Introduction

Colorectal cancer (CRC) is the third most commonly diagnosed cancer among both men and women worldwide including the United States (USA) and most European countries [1]. While improvement in cancer treatments have resulted in improved survival rates, survivors are more likely to experience long-term distress as a result of the disease process and/or the cancer treatments [2], which may continue throughout the course of treatment into survivorship [3, 4].

Cancer-related distress is defined by the National Comprehensive Cancer Network (NCCN) as “a multifactorial unpleasant emotional experience of a psychological (cognitive, behavioral, and emotional), social and/or spiritual nature that may interfere with the ability to cope effectively with cancer” [5]. It extends along a continuum from adjustment disorder to clinical depression and/or anxiety [5]. Cancer-related distress is frequently underdiagnosed, underestimated, and undertreated in patients with cancer [6]. If untreated, distress can impact recovery, satisfaction, health-related quality of life (HrQOL), and even increase medical costs [7].

This distress can have a complex relationship with other factors such as symptoms, treatment-related adverse effects, and sociodemographic factors [8, 9]. As demonstrated with other non-CRC malignancies, distress has been associated with symptoms, negative illness perceptions, poor coping after cancer treatments, and social-/behavioral-related factors in breast [10, 11], and with younger age, poor social support, and previous symptoms in lung cancer survivors [12].

Distress in CRC patients and survivors remains relatively understudied, and little is known about its severity, frequency, and associated risk factors [4, 13, 14]. One study assessed distress in CRC patients with the Impact of Event Scale-Revised (IES-R) and found 31.5% of 311 CRC survivors reported elevated distress after cancer treatments [3]. In particular, studies have found surgical experiences to be distressing to CRC patients due to surgery-related symptoms and functional changes such as change of sexual functions, having and managing stoma, and altered body image [15,16,17,18]. CRC patients also reported high stress levels caused by surgery, as well as other perioperative therapies such as anesthesia and analgesia [15, 19]. These results suggest that cancer-related distress should be assessed and managed during follow-up care visits for CRC patients.

Majority of previous research has focused on examining the associations of cancer-related distress with survivorship outcomes, such as symptom burden and recovery from cancer treatment including surgery, in patients with non-CRC malignancies. Thus, there is a need for prospective, longitudinal research on risk factors for distress in CRC survivors. This would help to identify those CRC patients at risk for distress and for whom interventions could be targeted earlier in the cancer continuum, such as shortly after surgery, to prevent or mitigate long-term distress.

To our knowledge, this study is the first prospective, longitudinal study to examine cancer-related distress severity and changes over time from 6 to 12 months after surgery and to examine risk factors that are associated with long-term distress in CRC survivors. Our first aim was to examine the severity of cancer-related distress at 6 and 12 months after surgery and changes in distress over time. Based on prior distress research, we hypothesized that cancer-related distress severity would decline over time, with some variation in the decline of the different subscales [20, 21]. Our second aim was to identify predictors of cancer-related distress at 6 and 12 months after surgery. Based on risk factors found in studies of psychological outcomes in CRC [22,23,24] and other cancer survivors [8, 9, 12, 25], we hypothesized that higher CRC distress levels at 6 and 12 months would be predicted by younger age, low socioeconomic status, worse clinical conditions (e.g., advanced cancer stage, more severe symptoms), worse social support, and worse HrQOL before surgery.

Methods

Settings and participants

The current study included patients recruited at the Heidelberg site as part of the ColoCare Study between October 2010 and December 2014. The ColoCare Study (ClinicalTrials.gov Identifier: NCT02328677) is a multicenter, international prospective cohort, recruiting newly diagnosed colorectal cancer patients (stages I-IV), and has been described in detail in prior publications [26,27,28], with the goal to investigate predictors of cancer recurrence, survival, treatment toxicities, and HrQOL. In the ColoCare Study, eligible patients fulfilled the following criteria: newly diagnosed with stage I–IV CRC (ICD-10 C18, C19, and C20), scheduled for surgery, 18 years of age or older at the time of diagnosis, German-speaking and able to provide informed consent.

Overall, 393 patients were recruited in the study. In this study, we included patients who completed the cancer-related distress measure at both time points, 6 and 12 months after surgery. A total of 212 CRC patients were included in these analyses. A total of 181 CRC patients (who did not complete the cancer-related distress questionnaire either at 6 months or 12 months after surgery) were excluded, because 34 died during the first 12 months after surgery, and 147 were lost to follow-up or did not complete the survey at the follow-up time points (Supplementary Figure).

