Video conference vs face‐to‐face group psychotherapy for distressed cancer survivors: A randomized controlled trial

This study assesses the effectiveness of face‐to‐face group positive psychotherapy for cancer survivors (PPC) compared to its online adaptation, online group positive psychotherapy for cancer survivors (OPPC), which is held via videoconference. A two‐arm, pragmatic randomized controlled trial was conducted to examine the effects of both interventions on emotional distress, post‐traumatic stress symptoms (PTSS), and post‐traumatic growth (PTG) among cancer survivors and analyze attrition to treatment.

1 | BACKGROUND Cancer incidence is expected to increase, as is the number of patients who are successfully treated. 1 However, higher survival rates do not necessarily imply greater well-being, since several studies show that cancer survivors have more functional limitations, risk of psychosocial problems and emotional suffering, work-related challenges, and fears about their health than noncancer patients. 2,3 All these consequences may lead to poorer quality of life (QoL), less adherence to oncological treatments, lower overall survival, via less adoption of healthy lifestyles, and self-care practices when it is not treated. [4][5][6] Although it has been shown that psycho-oncological interventions can improve psychological adjustment, health-related QoL, emotional distress, and prevent fear of recurrence, 7,8 access is far from universal. In addition to a shortage of psycho-oncologists in national health systems, several factors may limit availability, such as poor early detection, long waiting times and work, and mobility or time restrictions. 9 Manualized psycho-oncological treatments for emotional distress are structured as regular stage-oriented, face-to-face interventions.
They typically focus on the initial stages of cancer (ie, diagnosis and active oncological therapies), 10 palliative advanced phases, 11 or extended survivorship after treatment. 12 However, few studies have designed and tested specific interventions for early survivorship, at the end of oncological primary interventions and the return to everyday life. 13 Indeed, at this transition, chronic and delayed distress trajectories converge in a large number of survivors (30%-40%). 14 Consequently, this as a suitable moment to implement psychological treatments. 5,8,13 A study on positive psychotherapy for cancer survivors (PPC) 13,15 is one of the few that have analyzed and manualized psychological treatments to facilitate this transition. PPC is designed to offer accurate, proper psychosocial care combining stress management and emotional regulation (early-stage-oriented 10 ) with posttraumatic growth (PTG) facilitation via a meaning-making, existential approach (advanced-stage-oriented 11 ).Group PPC proposes that psychosocial treatment in cancer should be tailored and focus on the stress-growth balance. 13 Hence, the basic aim of PPC is to facilitate PTG and work on positive resources such as positive emotions, strengths, and personal meanings, to reduce emotional distress and post-traumatic stress symptoms (PTSS). The first clinical trial of face-to-face PPC groups vs treatment as usual and waiting lists showed promising preliminary results for its efficacy to reduce distress. 15 Similarly, a recent randomized controlled trial (RCT), 16 which compared PPC groups with cognitive-behavioral stress management groups, 10 also demonstrated better PPC results.
Considering the limited access to psycho-oncology services, and the broad availability of health-related resources on the internet, 17 online psychosocial treatments are a feasible, acceptable alternative that has the potential to solve some of aforementioned obstacles. 17,18 In psycho-oncology settings, treatments are normally focused on patients with high emotional distress and are preferably conducted through synchronous communication, although a wide range of options exists, usually through the adaptation of face-to-face treatments. 19,20 Some reviews have highlighted the advantages of online psychosocial resources, 17,18 including the possibility of overcoming geographical barriers, avoiding the interpersonal discomfort associated with attending face-to-face settings, or reducing the feeling of being overwhelmed that may be triggered by over-expression or interactions with therapists or other patients. However, disadvantages are also described, such as loss of nonverbal information 21 and high attrition rates. 21,22 Attrition has been underlined as a significant methodological problem in assessments of intervention effectiveness. It must be considered in all e-health trials to prevent reduction in their power. 21,22 Eysenbach 22 described two types of attrition processes: nonusage attrition (ie, low adherence), which describes the phenomenon of no longer using the application; and dropout attrition (ie, loss to follow-up or low retention), which refers to not completing the follow-up measures.
The introduction of psycho-oncological interventions through videoconferences is an example of contemporary solutions in practice, which offers most e-Health advantages while preventing communication limitations, increasing engagement, and even reporting slightly greater effectiveness than face-to-face intervention in outpatient settings. 20  This preregistered study (NCT03010371) included two consecutive clinical trials to compare (a) face-to-face PPC vs stress-management, 16 and (b) face-to-face PPC vs online PPC. The present paper covers the second trial, in which a two-arm, pragmatic RCT was conducted within the routine practice of public health centers. The pragmatic design was chosen for its potential to improve the internal-external validity balance, and its suitability within settings where strict randomization or concealment is not always clinically possible or acceptable. 26

