Feasibility of Medical Student Mentors to Improve Transition in Sickle Cell Disease

Abstract

Objective

Advances in medical care have resulted in nearly 95% of all children with sickle cell disease (SCD) living to adulthood. There is a lack of effective transition programming, contributing to high rates of mortality and morbidity among adolescents and young adults (AYAs) during the transition from pediatric to adult healthcare. This nonrandomized study evaluated the feasibility, acceptability, and preliminary outcomes of a novel medical student mentor intervention to improve transition outcomes for AYA with SCD.

Methods

Eligible participants were ages 18–25 years, either preparing for transition or had transferred to adult care within the past year. Twenty-four AYA with SCD (M_age = 20.3, SD = 2.6) enrolled in the program and were matched with a medical student mentor. Feasibility and acceptability of the intervention was assessed through enrollment rates, reasons for refusal, retention rates, engagement with the intervention, satisfaction, and reasons for drop-out. Dependent t-tests were used to evaluate the preliminary effects of the intervention on patient transition readiness, health-related quality of life, self-efficacy, SCD knowledge, medication adherence, and health literacy.

Results

Participants (N = 24) demonstrated adequate retention (75.0%), adherence to the intervention (M = 5.3 of 6 sessions), and satisfaction with the intervention components. Participants demonstrated significant improvements in transition readiness (p = .001), self-efficacy (p = .002), medication adherence (p = .02), and health literacy (p = .05).

Conclusions

A medical student mentor intervention to facilitate transition from pediatric to adult care for AYA with SCD is both feasible and acceptable to patients and medical students. Preliminary results suggest benefits for patients, warranting a larger efficacy study.

Introduction

Approximately 100,000 people in the United States have sickle cell disease (SCD), a red blood cell disorder accompanied by chronic complications leading to increased morbidity and early mortality (Hulihan et al., 2017). As a result of improvements in therapies and treatment, SCD has become a lifelong chronic illness rather than a disease of childhood in the United States, with over 95% of children with SCD living beyond 18 years (Quinn et al., 2010). Accordingly, the transition from pediatric to adult care has become an increasingly important clinical priority in this population. However, healthcare delivery models do not address the transition gap for the rising adult population with SCD. Inadequate transition care and low transition readiness contribute to poor health outcomes (Hankins et al., 2012; Lanzkron et al., 2013; Sobota et al., 2017).

Factors contributing to low transition readiness include poor knowledge of health history and future risks, low self-management skills and self-efficacy for self-management, negative expectations about adult care, concerns about relationships with new providers or reluctance to leave pediatric providers, poor communication with providers, and emotional concerns about transitioning (de Montalembert & Guitton, 2014; Kayle et al., 2016). Unaddressed, low transition readiness and poor self-efficacy in managing care can contribute to high rates of acute care encounters and rehospitalizations, as well as increased mortality for individuals with SCD who are transition age (Brousseau et al., 2010).

Transition programs have been developed to address the needs of adolescents and young adult (AYA) with SCD, but limited studies have reported on outcomes or feasibility of such programs (Allemang et al., 2016; Andemariam et al., 2014, Calhoun et al., 2019; Crosby et al., 2017; Hankins et al., 2012; Manwani et al., 2017, Melnick et al., 2017; Smith et al., 2011). These approaches included the use of health information technology (Crosby et al., 2017), patient education (Calhoun et al., 2019; Melnick et al., 2017; Smith et al., 2011), and patient navigators (Allemang et al., 2016; Manwani et al., 2017), but have been limited by small sample sizes, inconsistent outcome measurement, and lack of well-defined benchmarks for a successful transition. Program components including facility tours and tailored education have been found acceptable to AYAs with SCD (Hankins et al., 2012; Calhoun et al., 2019). Both electronic and in-person interventions also demonstrated high acceptability among participants. There is a noticeable gap in the literature in harnessing peer support to improve transition readiness. Peer support meets a developmental need of AYA (Bell et al., 2008; Domhardt et al., 2015) and offers a novel mechanism for targeting individual components of transition readiness. Peer support has been utilized in other populations of AYA with chronic illness and has demonstrated feasibility, acceptability, and efficacy in improving AYA self-management (Ahola Kohut et al., 2016; Stinson et al., 2016).

