Pilot testing the SUCCESS peer mentoring program for students with concussion: the role of personas in mobile technology development

Abstract Purpose College students with concussion experience academic, cognitive and psychosocial challenges, yet frequently lack supports necessary for successful reintegration into school. Success in College after Concussion with Effective Student Supports (SUCCESS) is a virtual peer mentoring program designed to provide education, support and connection through a mobile application. The purpose of this study was to describe use of personas as components of mobile app development and conduct preliminary testing of SUCCESS using personas. Methods Personas were developed from case studies and portrayed by college students trained as fictitious mentees. Mentors were blinded to use of personas. Eleven mentors completed measures pre and post a 4-week mentoring cycle. Mentors and personas interacted in the app via chat, video calls and sharing of educational materials. Measures included the Post-Concussion Symptom Scale (PCSS); PROMIS Self-Efficacy; Depression, Anxiety and Stress Scale (DASS); and a series of focus groups. Results Mentors suggested improvements to resolve instability of video calls, expand educational materials to address psychosocial functioning, and add structure to the mentoring relationship. Some preferences around communication, like groups chats and emoji keyboards, were not able to be addressed. As expected, PCSS scores were stable. DASS score (p = .04), especially depression (p = .03), decreased. PROMIS scores showed a trend towards growth (p = .057), although were not statistically significant. Conclusions Use of personas allowed technical challenges and program refinements to be addressed before including students with subacute concussion in testing. Although continued development will address enhancement of communication modalities preferred by students, future efficacy testing of SUCCESS is warranted. IMPLICATIONS FOR REHABILITATION Personas can be useful for the development of mobile applications, particularly those that include interaction between users, allowing for testing of complex usage scenarios to improve app functionality. Students who have recovered from concussion may benefit from providing peer mentoring to students with more acute injuries. Mentoring programs should have a clear structure that considers relationship closure as much as relationship building. Mobile apps can deliver just-in-time prompts that support both of these activities. Postsecondary students express a preference for augmenting written app-based communication with digital images such as emojis and gifs, considering these important towards relationship-building in virtual spaces.


Introduction
Estimates of the annual incidence of traumatic brain injury (TBI) in the United States range from 2.8 to over 3.4 million people.Mild traumatic brain injuries (mTBIs), also called concussion, account for approximately 80% of TBIs [1].Reports of previous history of concussion in college students range from 16 [2] to 28% [3], with the incidence of concussion highest in young adults [4].In addition, around one in 80 college students experience a head injury each year, often from non-sport-related mechanisms, such as falls [5].Students may experience lingering impacts on executive function [6], as well as acute symptoms that limit overall participation in academic and social life.Symptoms experienced can include physical effects, like headaches or light sensitivity; cognitive effects, like feeling slowed down or having poor recall; sleep effects, such as frequent fatigue or difficulty falling asleep; and emotional effects, including irritability, anxiety or depression.College students face fast-paced, high-pressure learning environments that put them at risk of rapidly falling behind [7], and may also have limited experience managing their own health care.

Concussion management and college students
Current guidelines for management of concussion suggest 24-48 h of rest, followed by a return to moderate levels of activity [8].For students, this means returning to the classroom about 2-3 days post injury, but with supports and accommodations in place to reduce workload to moderate levels [9].At the postsecondary level, concussion is a common injury in both athletes and non-athletes [5,10], although non-athletes tend to experience more symptoms and need more time to recover [10].For the majority of students who are not injured during varsity sports or school-sanctioned sports, supports are minimal [11].As healthcare entities on college campuses tend to operate independently from educational systems, students must initiate seeking both healthcare and academic support.
Timelines for symptom resolution are proposed to be about 30 days [12][13][14], but with variation depending on pre-injury factors such as a previous history of concussion, injury factors such as mechanism or severity of initial symptoms, and post-injury factors, including access to healthcare, adherence to management recommendations, and psychosocial support [15,16].As most symptoms resolve in 30 days, many students experience a temporary disability and therefore do not qualify for services under the Americans with Disabilities Act [17], limiting the utility of campus resources such as disability services.Instead, receiving accommodations to support recovery often depends on the student self-advocating for their needs and seeking support by contacting local resources [18].Modifiable post-injury factors, such as decreasing healthcare access wait times and increasing healthcare provider contacts, are linked to better outcomes [19].Thus, programs that focus on improving post-injury support are likely to benefit students with concussion.

Peer mentoring
Peer mentoring involves sharing knowledge and providing social support from a peer with more experience to a peer with less experience.Connecting people with disabilities to others who have had similar experiences provides a welcoming social environment.These relationships can also be a powerful educational experience, both during more didactic exchanges and by modelling successful re-entry to community participation [20].In fact, education delivered by peers has higher acceptance and adoption of behavioural change than that provided by health professionals [21].One recent randomized controlled trial of a peer-led concussion education program in college student athletes demonstrated that students are more receptive to education when it is provided by a peer [22], and other such studies are underway [23].Similarly, peer mentoring for disability management has also been shown to improve personal factors such as self-efficacy, coping, mood, behavioural control and adjustment to disability [21,[24][25][26][27][28].
Few programs have systematically evaluated the impact of peer support and mentoring following brain injury, but results from existing trials suggest similarly positive impacts to behavioural control, psychosocial processing of injury and disability, and mood [24].One previous study in high school students with concussion highlights the important role of social support in education and recovery [29].Participants in that study indicated that validation of shared experiences from a peer helped make sense of the recovery process and provided real life education in concussion management that could not always be provided by an adult.Similarly, college students may particularly appreciate a peer approach, as the developmental stage of early adulthood is marked by a strong desire for independence [30].Connecting students with acute or subacute concussion with other students who have successfully navigated management and re-entry to school could provide both an access point for education and local resources along with a role model who could relate to shared experiences.

