Surgery and Financial Toxicity are Risk Factors for Distress in Cancer Survivors: A Cross-sectional Study

Background: Patients with cancer often face some level of distress, regardless of disease stage. Distress in cancer survivors has a negative impact on their quality of life. The goal of this study was to identify risk factors for distress, under stand how treatment associated with distress and reveal the relationship between the psychological and nancial distress. Methods: This was a multi-center cross-sectional study of patients with cancer requiring surgery or chemotherapy. Patients completed questionnaires regarding their demographics, disease characteristics, psychological distress, and nancial toxicity. A multivariable logistic regression model was used to examine factors associated with distress in surgical versus chemotherapy treatment groups. Results: A total of 409 patients participated in the study. Patients treated with surgery (n = 172) were more likely to be female, unemployed, early stage compared with patients undergoing chemotherapy (n = 237). Multivariable analysis revealed that surgical patients tended to have a higher risk of distress compared with patients receiving chemotherapy (OR, 95% CI: 3.086, 1.854–5.137) due to higher rates of nervousness, pain, and diculty with bathing/dressing, and patients with high nancial toxicity had a higher risk of distress compared with those with low nancial toxicity (OR, 95% CI: 2.000, 1.278–3.130). The relationship between nancial toxicity and psychological distress was stronger in the chemotherapy group, with the correction coecient -0.294 and slope -1.196. Conclusion: Patients who underwent surgery and reported higher nancial toxicity were more likely to experience distress. Multidimensional distress screening and psychosocial interventions should be provided pre- and post-operatively for patients. on treatment methods: chemotherapy alone (n 237, 57.9%), and surgery (n 172, 40.1%), which included surgery alone or combined with chemotherapy. two groups similar in age (P marital status educational background (P and household income the signicant variations included sex 0.001), employment status cancer type and clinical stage


Introduction
Cancer diagnosis and treatment often lead patients to face some level of distress, regardless of disease stage. [1,2] Distress is conceptualized as a multifactorial, unpleasant experience of a psychological, social, spiritual, and/or physical nature that may interfere with the ability to cope effectively with the physical symptoms and treatment of cancer, [3] and has been considered the sixth vital sign, after pain, in cancer care. [4] Distress in patients with cancer may reduce adherence to treatment, decrease quality of life, and increase cancer-speci c mortality. [5][6][7] Early screening for distress leads to timely multicomponent intervention, which in turn improves quality of life.
Many studies of distress have focused on the emotional problems (e.g., anxiety, depression) and physical problems (e.g., fatigue) of distress, with little attention to treatment-related and nancial factors. [8][9][10] Financial toxicity has been de ned as objective nancial burden and subjective nancial distress experienced by cancer patients as a result of their treatment. [11] Assessment of nancial toxicity may help to build a framework for nancial counseling interventions on par with symptom management (e.g., for fatigue or pain). [12] The present study was conducted to identify demographic, clinical, and socioeconomic predictors of distress, understand how treatment associated with distress and reveal the relationship between the nancial toxicity and distress.

Study Design and Procedure
A cross-sectional observational study was conducted at three public cancer treatment centers: Cancer Prior to the survey, the trained researchers, the oncology nurses, explained to the patients the study purpose and that participation was voluntary. All participating patients provided informed consent. The patients completed the entire questionnaire except for the clinical information section, which was prepopulated from the electronic health record.

Sociodemographic and Clinical Characteristics
Demographic information and clinical characteristics were solicited in the rst part of the questionnaire, including age, sex, marital status, educational background, employment status, medical insurance status, cancer type, the time of diagnosis, clinical stage of cancer, and type of treatment.

Psychological Distress Assessments
Several instruments are available to identify the distress of patients with cancer [13]. The Distress Thermometer and Problem List (DT&PL) is widely used as a self-reporting tool for the screening of distress in patients with cancer. [14,15] The DT is a single-item, self-reporting instrument measuring the amount of distress experienced by patients within the last week, with a score ranging from 0 (no distress) to 10 (extreme distress). The PL groups various problems patients with cancer encounter after diagnosis into ve problem categories: practical, family, emotional, physical, and spiritual. Problems are selected by checking a corresponding ''yes'' or ''no'' on the survey. The Chinese version has been validated in various types of cancer patients. [16] A score of 4 or higher on the DT indicates signi cant distress. The prevalence of distress and speci c problems were analyzed in the current study.

