A 2-arm randomized trial design was used. Dyads were randomized to 1 of 2 possible study conditions: (1) iCanCope with SCD intervention or (2) attention-control education. Details on the trial methodology are available in an open-access protocol paper [26], and the trial has been registered on ClinicalTrials.gov (NCT03201874). This paper will focus on program engagement data from dyads who received the intervention condition. Comparative efficacy results from the broader trial will be reported separately in a future paper.

The study was approved by the locally responsible institutional ethics boards (Seattle Children’s Institutional Review Board, STUDY00001578; Emory University Institutional Review Board, 00092216; Connecticut Children’s Institutional Review Board, 17-115-CCMC; and The Hospital for Sick Children Research Ethics Board, 1000053724).

The main recruitment sites were Connecticut Children’s Medical Center, Children’s Healthcare of Atlanta/Emory University, Seattle Children’s Hospital, and The Hospital for Sick Children in Toronto. In addition, study referrals were accepted from University of Mississippi Medical Center, Boston Medical Center, and Northwestern University/Lurie Children’s Hospital. The study was centrally managed at Seattle Children’s Research Institute.

Youth were eligible if they (1) were aged between 12 and 18 years, (2) were diagnosed with any SCD genotype, (3) were able to speak and read English, (4) scored at least 4 (indicating moderate-to-severe pain interference over the past month) on the Sickle Cell Pain Burden Interview [6], and (5) were willing and able to complete online measures. Youth were excluded if they had significant cognitive limitations that would impair their ability to use and understand the iCanCope program, as per their health care provider or caregiver. Youth were also excluded if they had received more than 4 sessions of outpatient psychological therapy for pain management in the 6 months prior to the time of screening.

Caregivers were eligible if (1) they were able to speak and read English, (2) they were willing and able to complete online measures, and (3) their child was a study participant. Youth were permitted to enroll in the study even if their caregiver chose not to participate.

Dyads assigned to the intervention condition were given user-authenticated access to the iCanCope with SCD program for a period of 8-12 weeks. The program content, based on the principles of cognitive behavioral therapy, was delivered via a modular website and complementary smartphone app as summarized in Tables 1 and 2. Standardized orientation materials were used to guide participants on how to login and independently use the program. Separate websites with unique content were provided for youth and caregiver participants, while a smartphone app (iOS/Android) was provided for youth participants only. Website navigation can be found in Figures S1 and S2 in Multimedia Appendix 1. The youth and caregiver websites were each organized into 6 core modules in addition to 2 optional youth modules on insomnia and negative mood. Features of the smartphone app included a daily symptom tracker (ie, pain intensity, pain impact, mood, and sleep quality), a calendar to view historical symptom data, goal setting in relevant domains (eg, sleep and mood), a library of SCD education and pain coping strategies, and a community forum to interact with other users (Figure S3 in Multimedia Appendix 1).

Quantitative data were analyzed using STATA version 15.1 (Stata Statistical Software). Descriptive statistics were used to summarize background characteristics of the sample and program engagement data. An engagement interaction with the website was defined as a unique content page view. For the app, the following engagement interactions were captured: symptom check-in completed; goal created; library article viewed; and interaction within the community forum. For the youth-specific intervention, program content was distributed across the website and app such that an individual user might access topic-specific information (eg, “healthy sleep habits”) by viewing a website page or reading an app library article. Given this overlap in program content delivery, engagement data for youth participants were also mapped by content category (Table 1).

To address our exploratory aim regarding the relationship between parent engagement and child engagement, we conducted a chi-square test (significance was set at α<.05) and assessed the strength of the association using the phi correlation. The phi correlation can range from 0 to 1 and was interpreted as follows: no correlation or very weak correlation (0-0.19), weak correlation (0.20-0.29), moderate correlation (0.30-0.49), strong correlation (0.50-0.69), and very strong correlation (0.70-1.00) [30].

Audio recordings from the qualitative interviews were transcribed verbatim and analyzed by 2 team members (CL and FN) using Dedoose Version 9.0.17 (SocioCultural Research Consultants, LLC). Simple content analysis, a dynamic process that summarizes the informational content of data, was used [31,32]. Specifically, data for all participants were coded and organized into categories that reflected the emerging themes. The raw data were revisited on a regular basis throughout the analytic process to ensure that the codes were grounded in the data. Any disagreements were resolved through consensus.

Participant recruitment spanned periods before and after onset of the COVID-19 pandemic. For analysis purposes, March 2020 was used to differentiate between prepandemic and pandemic recruitment periods in North America.

Recruitment was carried out between January 1, 2018, and September 30, 2021. A total of 57 youth and 56 caregivers (56 dyads and 1 youth-only participant) were randomized to the iCanCope with SCD intervention condition and received instructions on how to access the program. The sample included dyads who enrolled in the study before (43/56, 75%) and after (14/56, 25%) pandemic onset. Demographic characteristics of the sample are provided in Tables 3 and 4.