Success in College after Concussion with Effective Student
Supports (SUCCESS) was designed to provide peer mentoring to college students with concussion through a mobile application.The program was developed using Patient-Centered Outcomes Research Institute (PCORI) guidelines [31,32] and with the input of stakeholders, including students with concussion, disability service providers, rehabilitation and medical professionals, university professors, peer mentoring leaders and parents of students with concussion [33].SUCCESS provides online screening and training of mentors, who must be at least 6 months post a medically diagnosed concussion that was sustained as a postsecondary student.The program is administered via an app designed to connect mentors and mentees through chats and videoconference calls, and also supports other mentoring activities like goal setting and sharing educational materials (see Table 1 for a summary of app features and Figure 1 or http://www.peersuccess.org/the-app.htmlfor more screenshots of app functionality).In addition, the app includes a matching algorithm that allows mentors and mentees to specify characteristics important to a successful match, and for administrators to prioritize matching based on those selections (e.g., university, major program of study, mechanism of injury, etc.).SUCCESS educational handouts uploaded to the app centre on four core topics:  2, and http://www.peersuccess.org/education.htmlto access the handouts) [34].The intervention is designed to move through four-week cycles, with each of the core educational topics being addressed during mentoring meetings each week.Mentor and mentee pairs can then request to extend the program for an additional four-week cycle or to close out the mentoring relationship.

Personas and technology development
Personas are fictional characters created based on research in order to represent different user types when developing and testing new products [35].The use of personas can be employed as a facet of the user-centred design process [36,37] to help deepen the understanding of user needs, behaviours and preferences in earlier stages of development [38,39].Personas have long been used to support development of a variety of product types, including automobiles [40] and everyday consumer technologies [41], and in recent years have been applied to the design and engineering of healthcare and assistive technology products [42][43][44].The use of personas may be especially helpful for uncovering usability and accessibility challenges [41], and, may be a safe and effective means of engaging vulnerable end users in early-stage healthcare related product development activities [39].
Although traditionally personas are descriptions of user archetypes that allow developers to anticipate user barriers rather than live actors, use of standardized patients who are trained to present as a person with a particular disease process or complaint is becoming more common in medical and rehabilitation training programs [40,41], including telehealth settings [42].Here, to allow for testing between dyads, research assistants were trained to serve as personas, interacting in the app with real mentors.These mentee personas were used in the development and testing of the SUCCESS peer mentoring program to allow technical challenges and program refinements of the SUCCESS intervention and mobile app to be addressed before including students currently recovering from concussion in testing.Furthermore, the use of personas facilitated pilot testing of the mobile app even as recruitment was limited by the global SARS-Cov-2 pandemic and reduced healthcare use during 2020 lockdowns.During this time period, the SUCCESS program similarly shifted the design from what was planned to be a combined face-to-face and virtual mentoring experience to a fully virtual environment, and several modifications were made to the program app, requiring additional user testing to ensure the program and the app would be ready for use with recently injured students.As with many outpatient facilities, the specialty concussion clinic from which participants were to be recruited temporarily closed.Even after reopening, evaluations of new clients were drastically reduced to meet social distancing and other mitigation guidelines.Persona testing allowed mentor user experiences to improve the app and program before enrolling students in more acute stages of concussion recovery.
As part of our development process to prepare the mobile app and program for delivery to mentee students with acute or subacute concussion, the purpose of the current study was twofold: (1) to describe app development and testing, including use of personas to create an authentic test environment to examine mentor perceptions of onboarding and preparedness to participate in the intervention and (2) to present results of preliminary testing of SUCCESS delivered via a mobile app with mentor participants who had been paired with trained personas.We present feedback from a series of focus groups and user surveys that provided recommendations for app and program modifications to prepare SUCCESS for future efficacy testing with newly injured students with concussion, along with results of mentor experiences with concussion symptoms, academics and psychosocial health pre-and post-intervention.

Participants
Participants were recruited through patients who had completed care at a specialty concussion clinic.Inclusion criteria were that mentor participants must: have sustained a concussion while in postsecondary studies, be at least 6 months post-concussion, be currently enrolled or within 1 year post graduation from undergraduate or graduate study, have been diagnosed with concussion by a medical professional, be over the age of 18, and speak  and read English fluently.In addition, exclusion criteria were that participants could not have a history of moderate or severe TBI or other neurological disorders.In addition, participants could not have a history of mental health condition that was not able to not be managed with outpatient care.Chart review identified candidates who met the inclusion and exclusion criteria, then each candidate was reviewed with a clinical care coordinator to confirm eligibility and that there had not been a known change in status since discharge that might impact eligibility.In addition, in preparation for study activities, protocols were established to assist any participants experiencing acute distress, as well as a mechanism to alert the clinical care coordinator if elevated levels of mental health complaints were present on the self-report measures, or if the participant reported such complaints elsewhere, such as in focus groups or direct contact with project staff.This clinical care coordinator could then review records to determine if the participant was receiving care either through the clinic or elsewhere (e.g., with a campus provider, which would have been noted in the medical chart at discharge) and elevate the concern to the clinical team as need be.
Recruitment materials were sent by email, and additional screening was conducted during the initial consent interview.All mentors confirmed that they had been discharged from care for concussion (although some had ongoing contact with behavioural health professionals due to pre-existing or co-occurring mental health conditions), their symptoms had resolved, and they had successfully returned to school.In a companion study, we found that many were uncomfortable with using the terminology of "recovered", as they viewed themselves as having been changed by the concussion, although in the long term this was largely viewed as being positive [45].Mentors received a stipend following completion of training and provision of program feedback, and again following completion of each mentoring experience.All human research activities were conducted with the approval and oversight of Shepherd Center's Institutional Review Board.
Eleven participants enrolled and completed pre-and postmeasures, as well as online training, to become a SUCCESS mentor.Participant characteristics are included in Table 3. Participants were primarily female (8/11) undergraduate students with a mean age of 21.64.Most had sustained more than one concussion (6/11), with their most recent concussion occurring on average 19.8 months prior to study enrolment.The most common mechanism of injury was sport/recreational activities, followed by hit to the head.

