Fear of Cancer Recurrence in Stomach Cancer Survivors: Exploring Socio-Demographic, Physical, Psychological and Quality of Life-Related Factors

Background: This study aimed to evaluate factors related to the fear of cancer recurrence (FCR) in stomach cancer survivors. Methods: A total of 363 stomach cancer survivors were divided into three groups according to the Korean version of the FCR Index (FCRI). We compared the socio-demographic and cancer-related factors, communication within the medical team during treatment, care coordination, social support, physical symptoms, psychological distress, and health-related quality of life (HRQoL), and health promotion and disease prevention activity, which was dened as second primary cancer screening, osteoporosis evaluation, supplementary drug intake, dietary pattern, regular exercise, and smoking cessation, according to the levels of FCRI using the linear regression analysis after adjusting for age, sex, stage, time since diagnosis, family cancer diagnosis, and comorbidities. Results: An average of 7.3 years have passed since diagnosis. The highest FCRI group was younger or had a high percentage of family cancer diagnosis. Satisfaction for communication and care coordination, social support, HRQoL, and physical symptoms, such as fatigue, pain, and insomnia, were negatively associated with the FCR of stomach cancer survivors. Anxiety and depression were positively associated with the FCR. However, FCR was not associated with sex, time since diagnosis, cancer stage, treatment modality, socioeconomic status, or health promotion and disease prevention activity. Conclusions: FCR in stomach cancer survivors was associated with physical, psychological, and HRQoL factors rather than socio-demographic or cancer–related factors. Understanding the association between HRQoL, physical and psychosocial factors, and FCR may advance survivorship care for stomach cancer survivors.

factors seem essential for cancer survivors' care [6]. Higher FCR is generally associated with younger age, higher education status, uncontrolled physical symptoms, and psychological distress [5,7]. However, there were mixed ndings for the associations with sex and marital status [5]. Furthermore, evidence of these associations was weak or inconclusive, depending on cancer type [6,8].
Stomach cancer is known for being signi cantly in uenced by behavioral and lifestyle risk factors [9].
Smoking, alcohol consumption, chemical exposure, obesity, and dietary factors (such as excessive salt intake) may be associated with the FCR of stomach cancer survivors and was considered a preventable condition [8][9][10][11][12]. Systematic reviews of 43 studies [8] and 130 studies [10] on the factors of FCR revealed that there are no stomach cancer-speci c FCR publications.
Therefore, we conducted this study to evaluate the association of FCR of stomach cancer survivors with a range of factors including cancer-related information, socio-demographic characteristics, physical symptoms, psychological distress, social support and quality of life in Korean stomach cancer survivors.
We further evaluated the effects of FCR on their health promotion and disease activity of Korean stomach cancer survivors.

Study participants
A total of 367 stomach cancer survivors with gastrectomy who visited a cancer survivorship clinic of a university-a liated hospital in South Korea due to health problems or post-treatment surveillance after ve years from initial cancer diagnosis was recruited from September 2014 to March 2017. The clinic has been involved in the care of a large number of Korean cancer patients. After excluding patients with metastasis at diagnosis (n = 4), 363 stomach cancer survivors were analyzed in this study.

