Social Support, Depression and Anxiety in Cancer Patient‐Relative Dyads in Early Survivorship: An Actor‐Partner Interdependence Modeling Approach Authors: Myriel Hermann, Ute Goerling, Charis Hearing, Anja Mehnert‐Theuerkauf, Beate Hornemann, Peter Hövel, Sabrina Reinicke, Hanna Zingler, Tanja Zimmermann, Jochen Ernst Published: 2024 ABSTRACT ABSTRACTObjectiveCancer places a psychological burden on both patients and their relatives. Perceived social support influences the extent of psychological distress. Our aim was to investigate associations between positive support, detrimental interactions, depression and anxiety in patient‐relative dyads in the initial period after diagnosis.MethodsPatients with a solid tumor and their relatives participated in this prospective, multicenter observational study. Participants answered validated measures including the Illness‐specific Social Support Scale (SSUK‐8), the Patient Health Questionnaire (PHQ‐9) and the General Anxiety Disorder Scale (GAD‐7). We analyzed cross‐sectional data from the initial time following diagnosis with paired t‐tests and actor‐partner interdependence models.ResultsA total of 347 dyads of patients (mean age 59.9 years) and their relatives (mean age 56.7 years) participated. Compared to their relatives, patients reported significantly higher levels of depression (patients: M = 6.31, SD = 4.94; relatives: M = 5.44, SD = 4.77) and lower levels of anxiety (patients: M = 4.40, SD = 4.10; relatives: M = 4.98, SD = 4.47) as well as more positive support (patients: M = 14.31, SD = 2.07; relatives: M = 12.46, SD = 3.29) and a lower frequency of detrimental interactions (patients: M = 3.21, SD = 2.97; relatives: M = 3.66, SD = 2.93). Intrapersonal effects: Positive support was associated with lower distress only for relatives, whereas detrimental interactions were associated with higher distress for both patients and relatives (all p < 0.05). Interpersonal effects: More positive support and fewer detrimental interactions experienced by relatives were associated with lower patient distress (all p < 0.05).ConclusionsBetter support for relatives may not only reduce their own distress, but also patients' distress. Relatives experience similar levels of distress and poorer social support than patients.Keywords: anxiety, cancer, caregivers, depression, oncology, psycho‐oncology, psychological distress, relatives, social support, survivorship Background The initial period after cancer diagnosis is especially difficult due to a variety of challenges, ranging from the shock of a potentially life‐threatening disease and uncertainty about the further trajectory to the first distressing side effects of therapy and suddenly altered social roles [1, 2]. Depression and anxiety after diagnosis are not only associated with depression and anxiety later on in the course of the disease [3], but can also reduce quality of life [4], lead to increased healthcare utilization [5] and are linked with higher mortality [6]. Importantly, cancer is not only affecting patients, but also those close to them, such as intimate partners, family members or friends. While earlier research focused on the impact of the diagnosis on patients, there is now a large body of evidence showing that relatives are at least as distressed. This finding has been replicated several times in studies comparing dyads of patients and their respective relatives [7, 8, 9, 10]. Relatives struggle with the existential distress of a loved one's potentially disabling and life‐threatening illness while providing many forms of assistance that require time and commitment. This often leads them to subordinate their own needs and neglect social activities and self‐care which may result in poor mental and physical health [11]. One resource in patients' and their relatives' process of adjusting to diagnosis and treatment is support from the social environment, which is theorized to buffer against the emotional distress of the illness [12]. Empirically, positive social support is adversely correlated with several measures of psychological distress, such as anxiety, and depression in both cancer patients (correlations ranging from r = −0.10 to r = −0.69) [7, 12, 13, 14] and their relatives (correlations ranging from r = −0.16 to r = −0.74) [7, 14, 15, 16]. While the buffering effect may work for all, the amount of support received may differ by role: patients tend to receive more positive support than their relatives [7, 17]. However, not all forms of support are beneficial. Actions intended to support are unhelpful, if they do not meet the needs of the recipient. The concept of unhelpful support, also labeled as detrimental interactions [18], is not conceptualized or researched as well as its positive counterpart. It includes perceived problematic interactions such as being pessimistic and overprotecting [18]. Studies show that problematic social support is associated with higher anxiety, fear of progression and depression in cancer patients (correlations ranging from r = 0.30 to r = 0.45) [13, 19, 20]. Studies on negative support in