Procedure
Mentor participants first completed pre-intervention measures (see below), including a semi-structured interview to confirm and expand injury and demographic characteristics.Next, participants completed a 2-h self-paced online training in mentoring that included content aimed at understanding concussion, understanding peer coping, building communication skills, encouraging selfadvocacy and knowing when and how to access the SUCCESS team for support.This training included open-ended reflective questions as well as closed-ended questions.Trained research assistants reviewed open-ended responses to ensure that mentors were able to adequately situate themselves in the care of a mentee with concussion, focusing on the mentee's needs above their own and recommending the use of medical and other services when appropriate.In addition, mentors were required to correctly answer at least 80% of closed-ended questions to advance to the second training module on the use of the app.This second training took approximately 20 min to complete and walked the mentor through downloading and installation of the app using program codes supplied by the research team, as well as completion of their app profile.The training then explained use of communication and mentoring features included in the app, and showed mentors where educational handouts covered in the first training were located (see O'Brien et al. [46] for more training details, including complete outlines of training topics).Mentors were then notified that they had been placed in the mentoring pool and would receive both an email and a push notification from the app when an appropriate mentee match was assigned to them.Additionally, mentors each attended at least one monthly focus group during which they responded to openended questions to provide feedback on program components, including the app, educational materials and training.

Personas
Five different mentee personas were created based on student interviews and focus groups conducted during discovery of needs activities that informed initial development of the SUCCESS program [33].Each mentor participant was paired with a mentee persona portrayed by a trained undergraduate research assistant.Details were provided to the research assistants portraying personas, such as major program of study, mechanism of injury that resulted in concussion, current symptoms and responsiveness to mentoring intervention (see Table 4 for persona example).
Mentors enrolled in the study were told that their mentee may have been a persona rather than an actual college student recovering from concussion.

The app
The SUCCESS program is housed in a 508 compliant app hosted by Qooper, a subscription-based platform that allows customization of their mentoring app to meet specific program goals and branding.Program administrators use an app admin portal to make minor customizations, upload video and written materials, post and schedule announcements, provide program participants with access to the app, connect mentors with mentees, and monitor participants' app usage metrics (see Table 1).Prior to the SUCCESS program, the Qooper app had been used primarily in corporate settings to support vocational mentoring, and it had not previously been used to support peer mentoring in people with disabilities or recovering from a medical condition.Initially, the SUCCESS study team met with a Qooper project manager several times to provide insight into the anticipated needs of SUCCESS program participants and to establish a mechanism for collaborating to address app functionality modification recommendations made by SUCCESS program participants.This resulted in the completion of several non-standard customizations to the app (e.g., removal of job skills and expertise from the user profile, addition of concussion symptoms and concussion story to the user profile) prior to testing the SUCCESS program with mentors and mentee personas.

Research and development activities
In this embedded mixed method design [QUAL(quan)] [47], we used a simple pre-post research design for quantitative measurement of mentor program feedback as well as concussion symptoms and psychosocial experiences to confirm that involvement in the mentoring program did not have an iatrogenic effect on mentors and to inform any related qualitative findings from focus groups.Development activities to shape the intervention focused on qualitative data that drove an iterative user-centred design approach [36,37], so that as mentors made recommendations for the app or program, those were implemented as the app developers or research team addressed each one.At least one focus group addressed changes following implementation to allow for ongoing feedback around acceptability of solutions.For recommendations not able to be addressed during the study period (i.e., app modifications not yet able to be supported by current coding architecture), subsequent focus groups instead focused on elaboration of needs and possible solutions for future continued development.

Qualitative
Focus groups.A series of seven focus groups were held over the course of six months, so all mentors had the opportunity to participate in at least one focus group.Most participated in at least three (at the beginning of mentoring, during mentoring and after mentoring).The first (KO) and last (TW) authors conducted all focus groups using Zoom videoconferencing software.
Extensive feedback on the app and program, including the educational materials, was collected through a series of focus groups held monthly throughout pilot testing.Each of the 11 mentors attended at least one focus group and provided general information about the mentoring experience while answering specific questions about the program app, educational handouts and structure.The focus groups also served as an opportunity for mentors to meet each other and to receive support for mentoring from the SUCCESS team.Following each focus group, facilitators (KO and TW) provided a brief summary of mentor feedback to the research team.In team meetings, research assistants serving as persona mentees then provided additional context or feedback as related to the mentee view within the app or other related mentor feedback provided during focus groups (e.g., "Mentors are saying that they cannot see your calendar availability, but you can see theirs.Tell us how that has been working"."Mentors said notifications of incoming calls have been inconsistent; is that something you've noticed?").This feedback from research assistants was provided either in a written format or verbally and then summarized in written meeting notes.

