The Dynamic Buffering of Social Support on Depressive Symptoms and Cancer Worries in Patients Seeking Cancer Genetic Counseling

Purpose: Social support is a crucial protective factor against psychological concerns in patients with cancer. However, there is limited knowledge regarding the differential impacts of social support on cancer worries and depressive symptoms in patients undergoing genetic counseling for hereditary cancer. The current study utilized a high-volume database from a multi-site cancer genetics clinic to assess the impact of perceived social support on depressive symptoms and cancer worries among patients of different age groups (young versus older patients) and diagnosis status (diagnosed survivors versus undiagnosed). Methods: 6,666 patients completed brief assessments of depressive symptoms, cancer worries, social support, and demographic questionnaires as part of routine clinical care between October 2016 and October 2020. Logistics and moderated regression were used to analyze the relationships between social support, depressive symptoms, and cancer worries. Results: Increased social support was associated with fewer depressive symptoms and fewer cancer worries across all patients. Social support mitigated depressive symptoms most signi�cantly for young adult patients with and without cancer. Social support mitigated cancer worries most signi�cantly for young adults with cancer and older adults without cancer. Conclusions While results were mixed, general �ndings upheld original hypotheses. Social support buffered depressive symptoms and cancer worries differentially for patients of different ages and different disease status. Implications for Cancer Survivors: Social support groups are bene�cial for all patients and should be emphasized by cancer clinics. However, increasing patient-tailored and age-appropriate support networks will be crucial for managing depression and cancer worries for high-risk survivors: young adults with cancer.


Introduction
Cancer genetic testing can identify gene mutations and cancer syndromes in individuals undiagnosed with cancer but have a known family history of hereditary cancer.Information received from genetic testing could help patients understand risk levels, aid decisions about reproduction, and in many cases, alleviate feelings of uncertainty [1,2].In addition, patients with a cancer diagnosis can receive cancer genetic counseling, as the results could assist patients with treatment, surgery decisions, and future cancer risk management.However, despite the known bene ts of genetic counseling, patients undergoing this process may experience many negative emotions (e.g., fear, cancer worries, loneliness), with some evidence suggesting that these emotions often arise pre-testing [1,3,4].Risk factors that could exacerbate these negative emotions include low social support, young age, and having a previous cancer diagnosis [4].As such, although the clinical populations of both undiagnosed and diagnosed patients seeking genetic counseling might experience similar stressors, young patients with cancer might endure a more signi cant psychological impact.Indeed, past research consistently demonstrates that young adults with cancer (hereafter referred to as YAs) are prone to much higher frequencies of mood disorders, suicidal ideation, and dysfunctional levels of fear of cancer recurrence when compared to older adult patients with cancer (hereafter referred to OAs) [5][6][7].By utilizing natural clinical data, the current study seeks to examine depression and cancer-related worries in various subgroups of patients seeking genetic counseling, including different age groups (young versus older patients) and diagnosis status (diagnosed survivors versus undiagnosed).

Depressive Symptoms in Oncology Patients
A 2013 large-scale systematic review revealed the global prevalence of depression in the general population to be approximately 5% [8].Concurrent research by Mitchel and colleagues reported the depression rate among cancer patients as 16.5% using the DSM-5 criteria [9], suggesting that the depression rate in cancer patients is approximately threefold that of the general population.Among patients with cancer, a substantial research body with high-volume samples has demonstrated depression to be more prevalent among those of younger age and those with limited social support [10][11].Although there is much debate as to whether chronological age is a predictor of depression among oncology patients, there is robust evidence to support that YAs are more prone to elevated depressive symptoms compared to OAs [12].For instance, among patients with breast cancer, younger patients consistently demonstrated more clinically signi cant depressive symptoms than their older counterparts [13].While clinically diagnosed depression in adult cancer survivors was prevalent up to 14.9%, the depression rate in YA survivors of childhood leukemia was two-fold, evident up to 28% [9,14].More recent research suggests that both YAs and long-term survivors of adolescent and young adult cancers report higher levels of depressive symptoms compared to OA patients [15].