relatives are scarce, but show a similar trend [19]. The negative effect of detrimental interactions on mental health is likely to be equal to or greater than the buffering effect of positive support [13, 18] and should be considered accordingly. Although there is increasing evidence that the mental and physical health of patients and their relatives are interdependent [7, 21, 22], and despite a growing body of research advocating the use of dyadic approaches to study interpersonal phenomena [21, 23], most studies to date investigated the associations between positive support, detrimental interactions and psychological distress at the individual level only. Exploratory analyses with a small dyadic sample found no interpersonal effects [24], but a study with a large sample of haemato‐oncological dyads of patients and their spouses suggests that the more positive support the relatives receive, the less depressed the patients are [7]. In line, another study found a similar interpersonal effect in advanced lung cancer patients and their family caregivers: higher levels of positive support perceived by caregivers were associated with lower levels of depression and anxiety in patients [14]. A recently published study among patients with advanced lung cancers and their caregivers corroborates the impact that the relatives' perceived support may have on patients' well‐being. This study found that relatives' positive support is positively associated with patients' level of hope [25]. Previous studies have mainly analyzed the associations between positive social support, detrimental interactions and psychological distress, such as symptoms of depression and anxiety, mostly in individuals and in samples that were very heterogeneous in terms of the time since diagnosis. Thus, the current study focuses on dyads of patients with solid tumors and their relatives enrolled within the first 2 months after diagnosis. In line with previous findings [7, 8, 9, 10], we hypothesize that patients and relatives will show comparable levels of anxiety and depression. We aim to explore potential differences in perceived support, since few studies have compared patients' and relatives' positive support and none have compared detrimental interactions as far as we know. In line with previous work [7, 12, 13, 14, 16, 19, 20] we assume that, at the individual level, positive support is associated with less depression and anxiety for both patients and relatives, whereas detrimental interactions show associations in the opposite direction (actor effects). While considering the intrapersonal effects, this study also examines how social support perceived by one member of the dyad, for example the patient, may affect the extend of depression and anxiety experienced by the other member, for example the relative—and vice versa (partner effects). Methods The current analysis was part of the study “Longitudinal analysis of mental disorders, psychosocial distress and psycho‐oncological care needs of patients and their relatives stratified by biopsychosocial factors (LUPE)” [26], using data from the baseline (enrollment within 2 months after cancer diagnosis). The study received ethical approval from the local ethics committees in the study centers in Leipzig (AZ: 207/18‐ek, approval date 28.05.2019), Berlin (the Charité Ethics Committee has approved the favorable vote of the Leipzig Ethics Committee, approval date 03.07.2019), Hannover (8533_BO_K_2019, approval date 01.07.2019), Dresden (SR‐EK‐536112020, approval date 12.01.2021) and Göttingen (14/4/21 U approval date 12.04.2021). Recruitment and Data Collection We recruited patients and their relatives between April 2020 and July 2022 at the university comprehensive cancer centers of the three main study centers in Leipzig, Berlin and Hannover, and at the cooperating university comprehensive cancer centers in Dresden and Göttingen as well as the cancer center Braunschweig. Patients were eligible if they met the following inclusion criteria: (1) at least 18 years of age, (2) a confirmed initial diagnosis of a malignant solid tumor, (3) sufficient German language skills and (4) cognitive, psychological and physical ability to participate. All patients were enrolled within the first 2 months after their diagnosis. The patients were asked to nominate a close relative from their immediate environment who had to be at least 18 years old. All participants had to provide written informed consent to participate in the study. We asked patients and relatives to complete a set of validated questionnaires within 10 days, either as a paper‐pencil version or online via a personalized link using the LimeSurvey software [27]. Participants were reminded by a telephone call after 2 weeks and again after 3 weeks if they had not completed and returned the questionnaires by then. Further details are reported elsewhere [26, 28]. Measures Sociodemographic and Clinical Data We extracted medical information from patients' medical records and assessed sociodemographic variables via self‐report. Depression Symptoms Depression symptoms were