All study participants provided written informed consent. The ColoCare Heidelberg Study is approved by the ethics committee of the Medical Faculty at the University of Heidelberg.

Data collection

At a clinic appointment several days before surgery, after informed consent, participants completed questionnaires on social support and HrQOL and returned them to the research team in-person or by mail if they did not have time to complete the forms in clinic after consent. Details on sociodemographic, medical, and treatment factors were abstracted from patients’ charts and medical records from the University Hospital of Heidelberg. Data on a set of multiple exposures among study participants as well as lifestyle characteristics, social support, and HrQOL variables were collected with the self-reported questionnaires. At 6 and 12 months after surgery, participants were mailed the cancer-related distress measure and asked to return it by mail in an included stamped self-addressed envelope.

Social support was measured pre-surgery by the ENRICHD Social Support Instrument [29], a 6-item patient-report measure that assesses support using a 5-point Likert scale (1 “none of the time” to 5 “all of the time”). Individual items are then summed for a total score (ranged from 6 to 30), with higher scores indicating greater social support [30]. Psychometric properties were measured in a representative German population sample (N = 2552). Internal consistency proved to be very good (Cronbach’s α = 0.93) [31].

Health-related quality of life (HrQOL) outcomes were measured pre-surgery by the European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire (EORTC-QLQ)-C30. The EORTC QLQ-C30 includes one global QOL status rating, 5 functional scales (i.e., cognitive, emotional, physical, role, and social), and 9 symptom scales (i.e., pain, constipation, diarrhea, appetite loss, nausea, dyspnea, fatigue, insomnia, and financial stress) [32]. In this study, all scales were transformed into a range from 0 to 100. Higher scores for the global QOL status and functional scales indicated better HrQOL. In contrast, higher symptom scores indicated more severe symptoms [32]. Overall means of 5 functional scales and 9 symptom scales were also computed for this study. The HrQOL scales have demonstrated strong internal consistency reliability (Cronbach’s α > 0.80 for all scales) and validity in German cancer patients [33]. Internal consistency was also good (Cronbach’s α = 0.81 for functional QOL scales, 0.72 for symptom QOL).

Cancer-related distress was measured by the Cancer and Treatment Distress (CTXD) measure. The CTXD has 22 items with 6 subscales (uncertainty, family strain, health burden, finances, loss of identity, and medical demands). Item examples include uncertainty (6-item), “not knowing what the future will bring;” health burden (3-item), “not being able to do what I used to do;” family strain (3-item), “wondering about the emotional toll on my family or other caregivers;” identity (4-item), “changes in my appearance;” finances (3-item) “cost of treatment;” and medical demands, “dealing with the medical system” [34]. The German version was translated and back translated, with cognitive interviewing to confirm content consistency, using standard measure translation procedures [21, 35]. The CTXD has demonstrated sensitivity and specificity to cancer distress and also captures nearly all cases of depression and anxiety symptoms in cancer survivors and is correlated with post-traumatic stress symptoms, including those with German cancer survivors [21, 34, 35]. The 4-point Likert scale (0 “none” to 3 “severe”) was used to assess distress or worry in the past week for each item. Using standard scoring, we computed 6 and 12-month total mean scores of the 22 items and means for each of the 6 subscales. The strong reliability and validity of the CTXD have been confirmed in multiple studies of hematopoietic cell transplantation (HCT) recipients before and after transplantation [20, 34, 36]. The reliability of the German version of the CTXD for the total mean score in the current study was Cronbach’s alpha = 0.85.

Data analysis

Means and standard deviations were used to describe continuous data, and frequency counts were used to describe categorical data. First, we examined the severity of cancer-related distress at 6 and 12 months after surgery and changes over time, from 6 to 12 months using a linear regression with covariates including mean of functional and symptom QOL subscales at pre-surgery and social support levels at pre-surgery. Second, we examined hypothesized predictors of cancer-related distress using stepwise linear regressions. Sociodemographic and lifestyle factors, clinical data, and pre-surgery measures of social support, overall mean of 5 functional QOL subscales (cognitive, emotional, physical, role and social), and overall mean of 9 symptom QOL subscales (pain, constipation, diarrhea, appetite loss, nausea, dyspnea, fatigue, insomnia, and financial stress) were a priori selected for the regression models, as these variables were the most frequently reported factors in relation to symptom burden and HrQOL in cancer patients [4, 6, 8, 15, 19, 25, 37,38,39,40]. We investigated distributions of overall data, and a normal distribution with no skew was observed for social support, functional and symptom QOL, and distress data. Before the regression analyses, candidate variables for the regression model as risk factors for cancer-related distress were identified through (i) Pearson correlations between the CTXD scores and the continuous variables, and (ii) one-way analysis of variance (ANOVA) between the CTXD scores and the categorical variables. For significant ANOVAs (P < .05) with more than two categories, post hoc Tukey HSD multiple comparison test was followed to determine which categories of participants reported higher distress. A predictor was considered to be a candidate for the final regression model if it was statistically significantly associated with cancer-related distress (P < .05).