Cancer Worries in Oncology Patients
Fear of cancer recurrence (FCR) is one of the most reported worries among cancer survivors [16].Elevated FCR can adversely affect daily functioning, quality of life, and social engagement among survivors [17].A 2013 systematic review demonstrated that age is a signi cant predictor of FCR, with younger cancer survivors endorsing much greater FCR than older survivors [18].A recent study examining FCR in YAs revealed that over 60% of YAs reported signi cant FCR, which, in turn, was associated with psychological problems and decreased quality of life [19].Notably, long-term survivors of young adult cancer also reported substantial levels of FCR, with up to 74% worrying about getting a new diagnosis and up to 20% worrying about disease recurrence [20].In sum, current literature indicates that YAs and long-term survivors of YA-cancer experience more signi cant FCR than older cancer survivors, which impairs their functioning across several domains.

Social Support as a Protective Factor for Depression and Cancer Worries
Previous research has consistently identi ed social support as a crucial buffer against psychological distress in cancer survivorship [21][22].Koch-Gallenkamp and colleagues reported that socially isolated survivors endorsed higher levels of FCR than their socially connected counterparts [23].Social support has also been shown to buffer against depressive symptoms in both YAs and OAs [24][25].Although current research substantiates the role of social support in buffering against both depressive symptoms and cancer-related worries, it is unclear how social dynamics might differ across age groups of patients.Therefore, this paper seeks to redress the dearth in the literature regarding the differential impact of social support on cancer worries and depression across patient groups seeking genetic counseling for hereditary cancer risks marked by age (young versus older patients) and by cancer diagnosis (diagnosed versus undiagnosed).

The Current Study
The current study used a high-volume sample from a cancer genetics clinic to evaluate social support, depressive symptoms, and cancer worries outcomes among various patient subgroups.In so doing, our aims were to: 1) examine whether social support buffers against depressive symptoms and cancer-related worry, and 2) examine how these buffering effects differ among the following patient groups: young adult patients with cancer (YAs), older adult patients who were survivors of youngadult cancers (OA/YAs), older adult patients with cancer (OAs), and their undiagnosed comparison groups: young adult patients without cancer (YAWOCs), older adult patients without cancer (OAWOCs).In Aim 1, we hypothesized that social support would be negatively associated with depressive symptoms and cancer worries.In Aim 2, we hypothesized that the buffering effects of social support on depressive symptoms and cancer worries would be the greatest for YAs when compared to OA/YAs, OAs, YAWOCs, and OAWOCs.

Participants and Data Collection Procedure
We derived our sample from a pool of 14,802 patients who sought genetic counseling through the Cancer Genetics Program at the University of Texas Southwestern (UTSW) Medical Center between October 2016 and October 2020.Data collection locations included academic medical center clinics, community clinics, and county safety-net hospital clinics.As part of routine clinical care, all patients received secure online questionnaires to complete before their genetic counseling appointments.Per UTSW-approved Institutional Review Board study (STU 062018-060), we identi ed 6,666 patients who fully completed data on all study variables, including patient sociodemographic, personal and family history of cancer, and several brief psychological and psychosocial assessments.Results were compiled into a singular database, with age at rst cancer diagnosis manually entered by genetic counseling assistants.The complete and de-identi ed database was then transferred to the psychology team for statistical analysis.
We established ve distinct patient groups based on their age at the genetic counseling appointment and age at the rst cancer diagnosis (if applicable).These ve groups were: (1) young adult patients diagnosed with cancer (YAs), (2) young adult patients without cancer (YAWOCs), (3) older adult patients diagnosed with cancer (OAs), (4) older adult patients diagnosed with cancer in their young adult years (OA/YAs), and (5) older adult patients without cancer (OAWOCs).In this participant sample, young adults refer to patients between 18 and 39 years of age, an age range consistent with past research [26].Older adults refer to patients of 40 years or older.

Measures Perceived Social Support
Perceived social support was measured using four items derived from the Interpersonal Support Evaluation List (ISEL), a widely reliable and validated instrument often used in health outcomes studies [27][28][29].Two of the four items constituted the Appraisal subscale of the ISEL, which measures the perceived availability of someone to discuss personal issues.The remaining items constituted the Belonging subscale, which captured the perception of being part of a social group.Responses were recorded on a 4-point Likert scale (0 = de nitely false, 1 = probably false, 2 = probably true, and 3 = de nitely true), with higher scores indicating higher perceived support availability.

