Feasibility of an mHealth self-management intervention for children and adolescents with sickle cell disease and their families

Abstract

Sickle cell disease (SCD) is an inherited hemoglobinopathy that leads to blood vessel occlusion and multiorgan complications, including pain, that may be experienced daily. Symptom management often begins at home, and tools are needed to support self-management strategies that can be implemented by children with SCD and families. The purpose of this study was to assess the feasibility of the mHealth self-management intervention (application) Voice Crisis Alert V2 for children with SCD and families. Feasibility assessment was guided by the Reach, Efficacy, Adoption, Implementation, and Maintenance framework. Data were collected with 60 dyads (children with SCD/caregivers) at four time points. Self-management data were collected via application use, and postintervention interviews were conducted. Analyses included descriptive statistics and constant comparison with directed content analysis. Recruitment was completed in 28 weeks, with 82% retention at end-of-intervention. Mobile Application Rating Scale scores and interview data indicated high satisfaction. From baseline to mid-intervention, 94% of dyads used the application (75% of total use); 45% used the application from mid-intervention to the end-of-intervention. Dyads made 2,384 actions in the application; the most commonly used features were recording health history and recording and tracking symptoms. Few reported issues with the application; most issues occurred early in the study and were corrected. After the intervention period was completed, 37% continued to use the application. Feasibility was confirmed by meeting recruitment and retention goals, high adoption of the application, and high reported satisfaction with the application. Challenges with sustained use were encountered, and areas for improvement were identified.

Children with sickle cell disease (SCD) may experience pain and other symptoms, such as fatigue and fever. Children with SCD often take medications and receive treatments at home to avoid negative outcomes from the disease, including pain, hospitalization, and more serious outcomes such as stroke. Children and families are required to self-manage the child’s medications, treatments, and other health care activities on a daily basis in the home, school, and community settings. Tools that help children and families develop self-management skills may help improve pain and other outcomes. Because self-management takes place in the home, school, and community environment, tools offered through smartphones and tablets may be ideal. This study assessed whether a self-management application (app) on smartphones and tablets would be used by children with SCD and their families and whether children with SCD and their families found the app satisfactory. The study found that nearly all children and families in the study used the app, though the duration and frequency of use decreased substantially over time. The parts of the app that were the most used were for keeping track of pain and symptoms, health history, and medications. Overall, children and families were highly satisfied with the app.

Implications.

Practice: An mHealth intervention for supporting the development of self-management strategies in children and adolescents with sickle cell disease (SCD) is feasible and acceptable.

Policy: Policymakers who are interested in improving health outcomes and decreasing avoidable health care utilization in children and adolescents with SCD should consider interventions that can be implemented in real-world settings to enhance self-management.

Research: mHealth self-management interventions for children and adolescents with SCD and their families should be further evaluated in adequately powered, randomized effectiveness trials.

INTRODUCTION

Sickle cell disease (SCD) is an inherited hemoglobinopathy characterized by alterations in the shape of red blood cells affecting approximately 100,000 individuals in the USA; of these, 90% are African American [1]. The altered red blood cells occlude vessels and lead to inflammation, infarction, organ damage, and pain [2]. As a result, individuals with SCD experience comorbidities, have high health care needs and utilization, and premature death [3,4].

Pain is among the first symptoms experienced by individuals with SCD, which often begins in early childhood and continues throughout the lifespan [5,6]. Children with SCD-related pain have poorer physical, psychological, and social outcomes than children in general and, as children age, the quality of life declines [3,7]. Because pain management in SCD frequently begins at home, it is critical that children and families develop effective self-management strategies for symptoms and prevent acute care utilization. Self-management is “the interaction of health behaviors and related processes that patients and families engage in to care for a chronic condition” [8]. Individuals with complex conditions requiring complicated treatment regimens, such as SCD, have a great need for self-management strategies and systems [9]. By developing self- and family-management skills, children with SCD and other chronic conditions may experience behavior change and improvements in symptom management, family functioning, individual and family quality of life, self-efficacy, and provider relationships, which may lead to decreased health care utilization [9,10]. To facilitate self-management skill development and behavior change, interventions should address barriers to self-management, such as gaps in knowledge, condition severity, busy home and work/school environments, and access to care [10]. Furthermore, among self-management intervention research, there remains a gap in studies on family-focused interventions [9] and on interventions to reduce caregiver burden and improve interactions with health professionals [10].

Mobile health, or mHealth, is a component of eHealth that includes the use of mobile phones, patient monitoring devices, and other wireless devices [11]. mHealth can overcome barriers to care and improve resource delivery in difficult-to-reach populations; thus, mHealth is useful for delivering self-management interventions [12,13]. Furthermore, mHealth interventions have been shown to improve symptom outcomes, quality of life, health care utilization, and medication adherence in chronic conditions [12].