Subsequently, we performed multivariate linear regression analyses to evaluate risk factors that contributed independently to variance in cancer-related distress at 6 and 12 months. To identify factors associated with the CTXD, stepwise eliminations were performed using p values of ≥ 0.05 as the limiting threshold. Correlation analyses between the independent variables and variance inflation factors (VIF) were calculated to assess multi-collinearity. Multi-collinearity is present if the mean VIF is greater than 5 for the current study [41]. Standardized beta coefficients (β) were used to quantify the strength of the associations. All tests were two-sided, and statistical significance was defined as a p value < 0.05. All analyses were performed using SPSS 24 (IBM Corp., Armonk, NY, USA).

Results

Baseline characteristics

Baseline sociodemographic characteristics are shown in Table 1. Included patients had a mean age of 61.2 (SD 12.2), and 35% were female. The majority was married (71%), non-Hispanic White (96%), and diagnosed with stage II (33%) or III (37%) CRC. The total sum score of social support items at pre-surgery was 26.8 ± 3.8. The global QOL score at pre-surgery was 57.1 ± 23.9. The mean scores of 5 functional QOL and 9 symptom QOL scales were 73.8 ± 20.5 and 20.6 ± 15.6, respectively (Table 1). Of note, the highest functional scale was cognitive functioning (84.2 ± 21.6) while the lowest functional scale was emotional functioning (55.1 ± 27.8) at pre-surgery (Supplementary Table 1). Among 9 symptom scales, nausea was the least severe symptom (4.4 ± 12.8), while insomnia was the most severe symptom (35.9 ± 35.4).

Table 1 Characteristics in ColoCare patients and cancer-related distress (CTXD) means at 6 and 12 months by characteristics
Full size table

Cancer-related distress over time

Among the 6 subscales of CTXD at 6 months, the mean distress for health burden (0.5 ± 0.7) was the lowest at 6 months, while scores for medical demands and identity were the highest (1.0 ± 0.8 for both; Table 2). Interestingly, these mean scores almost exactly reversed at 12 months, with health burden the highest score (1.0 ± 0.8) and medical demands and identity the lowest (0.4 and 0.5 ± 0.6, respectively), along with finances (0.5 ± 0.7) among the subscales at 12 months. Meanwhile, the mean distress scores for uncertainty and family strain scales remained at the same level across time from 6 and 12 months. The mean distress scores for finances (P = .004), medical demands (P = .001), and identity scales (P = .001) each decreased (i.e., improved), but the mean distress score for health burden increased (i.e., worsened, P < .001) from 6 to 12 months. Of note, within subscales as seen in Table 2, some items improved while others worsened over time even when the subscale mean may have remained unchanged (i.e., uncertainty, family strain).

Table 2 Changes in cancer-related distress (CTXD) in patients with colorectal cancer at 6 and 12 months post-surgery
Full size table

Predictors of cancer-related distress

We examined the associations of possible predictors with cancer-related distress at 6 and 12 months. Age, smoking status, alcohol consumption, income, cancer stage, tumor site, cancer treatment types, pre-surgery social support, and pre-surgery mean scores of HrQOL scales were evaluated as potentially associated with cancer-related distress by performing Pearson’s correlation or ANOVA tests (Table 1).

In the multivariate linear regression analyses for both distress outcome at 6 and 12 months, identified candidate variables showing significant associations with distress at both 6 and 12 months in Table 1 were included. The VIF for the predictor variables ranged from 0.5 to 2.1, except global QOL scale (VIF = 6.7). As a result of collinearity with functional and symptom QOL scales, the global QOL scale was removed from the regression model. After the stepwise selection for cancer-related distress at 6 months, the final model contained five significant variables: age, income, cancer stage, pre-surgery social support level, and overall mean of functional QOL scales at pre-surgery (Table 3). Similar results were observed with respect to associations of predictor variables with cancer-related distress at 12 months after surgery. Younger age, advanced cancer stage, lower income, lower social support levels, and worse mean functional QOL at pre-surgery significantly predicted higher cancer-related distress after surgery (Table 3). However, smoking status, alcohol consumption, types of cancer treatments, as well as the mean of 9 symptom QOL scales at pre-surgery did not predict cancer-related distress at either 6 or 12 months after surgery in the final model (Table 3).