Financial Toxicity Assessments
All patients also completed the COmprehensive Score for nancial Toxicity (COST) survey to assess for nancial toxicity. The COST measure was previously developed and validated by de Souza et al. to assess nancial toxicity in patients with cancer. [17,18] Brie y, the COST is an 11-item measure of nancial toxicity examining one nancial item, two resource items, and eight affect items. The patients were asked to respond on a ve-point Likert scale, from 0 (not at all) to 4 (very much). The total score ranges from 0 to 44 points. Lower COST values indicate sever nancial toxicity. Our team translated and adapted the Chinese version with high reliability (α = 0.89) among patients with cancer. [19] The COST scores were strati ed by high and low nancial toxicity, and we de ned high nancial toxicity as lower or equal to the median. [20] Statistical Analysis All sociodemographic and clinical characteristics of patients were summarized using descriptive statistics. Categorical variables were presented as frequencies and percentages. The chi-square test was used for comparing the group differences in categorical variables. Associations between factors were derived by multiple logistic regression models after adjusting for factors including occupation and insurance. Factors examined were age, sex, marital status, educational background, annual household income, type of cancer, type of treatment, and degree of nancial toxicity. Pearson correlation coe cient was used to test associations among the COST and DT scores. Statistical analyses were performed with SAS 9.2 software. All tests were two-sided and P values of 0.05 or less were considered statistically signi cant.

Characteristics of the study population
A total of 409 patients participated in the study. The demographic and clinical characteristics of patients are presented in Table 1. The median age was 59 years and more patients identi ed as female (55.3%) than male. Approximately 90.0% of patients identi ed as married, 24.4% as employed, and 17.6% as having high school education or more. In addition, 40.6% of patients were diagnosed with stage III cancer, and the most frequently reported type of cancer was lung (31.1%) followed by breast (27.4%). The patients were grouped into two categories for analysis based on treatment methods: chemotherapy alone (n = 237, 57.9%), and surgery (n = 172, 40.1%), which included surgery alone or combined with chemotherapy. The two groups were similar in age (P = 0.092), marital status (P = 0.936), educational background (P = 0.218), and household income (P = 0.671); the signi cant variations included sex (P < 0.001), employment status (P = 0.005), cancer type (P < 0.001), and clinical stage (P < 0.001). nancial toxicity (n = 217, 53.1%) had COST scores below or equal to the median of 18. Surgical intervention was associated with a higher degree of patient distress than chemotherapy alone (P < 0.001). There was no signi cant difference regarding nancial toxicity. The prevalence of the distress and nancial toxicity is presented in Fig. 1.

Factors Associated With Distress
In the unadjusted analyses, higher income and lower nancial toxicity were associated with lower prevalence of distress. A lung cancer diagnosis, early stage of disease, and surgical intervention were associated with higher prevalence of distress. In the nal multivariable model, treatment type and nancial toxicity were found to be signi cantly associated with distress when controlling for age, marital status, employment status, and educational background. These ndings are demonstrated in Table 2.
Patients in the surgery group reported more distress than the chemotherapy group (OR, 95% CI: 3.086, 1.854-5.137). Similarly, patients with a higher degree of nancial toxicity, re ected by lower COST scores, had a greater risk of distress than those with a lower degree of nancial toxicity (OR, 95% CI: 2.000, 1.278-3.130). Further analysis was to understand how surgery increased distress. The results showed that patients in the surgery group reported higher rates of nervousness (P = 0.001), pain (P < 0.001, and di cult with bathing/dressing (P = 0.008) compared with the chemotherapy group. Figure 2 outlines the top 10 factors causing distress in the treatment groups.

Financial Toxicity and Distress in Treatment Groups
The association between overall distress and nancial toxicity, respectively measured by the DT and COST, based on treatment is presented in Fig. 3. The correction coe cient between nancial toxicity and distress was − 0.188, and the COST score decreased by 0.572 points for every 1-point increase in the DT score (P = 0.014). This suggests that a higher degree of nancial toxicity is associated with greater distress in the surgery group. For patients in the chemotherapy group, the correction coe cient was − 0.294, and every 1-point increase in the DT score decreased the COST score by 1.196 points.

