A psycho‐oncological online intervention supporting partners of patients with cancer (PartnerCARE): Results from a randomized controlled feasibility trial

Partners of cancer patients are the primary supporters and burdened at the same time. Support for partners is hitherto scarce and existing offers are rarely used. The PartnerCARE online intervention was specifically developed to address the caregiving partners' needs. This two‐arm parallel randomized controlled trial (RCT) evaluates the feasibility, acceptability, and potential efficacy of PartnerCARE.


| INTRODUCTION
It is well known that a cancer diagnosis also places a heavy burden on partners of cancer patients. 1 Partners not only often take over the patients' household tasks and provide financial and emotional support, they also share the anxieties associated with the diagnosis. 2,3 The increased household responsibilities, the personal emotional burden, the existential uncertainty due to a lifethreatening disease exhibit a substantial impact on social activities as well as relationships. This may increase the risk for depression, anxiety and various somatic symptoms in partners. 4 Despite these conditions, partners are often unaware of their own needs and prioritize the patients' needs. 2,5 Therefore, partners rarely seek support for themselves. Furthermore, healthcare systems often neglect burden of caregiving partners because focus is on the patients.
Low threshold support, such as online interventions may overcome this barrier and help to introduce partners to support offers. 6 Additionally, online interventions are advantageous in being much more flexible in time and location, easily and quickly accessible and anonymous if preferred. 5,7 So far, online support offers especially for caregiving partners are rare 8 and the field of research associated with it is still in its infancy. 9 Existing online offers for partners include interventions addressing couples 10 or spouses, 11-13 a smartphone app 14 and a self-management application. 15 None of them are available in German. Statements about the efficacy of these offers are difficult because they are largely in the development or feasibility testing stage. 11,[13][14][15] Likewise the comparison of these studies is difficult as they targeted different outcomes like quality of life, 11 psychological distress, 12 emotional well-being 13 and caregiver burden. 15 The PartnerCARE online intervention has been the first online intervention in Germany, which was especially developed to address the support needs of partners of cancer patients. PartnerCARE includes psychoeducational information, practical exercises like other interventions. 12,14 This is additionally combined with exemplary partners (via image and text) who accompany the participants through the intervention and relaxing guided imagery exercises with the intention to increase well-being and improve selfcare in partners (cf. for a detailed description of development process and content of PartnerCARE in the corresponding study protocol 16 ). The herein presented development and feasibility study of PartnerCARE is only the first step of a consecutive development and evaluation process.
One important aspect of our study is, that it observes both possible positive and possible negative effects of PartnerCARE. This follows the CONSORT-Guidelines 17 as well as recommendations from Rozental and Colleagues. 18 The present study aims to explore the feasibility and acceptance of the newly developed online intervention PartnerCARE. It specifically focusses on recruitment, dropout, adherence, and satisfaction with PartnerCARE. Moreover, possible negative effects of the online training are examined. In addition, first findings on the potential efficacy of the intervention are discussed. With this feasibility study a consecutive development and evaluation process was initiated to provide a reliable support offer for partners.

| METHODS
This randomized controlled trial with a two-arm parallel design compares an intervention group (IG) receiving the guided online intervention PartnerCARE with a waitlist control group (CG). Intervention description and results of this clinical trial are reported according to the CONSORT-Statement for feasibility RCTs 17 and the template for intervention description and replication. 19 The study design is described in detail in the corresponding study protocol. 16 The study was approved by the Ethics Committee of Ulm University

| Study population and recruitment
To overcome the challenging recruitment of partners of patients with cancer, 20 a wide variety of recruitment routes was used, both online (e.g. social media groups, online support groups, links on clinic homepages) and offline (e.g. flyers, cancer counseling centers). Participants were recruited in German-speaking countries from April 2019 to March 2020. Interested partners registered for the study via a contact form on the PartnerCARE study homepage (https://esano. klips-ulm.de/de/trainings/tumor/partnercare/) and subsequently received an email with detailed study information and a detailed consent form. After providing written informed consent, participants were invited to the online baseline assessment (T0). Before questionnaires started, the following inclusion criteria were checked via self-report: (1) age 18 years and older, (2) being in a relationship and (3) the partner is diagnosed with cancer (any type, any disease stage, any treatment stage). Additional inclusion criteria ((4) having Internet access and a corresponding terminal device, (5) providing an email address for contact and (6) giving a signed informed consent) were checked through the previous study process. If inclusion criteria were not fulfilled, baseline assessment was canceled and the participant received an automated email with recommendations for already available support.
Since this was a feasibility study, a formal sample size calculation was not performed. A sample size of 60 participants (30 per group) BODSCHWINNA ET AL. was the predefined target, based on recommendations for sample sizes in feasibility studies. 21

| Randomization
After completing baseline assessment, participants were randomly allocated to the study conditions IG and CG (1:1). For this, a permuted block randomization with randomly arranged block sizes of two and four was used. An independent researcher (TD) conducted the randomization using an automated generated randomization list via an online program (www.sealedenvelope.com). Therefore, allocation of participants was not able to be forecasted by any other member of the study team. After randomization, all participants received an email including information about group allocation and the resulting study procedure.