Although peer mentoring has demonstrated positive changes for a number of health behaviors (Petosa & Smith, 2014), a significant challenge is identifying mentors who have the time to commit to a mentoring relationship. This can be particularly challenging for youth with chronic illness like SCD due to the unpredictable nature of sickle cell crises. Amongst youth with chronic illness, other models have defined “peer” more broadly beyond those with the same condition. For example, medical students were used as near-peer mentors for AYA patients with a variety of chronic conditions (Conatser & Babcock, 1993; Schaechter & Canning, 1994; Tess et al., 1997). Medical students have a unique combination of being integrated into the healthcare system and being similarly-aged to the patient, allowing them to provide specialized support. Developing a mentoring relationship with someone very familiar with the healthcare system offers a unique opportunity to guide patients through complex processes like transition Through communication with patients (mentees) and specific training and communication with the clinical team, mentors may help patients improve transition readiness, increase patient knowledge, increase rates of attendance at adult visits, and improve medication adherence.

To address the gap of peer support interventions for transition in SCD, we developed SHIFT (Students Helping Individuals Facilitate Transition), a medical student mentor intervention for AYA preparing to transition to adult care (for details, see Viola et al., 2021). SHIFT aims to improve transition readiness by enhancing knowledge, self-efficacy, self-management, and communication skills (Viola et al., 2021). The goal of this study was to assess the feasibility and acceptability of SHIFT. A priori benchmarks for feasibility and acceptability were set as a recruitment rate of >50% of eligible individuals, a retention rate of >80%, and an adherence rate of at least four of six calls. We also sought to explore the effect of the intervention on transition readiness, health-related quality of life, self-efficacy, medication adherence, health literacy, and disease knowledge.

Materials and Methods

Procedures

This study was approved by the Institutional Review Board at an urban academic medical center in the Northeastern United States and performed in accordance with the ethical standards in the 1964 Declaration of Helsinki and its later amendments. Study procedures followed the CONSORT extension for pilot and feasibility trials (Supplementary File 1). Patients with SCD were eligible if they were at least 18 years old, either preparing for transition to adult care within the next 2 years (as determined by the clinical team) or had transferred to adult care within the past year, had no documented or self-reported cognitive deficits, and were English-speaking. All pre-transition patients meeting these criteria were identified by the clinical team at the pediatric sickle cell clinic. All patients who had transitioned within the past 12 months (prior to enrollment initiation) were identified by the clinical team at the adult hematology clinic. Potentially eligible patients were then approached for participation following their routine clinical visit between January and December 2019. If a patient was interested in participating, their eligibility was verified and informed consent was obtained. For identified eligible patients who did not have or did not show for an appointment during the recruitment period, recruitment was attempted via mail and phone, with up to three calls made per patient.

Following informed consent, patient participants completed the baseline assessment using Qualtrics, a secure online survey service. Patients who were recruited at their clinic appointment completed the survey while in clinic. Those recruited by phone completed the survey in their homes. The assessment was repeated at 6 months after enrollment (i.e., the expected time for intervention completion). Participants received a $25 gift card for each assessment completed (up to $50 total). Postintervention assessment took place between June 2019 and June 2020.

Medical students were recruited via email and advertisements at the medical school from November 2018 to May 2019 to meet the needs of mentee recruitment. Eligible candidates needed to be a current medical student in good standing and between the ages of 21 and 29 years old. Potential mentors completed an application including two letters of recommendation. The candidates were interviewed to evaluate their interpersonal skills, level of commitment to the program, and relevant experience. Following recruitment, mentors attended a 4-hr training, which included education, expectations for the program, ethics, and role-playing with constructive feedback.

Measures

Feasibility and Satisfaction

Feasibility was evaluated by recruitment, retention, and adherence to the intervention. Based on prior feasibility studies in this population (Crosby et al., 2017; Green et al., 2017), a priori feasibility benchmarks were set at a recruitment rate of >50% of eligible individuals, a retention rate of >80%, and an adherence rate of at least four of six calls.