To address the gap in family-centered self-management interventions for individuals with SCD, the feedback was obtained from children and adolescents with SCD, parents/primary caregivers, and health care providers of affected children to develop and refine an mHealth application (app) to facilitate self-management behaviors in families of children with SCD [14]. The purpose of this study was to assess the feasibility of app use and implementation processes with children with SCD and parents/caregivers to inform the design of a future effectiveness trial.

METHODS

Sample and setting

The study was conducted at a Pediatric Sickle Cell Clinic in the Southeast United States that serves approximately 500 children and adolescents with SCD annually. Children of ages 0–17 years with SCD and their primary caregiver 18 years of age or older were eligible. Exclusion criteria were having cognitive disabilities or delays (child or caregiver) and not having access to Wi-Fi. Cognitive disabilities were documented in the child’s medical record and classified as severe neurocognitive deficits per neuropsychological evaluation. IRB approval (Pro00084400) was obtained prior to recruitment and enrollment. The study was registered in ClinicalTrials.gov number NCT03585543.

Recruitment and enrollment

Recruitment strategies included: (a) approaching child/caregiver dyads that participated in the initial development phase of the app [14] (Voice Crisis Alert V2) and agreed to recontact and (b) identifying potential participants by pediatric SCD clinic staff. Clinic staff were provided with inclusion and exclusion criteria and potential participants were prescreened for SCD diagnosis and documentation of cognitive disability or delay. Clinic staff introduced the study to potential participants and asked for permission to introduce the study team who then described the study in detail. Informed consent from caregivers and written assent from children 12–17 years of age were obtained by in-person consent or electronic consent via Research Electronic Data Capture [15] e-Consent.

Intervention delivery

The mHealth SCD self-management intervention, Voice Crisis Alert V2, was delivered via an app on the caregiver’s and child’s (when owned) mobile device during a 12 week intervention period. This multicomponent, web-based app consisted of (a) access to educational resources on the SCD process, treatment, home management strategies, and symptom prevention and management strategies; (b) tools for children and caregiver dyads to track and monitor symptoms; and (c) a mechanism for communication between the child or caregiver and provider. In the development phase of the study, we obtained feedback from children with SCD, parents/caregivers, and health care providers; subsequently, revisions were made to the app based on this feedback to best meet the needs of the population [14]. App development was informed by the Pediatric Self-Management model, specifically targeting processes for self-management behavior development [9]. Targeted self-management behaviors included symptom monitoring for pain and associated symptoms (e.g., fever and shortness of breath), disease-specific care, such as taking medications and preventive treatment (i.e., adequate hydration, cold avoidance), and parental support of treatment. Additional details on intervention development and how the app features were guided by the theoretical framework are described in a prior publication [14]. During the initial study visit, dyads were entered into the app’s backend database and were assisted with downloading the app onto the child’s device, caregiver’s device, or both. Caregivers chose whether they, the child, or both were entered into the app’s database; in cases where the child did not have a device, each dyad member could use their individual logins to access the app on a single device. The study personnel demonstrated the use of each app component and answered questions. Dyads were instructed to use the app in a “real world” environment for the 12 week intervention period; that is, to use all or any of the app components they felt best suited for the everyday care of the child. While dyads were encouraged to use the app frequently during the intervention period, they were also welcome to continue use after the intervention period ended. A printed instruction sheet was provided and dyads were instructed to contact the study personnel with questions or problems.

Education: “Sickle Cell Information”

The educational component was titled “Sickle Cell Information” and included materials organized by child age for those with SCD and caregiver specific materials previously developed, tested by authoritative sources, and publically available (i.e., National Institutes of Health, National Heart, Lung, and Blood Institute, and the Centers for Disease Control and Prevention). Links and citations for these original sources were included.

Tracking: “I’m in Pain,” “Pain History,” and “Crisis Care”

The symptom tracking and monitoring component included a customizable avatar for recording pain location, severity, and characteristics, titled “I’m in Pain.” Dyads earned achievement points to unlock bonus avatar accessories and backgrounds by reaching goals, such as the completion of the health history page. The avatar’s facial expression changed based on the child’s pain rating, which was measured using the validated Wong Baker Faces Pain Rating Scale [16] and a sliding bar that rated pain using the validated 0–10 numeric pain rating scale [17]. Dyads documented associated symptoms and pain relief measures with each pain entry, which could subsequently be viewed in a section titled, “Pain History” for the tracking of pain history and trends. Via a health history section, titled “Crisis Care,” dyads documented clinical information, such as type of SCD, medications, allergies, and other relevant data.