| Participants
Women with a range of cancer diagnoses were recruited between January 2016 and January 2019. This was due to the high prevalence of high emotional distress between female breast cancer survivors, thus favoring group homogeneity. In order to use the PPC protocol in their validated population and guarantee group homogeneity of the sample, the trial was focused only on female participants. They were referred, in clinical real-life settings and according to routine criteria, by medical oncologists or nurses to the psycho-oncology unit if they presented emotional distress at the end of their primary treatment.
The psycho-oncologist carried out a face-to-face interview and patients were asked to answer an online sociodemographic questionnaire and clinical instruments (eg, Hospital Anxiety and Depression Scales [HADS] administration). If patients met the inclusion criteria, they were invited to participate in the study. Medical information was selected from medical histories with permission. Inclusion criteria were: (a) age ≥ 18 years, (b) primary oncological treatment (ie, surgery, chemotherapy, radiotherapy) completed, (c) disease-free or clinically stable, (d) ≥10 on the HADS total score, (e) access to high-speed internet, and (f) competence to understand and read Spanish. We excluded patients if they (a) reported any current severe mental disorders or (b) any major concurrent medical disease that seriously affected their cognitive performance. Participants were assessed at baseline (T0), immediately after treatment (T1), and 3 months after treatment (T2).
The study was conducted according to the latest version of the Declaration of Helsinki. Approval was given by our hospital's Ethics Committee (PR104/13) and all participants signed an informed consent form.

| Procedure
Participants were recruited at the health care centers of a comprehensive cancer network in Barcelona. Two clinical psychologists, experts in PPC, conducted or supervised the group processes.
According to our pragmatic RCT design, which was conducted within the Spanish national health system, we aimed to provide a service to as many users as possible. Therefore, a computer-generated randomization table with random separate allocation was prepared.
After assessment, the two modalities were described to the patients and they were encouraged to accept randomization. However, those who showed a strong preference for one of the two options were allocated to their treatment of choice, while all other participants were randomized. This decision entailed some risk of bias, but may control for baseline differences in patient and clinician motivation and outcome expectations. 27 Moreover, it was decided that any betweengroup difference in sociodemographic or clinical characteristics at baseline would be statistically controlled in the analyses. points has been used as a cut-off point to assess the clinical change. 30

| PTSS
The Posttraumatic Stress Disorder Checklist-Civilian Version (PCL-C) 31 is a self-report that covers the diagnostic criteria for posttraumatic stress disorder from the Diagnostic and Statistical Manual of Mental Disorders. 32 In the current sample, the total score of the PCL-C was used, which obtained good reliability (α = .90), parallel to the Spanish validation (α = .90) 33 where a recommended cut-off score of 44 is used to detect clinical case. 33

| Treatment integrity
Adherence to the protocol and therapists' competence was assessed with an ad hoc questionnaire adapted and summarized from the Revised Cognitive Therapy Scale (CTS-R). 36 The ad hoc questionnaire adaptation is a summary of its more clinically relevant dimensions: agenda (sequence of modules and tasks), conceptual integration, appropriate positive feedback, application of positive change methods, and homework tasks. However, given the ViTAM constraints, the other main difference was group size, with OPPC clusters formed of five to six patients. In all cases, the information flows were encrypted to protect confidentiality. 37