Nearly all youth participants (55/57, 96%) were in middle school or high school. Most youth (47/57, 82%) reported using a smartphone multiple times per day. Caregiver participants (n=56) reported their highest completed education level as high school or less (11/56, 20%), vocational or trade school or some college/university (15/56, 27%), college or university (18/56, 32%), or graduate degree or professional school (11/56, 20%). Most dyad households (30/56, 54%) included 1 or 2 children under 18 years of age (range of 1 to 6 children). With regard to the total annual household income before taxes, participants reported less than US $24,999 (11/56, 20%), US $25,000-49,999 (16/56, 29%), US $50,000-74,999 (9/56, 16%), US $75,000-99,999 (9/56, 16%), and US $100,000 or more (5/56, 9%).

This is one of the first studies to characterize patterns of engagement with a web- and app-based pain self-management program among youth with SCD and their caregivers. The data demonstrate that most youth with SCD engaged with the program (40/57, 70%) while most caregivers did not (39/56, 70%). The youth program was delivered via a mobile app and a website. Youth were more likely to engage with the app than the website, and the most popular content categories were goal setting, introduction to the program, and symptom history. Among caregivers who engaged with the website, introduction to the program, behavioral plans, and goal setting were the most popular content areas. Exploratory analysis revealed a significant moderate association between caregiver and youth engagement with the intervention. The key insights from qualitative interviews were as follows: (1) nearly all youth intended to continue using the program; (2) caregivers would recommend the program to others living with SCD who were newly diagnosed or who frequently experienced pain crises; (3) caregivers cited the large amount of content and low relevance to their specific circumstance as reasons for low engagement; (4) youth preferred the app over the website because of perceived greater ease of use, acceptable time commitment, and interactivity; and (5) use of the youth program helped to facilitate SCD self-management behaviors such as goal setting, applying strategies for pain management, symptom trend monitoring, and learning through peer support.

There are several app- and web-based interventions designed for children and youth with SCD [33-36] that focus on enhancing SCD knowledge, medication adherence, quality of life, transition to adult care, and pain management. However, engagement or adherence metrics for these interventions were reported in only a few studies. Among 4 recent studies [34-36] that reported some engagement data, operational definitions varied, making direct comparisons difficult. First, Palermo et al reported high engagement with web-based management of adolescent pain (WebMAP), an internet-delivered cognitive-behavioral therapy intervention tailored for SCD [27]. Engagement was defined as participants completing at least one module in the WebMAP program. They found that 14 of 15 (93%) participants met this threshold, which was higher than the proportion of youth (40/57, 70%) who engaged with any iCanCope content. Second, Phillips et al reported engagement with an app-based program called Voice Crisis Alert V2 over a 12-week intervention period [35]. They characterized engagement based on the use of individual app components, similar to the content categorization approach used in our study. Among 60 dyads, symptom tracking and history were the most popular features (used by 90% of dyads), while a clinician messaging feature was least popular (total of 5 message threads generated during the study). In our study, we similarly found that symptom tracking and history were among the most popular content areas. Third, Saulsberry et al offered the Sickle Cell Transition E-Learning Program to 183 youth, of whom, 53 (29%) completed at least one of the 6 available modules [36]. In our study, a higher proportion of youth (70%) who were offered the iCanCope program engaged with the content. Fourth, Leonard et al reported engagement with a medication adherence app as percentage of days that each participant logged medication administration [34]. In a pilot sample of 11 patients, average engagement was 80%. Our study found a wide range of youth engagement with various program components, with the most popular feature being goal setting (35/40, 88%) and the least popular being education about SCD and treatment (6/40, 15%).

The iCanCope program has also been assessed in youth with juvenile idiopathic arthritis and chronic pain [24,25]. Usage levels were similar in the SCD population, although engagement was higher with the symptom history function in the juvenile idiopathic arthritis population [25] at 80% versus 60% among SCD users. A systematic review focusing on engagement and adherence to mHealth interventions in children and youth in clinical and nonclinical populations reported average adherence of 78% [37].

Several contextual factors may explain the differences in engagement patterns seen across studies in the SCD population, among users of other iterations of the iCanCope program, and generally among the pediatric population. These factors include differences in the intervention scope (eg, pain management versus transition readiness), distinct definitions of engagement, different study designs and clinical populations, and technology variation (eg, wearable technology vs app based) [38]. The literature suggests that treatment adherence among youth with SCD may be impacted by forgetfulness, side effects of pharmacological interventions (eg, nausea, heart racing, and taste aversion), questioning efficacy of treatment, and a desire to be “normal,” and some of these factors may similarly impact adherence to digital health interventions [39]. Previous literature also suggests that sociodemographic factors can impact engagement with digital interventions, highlighting lower rates of engagement among racial or ethnic minorities, and those with low socioeconomic status [40-42].