“Crisis Care” included a medication tracking system for the child or caregiver to enter routine home medications and the scheduled time. This prompted a reminder to appear on the mobile device home screen that would take the child or caregiver directly to the app. The child or caregiver then had an option to mark the medication as taken or snooze the reminder to prompt another reminder at a later specified time. Dyads tracked medication history through the app.

Communication: “Messaging”

Through an encrypted portal developed to provide direct access from the child/caregiver to the provider, a child or caregiver could send a message to the provider within the app to report pain or other symptoms. Providers could securely respond with an encrypted message via the portal from their computers.

Settings

In the “Settings” feature, the child or caregiver could modify the aesthetic features of the app (e.g., change the background color) and unlock additional bonus avatar features such as clothing, jewelry, accessories, and different backgrounds earned through achievement points.

DATA COLLECTION

To determine feasibility, the Reach, Efficacy, Adoption, Implementation, and Maintenance (RE-AIM) framework [18,19] was used to assess study implementation processes (Table 1). While the Efficacy component was addressed in the study by preliminarily investigating whether findings may support a future large-scale study, this component will be presented in a subsequent manuscript for a more complete description of methods and findings. Data were collected through surveys, app use, and key informant interviews. Children and caregivers completed age-appropriate assessments at baseline, mid-intervention (6 weeks from baseline), end-of-intervention (12 weeks from baseline), and 3 months after the end of the intervention (24 weeks from baseline). During the development of this project, we were unable to find a mobile health quality assessment instrument validated for use in pediatrics; therefore, the Mobile Application Rating Scale [20] (MARS) was delivered to caregivers only in this study. The MARS was administered at 6 and 12 weeks from baseline to assess the objective and subjective quality of the app. The MARS is a 23-item instrument using a five-point Likert scale (1—inadequate; 2—poor; 3—acceptable; 4—good; and 5—excellent) with subscales on objective engagement, functionality, aesthetics, and information quality and subjective quality. The MARS total score had a Cronbach’s alpha of 0.9 with subscales ranging from 0.7 to 0.8 and a total intraclass correlation coefficient (ICC) score of 0.79 [20]. As part of the MARS, an additional six app-specific items were administered to assess the app’s impact on awareness, knowledge, attitude change, intention to change, help seeking, and likelihood of behavior change. App-specific items use a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree. All self-report and proxy assessments were administered by phone or in person at the sickle cell clinic (by participant choice). Dyads were provided compensation in the form of a $50 gift card at each of the four data collection points.

Table 1.

Measures according to the Reach, Efficacy, Adoption, Implementation, and Maintenance (RE-AIM) Framework

RE-AIM component	Measures
Reach and recruitment	• Monitoring of sample representativeness
	• Recruitment activities
	• Rates of recruitment
	• Percentage eligible for the study of those who expressed interest
	• Number of consented participants
	• Number provided with baseline informational session
Adoption	• Adherence via the overall number of times each intervention component was accessed, and length of time spent within each component
	• Acceptability via caregiver satisfaction, number of reported problems, and types of problems reported
	• Instrument to assess engagement, functionality, aesthetics, and information
	• Education via the number of times and length of time the educational component accessed
	• Symptom monitoring and tracking and patient–provider communication via the number of times symptoms were recorded and the number and types of messages sent via the intervention
Implementation	• Technical challenges via the number of problems reported with the intervention and types of problems reported
	• Documentation of delivery of instructional sessions and follow-up questions
Maintenance	• Projection of sustained use via the number of dyads who report will continue to use intervention, perceptions of intervention, and perception of provider role

A postintervention interview was conducted to obtain data on perceived accessibility and usability and adherence to the intervention. A subset of dyads that completed the 12 week study visit was approached for postintervention interviews using purposive sampling based on child age to obtain perspectives from dyads with children within a broad age range. Semistructured interviews lasting approximately 30–45 min were conducted using a qualitative descriptive approach [21]. An interview guide included open-ended questions and prompts about what participants liked or disliked about each of the app components and additional questions on the adequacy of instructions on using the app, problems encountered with the app, satisfaction with the app, and whether they see themselves continuing to use the app and how. Dyads were also asked, “How did you use this part of the app to help manage your/your child’s health?” Interviews were audio-recorded and transcribed for analysis. Data saturation was met after 15 dyad interviews. Dyads were compensated with a $50 gift card.

DATA ANALYSIS

The purpose of this study was to establish the feasibility of implementing the intervention, obtain estimates of variability for the primary outcome measures, and obtain preliminary indicators of efficacy of the intervention rather than to confirm or refute hypotheses. Therefore, sample size considerations focused on the precision of estimates; the targeted sample size was 60 children/caregiver dyads. With this sample, we had a precision of ±0.11 to ±0.16 for estimated outcome proportions of the feasibility measures, including recruitment and dropout.

