Clinicians at MGH, DFCI, CONNECCS-affiliated sites, and external health care institutions could present and recommend the study to AYAs during a regular clinic visit. To facilitate provider support and referral, the study principal investigator (PI) presented the trial and solicited provider input at each of the Mass General Brigham and CONNECCS-affiliated sites. Interested patients either contacted the study clinical research coordinator (CRC) directly or provided permission to have the study CRC reach out to them.

The study CRC screened cancer survivors’ electronic health records (EHR) at MGH for demographic and clinical eligibility criteria. Following patient identification, the CRC requested permission from a member of the cancer care team (ie, oncologist, nurse practitioner) to approach the patient for study participation at their next scheduled clinic visit or pursue outreach via phone. DFCI study staff conducted a chart screen and transferred the names and contact information of potentially eligible participants to MGH using a secure study REDCap (Research Electronic Data Capture) database to allow for the CRC to assess eligibility and pursue outreach.

Recruitment flyers were distributed through social media; by providers at external, interested health care institutions and clinics, including the CONNECCS network; and at cancer and survivor-related conferences and organizations (eg, Stupid Cancer; Cancer Con).

The Bounce Back study appeared as a public page on the Mass General Brigham Rally Research Recruitment Portal, a tool that allows patients to express interest in ongoing research studies.

Social media advertisements were used to disseminate information about the Bounce Back study and to direct AYAs to contact the study team for more information. Both recruitment flyers and text posts were posted on a variety of social media outlets and forums including Reddit, Facebook, Instagram, and Twitter.

Feasibility metrics were modeled after resiliency studies led with survivors and other medical populations [36,37]. We evaluated program feasibility by examining several process variables, including rates of study eligibility (percent of individuals who were eligible), recruitment (number of eligible individuals who expressed interest in our study), enrollment (percent of eligible pool who consented and enrolled), and retention (percent of enrollees who completed the follow-up). Our primary measure of feasibility was determined by the proportion of patients who completed the program, defined as participating in 6 out of 8 sessions. We documented reasons for ineligibility and refusal as well as sociodemographic characteristics, medical history, and cancer characteristics of refusers.

Program acceptability was assessed at the postprogram data collection period (T1 for PG, T2 for CG) with 5 questions on an acceptability questionnaire rated on a 4-point Likert scale (1=not at all to 4=very much). Items prompted participants to rate the extent to which they found the Bounce Back program to be (1) enjoyable, (2) helpful, (3) applicable or relevant (ie, is it appropriate and applicable), (4) convenient (ie, in regard to delivery modality), and (5) likelihood of future use (eg, “Will you continue to use RR strategies in the future?”). Treatment satisfaction was assessed by items on the acceptability questionnaire, which asks participants to rate their level of satisfaction with the following items using a 4-point Likert scale (1=not at all satisfied to 4=very satisfied): (1) treatment structure, (2) treatment timing (ie, early survivorship period) and (3) treatment content. We qualitatively explored overall satisfaction by asking 3 open-ended questions regarding treatment likes, dislikes, and recommendations. Additional acceptability data are collected in the optional qualitative exit interview.

The Measure of Current Status Part A (MOCS-A) is a 13-item measure that assesses participants’ self-reported ability to deal with daily stresses. Composite scores range from 0 to 52, with higher scores demonstrating greater self-perceived confidence in handling daily stressors. The MOCS-A has 4 subscales that can be analyzed: relaxation, awareness of tension, assertiveness, and coping confidence [38].

Resilience was measured using 18 items from the Current Experience Scale (CES). The questionnaire reflects current self-perceived functioning in the domains of appreciation for life, adaptive perspectives, personal strength, spiritual connectedness, relating to others, and health behaviors. For each item, responses range from 0 (not at all) to 5 (a great deal). Composite scores range from 0 to 90, with higher scores indicating resiliency; greater scores on each of the 6 subscales indicate greater resiliency [39].

The visual analogue scale (VAS)-Stress is a 1-item scale asking individuals to rate their current level of stress. The VAS-Distress is a 1-item scale asking individuals to rate their current level of distress on a scale of 0 to 10. Higher scores on each scale indicate greater levels of the construct being measured [40].

Patient-Reported Outcomes Measurement Information (PROMIS) measures evaluate and monitor physical, mental, and social health in adults and children. The following subscales were utilized: PROMIS ED Anxiety – short form 4a, PROMIS ED depression – short form 4a, PROMIS ED anger – short form 5a, PROMIS fatigue - short form 7b, PROMIS sleep disturbance - short form 8a, PROMIS Social Isolation - short form 4a. PROMIS measures were scored by the HealthMeasures Scoring Service using response pattern scoring. PROMIS raw scores are converted into T-scores for each participant and compared to US population averages.

The Penn State Worry Questionnaire (PSWQ) is a 16-item measure used to assess worry. It is rated on a 5-point scale ranging from 1 (not at all typical of me) to 5 (very typical of me); select items are reverse scored. Total scores range from 16 to 80, with higher scores indicating greater worry [41].