assessed with the German version of the 9‐item depression module of the Patient Health Questionnaire (PHQ‐9) [29] for which participants self‐report the occurrence of each of the nine DSM‐IV diagnostic criteria in the past 2 weeks on a four‐point Likert scale from 0 (“not at all”) to 3 (“nearly every day”). The sum scores range from 0 to 27; higher scores represent higher levels of depression [29]. Internal consistency (Cronbach's α) in cancer patients is satisfactory with α = 0.84 [30]. In the current study the internal consistency was α = 0.84 and α = 0.86 for patients and relatives, respectively. Anxiety Symptoms Anxiety symptoms were assessed with the German version of the 7‐item General Anxiety Disorder Scale (GAD‐7). Participants are requested to evaluate the frequency of seven items representing the core symptoms of generalized anxiety disorder linked to the DSM‐IV criteria in the past 2 weeks on a four‐point Likert scale from 0 (“not at all”) to 3 (“nearly every day”). The sum scores range from 0 to 21; higher scores represent higher levels of anxiety [31]. Internal consistency (Cronbach's α) in cancer patients is satisfactory with α = 0.88 [32]. In the current study the internal consistency was α = 0.87 and α = 0.89 for patients and relatives, respectively. Positive Support and Detrimental Interactions Perceived support from the social environment was assessed with the German 8‐item short‐version of the Illness‐specific Social Support Scale (SSUK‐8) [18]. The two subscales Positive Support (e.g., “Talked about important decisions with you”) and Detrimental Interactions (e.g., “Tried to change the way you're coping with your illness in a way you don't like”) each consist of four items. Items are rated on a five‐point Likert scale ranging from 0 (“never”) to 4 (“always”), resulting in sum scores from 0 to 16 for each subscale. High scores on the positive support subscale indicate a higher frequency of perceived positive social support while high scores on the detrimental interactions subscale indicate a higher frequency of problematic interactions. Internal consistency (Cronbach's α) of the subscales is α = 0.88 for positive support and α = 0.68 for detrimental interactions [18]. In our samples internal consistency was partly lower, but still acceptable (positive support: α = 0.66 in patients, α = 0.83 in relatives; detrimental interactions: α = 0.70 in patients, α = 0.64 in relatives). Statistical Analysis We conducted data analysis using IBM SPSS Statistics 29 and for the Actor Partner Interdependence Models (APIMs) SPSS Amos 29. Missing values were imputed by an Expectation‐Likelihood‐Maximization (EM) algorithm [33]. The EM algorithm is an iterative two‐step technique estimating missing values. In the first step (expectation) missing values are estimated based on all relevant variables of the dataset. In the second step (maximization) new parameters, such as means and covariances, are calculated based on the dataset with the imputed values. The two steps are replicated until the parameters are not changing substantially by further iterations. Missing values were only estimated if at least 60% of the values of a scale were available for a person. To report sociodemographic and medical data we used descriptive statistics. Chi‐square and t‐tests were used to compare the patients with a participating relative to the patients without a participating relative. We calculated paired samples t‐tests (two‐tailed tests) for examining possible differences between patients and relatives in the levels of depression, anxiety, positive support and detrimental interactions. Cohen's d was calculated to evaluate the effect size of the differences—|d| = 0.8 represents large, |d| = 0.5, medium and |d| = 0.2 small effects. We computed Pearson correlations over all variables to check for non‐independence in the dyad. The correlations between the predictors were analyzed for possible collinearity. Only absolute values of the correlation coefficients > 0.50 indicate collinearity [34]. Taking the interdependency of the data into account, we analyzed the data using the Actor‐Partner‐Interdependence Model (APIM) applying a structural equation modeling approach. The APIM estimates intrapersonal actor effects and interpersonal partner effects in the dyad simultaneously while also considering the correlations between the predictors and the residual non‐independence in the outcome scores. Actor effects analyze the impact of predictor variables of one member of the dyad on the outcome variable of the same person. Partner effects are possible effects of predictor variables of one member of the dyad on the outcome variable of the other member of the dyad [23]. We estimated two independent APIMs (each with potentially four actor and four partner effects)—one for each dependent variable (depression and anxiety) with age and gender of patients and relatives as control variables. The Alpha level was set to 0.05. For standardized coefficients, we used the mean and standard deviation across the whole sample (patients and relatives). Results Out of 