| Data analyses
Baseline differences between participants randomized, and those who preferred allocation to each treatment modality were examined with chi-squared and Student's t tests. The same analyses were conducted between both therapy groups. In addition, we estimated treatment integrity using the T index, which allows the use of ordinal scales. In this case, agreement was defined as identical scores on an item using a 7-point scale. Logistic regressions were performed to test differential adherence to treatment and retention to follow-up. Missingness at baseline and variables that differed between groups at baseline were controlled for (ie, age, education, and work status).
Intention-to-treat (ITT) analyses were performed using general linear mixed models (LMMs) to test the effect of interventions (ie, PPC and OPPC) on distress (HADS total score, and anxiety and depression subscales), post-traumatic stress (PCL-C), and posttraumatic growth (PTGI) overtime (T0, T1, and T2). Little's MCAR test indicated that data were missing completely at random (χ 2 [248] = 207.67, P = .97). Since MCAR can be assumed and since maximum likelihood was the estimation method, no missing data imputation was applied. Akaike's Information Criterion and Likelihood ratio test were used, respectively, in non-nested and nested models, to guide the modeling process. Visual inspection of residual plots did not reveal any obvious deviation from homoscedasticity or normality.
The modeling process began with the most meaningful model close to the null. In the study, this was the unconditional model with time as a linear fixed effect and the intercept as a random effect.

| Participant characteristics
A total of 289 individuals were assessed for eligibility (see the flow diagram in Figure 1). Before randomization, 44 participants wanted to be allocated either to the PPC group (n = 28) or the OPPC group  Table 1. Individuals in the PPC and the OPPC groups differed in terms of their mean age again, but also in their education level and work status, and in the mean on the HADS depression subscale and the total scores. In contrast, treatment integrity was similar between interventions, with a mean T index of 5.6 for PPC and of 5.7 for OPPC.

| Treatment attrition
Analyses did not reveal significant differences between interventions in attrition, either for participants' adherence during the intervention

| Effect of PPC and OPPC over time
Given the significant baseline differences that were found, the estimated treatment effects of the final LMMs on the dependent variables were adjusted for age, education, and work status.
The LMM of the effect of treatment on the HADS total score showed a significant variance in intercepts across participants (Var

| Study limitations
Results from this RCT should be interpreted cautiously since, as a pragmatic RCT, there are some limitations that need to be considered.
The respect for patient treatment preferences may have partly biased the results, although it also brings them closer to real-world clinical practice. This study does not include a nontreatment control group because we focused on comparing treatment delivery modalities.
However, this intervention has been compared to a waiting list control group in a previous study 15 and to other evidence-based interventions 16 with better results. Differences found between-group at baseline were statistically controlled in the analyses. Other limitations were the difference in the number of individuals in each group, as there were 10 to 12 patients in PPC and 5 to 6 in OPPC groups.
These differences may affect the results as group sizes, not just modality, could influence group evolution, alliance, or commitment.

| Clinical implications
This study supports OPPC, which was found to be an efficacious psycho-oncological treatment for female cancer survivors. Our promising results should encourage the extension and adaptation of OPPC to men, mixed groups, adolescents, or young adults, especially considering the significantly younger age of those participants who requested to receive OPPC in our study. Furthermore, future research may also focus on describing online group factors, 41 or whether combined face-to-face and online modes are more suitable and effective. 42

ACKNOWLEDGMENTS
We thank CERCA Programme Generalitat de Catalunya for institutional support. We would also like to thank all patients and health professionals for their willingness to participate in the study, for whom several dissemination activities to share trial results are expected. This study has been funded by Instituto de Salud Carlos III through the project (FIS PI15/01278) cofunded by European Regional Development Fund. ERDF, "a way to build Europe" //FONDOS FEDER "una