Participants were asked to rate their satisfaction with the program, each of the monthly topics, the program duration, and the program format on a 5-point Likert scale from (Not at all) to (very). All participants also completed an exit interview to provide qualitative feedback on the program content, barriers to participation, and recommendations to improve the program.

Working Alliance Inventory- Short Form Revised

The mentor-mentee relationship was evaluated by a modified version of the Working Alliance Inventory-Short Form Revised (Hatcher & Gillaspy, 2006), which is a validated measure of the quality and strength of a therapeutic relationship. The inventory is comprised of twelve items in three domains (four items each): agreement about tasks, agreement about goals, and development of a bond. Question responses follow a 5-point Likert scale from (always) to (seldom), with possible scores ranging from 4 to 20.

Sickle Cell Transition Intervention Program- Readiness for Transition

The Transition Intervention Program- Readiness for Transition (Treadwell et al., 2016) includes 22 items assessing transition readiness in 4 domains: knowledge and skills in medical self-care (6 items), social support skills (4 items), independent living skills (8 items), and educational/vocational skills (4 items). The response format is a 5-point scale from (strongly agree) to (strongly disagree). Total scores are obtained by summing responses to the items, with higher scores indicating greater transition readiness (Treadwell et al., 2016). The overall scale demonstrated moderate internal consistency in this sample (α = .53–.66).

Adult Sickle Cell Quality of Life Measurement System

The Adult Sickle Cell Quality of Life Measurement System (ASCQ-Me) measure includes 25 items in 5 domains of quality of life: emotional impact, pain impact, sleep impact, social functioning impact, and stiffness impact (Keller et al., 2014). This measure yields a standardized T-score for each domain, with a mean of 50 and a standard deviation of 10. Higher scores indicate better health. T-scores more than 1 SD below the mean (i.e., <40) indicate clinically significant impairment in that domain (Keller et al., 2014). In this study, Cronbach’s alpha coefficients for all ASCQ-Me domains were adequate (α = .73–.93).

The Sickle Cell Self-Efficacy Scale

The Sickle Cell Self-Efficacy Scale (SCSES; Edwards et al., 2001) includes nine items related to patients’ perceptions of their ability to function on a day-to-day basis and to manage their SCD. Response choices range from (not at all sure) to (very sure). Total scores are obtained by summing responses, with high scores indicating higher self-efficacy (Edwards et al., 2001). In this study, the SCSES demonstrated acceptable internal consistency (α = .75–.81).

Morisky Medication Adherence Scale

The Morisky Medication Adherence Scale (MMAS-8) includes eight yes/no questions about adherence, such as forgetting, carelessness, reasons for stopping or starting, and situations related to adherence (Lam & Fresco, 2015; Tan et al., 2014). A total sum score of all items is computed and can range from 0 to 8. MMAS scores are categorized into three levels of adherence: high adherence (score = 8), medium adherence (score = 6 to <8), and low adherence (score < 6; Lam & Fresco, 2015; Tan et al., 2014). In this study, the MMAS-8 demonstrated acceptable internal consistency (α = .72–.78).

Health Literacy—The Newest Vital Sign

The Newest Vital Sign (Caldwell et al., 2018) evaluates health literacy and numeracy as well as the ability to locate and apply information. It is a six-item measure based on the ability to read and apply information from a nutrition label. Scores are calculated by receiving one point for each correct answer to yield a total sum. Scores are also categorized into three levels, including: high likelihood of limited literacy (score = 0–1), possibility of limited literacy (score = 2–3), and adequate literacy (score = 4–6; Caldwell et al., 2018). In this study, internal consistency ranged from α =.63–.76.

Transition Knowledge Questionnaire

This 24-item measure (Newland, 2008) assesses SCD knowledge in seven areas: (a) pathophysiology, (b) genetics, (c) physical manifestations, (d) treatment, (e) self-care, (f) psychosocial and developmental issues, and (g) healthcare delivery system. A total score is calculated as the sum of the number of correct answers, ranging from 0 to 24 (Newland, 2008). In this study, the Transition Knowledge Questionnaire demonstrated high internal consistency (α = .79–.81).