1702 eligible patients, 1003 patients answered the set of questionnaires at the first measurement time point; in 347 patient‐relative dyads both members answered. Further details on eligible patients and study participants are reported elsewhere [28]. Patients with a participating relative did not differ in age from patients without a participating relative (M = 59.9 vs. M = 60.9, p = 0.234). However, patients with a participating relative were more likely to have a high socioeconomic status (high socioeconomic status: 49.3% vs. 35.3%, low economic status: 10.5% vs. 17.8%, p < 0.001). Sample Characteristics The sociodemographic and medical characteristics of the patients and their relatives are shown in Table 1. The patients were 59.9 (SD = 13.9) years old on average and the majority were men (56.8%). The relatives were 56.7 (SD = 14.8) years old on average and the majority were women (66.6%). 78.7% of the relatives were partners/spouses. TABLE 1Patient and relative characteristics: sociodemographic and medical variables.Cancer patients (N = 347)Relatives (N = 347)Demographic characteristicsAge (years)Mean (SD)59.9 (13.9)56.7 (14.8)Range21–9119–89%(n)%(n)GenderFemale43.2(150)66.6(231)Male56.8(197)33.4(116)Educational qualificationUniversity degree48.8(157)50.7(176)High school/vocational training48.8(157)45.5(158)No qualification2.5(8)3.2(11)Other0(0)0.6(2)OccupationEmployed51.1(164)54.2(188)Unemployed0.9(3)2.6(9)Retired45.5(146)38.9(135)Other2.5(8)4.3(15)Socioeconomic status a Low10.5(36)7.8(27)Medium40.2(138)45.0(156)High49.3(169)47.3(164)Religious affiliationProtestant19.3(62)21.4(74)Catholic5.0(16)7.2(25)Other2.5(8)1.7(6)None73.3(236)69.7(241)NationalityGerman97.8(314)98.0(340)Other2.2(7)2.0(7)Residential areasUrban (≥ 20.000 inhabitants)55.3(178)56.7(196)Rural (< 20.000 inhabitants)44.7(144)43.4(150)Relative relationship to the patientSpouse/partner78.7(273)Parent3.7(13)Child13.0(45)Other family member2.9(10)Friend1.7(6)Medical characteristicsTumor siteProstate (C61)24.8(86)Digestive organs (C15–C26)14.1(49)Breast (C50)13.5(47)Female genital organs (C51–C58)13.3(46)Melanoma (C43–C44)8.9(31)Kidney/urinary tract (C64–C68)8.9(31)Head and neck (C00–C14)4.0(14)Lung (C34)2.0(7)Other10.4(36)Months since diagnosis b Mean (SD)1.5 (1.1)Range0–6≤ 2 months86.7(301)> 2 months13.3(46)UICCI41.9(144)II23.5(81)III19.8(68)VI14.8(51)Cancer treatment receivedNo1.4(5)Yes98.6(342)Type of cancer treatment c Surgery88.6(303)Radiotherapy17.5(60)Chemotherapy19.3(66)Other14.3(49) Note: (n) are valid answers only; the discrepancy between (n) and the full sample size represents the missing values; the percentages (%) are based on the valid answers.Abbreviations: SD, standard deviation. UICC, Union for International Cancer Control disease stage. a The socioeconomic status was calculated based on three indicators: (1) school education and vocational qualification, (2) income and (3) occupational status. More details are reported elsewhere [28]. b Months since diagnosis in relation to first questionnaire completion. > 2 months: deviation from the study protocol (maximum up to 6 months), since recruitment during COVID‐19 was only possible indirectly via mail, which extended the time required for the study inclusion process. c Multiple response possible; based only on patients who received a cancer treatment (n = 342). Patient and relative characteristics: sociodemographic and medical variables. Note: (n) are valid answers only; the discrepancy between (n) and the full sample size represents the missing values; the percentages (%) are based on the valid answers. Abbreviations: SD, standard deviation. UICC, Union for International Cancer Control disease stage. The socioeconomic status was calculated based on three indicators: (1) school education and vocational qualification, (2) income and (3) occupational status. More details are reported elsewhere [28]. Months since diagnosis in relation to first questionnaire completion. > 2 months: deviation from the study protocol (maximum up to 6 months), since recruitment during COVID‐19 was only possible indirectly via mail, which extended the time required for the study inclusion process. Multiple response possible; based only on patients who received a cancer treatment (n = 342). Levels of Depression, Anxiety, Positive Support and Detrimental Interactions Depression scores were significantly higher in patients (M = 6.31, SD = 4.94) than in their relatives (M = 5.44, SD = 4.77; p = 0.004, d = 0.16, 95% CI [0.05, 0.26]). Anxiety scores were significantly lower for patients (M = 4.40, SD = 4.10) than for their relatives (M = 4.98, SD = 4.47; p = 0.032, d = −0.12, 95% CI [−0.22, −0.01]). Patients reported significantly higher levels of positive support (M = 14.31, SD = 2.07) than their relatives (M = 12.46, SD = 3.29; p < 0.001, d = 0.51, 95% CI [0.40, 0.62]) and significantly lower levels of detrimental interactions (M = 3.21, SD = 2.97) than their relatives (M = 3.66, SD = 2.93; p = 0.032, d = −0.12, 95% CI [−0.22, −0.01]). Preliminary Analysis for the Actor‐Partner‐Interdependence Models Table 2 shows all