Participant flow
Figure 1 shows the participant flow, enrollment, intervention, and assessment schedule.
Recruitment
Oncologists in the core cancer hospitals will introduce this research to subjects by giving them a card containing a QR code linked to the research homepage. Those who are interested in participating in the research will apply for participation from the homepage, and an email requesting participation will be sent to the secretariat. We will start sampling at Nagoya City University Hospital and then extend recruitment to the other core cancer hospitals in Japan. In the second year of the study, we will seek collaboration from several other core cancer hospitals/clinics in Japan.
Research managing system
This is a fully decentralized individually randomized, parallel-group multicenter trial that the subjects could participate into the trial without involving in-person contact which is similar to our previous study40 (Figure 3). The study’s website (http://smileagain-project.org/) provides further information.
The website explains the purpose of the study and the eligibility criteria and methods used; it also features a video that briefly introduces the study as well as provides full written information about it. The clinical research coordinators (CRCs) at the central office will ascertain their eligibility by telephone.
Electronic informed consent and randomization at week 0
After screening, the CRCs will seek to obtain the subjects’ electronic informed consent (e-consent) via the e-PRO system at week 0. Participants will be asked to upload a picture of their identification materials (patients will be especially encouraged to attach a photo of the ID card from the hospital where they made regular follow-up visits for cancer). This e-consent procedure is in accordance with the guidance of the US Food and Drug Administration. The original informed consent material is shown in the online supplementary appendix.
After providing e-consent, completing the baseline investigation by the e-PRO system and
and completing the PE component, the participants will be randomly allocated to one of the combinations using the electronic data capturing (EDC) web program at the data management center (Figure 3). The random allocation will therefore be concealed.
If participants are allocated to the intervention group, which does not include all three components, they will be informed that the use of the remaining components can be resumed, if they wish, after week 24.
Trial period: Weeks 0-8
The research nurses will send weekly emails for eight weeks, encouraging the participants’ adherence to the application and asking them to record their responses on the e-PRO system. The research team can evaluate a patient’s progress using the application (e.g., the number of times and duration of use for each application) on the server.
Follow-up period: Weeks 8–24
At Week 24, the participants will receive an email encouraging them to provide their responses on the e-PRO system.
Termination assessment
If participants withdraw consent for assessment, they will not be followed up. Subjects will be excluded from the intention-to-treat (ITT) cohort of the trial if their characteristics are found to meet any exclusion criteria at baseline (e.g., aged under 20 years) after participation.
Assignment of interventions
Randomization and allocation concealment
The allocation will be stratified by the baseline PHQ-9 score (9 or less vs. 10 or more). We will use permuted block randomization to ensure balance in the number of subjects allocated to each combination because in the fully factorial design, imbalance in the allocated numbers among the factors would reduce the statistical power of the study. An independent data center will generate computer-generated random allocation sequences, which are maintained centrally, and the results of the assignment will determine which components are presented to the participants on their smartphone.
Blinding
Neither the participants nor the study personnel will be blinded to the intervention that each participant is receiving through the conduct of the trial. The assessment of all the primary and secondary outcomes will be self-reported by the participants and therefore not blinded. The statisticians will be blinded to the allocation through the statistical analyses by analyzing the datasets prepared by the study personnel in which all components are denoted only by a letter.
Plans to promote participant retention and complete follow-up
The outcome data will be collected even when a participant has not completed the allocated intervention unless he or she has expressed a wish to completely withdraw from the study. When a participant fails to fill in the scheduled assessments, reminders will be sent in 24 hours and 48 hours via an automated popup on the smartphone. If participants fail to provide their responses regarding the PHQ-9, a research nurse blinded to the assignment will send an e-mail to the subjects to elicit their answers. The participants will receive modest compensation for the time it requires to complete the questionnaire: 2000 yen when they complete the week 0 assessment, 1000 yen when they complete the week 4 assessment, 1000 yen when they complete the week 8 assessment, and 1000 yen for completing the week 24 assessment.
Data Management
A secure, web-based, password-protected database will be used to manage recruitment, eligibility assessments, randomization, scheduling and tracking, baseline and follow-up assessments, and the delivery of the allocated interventions. All the assessment data will be checked automatically for integrity by this platform. The security of the data transfer between the application and the server will be guaranteed through the Secure Socket Layer (SSL).
