	− 0.23699	0.776	0.789	0.155	4.022
Other primary cancers with liver metastases	0.571544	0.172	1.771	0.779	4.025
Neuroendocrine	− 0.52763	0.352	0.590	0.194	1.792
Treatment (TACE)		0.228
None	− 1.62455	0.027	0.197	0.046	0.834
Systemic chemotherapy	− 0.69315	0.195	0.500	0.175	1.426
90Y	− 1.46968	0.085	0.230	0.043	1.224
Radiofrequency ablation	− 0.87948	0.146	0.415	0.126	1.36
Resection or Whipple	− 0.4943	0.352	0.610	0.216	1.725
Quality of life	0.0001	0.988	1.000	0.968	1.033
Depression	− 0.22941	0.628	0.795	0.315	2.009
Fatigue	0.007968	0.712	1.008	0.967	1.051
Smoking	− 0.4385	0.292	0.645	0.286	1.457
Social support	− 0.81871	0.012	0.441	0.233	0.833
Daytime sleep dysfunction	0.239017	0.362	1.270	0.76	2.123
Sleep duration (linear)	0.133656	0.392	1.143	0.842	1.551
Sleep duration (quadratic)	− 0.001	0.787	0.999	0.991	1.007
Physically activity (inactive)		0.441
Moderately active	− 0.39008	0.423	0.677	0.26	1.758
Very active	0.193097	0.629	1.213	0.554	2.655

Discussion

Demographic, biological, and disease-specific factors have been extensively studied with regard to their role in tumor growth, development of metastases, and survival in cancer patients. In our study, low social support and older age were significantly linked to poorer survival in patients with advanced gastrointestinal cancers after adjusting for demographic, disease-specific, behavioral, and psychological factors. These findings are consistent with prior research [6, 26–29]. However, there is evidence that the association between social support and survival can differ based on cancer site [6], and there are some notable exceptions to this trend. For example, European studies have shown that social support was not linked to survival in patients diagnosed with breast cancer [30, 31].

Patients in our study who had higher perceived social support had low levels of stress and depression. Of note, lower levels of depressive symptoms have been associated with longer survival in patients with breast cancer [32]. Depression and stress are often associated, and the intensity of stress and severity of depression have a dose dependent relationship [33, 34]. Social support has been documented to alleviate the perceived stress through the use of alternative coping strategies that may be learned through others [35]. Depression may also erode social support as the negative self-image, complaints, dependency, and social inadequacy displayed by people who are depressed may estrange family and friends who provide support [36].

Level of education, having an income that meets the family needs, and marital status were all predictors of social support. Social support may have a more pronounced effect for patients who had a high school education or less, as one study has shown that patients who had limited literacy were less likely to ask a pharmacist questions, a form of informational support [37, 38]. As such, patients with a lower level of education are more likely to benefit from increased social-support particularly logistical types of support [39]. Similar to our findings with regard to perceived social support and income, individuals with lower SES had less available social support [40]. Married individuals were more likely to have social support available to them which is consistent with prior research.

It is important to mark the distinction between perceived and actual support. Uchino et. al showed that perceived support versus actual support had a more protective role with regard to both chronic disease development and acute disease susceptibility [41]. Interestingly, the availability of unwanted support has been associated with poor psychosocial adjustment in a setting of a mismatch between need and provision of support [42]. To illustrate, a study demonstrated that men with early-stage prostate cancer who had social support had worsening depression. However, those who had more advanced disease benefited from social support, which decreased the patients’ depressive symptoms [43].

Cutrona et. al found that specific types of social supports are most beneficial following different kinds of stressful life events [44]. Matching support with the individual needs for particular types is an effective strategy to properly allocate our resources. Health care providers may want to educate patients about the availability of social support and empowering them to acquire the necessary support that is needed at each of the stages in their disease [45].

The timing of social support with respect to the course of treatment is also pivotal in understanding its role on overall survival. Rottenberg et. al showed that diverse social networks present before a diagnosis correlated with survival after the onset of the disease [46]. After treatments, social support has demonstrated its importance in encouraging healthy behavior. In fact, a study demonstrated that cancer patients who perceived a better support system were less likely to continue smoking after diagnosis [47]. Meanwhile, social support during therapy has documented benefits and is potentially an area of intervention in our patients. It becomes a modifiable factor in that context and can even improve survival. For instance, Spiegel et al. has shown that women with metastatic breast cancer who had received weekly group-therapy for a year lived 18 months longer than the patients who did not receive the intervention [48, 49]. Likewise, Penedo et al. in a number of studies showed that social support in the form of a 10-week group therapy with 2-h weekly meetings had a significant impact on the quality of life of patients with prostate cancer [50–52].

A number of theories aim to explain the mechanisms by which social support affects survival. Pinquart et. al suggested that social support may influence cancer morbidity and mortality through a number of mediators such as biology, health-behaviors, access to health-care systems, adherence to cancer treatments, and/or improved psychological well-being [6]. In terms of biological factors, social support is hypothesized to mitigate the proliferative effect of stress-related endocrine changes [53]. For instance, cortisol, a known stress-hormone, has been shown to be suppressed in patients with social support [54]. Some studies identified a correlation between social support and lower blood pressure even in normotensive patients as well as a heightened activity of natural killer cells [55]. These cell types have an active surveillance role in cancer and were also noted to be higher in ovarian cancer patients who had social support [56]. Also, higher levels of social support were associated with an increased level of interferon gamma (IFN-γ) and decreased IL-4 levels, which imply a more robust immune response [57]

Study Limitations

Some notable limitations include that our study focuses on gastrointestinal cancers, which albeit aggressive are a specific disease subgroup. While the findings of this study could be applied to analogous cohorts, extrapolating our results to more cancer subtypes should be exercised with caution. Also, considering perceived support changes throughout the course of the treatment, our conclusions would mostly be limited to cross-sectional data.

This study also has several strengths. First, it included a large sample size and the power to be able to include demographic, disease-specific, psychological, and behavioral factors in the multivariable survival analyses. In addition to that, this study not only studied the link between social support and survival but also examined predictors of social support in itself thereby pointing to potential areas of intervention. Lastly, this study is the only study that addressed social support specifically in gastrointestinal cancer, which phenotypically affects patients differently than previously studied cancers.

Conclusions

There are several biological and psychosocial factors that predict cancer mortality. Social support appears to be a robust factor affecting mortality. Hence, the need of social support is critical, and a better understanding regarding the ideal type and timing of support is needed. Future research should focus on ways of identifying the best type of social support for a patient, the most beneficial timing for the support, and the most effective interventions to optimize support for gastrointestinal cancer patients.

Availability of Data and Material

The data that support the findings of this study are available from the corresponding author upon reasonable request.