intra‐ and interpersonal correlations indicating non‐independence and the appropriateness of the dyadic approach. The within‐ and between‐participants correlations of positive support and detrimental interactions were not exceeding |r| = 0.18, indicating that there was no risk of collinearity of the predictors [34]. TABLE 2Intra‐ and interpersonal correlations of depression, anxiety, positive support and detrimental interactions.PatientRelativeDAPSDIDAPSDIPatientDepression (D)0.74***−0.11*0.22***0.35***0.28***−0.17**0.21***Anxiety (A)−0.050.22***0.33***0.31***−0.14*0.22***Positive support (PS)−0.17**−0.09−0.13*0.15**−0.18***Detrimental interactions (DI)0.100.14**−0.17**0.12*RelativeDepression (D)0.81***−0.18***0.18***Anxiety (A)−0.19***0.21***Positive support (PS)−0.05Detrimental interactions (DI)*p < 0.05, **p < 0.01, ***p < 0.001. Intra‐ and interpersonal correlations of depression, anxiety, positive support and detrimental interactions. *p < 0.05, **p < 0.01, ***p < 0.001. Positive Support and Detrimental Interactions on Depression Figure 1 illustrates the results of the APIM explaining depression (controlled for gender and age). Actor effects: The association between patients' positive support scores and patients' depression symptoms was not significant (p = 0.759). More detrimental interactions in patients were associated with more depression symptoms in patients (p = 0.002). For relatives, more positive support was associated with fewer depression symptoms (p < 0.001) and more detrimental interactions were associated with more depression symptoms (p = 0.014). Partner effects: Both positive support and detrimental interactions in patients were not significantly related to the depression symptoms of their relatives (p = 0.938 and p = 0.399 respectively). More positive support reported by relatives predicted fewer depression symptoms in patients (p = 0.011) and more detrimental interactions predicted more depression symptoms (p = 0.008). The model for depression symptoms explained 15.2% of the variance in patients and 19.2% in relatives. Actor‐partner interdependence model of depression symptoms. Depression symptoms (PHQ‐9) of patients and relatives are controlled for gender and age of patients and relatives. The values on the arrows depict the unstandardized beta coefficients; the standardized coefficients are in parenthesis; *p < 0.05, **p < 0.01, ***p < 0.001. Positive Support and Detrimental Interactions on Anxiety Figure 2 illustrates the results of the APIM explaining anxiety (controlled for gender and age). Actor effects: Patients' scores of positive support and patients' anxiety symptoms were not significantly associated (p = 0.374). More detrimental interactions in patients were associated with more anxiety symptoms in patients (p < 0.001). The more positive support the relatives reported the fewer anxiety symptoms they had (p < 0.001) and the more detrimental interactions they reported the more anxiety symptoms they showed (p = 0.003). Partner effects: The positive support and detrimental interactions reported by patients were not significantly related to the relatives' anxiety symptoms (p = 0.618 and p = 0.096 respectively). The association between relatives' scores of positive support and patients' anxiety symptoms was significant (p = 0.039) and more detrimental interactions in relatives predicted more anxiety symptoms in patients (p = 0.005). The model for anxiety symptoms explained 18.5% of the variance in patients and 19.1% in relatives. Actor‐partner interdependence model of anxiety symptoms. Anxiety symptoms (GAD‐7) of patients and relatives are controlled for gender and age of patients and relatives. The values on the arrows depict the unstandardized beta coefficients; the standardized coefficients are in parenthesis; *p < 0.05, **p < 0.01, ***p < 0.001. Discussion We investigated how two different facets of perceived social support from the social environment—positive support and detrimental interactions—relate to symptoms of depression and anxiety in 347 dyads of cancer patients and their relatives within the first weeks after cancer diagnosis. We found small differences in the levels of psychological distress. In our sample, the relatives' depression scores were significantly lower than the patients' scores. Differences in activity levels may explain this finding. Activity restriction is known to be associated with depression in cancer patients [35]. While patients are forced to stop many pleasant, rewarding and meaningful activities directly after diagnosis, the relatives are likely to be more active with caregiving as an additional duty. Additionally, one might argue that patients directly experience the consequences of the disease through diagnostic examinations, immediate treatment and possible first side effects. This could lead them to focus more on the actual losses than their relatives resulting in more depressive symptoms. At the same time, relatives in our study reported significantly higher levels of anxiety. Perhaps patients focus more on current problems while relatives tend to think