Depressive Symptom Severity
The Center for Epidemiological Studies Depression Scale (CES-D) is a reliable and validated measure for assessing depressive symptoms [30].The 10-item version (i.e., CES-D-10) has also demonstrated adequate reliability and validity for screening depressive symptoms in patients with cancer [31][32].Total scores range from 0 to 30, with higher scores indicating greater symptom severity.Per previous clinical guidelines, a score of 10 was used as the clinical cut-off for indicating signi cant depressive symptoms.Only composite CES-D-10 scores were available in the de-identi ed database, preventing subsequent scale reliability analysis.

Cancer Worries
Previous Cancer Worries Scales have demonstrated adequate internal consistency and convergent validity [33][34][35].However, the original items were adapted for use in the current study.The phrasing of items and response anchors were slightly modi ed to improve clarity.Items were measured on a ve-point Likert scale (0 = Not at all, 1 = Rarely, 2 = Sometimes, 3 = Often, 4 = Almost all the time) and were as follows: (a) "How worried are you about getting or having a recurrence of cancer someday?",(b) "How much does your worry affect your mood?", (c) "How much does your worry affect your ability to perform daily activities?"Total scores range from 0 to 12, with higher scores indicating greater fear of getting cancer (for undiagnosed patients) or fear of cancer recurrence (for diagnosed patients).Only composite scores were available in the de-identi ed database, preventing subsequent scale reliability analysis.

Sociodemographic
Sociodemographic data included age at the genetic counseling appointment, age at rst cancer diagnosis, type(s) of cancer (if applicable), sex assigned at birth, race, ethnicity, insurance status, and zip codes (used to approximate household income based on the US Census Bureau).

Statistical Analysis
Before conducting statistical analyses, data were checked regarding assumptions of linearity and normality.Skewness and kurtosis were within acceptable bounds (i.e., ± 2 for skewness; ± 7 for kurtosis) and did not suggest the presence of outliers [36].All categorical variables were dummy coded with the following reference categories: no insurance for insurance type, female for sex, and YAs for group membership.Following this, we conducted two linear regression analyses to determine the direct effects of perceived social support and group membership on both depressive symptom severity and cancer worries.To determine the moderating effects of group membership on (1) the relationship between social support and depressive symptoms and (2) the relationship between social support and cancer worries, we conducted two moderation analyses.Regression and moderation models were adjusted for sex assigned at birth, income (in $1000s), and insurance (no insurance, private insurance, Tricare, Medicare, Medicaid, and county hospital insurance).Continuous measures -including perceived social support, depressive symptom severity, and cancer worries -were mean-centered for moderation analyses.All statistical analyses were conducted in IBM SPSS Statistics 28.0 [37], with moderation analyses performed using the PROCESS macro for SPSS (version 4.0) [38].Results were deemed statistically signi cant at p ≤ .05.

Sample Characteristics
The largest group in the sample was OAs (41.3%), followed by OAWOCs (30.0%),YAWOCs (19.2%),YAs (5.5%), and OA/YAs (4.0%).Participants predominantly identi ed as assigned female at birth (86.5%), non-Hispanic (86.5%), and White (72.3%).The average age of the sample was 50 ± 14.4 years, with an income (in $1,000s) of 80.97 ± 32.33.Most participants had private insurance (69.8%), and the most common diagnosis among cancer patients was breast (48.0%).The average level of cancer worries was 3.84 ± 2.62, while that for CES-D-10 scores was 6.75 ± 5.55, and for social support was 10.18 ± 2.60.Complete descriptive statistics for the full sample and by groups are in Table 1.Following this, a moderation analysis was conducted to examine the interaction between group membership and social support.Group membership emerged as a signi cant moderator of the relationship between social support and depressive symptoms (F (16, 6227) = 79.43,p < .001),adjusting for the same covariates included in the linear regression analysis.
Compared to YAs, all other groups, except YAWOCs, negatively moderated the relationship (OAWOC: B = -0.16,p = .028;OA: B = -0.36,p = .002;OA/YA: B = -0.28,p = .021),such that the inverse relationship between social support and depressive symptoms was lower in magnitude for OAWOCs, OAs, and OA/YAs in comparison to YAs, on average.No difference in magnitude existed in the social support-depressive symptoms relationship between YAs and YAWOCs (B = -0.01,p = .928).Full moderation results are in Table 3, with a graphical depiction of the moderating relationship provided in Fig. 1.