Discussion
In the literature, the prevalence of distress among patients with cancer varies by country, cancer type, sex, age, and other sample characteristics. [21][22][23][24][25] Prior studies identi ed worry on the DT&PL as the most distressing, possibly as a surrogate for the intensity of distress. [26] McFarland et al. reported that 40% of patients with breast cancer had fatigue, the most common physical problem associated with distress. [27] Unlike previous studies, [28,29] in this study nancial di culty was the highest-ranking single item associated with distress. The prevalence of psychological distress and nancial toxicity was 56.5% and 53.1%, respectively. Most patients experienced at least one practical, physical, or emotional problem, primarily nancial di culty (70.1%), worry (62.8%), and fatigue (49.8%). In this study, worry was a leading item in the emotional problem domain, but inferior to nancial di culty among all items attributed to distress. The prevalence of nancial concerns among patients with cancer may be due to the need to make nancially-based decisions throughout cancer treatment. [30] Cancer-related nancial problems have been associated with increased risk for depressed mood, a higher frequency of worry, and a signi cant and frequent source of distress among patients with cancer. [31,32] Several studies have examined distress among surgical inpatients. Basak et al. found that approximately half of surgery inpatients had depression and approximately one-quarter had anxiety. [33] Pastore et al. found that patients undergoing surgery for urological cancer had clinical levels of anxiety (9.8%) and depression (3.6%). [34] Furthermore, a signi cant correlation was observed between distress and esophagectomy among patients with esophageal cancer. [35] In the current study, 70.9% surgical treatment group reported a signi cantly psychological distress. One potential explanation was that patients underwent surgery were worried about preoperative preparation and postoperative pain. [36] The results of studies examined the effects of different treatment options on distress among cancer patients are inconsistent. Female patients who underwent chemotherapy were more likely to report fatigue and nausea, whereas surgical patients did not report these physical problems. [27] Patients with breast cancer who underwent mastectomy with reconstruction reported higher levels of distress compared with patients undergoing lumpectomy and mastectomy only. [10] Our data showed that statistically signi cant differences were noted between the surgery and chemotherapy group for nervousness, pain, and problems with bathing/dressing. Surgical treatment was a signi cant predictor of psychological distress with 3.09 times risk for psychological distress versus chemotherapy treatment. Further studies are needed regarding preoperative intervention and postoperative management for distress among cancer patients undergoing highly invasive procedures.
The literatures support the relationship between poor socioeconomic status (e.g., a low household income, nancial problems) and psychological distress. [37,38] Approximately 22% of patients with cancer were worried about paying medical bills. [39] Lung and colorectal cancer patients with limited nancial reserves reported increased pain. [40] One possible explanation was that cancer patients with poor nancial status encountered more barriers to timely diagnosis, optimal treatment, and survivorship care [41,42]. Carrera et al. suggested that nancial toxicity could be coupled with the use of DT&PL in screening for distress. [11] To our knowledge, this is the rst study identifying the relationship between COST and DT in hospitalized patients with cancer. In the current study, COST scores were negatively related to DT scores in two groups, suggesting a higher degree of nancial toxicity correlates with a greater severity of distress. Financial toxicity was signi cantly associated with distress, even after controlling for age, sex, and cancer type. This study has some limitations. First, the cross-sectional observational design could not evaluate dynamic changes of DT and COST with treatment, and did not provide interventions to patients with signi cant distress. Second, patients undergoing radiotherapy treatment were not included in this study, mainly because very few patients receive radiotherapy as rst-line or primary treatment. Finally, participations were solicited from three tertiary-level cancer centers from different cities, but all in Northeast China. Therefore, the application of the study ndings is limited to patients with cancer in China.
Patients with cancer experience distress caused by physical, emotional and nancial problems. Frequently, these problems overlap and exacerbate one another. This study demonstrates that a signi cant proportion of cancer survivors above the threshold for psychological and nancial distress, provides preliminary evidence for an association between treatment and nancial toxicity and distress, and evaluates predictors of distress in adults with cancer. The ndings con rm that surgical treatment and severe nancial toxicity are signi cant predictors of distress.

Declarations
Authors' contributions Huihui Yu participated in study conducting, data analysis and interpretation, drafting the manuscript and approval of nal manuscript. Tingting Zuo participated in data interpretation and approval of nal manuscript. Xue Bi, Hui Li,Haiyang Xing, Li Cao, Lijuan Cai, Zhen Zhang participated in collection and assembly of data and approval of nal manuscript. Yunyong Liu participated in study designing, critically revising the manuscript, approval of nal manuscript and is the guarantor of the study.