more about potential difficulties in the future, leading to more anxiety in the later. Consistent with this cautious explanation, relatives seem to experience more fear of recurrence than the patients themselves [8]. However, the effect sizes of both differences in psychological distress were very small. Therefore, the hypothesis still holds that the levels of psychological distress of patients and relatives are quite comparable. Patients experienced more positive support from their social network than their relatives did. Several authors suggest that the social system of family and friends focuses on providing positive support and care for the cancer patients while neglecting the needs of the close relatives [7, 9, 17]. In addition, patients—as an important part of their relatives' social network—may be preoccupied with their own difficulties in the early survivorship stages. They might not be able to provide as much support as they did before the diagnosis. The slightly higher frequency of detrimental interactions perceived by the relatives may be due to relatives trying to keep patients away from distressing interactions and acting as gatekeepers [36] for them. As in previous research [7, 10], patients' and their relatives' depression and anxiety scores were moderately correlated, indicating interdependence of psychological distress within dyads: the more distressed one member of the dyad, the more distressed the other member too. Individual level (actor effects): Positive social support in relatives was reversely associated with their symptoms of anxiety and depression: the more supported they felt, the fewer symptoms of psychological distress they reported. This was consistent with previous findings [7, 14, 15, 16]. Interestingly, no such associations were found between positive support and psychological distress in patients which was inconsistent with our hypothesis formulated based on the existing literature [12, 13, 14]. A possible explanation could be that positive support has less (or no) impact on the newly diagnosed patients because they first struggle with the existential threats of death and health impairments. Narrative accounts of interviewed cancer patients describe an initial phase after diagnosis of isolation, introversion and “switching off” to cope with the life‐threatening situation [1]. Maybe the positive support unfolds its buffering effects only after this phase of the diagnostic shock. As in previous studies that focused mainly on patients [13, 20], detrimental interactions were positively associated with symptoms of anxiety and depression in both patients and relatives. This means that a higher frequency of negative interactions, such as perceived overprotection by another person, was associated with more symptoms of psychological distress. Interpersonal level (partner effects): We did not find any partner effects from facets of perceived social support experienced by the patients to the depression and anxiety scores of the relatives. The positive support that relatives perceived for themselves was negatively associated with symptoms of depression and anxiety in the patients. In other words, the better supported, helped and encouraged the relatives felt by their social network, the less depressed and anxious the patients were, in line with two previous studies examining patient‐relative dyads [7, 14]. It is possible that patients are afraid to burden their relatives with the disease and feel relieved if their relatives get good social support [7]. Another explanation might be that relatives with more positive support and less detrimental interactions are more likely to functionally cope with the stressful situation [37] which may enhance constructive communication within the dyad and therefore lead to less distress in patients. Furthermore, detrimental interactions reported by the relatives were associated with patients' scores of depression and anxiety—the more negative interactions, like criticizing, the relatives experienced, the more depressed and anxious the patients felt. To our knowledge, this is the first study showing partner effects of negative interactions. Both actor‐partner interdependence models explained only a small portion of the variance in depression and anxiety in patients and their relatives, suggesting that many other factors besides social support, age and gender may influence these measures of distress. Study Limitations Our results must be interpreted with caution due to several limitations of the current study. The study sample consisted of dyads with a relatively high socioeconomic status. Patients with participating relatives were more likely to have a higher economic status than patients without participating relatives. Furthermore, former analysis of our sample [28] showed that participating patients were of higher socioeconomic status than non‐responders. While we assume that support influences the level of distress in line with the well‐described buffering effect [12], it could also be that distress leads to less positive support from close ones and more negative