Intervention

SHIFT was developed following a formative needs assessment of the patient population and using the SMART framework (Viola et al., 2021). The 6-month program consisted of monthly medical student mentor video calls using Doxy.me, an encrypted HIPAA-compliant video conference system, supplemented by weekly text messages using WhatsApp, an encrypted text messaging service. Each patient participant was matched with one of nine trained medical student mentors based on gender and availability. Mentors made initial contact by sending a secure text message to their mentee to set up a time to video chat. The objective of the first video call was to establish rapport and discuss the mentee’s biggest challenges with regards to transition. Each month, the mentors addressed specific content related to transition in calls that lasted from 25 min to 1 hr (see Supplementary Table 1). In addition to the monthly video calls, mentors sent text messages to their mentee at least once a week to offer brief encouragement, support, or to answer any questions. This dose of communication was selected based on feedback from medical students and patients, as well as prior peer mentoring interventions (Ahola Kohut et al., 2016; Stinson et al., 2016). Mentors were required to audio record their video calls with their mentees (with patient consent) for use for fidelity checks (using structured checklists tailored to each of the six sessions), supervision, and content analysis. For the first mentee-pairing for each mentor, the first of the six education calls was reviewed, as well as 20% of the remaining calls for the pairing (∼1 additional call). Mentors were also required to log each time they sent a text message to their mentee, monitored by the study team. Mentor feedback was provided on (at least) a monthly basis by the PI. Mentors were paired with between one and three total patients.

Statistical Analyses

Statistical analyses were conducted using SPSS version 26. Feasibility and satisfaction were evaluated using descriptive analyses (means, SDs, and frequencies). Potential differences between study participants and nonparticipants (those who opted to not be in the study) were evaluated with t-tests and Chi-square analyses. Intervention effects were explored using dependent t-tests to assess the changes in outcomes between the two time points. Cohen’s d effect size was calculated, with an effect size of 0.2 considered small, 0.5 considered medium, and 0.8 considered large (Cohen, 1988). All analyses were assessed for skewness and kurtosis. Preliminary analyses identified an outlier in SCD knowledge score postintervention. The outlier was determined to be invalid based on marking the same response for all items, and was eliminated from the knowledge score analysis. There was no other missing data and all other responses were included in the analysis. Responses to the open-ended satisfaction questions and the exit interviews were thematically coded by two members of the research team, with any discrepancies resolved through discussion.

Results

Enrollment

Of a total of 44 potential participants identified, three were excluded due to cognitive impairment, and three were excluded due to transferring out of the local hospital system. Of the remaining eligible patients (N = 38), seven (19.4%) declined to participate (due to time constraints [N = 3], lack of interest [N = 2], or did not feel like it would benefit them [N = 2]) and seven (19.4%) did not respond to recruitment efforts. Twenty-four (63.2%) consented to participate in the study (21 were current pediatric patients and 3 had transferred to the adult clinic within the past year). The consort flow diagram is presented in Figure 1. There were no significant differences between participants and nonparticipants in age, race, sex, and hemoglobinopathy status (all ps > .05). Participant characteristics are presented in Table I. Nine medical students participated as mentors. The mentors were predominantly female (N = 7, 77.8%), Black (N = 5, 55.6%), and were a mean age of 23.8 years (SD = 2.0).

Figure 1.

Consort diagram.

Table I.

Patient Demographics as N (%) and M (SD)

Demographic	N (%) or M (SD)
Sex
Female	11 (45.8)
Male	13 (54.2)
Age	20.3 (2.6)
Hemoglobinopathy status
HbSS	18 (75.0)
HbSC	4 (16.7)
HbSβ+	2 (8.3)
Racial background
Black or African American	21 (87.5)
White	1 (4.2)
More than one race	2 (8.3)
Hispanic or Latino	3 (12.5)
School/employment status
Part-time student	2 (8.3)
Full-time student	13(54.2)
Working part-time	3 (12.5)
Working full-time	4 (16.7)
Unemployed	1(4.2)
Other	1 (4.2)
Highest grade completed
10th to 11th grade	3(12.5)
High school graduate	14 (58.3)
2-year college	3 (12.5)
4-year college	2 (8.3)
Graduate degree	2 (8.3)
Marital status
Single/never married	24 (100.0)
Current living situation
At home with parents (and siblings)	23 (95.8)
On own, with roommate(s) or partner	1 (4.2)
Health insurance statusa
Yes, through employer/school	2 (8.3)
Yes, through parents	10 (41.7)
Yes, covered through Medicaid or public assistance program	4 (16.7)
Yes, covered through another state program	3 (12.5)
Other	3 (12.5)

^aTwo participants did not respond to this question.