Study variables
We obtained data on Fear of Cancer Recurrence Index (FCRI) [6], health-related quality of life (HRQoL) [13], clinical and socio-demographic characteristics, satisfaction with communication within medical team and care coordination [14], functional social support [15], and health promotion and disease prevention activity using a self-administered questionnaire. If necessary, a trained research assistant completed incompletely answered questions through an additional face-to-face interview.
FCRI is a multidimensional questionnaire composed of 42 items with seven subscale components of FCR [16]. Potential stimuli activating FCR (triggers), presence and severity of intrusive thoughts associated with FCR (severity), emotional disturbance associated with FCR (psychological distress), impact of FCR on important areas of functioning (functional impairments), self-criticism toward FCR intensity (insight), and reassurance seeking such as thorough self-examination or repeated medical consultations (reassurance) and other strategies to cope with FCR (coping strategies) were evaluated. A total score of FCR ranges between 0 and 168. Cronbach's alpha coe cient for the FCRI-Korean version was 0.85 for the total scale and 0.77-0.87 for the subscales [6].
The EuroQoL-visual analog scale (EQ-VAS) and the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire version 3.0 (EORTC QLQ-C30) were used to assess HRQoL. The EQ-VAS is a standard vertical 20 cm visual analog scale for rating current HRQoL, ranging from zero (worst imaginable) to 100 (best imaginable). EORTC QLQ-C30 is a 30-item questionnaire developed to assess HRQoL of cancer patients incorporating ve functional scales (physical, role, emotional, cognitive, and social), a nine-symptom scale (fatigue, nausea and vomiting, pain, dyspnea, insomnia, appetite loss, constipation, diarrhea, and nancial di culties), and a global quality of life scale. Participants responded using a four-point or a seven-point Likert scale. The score of each scale was calculated and ranged from zero to 100. Except for the cognitive functioning subscale, the Cronbach's alpha coe cients of the Korean version of the EORTC QLQ-C30 subscale are greater than 0.70 [13].
Fatigue severity scale (FSS) was developed to assess fatigue in patients with chronic diseases and included 10 questions that are rated from one (strongly disagree) to seven (strongly agree) [17]. A higher score denotes more severe fatigue. The Cronbach's α for total FSS was 0.935, ranging from 0.925 to 0.932 for subscales [18]. Sleep problems were surveyed by a self-administered questionnaire. We assessed the level of sleep problems in terms of mean frequency: 1) no problems, 2) problems 1-2 nights per week, 3) problems 3-4 nights per week, and 4) problems every night. Anxiety and depressive mood were evaluated by the Hospital Anxiety and Depression Scale (HADS, range 0-21). The Cronbach's α of the Korean version of HADS was 0.89 for the anxiety subscale and 0.86 for the depression subscale [19].
Socio-demographic characteristics were obtained using a self-administered questionnaire: monthly income level (≥ 4,000,000 won per month, 2,000,000-3,999,999, and < 2,000,000), achieved education level (≤ 9 years, 10-12, and ≥ 13), marital status (married/with a partner and unmarried/without a partner), family cancer diagnosis (spouse, 1st degree relatives, or none), and religion (yes or no). We reviewed their medical records for cancer-related information, such as time since cancer diagnosis, cancer stage, additional treatment modality (chemotherapy, and/or radiotherapy), and Carlson's comorbidity index, which was calculated using 18 conditions, excluding age [20].
Satisfaction with the level of communication on harmony, interaction, and role responsibility within the medical team was assessed with ve points each for a possible total of 15 points. A higher score equates to a higher level of dissatisfaction with communication within the medical team. Care coordination assessment was based on this question: "Did you receive all care services that were necessary for dealing with your health concern during the cancer treatment period?" [14]. Scoring ranged from one to ve, and a higher score equated to a higher level of dissatisfaction with care coordination. We measured the person's perceived need for a social support network by the Duke-UNC functional social support questionnaire (FSSQ) [15]. The questionnaire was composed of eight-items from two subscales, con dant and effective support, with ve answer choices ranging from ve ('As much as I would like') to one ('Much less than I would like') [21]. We presented the average score of the eight responses. The Cronbach's α of the Korean version of the FSSQ was 0.89 [22].
To evaluate health promotion and disease activity, data on receipt of the second primary cancer screening test (colonoscopy, mammography, and Pap smear), patterns for smoking, drinking, and exercise, bone mineral density evaluation within two years, supplementary drug intake, and dietary pattern changes after cancer diagnosis were collected by self-administered questionnaire. Quitting of smoking and drinking was evaluated for occurrence after a cancer diagnosis or treatment. Regular exercise was de ned as at least one time per week for 30 minutes. Supplementary drug intake was de ned as consistently (more than 2 weeks) taking supplements such as vitamin B complex, vitamin C, vitamin D, multivitamins, calcium, and iron after a cancer diagnosis. Dietary pattern changes after cancer diagnosis were evaluated by asking whether the frequency of intake increased or decreased for organic food, fruit or vegetables, pork or beef, processed meat, and salty food.