Quantitative
User Feedback and Satisfaction Questionnaire.Following completion of mentor training, each mentor rated their confidence in using the following six key components of the SUCCESS program on a scale of 0-100, with 100 indicating feeling "very confident": (1) using the app; (2) meeting with their mentee; (3) communicating with their mentee; (4) using the educational handouts; (5) completing mentoring logs; and (6) knowing when and how to contact the SUCCESS team for support.After completing a mentoring cycle, the mentors then rated their perception of their success in using the same six components of the program on a scale of 0-100, with 100 indicating "very successful".
Similarly, following completion of training, mentors rated their likelihood of using the following five key app features with mentees on a scale of 0-100, with 100 indicating "definitely will use": (1) chat; (2) video meetings; (3) goal setting; (4) educational handouts; and (5) group discussions.Following completion of a mentoring cycle, the mentors then rated how often they used the five key app features on a scale of 0-100, with 100 indicating "very frequently".Mentors also provided program satisfaction feedback by answering 15 additional closed and open-ended questions about the usefulness of program materials, what they liked most and least about the program, what they suggested be changed, and how likely they would be to recommend the program to a friend.
Post-Concussion Symptom Scale (PCSS) [48].The PCSS is a 22item self-report measure of concussion symptom presence and severity.Participants rate symptoms on a scale of 0-6, with higher ratings representing greater severity of symptom presentation.
Patient Reported Outcome Measurement Information System General Self-Efficacy (PROMIS-GSE) [49].The PROMIS-GSE is a 10item patient-reported measure of confidence in managing various situations, problems and events.Participants rate their degree of confidence in performing tasks using a five-point Likert scale.
Depression, Anxiety and Stress Scale (DASS 21) [50].The DASS 21 is a 21-item patient-reported measure of mental health.Participants rate how often they experience situations that indicate depression, anxiety and stress problems.

Qualitative
Focus groups were recorded and then transcribed using Otter.aisoftware.Subsequently, trained research assistants reviewed all transcripts to ensure precision of transcriptions and speaker assignment (and consistency in speaker assignments across focus group sessions).Each focus group began with researcher introductions, review of the purpose of the research study and focus group, collaborative setting of ground rules, and participant introductions.Interviews were conducted using a discussion guide that was presented both visually and orally to participants (see Supplemental Materials for complete focus group guides).Subsequent questioning was based on participant responses and focused particularly on generating solutions or suggestions for improvements.Following recommendations of Hurworth [43], a broad question initiated discussion ("What are your overall impressions so far of the SUCCESS Peer Mentoring Program?"), then moved through transitional narrowing questions (e.g., "Tell us about getting to know your Mentee and their needs") to key follow-up questions related to the app (e.g., "Tell us about your experience using the app") and program (e.g., "What resources have you used the most to help with mentoring?").No more than nine questions were planned in any focus group, allowing for follow-up questions around ideas introduced by participants.
Frequent summarization allowed participants to confirm that their responses had been accurately understood and reported by the research team [44].As participants attended multiple focus groups, member checking occurred both within a focus group and across sessions, and all solutions to app or program limitations were presented to participants for review.
During each focus group, the first (KO) and last (TW) authors took written field notes, marking common themes and points of congruence and divergence across participants.Following transcription, all transcripts were reviewed and coded following standard qualitative analytic procedures [51].As this study was designed to shape and improve the intervention package rather that make generalizable statements about the experience of providing mentoring after concussion, content analysis focused narrowly on summarizing common themes within the broad categories of feedback and recommendations related to (1) use of the mobile app and (2) the intervention program itself, including program materials.

Quantitative
Wilcoxon's signed-rank test was used to demonstrate pre-mentoring readiness and success in implementing SUCCESS components.The same analytical approach was used to assess the likelihood and actual usage of main app features, as well as to assess the level of change in concussion symptoms, general self-efficacy and mental health conditions pre-and post-mentoring.

App feedback
Overall, participants found the app intuitive to use and engaging.A common theme was that mentors appreciated having all program materials housed in the app, allowing them to go to a single source for everything related to SUCCESS.MR2 said, "Having everything on the app I think is really good because I know when I was concussed and I had to go to so many different places to like find things it was just very frustrating.And made my headaches worse.So, kind of having a lot of things in the one place is really good".MR4 agreed, saying, I do like the app.I like the way it's set up.I feel like it's very user friendly.And like intuitive of where to go to get your resources, where to go to check on your mentee, where to go to see your own settings, and stuff like that.I feel like it's very intuitive and I get I don't know if fun is the right word, but it encourages you to interact with it I guess with the way it's set up.I feel like it's not harsh or anything.So, I like the setup of it a lot from what I've experienced with it so far.
Participants also provided feedback on the app that revealed opportunities for enhancement or improvement (see Table 5 for summary).The most important and frequently described issues were around quality and stability of video call functionality.The mentors reported that calls frequently dropped or froze, had difficulty connecting, and did not notify users of incoming or missed video calls.
Next, most frequently cited were problems sorting through the educational handouts and assigning these to mentees.As MR8 explained: The worksheets that y'all made are awesome, I think there's a lot of information on them.I don't remember which is in which subgroup so I'm like okay core resources and then I look at all five and then like some of them have sub [folders], you click on them and then there's three more and then I'm like okay, that's not right, let me go back out.And then I'm like okay, where's the one that I'm looking for?So, if we could just have like all of them listed?I don't know maybe you could still have the categories in some way but that might be helpful.Or like a search function?
In the same vein, mentors reported that they wanted to prevent mentees from being overwhelmed by being able to recommend or send just one handout at a time, but were unable to find a way to link an educational handout in the chat function.
Also described as needing adjustment was the mentor and mentee meeting scheduler functionality.Both mentors and mentees could provide general availability windows to make scheduling easier.However, mentees were able to view mentor availability, but mentors could not view mentee availability when scheduling meetings.Instead, mentors wanted to be able to lead this task for mentees to reduce burden and effort on their part, and recommended that that both mentors and mentees be able to view one another's availability for meetings.Similarly, mentors pointed out that goals were currently structured so that mentors assigned goals to mentees, rather than mentees being able to establish and enter their own goals.Although all mentors were happy to help their mentee enter goals, they felt that mentees would rather have autonomy over this important aspect of their recovery and participation in the program.
In less pressing functional concerns, but which would likely drive app engagement, mentors expressed strong preference for having access to a variety of keyboards to enhance communication, including providing access to emojis and gifs.All mentors stated that they used the chat feature most often and that another important feature might be the ability to set up group chats so that mentees could benefit from the general conversation even though they may not always have questions of their own or know what to ask.Although the app does include group discussions, these are more traditional discussion boards in which replies are posted to specific queries.Such formats were associated with school and formal learning, whereas mentors described group chats as more informal and likely to encourage participation; in fact, participation in group discussions was poor even with frequent seeding of discussion topics by the research team.In addition, mentors did not describe any screen sensitivities in using the app, but recommended that the ability to enable dark mode would be an important feature for acutely injured mentees.Lastly, mentors reported wanting enhanced functionality of meeting logs.Meeting logs were posted as links to brief surveys so that mentors could report any concerns to the research team and note that a meeting had occurred.However, mentors reported that they would prefer to have meeting logs as part of their dashboard so that they could make notes during a meeting and then review them before subsequent meetings.