interactions. Alternatively, distressed individuals, especially if they are depressed, might rate their social support worse (with less positive support and more detrimental interactions) than non‐distressed ones due to the well‐known negative interpretation bias [38]. The social support questionnaire we used (SSUK‐8) does not evaluate who provides the perceived support; it can include family, friends and partners. We can only assume that, among others, the patient is a relevant caregiving person in the support system of the relative and vice versa. We collected data from spousal (partners) and non‐spousal caregivers. Due to the small number of the non‐spousal caregivers, the analysis was done for the whole sample without subgroup analyses. Perhaps the pattern of the interpersonal partner effects is different or at least of different size for the spousal versus non‐spousal relatives. We collected data between April 2020 and July 2022 during the Covid‐19 pandemic. The restrictions to fight the pandemic had a major impact on social interactions, especially for cancer patients and their relatives [39]. We would argue that although the restrictions made the provision of concrete support more difficult, they did not have as great an impact on the perceived social support as conceptualized in the SSUK‐8, which does not rely on the face‐to‐face interactions. Future research will show whether similar result patterns appear under non‐pandemic conditions. Clinical Implications Although we found small differences in the anxiety and depression scores for patients and relatives, the levels are comparable. Like many studies have shown before, professionals need to take care of the relatives as much as of the patients. Interventions aimed to promote positive support and, perhaps even more importantly, to reduce negative interactions could contribute to the reduction of depression and anxiety. As our results show that relatives experience less positive support and more detrimental interactions than patients, they may be even more in need of such interventions. The partner effects in this study suggest that the social support perceived by relatives is related to patient distress. Therefore, supporting the relatives may not only benefit them, but also the patients. Social support interventions and their outcomes have been reviewed for cancer patients [40]. More research on interventions for relatives is needed. Conclusion Our results suggest that cancer patients with solid tumors and their relatives experience comparable levels of depression and anxiety within the first weeks after cancer diagnosis—with patients experiencing more depressive symptoms and relatives experiencing more anxiety symptoms. Relatives appear to feel less supported by their social network than patients, experiencing less positive support and more detrimental interactions. It seems promising to support relatives directly and promote support in their private social networks as the support they experience may not only reduce their own psychological distress, but also the distress of the patients. Author Contributions Ute Goerling, Anja Mehnert‐Theuerkauf, Tanja Zimmermann and Jochen Ernst acquired the financial support for the study, provided the resources for the data collection and analysis, were responsible for the methodological realization and coordinated and supervised the research activity. Myriel Hermann, Charis Hearing, Peter Hövel, Sabrina Reinicke and Hanna Zingler were responsible for the data collection. Myriel Hermann analyzed the data, interpreted the findings, and drafted the manuscript including the tables and figures. All authors revised the manuscript critically for intellectual content and approved the final version. Conflicts of Interest The authors declare no conflicts of interest. Acknowledgments The study was funded by the German Cancer Aid (Deutsche Krebshilfe). In addition, our study was financially supported by the Department of Medical Psychology and Medical Sociology at the University Medical Center of Leipzig, by the Charité Comprehensive Cancer Center, Berlin, and the Department of Psychosomatic Medicine and Psychotherapy, Hannover Medical School, Hannover. We thank all participants of our study as well as the study team. Furthermore, we would like to thank Dr. Michael Friedrich from the Department of Medical Psychology and Medical Sociology, University Medical Center Leipzig (Leipzig, Germany) for answering questions regarding the actor‐partner interdependence models. Open Access funding enabled and organized by Projekt DEAL. Funding: This study was funded by the German Cancer Aid (Deutsche Krebshilfe). In addition, our study was financially supported by the Department of Medical Psychology and Medical Sociology at the University Medical Center of Leipzig, by the Charité Comprehensive Cancer Center, Berlin, and the Department of Psychosomatic Medicine and Psychotherapy, Hannover Medical School, Hannover. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. References Associated Data Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.