Statistical analysis
Descriptive statistics of study participants were expressed as mean value ± standard deviation for continuous variables and number (proportion) for categorical variables. Because a cut-off point score for identifying people with abnormally high FCR has not been determined, the level of FCR was divided into tertile (lowest/middle/highest groups) of the FCRI summed score of triggers, severity, distress, functional impairment, and insight subscales. The differences in cancer-related information and socio-demographic variables according to tertile of FCRI score were compared by one-way ANOVA analysis, chi-square test, or Fisher's exact test. The satisfaction of communication with a medical team and care coordination and the mean levels of FSSQ were compared using a linear regression analysis with an adjustment for selected covariates [6]. We presented the mean scores of HRQoL measured by EQ-VAS and EORTC QLQ-C30 and the level of fatigue severity, sleep problems, anxiety, and depression according to the tertile of FCR. We evaluated the relationship between health promotion and disease prevention and the level of FCRI using a multiple logistic regression analysis after adjusting for age, sex, cancer stage, Carlson's Comorbidity Index, time since diagnosis, family cancer diagnosis, Duke FSSQ, and satisfaction with communication within a medical team and of care coordination. All statistical analyses were performed using IBM SPSS Statistics for Windows, version 24.0 (IBM Corp., Armonk, NY, USA). Table 1 shows the general characteristics of study participants (n = 363). The mean age of subjects was 56.9 years old, with an average of 7.3 years since diagnosis. Two-thirds of subjects were on stage I and 61.7% of subjects performed surgical treatment only. Table 2 shows the related factors with FCRI in stomach cancer survivors according to the tertile level of FCRI. Younger survivors were more likely to be classi ed into the highest group of FCRI. A spouse or 1 st degree relatives was diagnosed with cancer in the highest FCR group. A higher FCRI level was associated with unsatisfaction of communication and care coordination and less social support. There were no signi cant differences in the time since cancer diagnosis, sex, income, education, marital status, religion, cancer stage, treatment modality, and comorbidity according to the level of FCRI. Figure 1 shows the distribution of HRQoL, physical, and psychosocial factors according to the level of FCRI. Compared to subjects in the lowest tertile, those of the highest tertile of FCRI showed worse HRQoL. The trend of EQ-VAS and the global health status and functional scales of EORTC QLQ-C30 were signi cantly lower in the highest tertile group (Figure 1-(A)). The symptom scales of EORTC QLQ-C30 (fatigue, nausea and vomiting, pain, dyspnea, and insomnia) and nancial di culty were signi cantly higher in the highest tertile of FCRI ( Figure 1-(B)). A similar direction lacking statistical signi cance was observed for appetite loss, constipation, and diarrhea ( Figure 1-(B)). Fatigue severity scale, sleep problems, anxiety, and depression were also higher in the highest FCRI group (Figure 1-(C)).