Feedback about SUCCESS
The most prominent program feedback expressed by mentors in focus groups was concerns about ending the mentoring relationship as a mentee's symptoms resolved.Mentors reported that it would seem awkward and abrupt to end the relationship entirely and on a particular date.Instead, mentors described that genuine friendships and closeness were likely to have formed over the weeks of mentoring, and that perhaps a period of time when mentees and mentors remained connected in the program app, but with different expectations about contact and mentoring activities might allow for a more gradual and natural transition.As MR6 explained, "We don't have to meet every week but just kind of having someone there if you have just a simple question you need to text them on the chat or something.Like, that might be nice just so you're not totally cutting them off you know?" Mentors also provided feedback on educational handouts they reviewed with mentees.Several mentors expressed that there was too much information in the Concussion Basics handout, and one suggested adding an index or directory listing of all the handout topics to support ease of use.Mentors also recommended creating additional content on active recovery, coping and advocacy.For example, MR1 shared, "Returning to school also involves how to interact with classmates, peers and professors, how to talk about your injury in ways that are comfortable for you, and let others know what is going on so that they can support you to the best of their ability", emphasizing the need to add content on advocating with people who play various different roles in a student's life.MR6's comments on managing mental health needs after concussion were also representative of the discussion, and mentors describing the need for more educational content in this key domain: I think maybe some more emotional and mental support handouts might be good.Cause I know for me, specifically I mean I'm pretty much an open book but I struggled a lot with my mental health with my concussion just because … I feel like my physical state affected how I felt mentally, and I got frustrated which in turn made me feel even worse physically.So, it was just like this big ugly cycle.So, maybe some ways to kind of cope with frustration or stress or whatever it is just because I know that can make you feel worse.Some comments bridged both the app and the program, explaining how delivery through the app or in a virtual space could be enhanced.MR11 explained how using screenshares and the training videos might anchor conversations with her mentee, and would allow them to develop solutions together.
I felt bad sending my mentee like a ton of handouts, so I just had to pick specific ones, even though a lot of them could be applicable to her situation.So maybe if we put forth more of an effort to be like hey, there's this video or, like, if we do Zoom, share a screen and just watch Although not directly commenting on recommendations, the mentors spoke at length throughout the focus groups about feelings of capability and appreciation for giving back to others as a result of mentoring, and that discussing a path through injury with students currently recovering served to remind them of their own progress.Mentors described how this allowed them to view themselves more compassionately, and to accept that they may have long-term consequences from their injury, or that recovery took much longer than they had anticipated, and yet they still could now appreciate all of the progress that had been made.

App modifications
Of greatest concern to mentors was the considerable difficulty with the video calls, including poor stability and inconsistent availability or notification of calls.The app developer responded by updating the video call function to use either a Zoom or Microsoft Teams plugin.Mentors reported that this improved the call quality and functionality overall, but there was a learning curve to ensure that all mentors were using the updated app and the plugins rather than the older version.Mentors also expressed appreciation that both of these videoconferencing platforms allow screen sharing so that mentors could share educational handouts and other resources directly with mentees during their meetings.
In addition, the app development team created a "Steps" curriculum feature in which handouts could be assigned and push notifications sent to mentors and mentees to prompt specific mentoring activities on a pre-programmed schedule.For example, during week 1, mentors received guidance around "Getting to Know Each Other" and sharing information from the Concussion Basics handout.In week 2, a notification prompted mentors to review the Success in School Part 1: Advocacy, Accommodations, and Resources handout with mentees.And in week 4, mentors received the guidance that their mentee may be ready to reduce the level of support provided and move to following status (as described in the closure related program modifications elaborated below).These program "Steps" were customized to both mentors and mentees, providing structure to each of the weekly mentoring activities, including both initiating and concluding the mentoring relationship.Figure 2 provides an example of a "Steps" push notification.
Also added to the app was a "Mentoring Agreement", a form in which mentors and mentees specified the expectations for the relationship during their first meeting.Specifically, the form guided mentoring pairs to discuss the anticipated length of the relationship and responsibilities of each party.This agreement also provided a reminder to mentors to explain to mentees that while conversations were confidential, concerns about the mentee's health or safety would be shared with the SUCCESS program administrators.The mentoring agreement was set to a four-week default length, but the mentoring relationship could be shortened or extended at any time and the agreement modified.This process allowed pairs to discuss the larger flow of the program early, so that weekly activities and the eventual closure of the relationship could be anticipated more easily.Some recommendations made by mentors, such as implementing "dark mode" and adding emoji keyboards, were unable to be addressed during the study period (see Table 5).Future development can be targeted towards these identified needs.