| Intervention
PartnerCARE is a guided individually psycho-oncological online intervention delivered on the online platform minddistrict (www. minddistrict.com). It aims to improve quality of life, self-efficacy, and caregiver burden in partners of patients with cancer by creating awareness of one's own needs and learning to meet them. The intervention is based on psychoeducation, cognitive behavioral therapy, supportive therapy and guided imagery elements, which have previously been proven to be beneficial concepts for cancer caregivers. 22 PartnerCARE consist of an introductory session, six weekly main sessions (topics: specific burdens, inner drivers, partnership communication, handling negative feelings, control and acceptance, path and goals), four optional additional sessions (topics: support of own children, healthy sleep, closeness and sexuality, existential burdens) and one booster session (2 weeks after the last main session). Detailed information about structure and content of the intervention is provided in Table S7 and in the corresponding study protocol. 16 Sessions were multimodally and interactively conceptualized through text, visual and auditory material, practical exercises, and homework assignments, as well as three exemplary partners. Each session has a duration of 30-60 min. In addition to the intervention, participants had the option of choosing an automated Short Message Service-(SMS) Coach, who sent two motivational text messages per week.
All participants of the IG received a personal account for the online platform minddistrict and were assigned to an e-coach (trained and supervised postgraduate psychology student and psychologists). After initial login, participants were able to start with the introductory session of the PartnerCARE intervention. Each completed session had to be sent to the e-coach who replied within two weekdays after session completion by a partly standardized and partly individualized feedback for the respective session.
Additionally, participants had the possibility to message the e-coach via minddistrict, for example, to report technical problems or to inform the e-coach in case they were not able to work on the session at the agreed time. The e-coach also reminded the participants when they missed a session (three times in intervals of two weekdays). All communication and feedback between participant and e-coach was asynchronous.
Participants of the CG received access to the unguided Part-nerCARE intervention after completion of the follow-up questionnaire (4 months after randomization). All participants had unrestricted access to standard care treatments.

| Outcomes
Self-report measurements took place at baseline (T0), post treatment (T1, 2 months after randomization) and at follow-up (T2, five "very much"). Perceived self-efficacy was rated on the Generalised Self-Efficacy scale. 35 The total score of the 10 items ranged from 4 to 40 and the scale reached excellent reliability (T0: Orientation to Problems Experienced) Inventory was used to measure coping. 36 The three subscales achieved up to acceptable reliability

| Data analyses
All statistical analyses were performed using SPSS 26 (IBM SPSS Statistics). All tests were two-tailed with a significance level of p ≤ 0.05. McDonald's omega (ω) was calculated to estimate the reliability of outcome scales. 37 To investigate significant differences at baseline between IG and CG as well as differences between intervention completers and non-completers, t-tests, Mann-Whitneytests and χ 2 tests were conducted.
All statistical analyses were conducted following the intentionto-treat principle (ITT). Additionally, per-protocol analyses were conducted based on participants who adhered to treatment protocol (=completers; completing at least 80% of the intervention, five out of six main sessions). Multivariate imputation was performed by fully conditional specification with Markov Chain Monte Carlo algorithm. 38 Missing data was assumed to be missing at random. 39 Predictive mean matching was applied as imputation method. 40 Based on recommendations m = 20 data sets were imputed. 39 The imputed data sets were each analyzed and results were pooled using Rubin's Rule. 41 Feasibility and acceptability outcomes were analyzed descriptively (participants' adherence, feedback from participants and ecoaches, side effects, attitude towards online interventions) and with t-test (satisfaction). Efficacy outcomes for testing a potential   Table 2).

| Attitude towards online interventions and satisfaction with PartnerCARE
Overall, participants of the study displayed a positive attitude to-

| Potential intervention effect on efficacy outcomes
The main effect of time was significant for some outcomes (   Table S6. In contrast, there was no statistically significant main effect of group (IG vs. CG) and no significant interaction for all efficacy outcomes. In PPanalyses similar effects were found with slight differences in effect sizes (Table S3).