There is limited research on caregiver engagement with digital health interventions for SCD. Many interventions designed for youth do not have a caregiver component. However, 3 recent digital health studies have reported parent engagement in pediatric SCD, chronic pain, and cancer populations, respectively [43,44]. The WebMAP program was designed for pediatric chronic pain patients and adapted for the SCD population. In the chronic pain population, caregiver engagement was higher compared with youth (82% vs 74% completed at least one module) [43]. Similarly, caregiver engagement was also higher (100% vs 93% completed at least one module) than youth engagement in the SCD population [27]. In contrast, the Teens Taking Charge Cancer web-based self-management program [44] reported lower caregiver engagement compared with youth (8% vs 28% completed at least one module), which is consistent with our results. One study on the psychology of eHealth use among caregivers with children having complex health conditions found that factors, such as poor caregiver psychosocial health, high eHealth literacy, and high acceptability of eHealth, were associated with increased engagement [45]. Furthermore, caregiver engagement with nondigital or analog health interventions that report high levels of engagement tend to include methods, such as motivational interviewing and teaching-learning processes (eg, role play and interactive problem solving) [46]. Our program was self-guided, which may have contributed to lower levels of engagement from caregivers.

Our study had several strengths. First, a phased evidence-based approach was used to develop the iCanCope with SCD program. A qualitative needs assessment involving youth with SCD, caregivers, and health care providers was completed to determine essential components of the program [12]. Next, design sessions were completed with patient partners with SCD to adapt the platform and content for a population with SCD [26]. The changes that were implemented (eg, content tailoring and app design) were cited by participants and their caregivers as positive aspects of the iCanCope program. Second, this study used both qualitative and quantitative methods, which enhanced our ability to understand and contextualize the diverse set of engagement experiences. Furthermore, by including youth and caregiver perspectives, we were able to distinguish their unique experiences and understand how the family unit engaged with the program. Third, this study was a multi-center randomized controlled trial, and this increases the generalizability of the findings to SCD patients in North America. Moreover, the study sample was inclusive of families from a variety of socioeconomic backgrounds.

These findings should be considered in the context of a few limitations. First, caregivers were not heavily included in the design of the iCanCope program, which may partially explain their reduced engagement with the content. As per the user design approach, the creation of iCanCope was focused on meeting the needs of youth, as the primary users were envisioned to be youth, rather than caregivers. However, our study findings indicate that parent engagement with SCD program content is important as it is positively associated with child engagement. Unfortunately, poststudy interviews with caregivers revealed that the program was too content heavy and time-consuming. One strategy to improve caregiver engagement would be to account for the individual need for cognition, which refers to the enjoyment of deep thinking [47]. For those with a low need for cognition, information could be provided in short easily accessible excerpts, such as videos or infographics. In contrast, long-form content would be more suited for individuals with a high need for cognition. Notably, once the program is publicly released, caregivers will be able to review content at their own pace rather than within the time constraints of a research study. In addition, consultations with parents may be completed to fine tune the program to meet their needs prior to public release. Second, differential use of the app versus website among youth suggests a need for changes to the website to increase engagement with the overall program. A specific technical limitation of the website was that it was unable to store the responses of participants, making it difficult for them to gauge their progress. Finally, we were only able to interview 1 youth participant who did not engage with the program at all.

Prior to public release of the program, there are opportunities for refinement using the elements of persuasive design [48]. Wen et al found that more frequent prompts were associated with higher engagement in studies with clinical pediatric populations [37]. Although the iCanCope app did incorporate reminders to prompt participants to engage (eg, push notifications to complete check-ins or set a goal), the frequency and timing of the reminders could be customized. However, it is important to balance the number of reminders while minimizing interruptions to the user. As such, identifying ideal opportunities to use the iCanCope program individualized to each participant is important. A novel approach includes using algorithms to determine these opportunities, which have been developed for adult populations [45,49]. In tandem, providing praise (eg, congratulatory messaging) and rewards when participants complete a check-in, read an article from the library, or complete a goal may influence engagement positively [48]. Rewards can include elements of gamification where participants gain points or virtual rewards for completing certain tasks, which they can use to promote engagement with self-management content. Finally, real-world use of the iCanCope program can also be studied in the future, after public release.

In terms of program dissemination, using strategies to target the family unit, rather than individual members, may increase engagement as SCD typically impacts multiple members of a family. This approach is supported by youth participants who reported ad hoc sharing of program content with their siblings and the finding that parent and child engagement was positively associated. Furthermore, by partnering with SCD clinics, the program could be introduced to youth and families when the SCD pain becomes more frequent to help cope with the corresponding increase in pain burden.

This is one of the first studies to apply digital health analytics to characterize patterns of engagement with SCD self-management among youth and caregivers. Differential engagement patterns were found, with more youth engaging in the program than their caregivers; however, youth were more likely to engage with the app than the website. These findings will be used to optimize the iCanCope with SCD program prior to release, with the potential to improve the accessibility and acceptability of pain self-management support for families affected by SCD.