Program modifications
Revisions were also made to the structure of the training and mentoring program to provide additional support for the closure of mentoring in response to mentor concerns.The primary change entailed further structuring SUCCESS mentoring to occur in two distinct phases: an active phase and a following phase.The active phase is the initial mentoring period previously outlined as part of the SUCCESS program, when mentors and mentees are expected to meet and chat regularly, and discuss the SUCCESS educational handouts.This period is expected to last about four weeks, and the training now provides guidance on how to determine whether the mentee is ready to move to the next phase at that time or whether greater than four weeks in the active phase may be beneficial for a particular mentee.When ready, the mentoring relationship then transitions to the following phase.During this phase, the mentor and mentee stay connected in the SUCCESS app and can chat or meet at any time, but are no longer expected to do so regularly.Once a mentee moves to the following phase, their mentor may be placed back in the mentor pool and matched with a new mentee.
Several changes were also made to the educational handouts in response to mentor feedback.Although already addressed in mentor training, additional support was added to the Mentoring & Communication Strategies handout to initiate and conclude relationships, including specific conversational prompts.Concussion Basics was modified to reduce the amount of information provided overall, and an index was added to list and hyperlink to other program handouts.The Return to Life handout was renamed Return to Activity, and information on active recovery was added to the handout that was previously included in Concussion Basics.
Five additional handouts were created: Healthy Coping, Self-Advocacy, Concussion Pitfalls, Concussion Myths and What I Want You to Know About Concussion handouts.Healthy Coping covers information on emotion-focused and problem-focused coping.Self-Advocacy covers two models for communicating needs and provides examples for use of each.Concussion Pitfalls covers common actions people take during their recovery period that have the potential to negatively impact their recovery.Concussion Myths aims to help clarify common misinformation about concussion, with content based on misinformation SUCCESS mentors reported encountering during recovery.What I Want You to Know About Concussion provides basic information on concussion and concussion recovery.It is formatted as a fillable pdf so that mentees can personalize the information in the handout to communicate about their needs and experience related to their concussion.The completed handout can then be shared with family, friends, professors and others to support self-advocacy.
Additional program materials were developed and made available via the program app.First, the mentor training was transformed into an interactive manual, hosted by Qualtrics, but directly linked within the program app.A table of contents was added to help mentors efficiently locate supportive training resources.Later, given that several mentors responded positively to the inclusion of educational video testimonials in the training, a selection of those videos was made available within the app for both mentors and mentees to access.The videos contain brief testimonials of real students who have recovered from concussion talking about their recovery or needs, advice from their experience, how to find strategies that work, and how following recommendations can be hard (e.g., https://www.youtube.com/watch?v= L5Ol_XKeTp8).Mentors were later invited to record additional testimonials to add to the video library within the app.

Training and App functionality
Pre-training, mentors rated their confidence and likelihood in using key components of the SUCCESS program and app.Mentors reported high confidence in using the app, meeting with their mentee, communicating with their mentee, using the educational handouts, completing meeting logs, and contacting the SUCCESS Team, demonstrating the greatest confidence in communicating with their mentee and the least in using the educational handouts (see Figure 3).Several mentors shared that the videos were what they liked most about the training, while one mentor commented that the videos were what they liked least (because it slowed the pace of the training).Other feedback provided by the mentors included wanting "to know a bit more about best ways to start off our first meeting with our mentee" (MR5), as well as wanting more support to address the emotional aspect of concussion and mentee coping.One of the mentors indicated interest in accessing the training manual for ongoing support, while MR4 stated, "Walk through was great, I just need to take the time to look through some of the resources so I am better prepared to know what my mentee may need in the future!" On average, mentors sent 53 messages and mentees sent 49 messages during mentoring.However, the number of messages exchanged varied widely between mentor-mentee pairs, from a minimum of 6 to a maximum of 112.After completion of their first mentoring cycle, there was a significant change only in app usage (V ¼ 2, p ¼ .030);mentors reported feeling less successful using the app, likely due to some of the technical challenges that were discovered.Apart from using the chat feature in the app (V ¼ 0, p ¼ .036),all other ratings of app features decreased from pre to post mentoring, indicating that mentors used those features less than they had expected prior to mentoring (see Figure 4).In contrast, mentors reported feeling very successful implementing most program components (see Figure 3).On a scale of 0-100, mentors indicated they would be very likely to recommend SUCCESS to a mentee (M ¼ 95.91, SD ¼ 8.87) and to a mentor (M ¼ 92.73, SD ¼ 10.30).

Concussion and psychosocial measures
Out of the 11 mentors, seven showed low baselines of concussion symptoms, ranging from 0 to 3. As expected, PCSS total (V ¼ 28, p ¼ .18)and subscale scores were stable, indicating that providing mentoring did not exacerbate previously resolved concussion symptoms.PROMIS scores showed a trend towards an improvement (V ¼ 6, p ¼ .057),with seven participants increasing scores from pre to post, two stable, and two slightly decreasing (see Figure 5).DASS scores decreased significantly (V ¼ 48, p ¼ .041),indicating improvement in mental health.However, significance appeared to be driven by one of the mentors who showed high emotional symptoms at baseline that decreased after the intervention.Removal of that participant resulted in change becoming non-significant (V ¼ 38, p ¼ .07).Even so, eight of the remaining 10 mentors demonstrated decreases in pre to post total DASS scores.Post hoc analysis of DASS subscale scores showed similar significant changes in depressive symptoms (V ¼ 21, p ¼ .03)that became non-significant after removal of the possible outlier (V ¼ 15, p ¼ .05),while there was no change in anxiety (V ¼ 11, p ¼ .41)or stress (V ¼ 27.5, p ¼ .20)levels reported by mentors (see Figure 6).