Retention and Adherence

Of the 24 enrolled, 18 completed the final assessment (75.0%). One participant completed all calls but did not complete the final assessment, whereas the remaining five were lost to follow-up. Participants completed an average of 5.3 out of 6 video calls (SD = 1.2, range = 0–6). Of note, one participant did not complete any of the video calls but communicated via text-message with his/her mentor for over 6 months. The most common time of dropout was between completing baseline and the first phone call.

Those who did not complete the study (i.e., did not complete the final assessment, N = 6) were younger (M = 18.3; SD = 0.6) than those who completed the study (M = 20.8; SD = 2.3; t [23] = −2.65, p = .01), but did not differ on other demographic variables. Those who did not complete the study demonstrated lower baseline health literacy (M = 1.2; SD = 1.0) than those who completed the study (M = 3.0; SD = 2.4; t [23] = 1.8, p = .001), but did not differ on any other outcome measure at baseline. Reasons for dropout included social issues (becoming homeless), loss to follow-up during COVID, and severe illness. No adverse events took place during the intervention period.

Satisfaction and Acceptability

Participant satisfaction is reported in Table II. Participants reported greatest satisfaction with the quality of the video chat calls with their mentor (M = 4.59, SD = 1.00), the content of the text messages with their mentor (M = 4.39, SD = 0.92), and the program overall (M = 4.33, SD = 0.84). When asked about topics that they found particularly helpful, healthy lifestyle (M = 4.82, SD = 0.95), differences between pediatric and adult care (M = 4.76, SD = 0.75), and self-advocacy (M = 4.47, SD = 1.18) emerged as the three highest-rated topics. Over 88% of patients reported that they would be quite a bit or very likely to recommend this program to other AYA who are preparing to transition.

Table II.

Mentee Satisfaction

How satisfied were you with the…	Not at all (%)	Slightly (%)	Somewhat (%)	Quite a bit (%)	Very (%)	M (SD)
Program overall	0	5.6	5.6	38.9	50.0	4.33 (0.84)
Content of mentee manual	0	5.6	16.7	33.3	44.4	4.17 (0.92)
Content of text messages	0	5.6	11.1	22.2	61.1	4.39 (0.92)
Frequency of text messages	0	11.1	16.7	16.7	55.6	4.17 (1.10)
Quality of video chat calls with your mentor	5.6a	0	0	16.7	72.2	4.59 (1.00)
Frequency of video chat calls	11.1a	5.6	0	27.8	50.0	4.12 (1.32)
Length of the program	11.1	11.1	0	27.8	44.4	4.00 (1.32)
Which topics did you find particularly helpful?
Goal setting for transition	0	16.7	5.6	16.7	61.2	4.28 (1.23)
Healthy lifestyle	0	5.9	11.8	17.6	61.2	4.82 (0.95)
Pain management	0	11.8	17.6	23.5	47.1	4.35 (1.22)
Medication adherence strategies	0	11.8	23.5	23.5	41.2	4.24 (1.25)
Information about SCD	0	5.9	29.4	29.4	25.3	4.24 (1.15)
Physician communication	0	5.9	11.8	29.4	52.9	4.41 (1.00)
Self-advocacy	0	5.9	23.5	11.8	58.8	4.47 (1.18)
Differences between pediatric and adult care	0	0	11.8	23.5	76.5	4.76 (0.75)
How likely are you to recommend this program for other young adults who are preparing to transition?	0	0	11.1	22.2	66.7	4.56 (0.70)
How much did the program help you improve your own management of your healthcare?	0	11.1	16.7	55.6	16.7	3.78 (0.88)
How much did the program improve your confidence for transition to healthcare?	5.6	0	16.7	44.4	33.3	4.00 (1.03)

^aParticipant who did not complete any of the six calls.