The Intolerance of Uncertainty Scale (IUS-12) is a short version of the original 27-item Intolerance of Uncertainty Scale [42] that measures responses to uncertainty, ambiguous situations, and the future. The 12 items are rated on a 5-point Likert scale ranging from 1 (not at all characteristic of me) to 5 (entirely characteristic of me). The IUS-12 is scored on a scale from 12 to 60, with greater scores indicating greater intolerance of uncertainty. IUS prospective and inhibitory subscale scores will also be examined [43].

The Interpersonal Reactivity Index (IRI) perspective-taking subscale is a 7-item measure that assesses the tendency of an individual to take on the perspective of another in daily life. Items are rated on a scale of 0 (does not describe me well) to 4 (describes me very well). We used 6 of the 7 items in the subscale. Total scores range from 0 to 24, with higher scores indicating greater perspective-taking ability [44].

The Coping Self-Efficacy Scale (CSES) is a 26-item measure of self-perceived efficacy for coping with challenges and threats. Respondents are asked to rate their confidence performing adaptive coping behaviors (ie, talking positively to oneself) on a scale of 0 (Cannot do at all) to 10 (Certain can do). Scores range from 0 to 260, with higher scores indicating greater coping self-efficacy [45].

A single-item, investigator-developed measure was administered to assess frequency of self-guided RR exercise practice. Participants were asked to describe the frequency of their RR exercise (eg, mindfulness, guided imagery, deep breathing) practice on the following scale: Daily, A few times per week, Once or twice a month, or Never.

The Health Behavior Questionnaire is an investigator-developed questionnaire designed to assess habits related to substance use, exercise, and nutrition.

This measure was added mid-study to account for any COVID-19–related stressors that occurred during study participation that may have influenced prior survey responses. Participants were asked to report on their COVID-19–related concerns, COVID-19–related lifestyle changes (ie, diet, sleep), changes in stress level, changes in cancer-related concerns, and more broadly how the virus was impacting their life.

Participants were asked to provide hair samples to measure potential changes in cortisol (ie, a stress hormone). We found that hair cortisol sampling was feasible in a similar behavioral trial conducted with posttreatment lymphoma survivors [46]. Participants were instructed to provide 1 hair sample preprogram (T0 for PG, T1 for CG) and 1 sample at the end of the program (T1 for PG, T2 for CG). The CRC sent detailed hair sampling instructions and stamped, pre-addressed envelopes to facilitate returns. Participants were instructed to cut a small sample of hair (approximately 150 strands, about the diameter of a pencil eraser) from the back of their head, as close to the scalp as possible. They were asked to tie the strands near the scalp end, place the sample in aluminum foil, and mail back to the research team. The hair sampling instructions also included 6 questions about hair care, exercise, and glucocorticoid use, as these can affect hair cortisol measurements. Hair samples were not collected from participants who had taken glucocorticoid medications (eg, prednisone) within the past 3 months, as these medications can suppress endogenous cortisol levels or cause cortisol measurements to be inaccurate. We tracked the reasons why any hair samples were not collected to inform the feasibility and acceptability of hair cortisol collection and analysis for this population. We also collected feedback and perceptions of hair sampling measures at study completion. Prior to processing, samples remained wrapped in aluminum foil, labeled with a study ID, and stored at room temperature in a padded envelope. Hair samples were processed by Dr. Jerrold Meyer’s laboratory at the University of Massachusetts, Amherst using an ELISA assay kit.

Qualitative data collection for this trial is ongoing. A randomly selected subset of 20 participants will be invited to participate in one-on-one exit interviews after study completion. To ensure inclusivity and take into account the effects of the COVID-19 pandemic on the program delivery, the sample will be stratified based on the following characteristics: (1) gender, (2) race (ie, non-Hispanic White; survivors of color), and (3) time of study participation (ie, before, during, or after onset of COVID-19 pandemic). Exit interviews may be completed via Zoom videoconferencing to explore additional barriers or facilitators to study participation, treatment adherence, program engagement, and study completion. Participants will be asked more detailed information about perceptions of the treatment and preferences for further adaptation after having participated in the program. A series of questions will be asked about using social media outreach for future research recruitment. We will also ask participants to report on how COVID-19 may have impacted their stress levels or ability to participate in the program. These interviews will be audio-recorded and qualitatively analyzed for themes that will help to determine whether treatment modifications are needed in future work. It is estimated that the interviews will take approximately 45 minutes to complete. Participants will be informed that the qualitative exit interviews are an optional portion of the study but if completed, will result in additional compensation (US $30).