Discussion
The SUCCESS Peer Mentoring Program was designed to connect acutely concussed college students with students who had successfully returned to learn following concussion, allowing mentors to connect mentees to education, resources and support through a mobile app.The program was developed based on stakeholder input [33], and the current report describes testing using mentee personas rather than acutely injured students to ensure that the app and program were appropriate for use with a symptomatic group of users.Findings in the current study support the use of personas for pilot testing to refine both the program and the app.These recommendations for improvements included adding group chats to enhance meaningful discussion with mentees, improve video call stability with enhanced features to more easily share educational handouts, refine educational handouts and increase information on specific resources, add dark mode accessibility features, and enhanced education on mentoring techniques.
Mentors anticipated they would use more features in the app following training than were actually accessed during the trial, with the notable exception of the chat feature.As the chat features aligned with smartphone texting functionality and texting tends to be employed by college students much more frequently than other forms of smartphone communication [52,53], it is perhaps unsurprising that this was accurately rated as very likely to be used during mentoring.There was considerable variability in chatting frequency and number, and while the current study design was unable to assess the role this may have played in   Following training, mentors rated likelihood of using key app features with mentees on a scale of 0-100, with 100 indicating feeling "definitely will use".Following completion of a mentoring cycle, mentors rated how often they used key app features on a scale of 0-100, with 100 indicating "very frequently".outcomes, future work may need to consider chat behaviours as moderators of outcomes (perhaps as a proxy for intervention dosage).As high rates of texting are associated with greater psychological distress [54] and sleep problems [55], continued monitoring of the chat in future trials may be important to identify students at risk for complicated recovery.
The most pressing issue that mentors described related to the app was the quality and stability of the video calls.Video and audio quality are important moderators of uptake of telehealth interventions for both healthcare providers and users [56,57], with higher quality improving interactions and reducing attrition from care [58].Switching to plug-in features for use of existing video conferencing software was both a straightforward programming solution and one that the mentors found to be a relatively easy adjustment.In contrast, adding a dark mode to reduce screen sensitivity was not able to be addressed during the study period.Although recent evidence suggests that people with concussion are only susceptible to complicating effects on recovery from screentime use in the first few days following concussion [59], photophobia and poor screen tolerance are still fairly common complaints [60,61].Despite this, videoconferencing and teletherapy services provided to people with concussion have been shown to be as effective as in-person services [62,63], or even favoured for ease of access [64,65].Regardless, dark mode has become an increasingly popular choice for both mobile and laptop platforms and would likely reflect preferences of many people with concussion.
Similarly, students expressed a preference for keyboards that allowed the use of emojis and gifs.Both have become common in the lexicon of texting and social media [66] as a way to supplement written linguistic text with supralinguistic features provided in face-to-face verbal communication through tone, prosody, facial expression and gesture [67].Emojis allow for specification of emotional intent or softening of requests or feedback [68].Although college students generally judge emojis as "unprofessional" when coming from an institution [69], they are instead viewed as important for building trust and relationships between peers [69,70].This further supports the role of using peers as compared to authorities to promote concussion reporting in student-athletes [22].Further development of the SUCCESS app should include addressing these student recommendations around preferred communication supplements.
Regarding the SUCCESS program itself, the mentors were concerned about feeling prepared to close out a mentoring relationship, requesting much more guidance and structure.These concerns regarding closure were similar to those expressed by peer befrienders (i.e., mentors) in Northcott et al.'s [71] description of peer befriending of people with post-stroke aphasia.In that study, mentors described relationship closure as easier when the mentee had other supports in place, there was a possibility of future meetings, and when the befriender had kept in mind that the relationship had a time limit.The SUCCESS program additions addressed closure needs in a similar manner by developing a step-down model towards ending the mentoring period, with pairs moving from active (weekly contact) to following (contact at request), which kept open the possibility of future meetings, but reduced expectations around maintenance of the pairing.The mentoring agreement provided in the app was modified to clearly spell out the nature of the relationship and that it was time limited, allowing mentors to clearly discuss expectations at the outset so that they could be assured that relationship closure would not be unexpected to mentees.In addition, the "Steps" curriculum feature was used to provide guidance for both mentors and mentees as to when a mentee may be ready for following.Feedback from mentors indicated a need for more mental health resources which resulted in the development of additional educational handouts, underscoring the importance of engaging end users in design of health products and services to ensure person-centeredness, acceptability and feasibility [36,[72][73][74].Testing with real mentees may identify a need to develop other mental health resources, given the known impact of depression, anxiety and stress on mTBI outcomes [75][76][77] as well as the challenges in meeting needs of patients with a history of mTBI and mental health comorbidities reported by first-line healthcare providers [78].Further, future testing of SUCCESS with real mentees may uncover benefits of virtual, app-delivered peer support as a mental health intervention in this population, as peer mentoring has been shown to lead to improvements in mood and coping in other populations of people with developmental and acquired disabilities [24,28,79,80].
In preparatory work, we found that students with a history of concussion felt they may benefit from participating as a peer mentor by reflecting on their progress, giving to others, and growing in their confidence in their own ability to overcome challenges [33].Other studies investigating peer mentoring in college students and in neurorehabilitation have examined the impact of such programming [23], as well as shown benefits to peer mentors include psychosocial growth, enhanced mood and quality of life, and improved self-efficacy [81,82].While our quantitative results did not show significant improvements in mentor psychosocial functioning, it also appears that providing mentoring did not exacerbate previously resolved concussion symptoms or increase mental health complaints during the study period.In addition, while we did not undertake a qualitative approach designed to generate generalizable knowledge about the experience of providing mentoring after concussion, statements from focus groups aligned with individual improvements in psychosocial functioning, particularly self-efficacy.These findings suggest that changes in psychosocial function should be measured in future trials to determine if mentors may benefit from the intervention alongside their peer mentees, as has been observed in other studies of peer mentoring in neurorehabilitation and chronic disease management [21,[83][84][85][86][87][88][89].