Responses to the open-ended item about their favorite part of the intervention indicated two themes: education and social support (see Table III). Fourteen participants (77.8%) reported that they valued the relationship they developed with their mentor. In response to the open-ended item in what ways they changed through participation in the program, three themes emerged: self-management skills, self-confidence, and self-advocacy. Suggested improvements for the program included increasing the frequency of the video calls, adding a group component where the mentees could interact with other AYA with SCD, and having the mentor sessions occur in person during the mentee’s clinic visit.

Table III.

Themes of Mentee Responses to Open-Ended Questions

Favorite part of the program	Example quotes
Education	“Being informed on things I did not know about sickle cell”
	“I learned how to speak up and advocate in a respectful way”
	“I got to ask a lot of questions especially about insurance and now know how to write a check and about co-pays”
Social support	“I was able to talk to him like I knew him for years. I treated him like a brother, and he did the same to me”
	“Getting to talk to someone who won’t judge me and learn things that I did not know before”
	“Talking about my concerns and venting about them”
	“My mentor was around my age. It always felt like a comfortable conversation and not like she was teaching/talking at me”
Ways mentees changed
Self-management skills	“Helped me grow up a little bit. I started taking better care of myself and staying on my meds”
	“I became more responsible for my own care, and I adapted to taking my medicine at different times”
	“Stayed on top of my prescriptions a little more”
Self-confidence	“I feel more confident about my transition”
	“I am not as nervous to transition to adult care”
	“I was more confident in coming to my appointments”
	“I have more insight in the adult care aspect of things. I am not as nervous to transition to adult care”
Self-advocacy	“I learned how to advocate for myself when I am sick in a respectful way”
	“I learned how to speak up for myself”
Suggestions/improvements
More frequent communication	“I think the phone calls should be every other week”
	“Call should be more like two times per month”
	“Didn’t talk too much during text, would rather hear her voice. More frequent calls would help”
	“More communication with the mentors”
Meet others with SCD	“If people feel comfortable, maybe meeting others around the same age in this program so that we can support each other in the transition”

In response to the item about their thoughts on working with a medical student as the mentor, participants indicated that being similar in age made the mentor relatable, that it was a great opportunity to learn from each other, and that the mentors were knowledgeable and provided helpful insights. When prompted, participants did not report any negative responses to working with medical students. Overall, patients had high ratings for each of the three domains of working alliance (goal, M = 17.9, SD = 2.1; task, M = 18.3, SD = 1.9; bond, M = 18.7, SD = 1.7).

All medical students who entered the program completed it fully. Medical students reported high satisfaction with the program, including the mentor training, program materials, and relationship with their mentees. Medical students reported that they felt that this experience deepened their understanding of barriers to health and enhanced their didactic medical education about topics like managing a chronic illness and transitions of care.

Secondary Outcomes

Transition Readiness

Participants reported statistically significant improvements in overall transition readiness (t [17] = 5.21, p < .001), healthcare knowledge and skills (t [17] = 3.81, p = .001), and independent living skills (t [17] = 2.98, p = .008; see Table IV). Cohen’s d were consistent with large effects (ds = 0.80–1.18). No significant changes were demonstrated in education and vocation planning skills or social support skills.

Table IV.