| DISCUSSION
Findings from this feasibility trial indicate that the newly developed psycho-oncological online intervention PartnerCARE is both feasible and acceptable. Feasibility was particularly evident in high intervention completion rate, less drop out and choice of recruitment strategy.
The high intervention completion rate (73.3%) indicates that once participants begin the intervention, they are likely to finish it. This treatment adherence is comparable to other online intervention studies with caregivers, but substantially higher than average intervention adherence in online intervention studies with patients. 12,45,46 This rather high intervention adherence rate might result from the intervention being guided by e-coaches, 47  As highly recommended, our study assess possible negative side effects. 18 The advantage of this assessment is that one can only react to negative side effects when aware of them. Participants of Partner-CARE attributed a small number of the negative side effects to the intervention, which were mainly related to the category "intrapersonal change", for example, by reporting "feeling dependent on the online program" or "having longer periods of feeling bad". As the intervention encouraged participants to reflect on personal problems and conflicts, negative feelings can commonly increase initially with an increase of self-awareness. 52 This common fluctuation of symptoms is also indicated by the decrease of these negative side effects again from T1 to T2. All reports of suicidal ideation could be clarified by using our and showing them that they can effectively help themselves. For the IG these effects seem to be stable or even increase over time (like for distress, depression, mental quality of life, fear of progression), which may reflect a potentially delayed effect of the intervention. 22 In the pre-follow up comparison, also participants of the CG experienced some improvements over time. Since that effect only occurs at follow up, it may be, that through their repeated involvement with the topics of the questionnaires, they were encouraged to reflect and changed their behavior in the meantime. It should be emphasized, that all these results are only interpretable to a limited extent and a larger study sample is needed to make reliable statements about group differences as well as to verify the potential intervention effects.

| Study limitations
The first limitation of our study concerns the sample population. As has been seen in comparable other studies 12,20 the majority was female and highly educated, which does not reflect the population of partners in real-life. In addition, generalizability of the results is limited, because it can be assumed that only partners with an existing affinity for online tools have expressed their interest in the study via homepage. A reason for non-significant group effects may be the small sample size, which was however appropriate for the feasibility study. 21 A further limitation of the present online assessment is the uncertainty regarding the diagnosis of the ill spouse as this was recorded via self-report. The additional collection of medical data BODSCHWINNA ET AL.
(through a physician's report) is more costly and may increase the barrier for participation, but this should be weighed against possible certainty gains in future studies. Likewise, the different time periods of questionnaires (last 2 weeks, last 4 weeks, current) should be viewed critically and there is also uncertainty whether participants take these periods into account in their answers. In addition, although we recorded the number of participants undergoing psychotherapy treatment, subgroup analyses were not possible due to the small group size. Moreover, subgroup analyses with high and low burdened participants were not practicable, but it would be as well interesting for the upcoming efficacy study to investigate whether the level of burden has an influence on efficacy of the intervention.
As dropout rate in IG was higher than in CG, identification of reasons for not attending post-assessment or follow-up will be a future interest. One reason may be that participants from the waitlist-group were more motivated to complete assessments, because afterward they received access to the unguided version of PartnerCARE. Concerning the overall dropout, participants might also be very busy and prone to forgetting the questionnaire, which we tried to prevent with sending three reminders to complete the questionnaires. Further studies should consider asking for dropout reasons, thereby taking incentive strategies into account for receiving this important information from those not willing to further participate.

| Clinical implications
Partners of patients with cancer frequently experience low quality of life as well as depression and anxiety at a subsyndromal level. 4 Therefore, it is necessary to provide support offers for partners as an integral part of the healthcare routine. Online interventions may be a suitable offer to support partners in a flexible way. 6 But there are two barriers that must be overcome before partners use such support offers: partners' lack of awareness of their own needs 2 and how to reach burdened partners. 22

| CONCLUSION
The present study demonstrated feasibility and acceptance of the online intervention PartnerCARE and the corresponding study process. Online interventions can overcome barriers regarding the use of psychosocial health care offers and improve the care for partners of patients with cancer. 20 Based on the received feedback, PartnerCARE is currently being further developed to improve content, modify how content is displayed (in form of short videos), solve technical problems and enhance the audio quality of the guided imagery exercises. Subsequently, the promising preliminary results and implementation of PartnerCARE into standard care will be further investigated in a randomized controlled efficacy evaluation trial. The process was initiated to provide a reliable, long-lasting support offer for partners.

ACKNOWLEDGEMENT
The authors would like to thank all partners for their participation and their adjuvant written feedback. Furthermore, authors thank Sabrina Heil for her support in the study process and her and Daniel Neveling for e-coaching. The authors also thank Tim Dretzler (TD) for conducting the randomization of the study participants.