Discussion
In this study, the presence of uncontrolled physical symptoms, psychological distress, and poor HRQoL was associated with high FCR in stomach cancer survivors even after adjusting for covariates. Functional social support and satisfaction for communication within the medical team and care coordination were negatively associated with the FCR. However, FCR was found to have no relationship with cancer-related or socio-demographic factors, except age and health promotion or disease prevention activity of stomach cancer survivors.
Like other cancer patients, stomach cancer survivors who are younger, with uncontrolled physical symptoms, psychological distress, and lower HRQoL were associated with higher FCR [10]. In a previous study with 342 Chinese breast cancer survivors who were diagnosed less than 2 years (67.3% of the total number of subjects), between 2 and 3 years (14.0%), and more than 3 years (18.7%), no religious beliefs, lower family income, and treatment modality (breast-conserving surgery) were associated with the higher level of FCR, unlike our result [23]. The differences in our FCR effects may be due to the diagnosis period; this study has a long time (average of 7.3 years) since its diagnosis. In general, breast cancer patients are often easily affected by psychological and social factors and have a higher level of FCR [24]. We con rmed the relationship between social supports, social and emotional quality of life, and levels of FCR in stomach cancer survivors. Because social support was related to predicting mood, coping strategies, and positive adjustment in cancer survivors [25][26][27], it should be considered in understanding FCR. But it has to be interpreted with caution as different results may be arisen depending on cultural differences of a family-dependent nature, age, perceived support level, or family roles [27,28].
Null associations were identi ed for gender and education level in this study. In Canadian patients with breast, prostate, lung, and colorectal cancer, the FCRI for female cancer survivors was high [29]. However, this nding did not hold after controlling for the cancer type. Although further research is needed when there is not enough gender-speci c research, the systematic review of 43 studies reported that gender appeared unrelated to FCR [8]. However, recent studies of hematologic cancer patients, including 467 Korean lymphoma patients, reported that female patients had a greater FCR than male patients [28,30].
Stomach cancer is one of the cancers with a good survival rate. Since 85% of the study participants were stage 1 or 2, the ve-year survival rate is increased up to 97.4% in 2017 [3]. On the other hand, the veyear survival rate of the Korean non-Hodgkin lymphoma patients is 63.8 % in the same year [3]. We thought that there may be a gender gap of FCR because a relatively poor prognosis can be vulnerable to emotion [28,31]. Regarding the relationship between education and FCR, FCR was associated with a lower education level in 89 American breast cancer patients [32]. There was a suggestion for the role of education that highly educated patients may have a greater understanding of the implications of cancer diagnosis [8]. However, the lack of relevance between FCR and education levels in this study is that the role of education may have been relatively insigni cant as long-term survivors who do not need the understanding of cancer diagnosis.
Several studies revealed no association between FCR and the in uence of treatment type and duration, and stage of the disease [10,28]. However, physical and psychological symptoms, such as fatigue, pain, sleep problems, distress, anxiety, and depression, were strongly associated with FCR regardless of cancer type or time since cancer diagnosis [10]. We suggest that this is also due to the nature of FCR in longterm survivors of cancer on the effect of uncontrolled physical and psychological symptoms [6, 10,24]. Therefore, continuing interest to control their physical and psychological problems is important in aspects of both FCR management and care coordination [14].
It was a noticeable nding that FCR was associated with the satisfaction of care coordination in stomach cancer survivors, which is compatible with the ndings of a previous study in 2,290 non-metastatic multi-ethnic breast cancer patients [33]. Well-coordinated care may re ect a good relationship between patients and healthcare providers. A provision of proper coaching or strategies for FCR from the healthcare providers would have lowered the level of FCR [14]. Additionally, whole-person care and tailored patient education for lifestyle modi cation may reduce the development of comorbidities and prevent late adverse effects after cancer diagnosis [34].
However, survivors' health promotion and disease prevention activities were not associated with the level of FCR in this study. Even in survivors with high FCR, lifestyle modi cation, secondary primary cancer screening, and surveillance of comorbidities were unaffected. These ndings do not support the 'teachable moment theory,' suggesting that cancer survivors who experience greater FCR are motivated to engage in health behaviors [11,35]. This theory has worked in the case of smoking cessation [36].
However, the theory does not apply to the health behaviors of cancer survivors because FCR severity may neither encourage nor hinder most survivors [11]. Cancer survivors with high FCR may experience high levels of distress and may not adhere to healthy lifestyle recommendations [37]. For example, a study in the United States found that survivors of breast, gynecologic, colorectal, and non-melanoma skin cancers with high FCR were twice more likely to smoke as those with no FCR and didn't do enough exercise (half an hour or more at least ve times a week) [38]. To explain this, we propose the following: that most cancer survivors try to improve their health behaviors regardless of the level of FCR. The practice of healthier habits de ned as any positive activities concerning cancer or treatment was found to have doubled after cancer diagnosis regardless of the level of FCR in US survivors [38]. Additionally, the rate of unhealthy behaviors in our study was too low to compare for signi cant differences across the different FCR groups; the proportion of current smokers in our study was much lower (0.4%) compared to that (14%) of the United States study [38]. Last, the effect of FCR on health behaviors may be motivational for a certain period of time only or during a treatment period. It was known that FCR stabilizes over time [10]. Therefore, there was no association between FCR and health promotion and disease prevention activities for long-term survivors [39]. Unfortunately, we did not nd the variance between time since diagnosis and FCR in this study. Nevertheless, this study is the rst to study to evaluate the related factors with FCR in stomach cancer survivors and has the strength to consider a wide range of factors, including clinical and sociodemographic characteristics, physical symptoms, psychological distress, and quality of life.

Conclusions
The present study con rmed the association between FCR of stomach cancer survivors and physical symptoms, psychological distress, social support, HRQoL, and satisfaction for communication within the medical team and care coordination. These ndings will help better understand the FCR of stomach cancer survivors and make a survivorship care plan. The written informed consent was obtained from all individual participants included in the study.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
None of the authors have nancial or non-nancial competing interests to declare.