Statistical analysis and power calculation
Primary analyses
We will use mixed-effects repeated-measures analysis to estimate the mean difference in change scores on the PHQ-9 for each component. The model will include random effects for subjects (intercepts) and fixed effects of treatment (main effects and second-order interaction effects of the three components), visit (as categorical), and treatment-by-visit interactions, adjusted for age, sex, and baseline PHQ-9 scores. Each of the experimental factors will be coded at two levels (presence coded as on and absence coded as off) using effect coding. The primary outcome is expected to change for PHQ-9 scores from baseline to week 8. The estimated mean differences will be converted into standardized mean differences using the root if the variance is taken from the covariance matrix. No adjustment for multiple testing will be applied in the estimation of the statistical significance of the main and interaction effects in this model, and the conventional threshold for statistical significance (p < 0.05, two-sided) will be used because of the following: in the optimization phase of the MOST framework41, the emphasis is on making a decision about what components will make up the optimized intervention, and factorial designs usually evaluate multiple, completely different interventions that could have been assessed in separate trials and, conventionally, the multiple hypotheses have been tested independently for these trial designs42 43.
Secondary analysis
The secondary analysis will use the same models as those used in the primary analysis. Mixed-effects repeated-measures analysis will be used to estimate the mean differences in changed scores on the GAD-7 questionnaire for each component. The secondary outcome is expected to change for GAD-7 scores from baseline to week 8.
Interim analysis
We do not plan any interim analysis.
Sample size estimation
In the smartphone psychotherapy intervention used in this study, CBT, including behavioral activation and cognitive restructuring, showed a statistically significant effect among 164 depressive patients, with an effect size of 0.3–0.4. In this study, to determine an effect size of 0.3, for each intervention factor (problem-solving skills, behavioral activation, assertive communication) and its interaction at an alpha = 0.05 and a beta = 0.20, we need a total sample size of 350. We estimated the effect size as 0.3 because problem-solving skills or behavioral activation had the same effect as full package CBT in a recent comprehensive meta-analysis44. To achieve this, the sample size needs to be 22 for each of the 16 combinations, and the total sample size is therefore 16 × 22 =352.
Although we expect a dropout of approximately 20%, we will use mixed-effects repeated measures analysis.
As a study with more than five measurement points is known to require 30–50% fewer participants for the same power compared to a pre-to-post-only assessment, we set the final target sample size of our study conservatively, as estimated above.
If the case accumulation is smoother than expected, we will continue the case accumulation aiming at a sample of 787, which will enable an effect size of 0.2.
Monitoring
Data monitoring
Data integrity will be monitored centrally, first through the built-in data check system on the server program and second by the data management center on a monthly basis. The data management center will prepare the monthly summary (the number of participants entering the study, the number of participants completing the study, and serious adverse events) to be presented. Since the psychological intervention provided by the application is not invasive or expected to cause serious harm, a data monitoring committee will not be organized.
Harm
No specific or serious adverse events are expected for the participants who use this smartphone app. However, using it might lead to psychological distress in some participants, depending on their psychological state. We evaluated these potential adverse events by qualitative evaluation of the intervention in a previous study.
Compensation
Our previous trials suggest that a little harm occurred in this trial. However, if any health hazards occur, they will be covered by the National Health Insurance system.
Auditing
Because the intervention can be classified as a “minimally invasive intervention”, no formal auditing will be conducted.
Ethics and dissemination
Research ethics approval
The protocol was approved by the Institutional Review Board of Nagoya City University (ID:46-20-0005) and will be approved by the ethics committees of other collaborating universities.
Protocol amendments
If important protocol modifications are needed, the investigators will discuss and report them to the Institutional Review Board of Nagoya City University for approval. Once approved, they will be reported to all the study investigators and, when necessary, to the study participants. Further approval of the ethics committees of the participating universities will also be sought.
Confidentiality
Each participant will receive an identification number. All records will be managed using the identification numbers. The security of the data transfer between the application and the server will be guaranteed by SSL, and the data will be stored on the secure server. The data management center will download the data regularly and store them using a medium that is not connected to the internet and is kept in a locked drawer. Once the trial is completed, the data on the server will be erased permanently. The medium storing the downloaded data will be kept in the locked drawer in the student health center for 10 years after the publication of the primary findings.
Access to data
All members of the steering committee will have full access to the final trial dataset.
Ancillary and posttrial care
All participants will receive the standard care, except specialized cognitive behavioral therapy, provided by the participating hospitals and corresponding facilities both throughout and after the study.
Dissemination policy
The protocol paper and the study results will be submitted to peer-reviewed journals. The first author of the main paper will be a member of the steering committee (authors of the protocol paper). If approved by the steering committee, another person could be the first author. The list of coauthors will be determined before submitting each paper. The main and relevant findings will be presented at conferences.
Patient and public involvement statement
The study protocol was designed with a patient (a breast cancer survivor) who participated in the study as a researcher. She appropriately discussed the protocol with other patients when a patient’s preferences and/or opinions were considered. She will play the same role in implementing the study. Therefore, patients have been and will always be included in the study. The results of the study will be made public on the study homepage.