Social Support and Cancer-Related Worry
Regression analysis in which cancer worries speci ed as the outcome was statistically signi cant (F [12,6215] = 49.020,p < .001).Together, the model covariates explained approximately 9% of the variance in cancer worries (R2 = .086).Similar to the rst regression analysis, the main effect of social support was statistically signi cant (B = -0.149,p < .001),such that greater social support was associated with fewer cancer worries.Relative to YAs, all other levels of group membership had signi cantly less cancer worries, on average (OAWOCs: B = -1.947,p < .001;YAWOCs: B = -1.392,p < .001;OAs: B = -0.869,p < .001;and OA/YA: B = -0.928,p < .001).Additionally, those identifying as male at birth (B = -0.755,p < .001)tended to have fewer cancer worries.Compared to having no insurance, having private insurance (B = -0.6116,p < .001) or Medicare (B = -1.059,p < .001) was associated with lower cancer worries, on average.The results for the second regression analysis are in Table 2.
After adjusting for covariates, the moderating effect of group membership on the relationship between social support and cancer worries was signi cant (F [16, 6211] = 37.48, p < .001).Contrary to the moderation ndings with depressive symptoms speci ed as the outcome, there was no signi cant difference between OAWOCs and YAs (B = -0.05,p = .154).Compared to YAs, signi cant interactions between social support and group membership emerged for YAWOCs (B = -0.06,p = .040),OAs (B = -0.14, p = .014),and OA/YAs (B = -0.13,p = .032).Together, these results suggest that while the magnitude of the inverse relationship between social support and cancer worries signi cantly differed between YAs and YAWOCs, OAs, and OA/Yas, no differences existed between OAWOCs and YAs.That is, the effect of social support on cancer worries was less signi cant for YAWOCs, OAs, and OA/YAs compared to YAs. Results from the second moderation analysis are in Table 3, with a graphical depiction of the effect presented in Fig. 2.

Discussion
The present study evaluated the relationship between social support, cancer worries, and depressive symptoms among a large sample of patients seeking genetic counseling for hereditary cancer.Our participants included YAs, OA/YAs, OAs, and their two undiagnosed comparison groups: YAWOCs and OAWOCs.Our primary aims were 1) to examine whether social support buffers against depressive symptoms and cancer worries and 2) whether these buffering effects would be the strongest for YAs.Our ndings for Aim 1 supported our hypothesis, suggesting that regardless of a patient's age, diagnosis status, sex, income, and insurance status, social support ameliorates the severity of both depressive symptoms and cancer-related worries.To our knowledge, our study is the rst to account for several demographic variables while assessing the clinical signi cance of how social support mitigates depressive symptoms and cancer worries in patients undergoing genetic counseling.With carefully controlled confounds, our results underscore the importance of social support in managing the psychological burden associated with cancer survivorship, in the context of seeking genetic counseling for hereditary cancer [21][22][23][24][25][39][40].
For Aim 2, our ndings on depressive symptoms partially supported our hypothesis, indicating that although social support mitigates depressive symptoms signi cantly more for YAs than OA/YAs, OAs, and OAWOCS, these effects do not differ between YAs and YAWOCs.As such, YAs and YAWOCs struggling with depressive symptoms would particularly bene t from high levels of social support compared to the rest of the age and cancer groups.To interpret this nding, we considered previous research revealing that depression was more prevalent among young cancer patients with limited social support, suggesting that the experience of surviving cancer and/or hereditary cancer testing for YAs is often tremendously isolating [41][42].Combining such experiences with young age (a predisposing factor to loneliness [43]), it is unsurprising that strong support networks are crucial for relieving depressive symptoms in these two groups.As part of Aim 2, the ndings on cancerrelated worries partially supported our hypothesis, suggesting that social support would be most effective at mitigating fears of cancer recurrence for YAs and fears of incurring cancer for OAWOCs compared to the rest of the groups.Although an unexpected outcome, our sample characteristics might offer a potential explanation.Despite reporting the lowest CES-D-10 and cancer worries scores, OAWOCs also demonstrated the lowest level of social support relative to the rest of the group.Wang and colleagues (2014) discovered that if high-stress individuals also reported high support, the impact of stress on their depression was much smaller than those of low-stress and low-support [44].This research provides some evidence to suggest that an inadequate baseline social network for OAWOCs means receiving more social support may bring signi cant improvements in their cancer worries levels.Such ndings also imply that for OAWOCs, cancer worries might be a particularly burdensome psychological concern as they undergo genetic counseling for hereditary cancer risks.
In sum, our ndings underscore the differential buffering effects of social support on depressive symptoms and cancer-related worries among various groups.By demonstrating the speci c dynamic of these effects among patient groups across different ages (young adult versus older adult patients) and by cancer diagnosis status (diagnosed versus undiagnosed), we found that social support buffers depressive symptoms most signi cantly for YAs and YAWOCs, and respectively, cancer worries for YAs and OAWOCs.
Finally, in addition to the above ndings, our results indicated that patients with elevated depressive symptoms and cancerrelated worries included those who were YAs, those identi ed as female, those with low household incomes, and those with no insurance.These ndings are consistent with previous research on the socio-demographic factors predisposing patients to increased risks for depression and cancer worries, including age, sex assigned at birth, and socioeconomic status [45][46].

