Participants
Patients with cancer of any type or relatives of cancer patients were included in the intervention group (IG). Requirements included that they had to visit one of two cancer counseling centers of the German Cancer Society between December 2018 and August 2021 for the first time, had sufficient German language skills, were at least 18 years old and were physically and mentally able to participate in the study. The IG was compared to patients with cancer or relatives of cancer patients who did not seek support at one of the cancer counseling centers (control group, CG). For the CG, participants were approached in seven different outpatient oncological clinics and practices in the same area. Patients and their relatives receiving psychotherapy were not eligible to participate. All eligible candidates gave their informed consent before participation.
Participants of the IG received the baseline questionnaire (t0) before their first support session. The CG received the same questionnaire after agreeing to participate in the study. The second questionnaire (t1) was sent by post two weeks after the first support session. The follow-up questionnaire (t2) was sent 12 weeks after the first session.
Further information on the study procedure and design is available in the published study protocol [17].
Intervention
The support session was administered face-to-face by trained psycho-oncologists, social workers, or oncologists at one of the two cancer counseling centers. The type of health care personnel and content of the session depended on the main concerns of the clients, which could be of psychological, social, legal, or medical nature. The duration of one session ranged from 0.5 to 1.5 hours. Between t0 and t1 only one session was administered. Between t1 and t2 about 37% of the participants received additional counseling sessions of which n = 49 received one additional session, n = 24 received two additional sessions, n = 12 received three additional sessions and n = 9 received between four and eight additional sessions.
Outcomes
The primary outcome was distress, which was assessed with the Distress Thermometer (DT) [18]. The secondary outcomes were depressive and anxiety symptoms, well-being, generic and cancer specific quality of life (QoL), self-efficacy and fatigue. Cancer specific QoL and Fatigue were only assessed among patients with cancer but not their relatives.
The secondary outcomes were assessed as follows: depressive (Patient Health Questionnaire, PHQ-9) [19] and anxiety symptoms (General Anxiety Disorder questionnaire, GAD-7) [20], generic QoL (Subscales Mental and Physical Component Summary of the Short Form 8, SF-8) [21] [22], well-being (World Health Organization rating scale, WHO-5) [23, 24], self-efficacy (Sense of Mastery Scale, SOM) [25, 26], cancer specific QoL (European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire, EORTC-QLQ-C30, subscales functional, symptom and global) [27] and fatigue (Fatigue Assessment Scale, FAS) [28].
Additionally, satisfaction levels with the support and associated concerns, and barriers to the utilization of psychosocial support were evaluated by self-developed questionnaires. Satisfaction with the support services was assessed by the German patient satisfaction questionnaire (ZUF-8) with sum scores ranging from 0 to 32, where higher scores indicate greater satisfaction [29]. The helper-client alliance was evaluated with the Helping Alliance Questionnaire (HAQ) with sum scores ranging from 6 to 54, where higher scores quantify a stronger alliance [30].
Covariates
All analyses were adjusted for previously defined covariates assessed at baseline and outlined in the study protocol [17]. The following covariates were included in the analysis:
Sociodemographic factors: Age (in years), sex (female/male), patient or relative, migrant background (yes/no), educational background (high school or higher: yes/no), current occupational status (employed: yes/no), source of income (occupation/pension/governmental support), family status (single: yes/no), children (yes/no).
Clinical factors: Included during Covid-19 pandemic (yes/no), cancer type (breast cancer/other), metastases (yes/no), other diseases e. g. cardiovascular disease (yes/no), time since first cancer diagnosis (in months), treatment (radiation/chemotherapy/radiation and chemotherapy/other e. g. hormone therapy), treatment status (current: yes/no), medication influencing the mood (yes/no), contact with other psychosocial support functions e. g. psychiatrist, psychologist etc. (yes/no).
Statistical Methods
A detailed description of the sample size calculation is available in the study protocol.
For the descriptive analyses, frequencies, means, and standard deviations were calculated. To answer the main research question, linear mixed model analyses (LMM) were conducted with the primary and secondary outcome measures. Baseline differences between IG and CG were calculated with chi2-tests and independent t-tests, where the effect sizes phi (Φ) and Cohen’s d were reported. The analyses were adjusted for covariates to account for baseline differences between the IG and CG. The adjusted means and 95% confidence intervals were reported for each group at all time points. Furthermore, Cohen’s d effect sizes were calculated for variations from t0 to t1 and t0 to t2 for each group separately by dividing the mean difference between time points for each group by the pooled standard deviation of the two time points for each of the two groups [31, 32]. For the Cohen’s d effect sizes of the interaction between group and time, the effect sizes of the time point differences for IG and CG were subtracted [31, 32]. Missing data was imputed at baseline with the expectation maximization method at score level. Missing data at t1 and t2 was estimated by the LMM. Participants with no data at t1 and t2 were regarded as lost to follow-up. To account for possible differences between participants lost to follow-up and the study population as well as IG and CG, chi2-tests, and univariate analyses of variance (ANOVA) were calculated. Effect sizes were reported as phi (Φ) and partial eta2, respectively. Furthermore, a sensitivity analysis was performed to allow for a comprehensive interpretation of the results and account for possible attrition bias. Hereby, the main analyses were repeated with the intention to treat (ITT) population, which constituted of the study population and lost to follow-up. For all analyses, an effect was considered significant at a two-sided p < 0.05. Regarding the exploratory analyses (e.g., evaluation of support) no imputation of missing data was performed and they were solely of descriptive nature.
All analyses were conducted using the IBM Statistical Package for Social Sciences (SPSS) software version 25 or higher.