Table 3 Linear regression final model of predictors of cancer-related distress (CTXD) at 6 and 12 months post-surgery in patients with colorectal cancer
Full size table

Given the observed significant associations of cancer-related distress at both time points (6 and 12 months) with several sociodemographic and clinical characteristics, social support, and mean functional and symptom QOL scales as described above, we conducted further analyses to examine possible predictors of each of the 6 subscales of CTXD (i.e., uncertainty, family strain, health burden, finance, medical demands, and identity). We used the same analytic approaches for the primary analysis.

Several sociodemographic and clinical factors at pre-surgery played a role in cancer-related distress after surgery (Supplementary Tables 2 and 3). Overall, age, income, BMI, social support, and mean of functional QOL scales were significantly associated with and predicted some of the 6 cancer-related distress subscales at both time points (Supplementary Tables 2 and 3). However, some variables were unique predictors of each of the 6 distress subscales. For example, higher BMI was a predictor of higher distress in the health burden subscale at 6 months (Supplementary Table 2). Younger age, being female, and having an ostomy were significant predictors of greater distress for the identity subscale. Advanced cancer stage was associated with greater distress for uncertainty and medical demands. Lower income predicted greater distress for the uncertainty subscale at 6 months (Supplementary Table 2). For the 6 subscales of distress at 12 months, higher BMI and lower income predicted greater distress for the family strain subscale, while younger age and lower income predicted greater distress for the finance subscale (Supplementary Table 3).

Of the 6 subscales of CTXD, social support levels and functional QOL at pre-surgery significantly predicted most of the subscales of cancer-related distress, while symptom scales did not predict any of the 6 cancer-related distress subscales at both 6 and 12 months after surgery.

Discussion

In this prospective study, cancer-related distress related to finances, medical demands, and identity was higher at 6 months compared to distress levels at 12 months. Distress in health burden became more severe at 12 months compared to 6 months. In this study, younger age, lower income, advanced cancer stages, poor social support, and poor functional QOL at pre-surgery did predict cancer-related distress after surgery in CRC survivors. These findings provide strong evidence that cancer-related distress is persistent in CRC patients after surgery and suggest early assessment and management of vulnerable risk factors in subgroups at high-risk of cancer-related distress before the surgery in this population.

Our findings showed that the cancer-related distress on the finances, medical demands, and identity subscales improved, while health burden distress became worse from 6 to 12 months after surgery. This could be because patients were more likely have to deal with financial issues such as loss of work, financial burden due to surgery, and cost for post-surgical care, including frequent interactions with medical people to get some advice for post-surgical management [15, 19]. Although the unique reasons for financial distress with the German population in third-party payer healthcare systems is not clear, they may differ from the USA as insurance costs vary. For example, there are burdens for co-pay, out-of-pocket expenses in the USA, while patients in Germany have uniform healthcare coverage by social insurance. Therefore, further studies are warranted to understand the mechanisms of financial toxicity in cancer patients in different healthcare context.

Having an ostomy was associated with higher distress of identity subscale at 6 months, but not at 12 months. This could be due to the negative impact of the experience of stomas on impaired self-esteem and poor body image in patients with CRC [16, 42]. Our finding indicates that younger female, having an ostomy, may be the most vulnerable to the distress of identity subscale. Thus, the identity distress should be screened and managed for this group, specifically at 6 months after surgery compared to 12 months. On the other hand, patients are more likely to suffer from distress of health burden, and they may be exhausted in managing their symptoms and disease, as well as maintaining usual energy and daily activities for the long term at 12 months after surgery. Thus, the distress in health burden may increase over time during the long-term follow-up. This finding supports the current concerns of the long-term consequences of cancer treatment. Approximately one-third of adults with cancer face functional and emotional exhaustion and poor health or disability after cancer treatment [43]. Thus, healthcare professionals should ensure CRC patients at high-risk of cancer-related distress are identified and approached with timely management.

We observed that younger age was positively associated with greater cancer-related distress. This finding was similar to other studies in breast [11], lung [12], and heterogeneous [38] cancers. Numerous plausible explanations may exist for different symptom experiences by age. In previous study in patients with heterogeneous cancers [38], the lack of family support (unmarried, single), job insecurity, financial concerns, and lack of coping for emotional, social, and physical problems, were more prevalent in young adults with cancer, than senior adults with cancer. A previous CRC study demonstrated that older patients experienced less severe, less frequent, and less prolonged adverse symptoms in general, and symptoms in older patients were less likely to be associated with cancer treatments as compared to younger patients [44]. These factors may influence the positive association between younger age and increased distress level in our study. Furthermore, Rasumussen et al. (2015) [44] highlighted the fact that younger-aged survivors were more likely to undergo surgery and chemotherapy than older-aged survivors. Thus, younger-aged CRC survivors may report more acute and treatment-related symptoms and cancer-related distress, compared to older-aged cancer survivors.