The PI monitored the safety of this trial and complied with reporting requirements. All adverse events were reported to the IRB within 24 hours. Study recruitment, enrollment, and retention were reviewed by the PI and CRC weekly. The PI’s mentor, co-mentors, consultants, and scientific advisors functioned as a Data Safety and Monitoring Board. This group convened on a semi-annual basis to monitor study participant safety and to review study progress and other study-related events (including, but not limited to, enrollment, recruitment, retention, and adverse events). During these meetings, any study-related concerns were reviewed, and as needed, an action plan was established. The outcome of these meetings and proposed action plans were summarized and distributed to all mentors, consultants, and scientific advisors. The PI and her team also met quarterly with collaborators within the CONNECCS network to review study progress, request referrals, and discuss other study-related activities and events. Study updates were summarized and distributed to all CONNECCS collaborators following the quarterly meetings.

We instructed participants to maintain the confidentiality of the group by not discussing anything that occurred in the group with anyone outside of the group. Group privacy and confidentiality were discussed at the first session and in the subsequent session to reinforce practice. Careful attention was taken during the informed consent process to explain the limits of confidentiality. Participants were advised to wear headphones and sit in a quiet place to protect their own and other group members’ privacy. All data and personal information created by this research study were stored in password-protected computer files accessible only to study staff and stored on a secure drive only accessible by members of the research team.

Consistent with best practices in treatment development, the aim of this pilot is to establish the feasibility and acceptability of a stress management and resiliency program for early posttreatment AYAs [37,47]. We consider 75% session completion rate (approximately 6 of 8 sessions) as a threshold for program completion. As such, we consider 60% of participants reaching the threshold to establish program feasibility. With a sample size of 60 participants, we would have 80% power to demonstrate a difference of 15% from our preset criterion with a one-sided significance level of .05. Therefore, we believe our sample size of 72, accounting for 10%-15% attrition based on prior trials [46], will be sufficient to answer our questions about feasibility and acceptability.

Descriptive statistics, including means, frequencies, and ranges, will be used to describe the sample and to summarize feasibility, acceptability, and program satisfaction. Feasibility outcomes will be assessed by determining the proportion of individuals who were recruited, screened, and enrolled in the study. Response frequencies will summarize reasons for ineligibility and refusal. We will also determine the proportion of enrolled participants who complete the program. Participants who complete at least 75% of the treatment sessions (6 of 8 sessions) will be identified as treatment completers. We will examine the proportion of individuals who attend each session. For acceptability, response frequencies will summarize quantitative feedback on the acceptability questionnaire. Together with qualitative feedback from the exit interviews, this information will be used to inform the feasibility and acceptability of the program.

Preliminary outcome data may be used to inform future assessment instruments and methods. We may also conduct exploratory hypothesis testing to examine preliminary changes in our proposed program targets (changes in psychosocial outcomes, including mindfulness, depressed mood, anxiety, and stress). A priori statistical tests of program-related changes will be planned for a future efficacy trial of this program. First, we will examine the frequency distributions of all variables. Potential variables of interest (eg, gender, history of RR practice) will be included as covariates if they are significantly correlated with each outcome of interest at P<.25. We will also compare the baseline characteristics of completers versus study noncompleters. The primary analysis will be a completer analysis limited to those with complete data, and we will conduct a sensitivity analysis using multiple imputation for missing data. For our exploratory psychosocial outcomes, we will examine between-group differences in change scores from enrollment (T0) to T1 (posttreatment for PG, 3 months postenrollment/baseline #2 for CG). To further explore preliminary efficacy, we will evaluate within-group changes from pre- to postprogram (using T0 to T1 data for the PG and T1 to T2 data for the CG) for each condition separately and then for both groups combined. Finally, within the PG only, we will explore potential maintenance of program benefits with a repeated measures analysis of variance (ANOVA), including the 3 survey time points. Exit interviews will be audio-recorded and transcribed; NVIVO software will be utilized in the thematic analysis, which will be led by members of the study staff under the mentorship of the study PI. Coders will meet on a weekly basis to discuss the coding framework, categories, and coding plan. To ensure coding reliability, coding discrepancies will be resolved through discussion and comparison of raw data. Coding will continue until a high level of reliability (kappa ≥0.80) is established.

We will examine the feasibility, acceptability, and preliminary effects of collecting hair samples to examine changes in stress reactivity. Feasibility metrics for the hair sampling include hair return rates. For measures of acceptability, response frequencies will summarize quantitative feedback from the acceptability questionnaire about the acceptability of hair collection procedures. Hair cortisol samples will be analyzed in a laboratory, and group differences in hair cortisol concentration at T1 will be examined using independent samples t tests. Pearson correlation or Spearman rank correlation will examine the association of hair cortisol concentration with each of our psychological outcomes, controlling for potential confounders

We will assess whether the mechanism of missing data is missing at random. We will explore differences between study completers and noncompleters on participant demographic and other relevant variables to inform the next phase of this trial. We will perform sensitivity analysis using (1) a completer analysis limited to those who have complete data and (2) multiple imputations for missing data [48].