Changes in Psychosocial Variables From Baseline to Follow-Up

	Baseline M (SD)	Follow-up M (SD)	Mean difference (95% CI)	p	d
Overall transition readiness	97.94 (6.62)	104.67 (4.61)	6.72 (4.15, 9.29)	<.001	1.18
Healthcare knowledge and skills	26.00 (3.44)	28.89 (1.53)	2.89 (1.29, 4.49)	.001	1.09
Education and vocation planning	18.17 (1.72)	18.94 (1.16)	0.78 (−0.08, 1.64)	.74	0.52
Social support skills	16.28 (3.35)	17.56 (2.62)	1.28 (−0.25, 2.80)	.09	0.43
Independent living skills	37.50 (2.96)	39.28 (1.07)	1.78 (0.52, 3.04)	.008	0.80
Emotional impact	54.81 (5.19)	57.12 (5.72)	2.31 (−1.19, 5.82)	.18	0.42
Pain impact	54.22 (9.71)	54.54 (8.05)	0.33 (−5.09, 5.74)	.90	0.04
Sleep impact	52.43 (5.93)	54.98 (6.65)	2.54 (−1.10, 6.19)	.16	0.40
Social function impact	54.82 (7.30)	58.10 (8.70)	3.27 (−0.57, 7.11)	.09	0.41
Stiffness impact	53.79 (9.40)	55.13 (6.97)	1.34 (−4.37, 7.06)	.63	0.16
Medication adherence score	3.15 (1.56)	4.70 (1.84)	1.55 (0.30, 2.79)	.02	0.91
SCD self-efficacy score	33.39 (5.04)	36.78 (6.83)	3.39 (1.40, 5.37)	.002	0.56
SCD knowledgea	19.12 (2.31)	19.59 (3.35)	0.47 (−0.70, 1.64)	.40	0.16
Health literacy score	3.00 (2.35)	3.89 (2.14)	0.89 (0.02, 1.75)	.05	0.40

^aOutlier excluded from analysis.

Health-Related Quality of Life

There were no statistically significant changes in any health-related quality of life domains (see Table IV).

Sickle Cell Self-Efficacy

Participants demonstrated a statistically significant increase in SCD self-efficacy, t (17) = 5.38, p = .002, d = 0.56 (see Table IV).

Medication Adherence

Fifteen patients reported being prescribed daily medications/vitamins. The average medication adherence score improved significantly, t (17) = 2.66, p = .02, d = 0.91 (see Table IV). In terms of adherence categories, 12 reported low adherence at baseline. Despite overall average improvement, there was little change in the distribution of medication adherence levels.

Health Literacy

At baseline, participants had a mean health literacy score of 3.00 (SD = 2.35), with just over a third of the sample (N = 7) demonstrating a high likelihood of limited literacy, two (11.1%) demonstrating the possibility of limited literacy, and nine (50.0%) demonstrating adequate literacy. At follow-up, participants had a statistically significant higher mean score of 3.89 (SD = 2.14; p = .05), with a greater proportion demonstrating adequate literacy (N = 13, 72.2%).

Disease Knowledge

At baseline, participants answered an average of 19 out of 24 items correctly. Participants demonstrated a non-significant increase of 0.47 points in knowledge from baseline (see Table IV).

Discussion

Feasibility of this novel medical student mentor intervention was demonstrated through adequate enrollment of eligible participants (63.2%), retention (75%), and adherence to the intervention (5.3 out of 6 calls). Of note, recruitment and adherence to the intervention exceeded our a priori benchmarks, but the retention rate of 75.0% fell just short of our 80% benchmark. Published studies show that interventions delivered in conjunction with clinic appointments generally have higher rates of retention (range: 50–100%; Calhoun et al., 2019; Crosby et al., 2017; Melnick et al., 2017; Smith et al., 2011), whereas interventions that required participant involvement outside of dedicated clinic time have lower rates (40–45.8%; Hankins et al., 2012; Melnick et al., 2017). Our rate of 75% surpassed any of the current published studies of interventions delivered outside of the clinic setting and was in line with interventions delivered in conjunction with clinic appointments. Despite our intervention taking place outside of clinic time, the use of digital methods (video calls and text messages) may have offered greater flexibility for completing sessions and reduced geographic and transportation barriers.

Adherence to the intervention was high (on average, 5.3 out of 6 calls). The most common time of dropout was between completing baseline and the first phone call, as most participants who completed one call went on to complete all six calls. Mentors were required to send a text message to their mentees on the day of mentee recruitment, but there was no deadline for scheduling or completing the first call. The consistent drop out at this time point may suggest that a text message contact is not enough to initiate the relationship. The intervention could be modified to set the expectation that the first call occur within one week of enrollment or potentially offer an in-person meeting of the mentee and mentor at baseline. Although some reasons for dropout were unavoidable and reflect the complexity of the disease and issues faced by young adults (i.e., severe illness and becoming homeless), it was notable that participants who did not complete any calls were younger on average than those who completed the program. Although current recommendations for transition planning suggest beginning preparations as early as 12 years old (American Academy of Pediatrics, 2002) younger patients may not feel a sense of urgency about transition preparations, resulting in poorer engagement with transition programs.