Our findings also suggest that the social support and stable economic status at pre-surgery are important factors to prevent and alleviate cancer-related distress over time. Our results are similar to previous studies in other cancer types [39, 45, 46]. In a cross-sectional study in 235 patients with breast cancer, poor social support was associated with higher cancer-related distress measured by distress inventory cancer [39]. Barber and colleagues reported in 101 adult cancer survivors with heterogeneous cancer diagnoses that pre-surgery higher social support levels were associated with better self-efficacy and HrQOL [45]. Social support mediated the relationship between optimism and cancer distress in 69 women with breast cancer [46]. Current findings do suggest that strategies to enhance social support at diagnosis, as well as during the follow-up periods, may be useful in alleviating cancer-related distress after the surgery.

For the current study, pre-surgery functional QOL scale predicted levels of cancer-related distress in CRC survivors at both 6 and 12 month follow-up. Previous studies in CRC [40, 47] support our findings showing that a poor HrQOL score at pre-surgery was positively associated with higher long-term symptom burden and poor survival rates. Functional health is associated with a self-coping and adaptation to changes in cancer patients [48]. CRC patients with poor functional QOL at pre-surgery may have lack of self-coping and adaptation to the changes of their physical functions and emotional resilience after surgery. Thus, routine assessment of HrQOL at pre-surgery may be important for survivorship care, and treatment strategies should correspond to individual patients’ functional ability in designing long-term survivorship care in cancer.

Our findings may help clinicians provide an early screening before surgery and tailor long-term follow-up of cancer-related distress in post-surgical CRC patients. Understanding and recognizing individuals who are at increased risk for developing high levels of cancer-related distress is an important aspect in the screening and assessment of patients. Furthermore, our findings provide information for clinicians to better understand about specific subtypes of cancer-related distress such as health burden and predictors of each subtype of distress. This information is useful to provide different clinical interventions to each different distress subscale. Although patients can have similar severity of cancer-related distress, they may have different needs and services tailored to the specific subscales of CTXD. For example, younger patients, women, and patients having an ostomy are more vulnerable for high distress of identity. Thus, interventions for self-esteem or body image may benefit to this patient group.

There are several limitations to this study. A possible selection bias can exist, because we only included participants who completed both pre-surgery and follow-up questionnaires. In patients not included for the current study (N = 181), more patients were diagnosed with stage IV and had multiple chronic diseases, and they reported lower social support levels and poor HrQOL at pre-surgery, as compared to patients included for our analyses (Supplementary Table 4). It is possible that the cancer-related distress in patients who died or did not complete follow-up CTXD could be more severe. Our results in the German population may not generalize to CRC patients in other countries due to different healthcare systems. In addition, we did not assess cancer-related distress at pre-surgery; thus, we did not control the effect of pre-surgery cancer-related distress on our analyses.

Strengths of our study include a prospective design, utilizing a validated measure, CTXD, to assess cancer-related distress. There are several screening tools for cancer-related distress such as the Hospital Anxiety and Depression Scale (HADS), the Distress Thermometer (DT), and the Psychological Distress Inventory [49]. However, some measures focusing on psychological distress are limited to fully capturing cancer-related distress, not specific to cancer patients, and are intended for general populations [50]. The DT also has a high rate of false positives [49]. The CTXD has a high sensitivity and high specificity that capture the multiple dimensions of distress in cancer patients receiving hematopoietic cell transplant [34]. Our study also assessed each subscale of CTXD, which underscores the value of including assessment of specific distress after surgery.

Conclusions

Cancer-related distress was detected at 6 and 12 months after surgery in CRC patients. There are several risk factors of cancer-related distress after surgery, including younger age, low income, advanced cancer stage, poor social support, and poor functional QOL at pre-surgery. Each of the 6 subscales of cancer-related distress (uncertainty, family strain, health burden, finances, medical demands, and identity) presents a unique predictor, which can increase an individual’s vulnerability to cancer-related distress. Future research is suggested to utilize the CTXD tool to understand the patients’ perspectives in relation to their physical, psychological, social, or spiritual needs and to design and implement management of personalized cancer-related distress, such as self-symptom management, mental health counseling, and social worker services in CRC patients.