Clinical Recommendations Future Directions
First, although cancer clinics should emphasize social support groups for all patients, a more tailored and age-appropriate support network may be most important for managing depression and fear of cancer recurrence for YAs.Second, cancer genetics clinics consider collaborating closely with mental health providers to implement a specialized support network for managing depression for YAWOCs and cancer worries for OAWOCs.Third, oncology and cancer genetic clinics should continue incorporating screeners for depression, cancer worries, and social support into routine medical care.This research would bene t from future studies that attempt to replicate the above buffering dynamics of social support on depressive symptoms and cancer worries in patients with other cancer types, as many of our patient participants were diagnosed with breast cancer.Studies with analyses by the duration of time living with cancer (for diagnosed patients) and by risk levels (for undiagnosed patients) would also further our understanding of the differential impacts of social support.Lastly, it may be clinically useful to assess the actual impact of increasing social support activities (e.g., through attendance in group therapy or peer support activities) on depressive symptoms and cancer worries severity in high-risk patients (e.g., YAs).

Limitations
While ndings from the present study provided a deeper understanding of social support as a protective factor, there are limitations to note.We categorized diagnosis status as a binary variable (yes/no), which omitted the inherent heterogeneity in cancer diagnoses.We generated the income variable based on zip code data -a proxy sometimes unrepresentative of actual income.Additionally, patients referred to genetic counseling might already be at increased risk for cancer regardless of diagnosis status, combined with the fact that our participant samples were predominantly female-identifying patients with breast cancers, could pose limits to the external validity of our results.Our outcomes might be affected by selection bias to some degree; and thus, might not be generalizable to cancer patients who are not seeking genetic counseling or those without a family history of cancer.Moreover, cross-sectional data analyses were available for patients who fully completed the preappointment questionnaires only, introducing nonresponse bias and limiting our ability to infer strong causality.Finally, our total scores on CES-D-10, cancer worries, and social support were pre-computed by the cancer genetics system before statistical analyses, preventing scale reliability analyses.Nevertheless, the high-volume sample size (> 6000) and naturalistic clinical data collected at various locations bolster the ecological validity of our ndings.

Declarations
Funding and Competing Interests: We have no funding source and competing interest to disclose.
Author Contributions: All contributed to the study conception and design.Material preparation and analysis were performed by Sally Ho and Jayme M. Palka.The rst draft of the manuscript was written by Sally Ho.All authors commented on previous versions of the manuscript, read, and approved the nal manuscript.The moderating role of group on the relationship between perceived social support and cancer worries.OAWOC = Older adults without cancer, YAWOC = Young adults without cancer, OA = Older adults with cancer, OA/YA = Older adults who were survivors of young-adult cancer, YA = Young adults with cancer.

Figures
Figures

Figure 1 The
Figure 1