Satisfaction with the intervention was very high. Although participants were largely satisfied with the content, they suggested adding more information about health insurance. This is a critical topic, as patients often report changing or losing insurance during the time of transition (Bemrich-Stolz et al., 2015). The most common suggestion to improve the intervention was to increase the frequency of calls (e.g., two times per month vs. one time), which may facilitate relationship-building and shorten the duration of the intervention.

One important aspect of this study was to evaluate whether patients would find medical student “peer” mentors acceptable. Patients reported high satisfaction with their relationship with their mentor, demonstrated by high scores on the working alliance scales and remarks during exit interviews. Patients found the medical students to be “relatable” and appreciated the medical students’ perspective and knowledge about the healthcare system. Despite high satisfaction and acceptability, patients suggested adding an in-person or digital support group for peers with SCD. Consistent with prior research (de Montalembert & Guitton, 2014) many reported that they did not know anyone else with the disease, so this intervention did not fill this support gap. Adding virtual meetings using social media or other digital tools (e.g., Facebook, text messaging) may be a feasible way to add additional peer support along with the mentorship relationship.

Finally, participants demonstrated promising improvements in transition readiness, self-efficacy, medication adherence, and health literacy. This is one of the first studies of youth with SCD to demonstrate improved transition readiness after a program. A patient navigator intervention (Manwani et al., 2017) also found improved transition readiness among participants, suggesting a one-on-one supportive relationship with a mentor or navigator can be helpful in guiding young adults through transition. Improvements in self-efficacy are promising, as self-efficacy has been associated with reduced healthcare use and improved health outcomes including health-related quality of life (Goldstein-Leever et al., 2020). Finally, the demonstrated improvements in average medication adherence are consistent with results of another patient navigator intervention (Allemang et al., 2016). The common focus on social support is consistent with the broader literature demonstrating improved adherence with social support interventions across multiple disease populations (Boardman et al., 2014; Mondesir et al., 2018).

Despite promising results, this study had several limitations. Although a small sample from a single medical center was appropriate for feasibility testing, our findings may not be generalizable. The study was not randomized and results could reflect spontaneous or developmental improvement. Future studies should include a control group and an adequate sample size to evaluate the efficacy of the intervention. Since participants self-selected into the study, it is possible that more motivated patients were more likely to participate in the program, which could lead to better overall adherence to the program and improvements in transition readiness indicators. However, we feel the risk of this bias is low, as participants and nonparticipants did not vary significantly in demographic factors but may have differed by some unmeasured factors. This study excluded patients with SCD with known cognitive impairments; future studies should consider whether a mentoring program could be tailored to work with young adults with cognitive impairments and their caregiver. Finally, medical students only represent one pool of potential mentors; other allied health professional students including psychology, nursing, physician assistant, or social work students could also serve as mentors to provide support, educate patients, and facilitate positive behavior change.

Although we found medical students to be acceptable “near-peers” in this study, it is important to recognize that there is a potential power differential between patients and any professional student connected to the patients’ medical care. As a community that has faced many health disparities, there is also significant potential for medical mistrust and stigma (Haywood et al., 2014) that may make it difficult to build trusting and productive relationships. Thus, it is important to carefully design these programs, preferably with the input of community members, and to train and supervise mentors to respectfully interact with patients in culturally competent ways.

In conclusion, a medical student mentor intervention to improve transition from pediatric to adult care is both feasible and acceptable to AYA with SCD. Given the recent unprecedented need for virtual healthcare services, this is a promising virtual program that offers additional support for a vulnerable population. These promising results warrant further study in an adequately powered randomized controlled trial to examine the efficacy of the intervention on psychosocial outcomes. Pediatric psychologists might take the lead in designing and implementing these types of programs given their knowledge of self-management techniques and communication skills to help build relationships. Partnering with patients and other key members of multidisciplinary teams may lead to success in identifying acceptable mentors and providing adequate training to facilitate successful mentoring relationships.

Supplementary Data

Supplementary data can be found at: https://academic.oup.com/jpepsy.