Limitations
Although the use of personas helped to improve the acceptability of the program by mentors, the application and program design have not yet incorporated the input of mentees.As such, further revisions are anticipated following future testing with mentees.Although mentors did not show a return of previously resolved symptoms but did show some benefits after providing mentorship, these results have limited generalizability due to the small number of participants.It is also possible that mentor experiences will differ when paired with real mentees, and as such, the benefits and drawbacks to program participation may differ with real mentees as well.Further, the study design did not include a control condition and, therefore, the demonstrated benefits cannot definitively be ascribed to participation in the peer mentoring program.This study also included primarily mentors who identified as female.Future larger studies should include a more diverse cohort of mentors and mentees.

Future directions
Future studies should include testing with mentees, resulting in additional app and program revisions based on mentee feedback.
Once mentee input has been incorporated, a larger controlled trial should be completed that includes mentors and mentees with diverse demographic and injury characteristics.Finally, future directions may also include comparative effectiveness trials to further understand the impact of mode of delivery (e.g., in person vs. virtual) and efficacy of particular aspects of the program itself (e.g., specific educational handouts or mentor training components).Other adaptations to the technology that would maximize use of the mobile app delivery might also include further development of the communication monitoring capabilities towards supporting high quality mentoring relationships and screening for possible issues.Specifically, this development might target text mining to monitor chat interactions for words, phrases or frequency of communication that can be automatically flagged for evaluation by the study team (e.g., words or phrases related to self or other harm might automatically trigger an alert to the team, rather than relying on the mentor to report or waiting for the team to review transcripts).In addition, more access to mentor and mentee communication exchanges would also allow for evaluation of communication effectiveness as a component of response to intervention.

Conclusions
Results support the feasibility of delivering SUCCESS via an app and indicate readiness of the program for testing with real mentees.The use of personas during pilot testing allowed technical challenges and program refinements to be addressed before including students with subacute concussion in testing.Incorporating personas as part of the user-centred design process resulted in improved quality of video calls and app features, optimization of educational materials, and improvements to the SUCCESS program design and supports for participants.The use of personas also helped confirm the feasibility of the program and ultimately led to readiness for testing with real mentees.Additionally, trends observed in mentors towards reduction of depressive symptoms and increases in self-efficacy suggest that future, larger trials should examine whether this app-based program might benefit mentor as well as mentee psychosocial health.Personas can serve an important role in ensuring technological solutions are appropriate for targeted populations of people with disabilities and can serve as a beneficial component of the user-centred design process.

Figure 1 .
Figure 1.Example of app chat feature.

Figure 2 .
Figure 2. Sample of mentoring steps.Note.Examples of mentee steps are provided on the left, and mentor steps on the right.Steps appear as push notifications on a timed schedule and then are stored in the steps section of the app for review.

Figure 3 .
Figure3.Pre-and post-mentoring ratings of SUCCESS program components.Note.Following training, mentors rated likelihood of using key program features with mentees on a scale of 0-100, with 100 indicating feeling "definitely will use".Following completion of a mentoring cycle, mentors rated how often they used key program features on a scale of 0-100, with 100 indicating "very frequently".

Figure 5 .
Figure 5. Change in symptoms, self-efficacy and mental health from pre to post mentoring.Note.Summary of group and individual results from pilot study of impact of SUCCESS on peer mentors.Boxes provide the minimum, first quartile, median, third quartile and maximum of the data set, and lines indicate positive or negative movement from pre to post measurement for each individual.For PCSS and DASS, lower scores indicate few symptoms, whereas higher scores indicate greater selfefficacy for PROMIS.

Figure 4 .
Figure 4. Pre-and post-mentoring ratings of app features.Note.Following training, mentors rated likelihood of using key app features with mentees on a scale of 0-100, with 100 indicating feeling "definitely will use".Following completion of a mentoring cycle, mentors rated how often they used key app features on a scale of 0-100, with 100 indicating "very frequently".

Figure 6 .
Figure 6.Change in DASS subscales from pre to post mentoring.Note.DASS Subscale scores with cut-offs for bands for possible indication of clinically significant mental health concerns indicated by dotted lines.
Concussion Basics; Success in School Part 1: Advocacy, Accommodations, and Resources; Success in School Part 2: Study Skills and Learning Strategies; and Returning to Activity (see Table

Table 3 .
Participant demographic and injury descriptors.

Table 4 .
Sample persona used for pilot testing.
Moderate -one semester long symptoms (decrease in symptoms every 2 weeks) Academic difficulties Don't realize when you are not paying attention until the professor has gone through 3-4 slides, it takes double to triple the amount of time to complete homework (go on social media, start doing other tasks, etc.), cannot focus while completing online classes this semester, hard to be in room to study, when the professor speaks or even a classmate, it sounds like they are the Charlie Brown teacher.a Gender, race and ethnicity were aligned with those of the trained research assistants.

Table 5 .
Recommended app features and modifications.itreal quick and be like, how do you feel about what they said?And then maybe instead of trying to ask like twenty questions with them or wait for them to kind of bring something up, maybe that will spark something.It might help us actually more or equally.