Demographic and clinical characteristics of the study sample and feasibility outcomes were analyzed via descriptive statistics using frequencies, proportions, and measures of central tendency (means and medians) and variance (standard deviation) as appropriate using SPSS version 25 [22]. Metropolitan versus nonmetropolitan status was determined by comparing the participant’s current county of residence to the Centers for Disease Control and Prevention NCHS Urban–Rural Classification Scheme for Counties [23].

Qualitative analysis of key informant interviews

Data collected from postintervention key informant interviews were analyzed using directed content analysis and nVivo qualitative data analysis software version 11 [24]. Codes were developed and themes emerged around the acceptability of the intervention (adoption), satisfaction with the intervention (adoption), and descriptions of planned sustained use (maintenance). Further themes were developed around variations in uses of the app (adoption) and suggestions for further improvements.

RESULTS

Approximately half of the child participants were males, and all were African American, Non-Hispanic with a mean age of 7.8 years (Table 2). The majority of caregivers were employed, had some college or a college degree, and were single. Nearly all were the mother of the child. Most dyads lived in a metropolitan area with a mean of 3.4 people living in the same household.

Table 2.

Participant demographics

Variable	Sample
Child participants (N = 60)
Age (years), mean (SD), range (years)	7.8 (4.8), 0–17
0–7 years (n = 30), mean (SD)	3.8 (2.0)
8–17 years (n = 30), mean (SD)	11.8 (2.9)
Gender, n (%)
Female	28 (47)
Male	32 (53)
Parent/caregiver participants (N = 60)
Caregiver type, n (%)
Mother	55 (91)
Father	1 (2)
Grandmother	4 (7)
Employment status, n (%)
Employed	40 (66)
Caregiver/stay-at-home parent	3 (5)
Unemployed	10 (17)
Student	1 (2)
Disabled	1 (2)
Unknown	5 (8)
Education level, n (%)
9th–12th grade or less	9 (15)
High school graduate or GED	11 (18)
Some college	22 (37)
College graduate or above	18 (30)
Marital status, n (%)
Married	17 (28)
Living with partner	2 (3)
Single	38 (63)
Divorced	3 (5)
Household size (number of people), mean (SD)	3.4 (1.5)
County type, n (%)
Metropolitan area	53 (88)
Nonmetropolitan area	7 (12)

SD standard deviation.

Reach and recruitment

Recruitment goals to enroll two dyads per week were met; 60 dyads were enrolled over the 28 week recruitment period. Sixty-eight dyads were approached by clinic staff; 64 expressed interest and were willing to speak with the study staff. All interested dyads met the eligibility criteria. Four dyads opted not to initiate the trial prior to consent. The baseline informational session was completed with 60 dyads; 48 (80%) completed the 6 week mid-intervention visit, 49 (82%) completed the 12 week end-of-intervention visit, and 37 (62%) completed the 24 week postintervention follow-up visit.

Adoption

Adherence to the intervention was determined by calculating the number of dyads that used the app between each study visit time point. From baseline to mid-intervention (6 weeks from baseline), 56 dyads (94%) used the app; from mid-intervention to end-of-intervention (6 weeks from baseline to 12 weeks from baseline), 27 (45%) used the app; and from end-of-intervention to follow-up (12 weeks from baseline to 24 weeks from baseline), 22 (37%) used the app. Overall, participants made 2,384 individual actions within the app.

Approximately, 75% of app use occurred from baseline to 6 weeks (1,805 actions); 90% of dyads used the tracking components: Crisis Care, I’m in Pain, and Pain History (Table 4). The highest number of actions occurred in the I’m in Pain component, followed by the Crisis Care component. From 6 to 12 weeks, participants made 216 individual actions within the app, representing 9.1% of all actions. I’m in Pain was used by the highest number of dyads (40%), and the I’m in Pain and Crisis Care components had the greatest proportion of individual actions. From 12 to 24 weeks, participants made 279 individual actions (11.7% of total). Again, I’m in Pain was used by the highest number of dyads (33%) and had the highest number of actions. After the close of the study, participants had 65 total actions within the app; Sickle Cell Information was used by the greatest number of dyads (10%), but the highest number of actions took place in the Settings component, closely followed by the I’m in Pain component. The messaging component was the least frequently used overall, with only five message threads between participants and the clinic nurse practitioner during the study.

Table 4.

App use by component and study timeframe N = 60 dyads

	Crisis care na (%)	I’m in Pain! na (%)	Messaging na (%)	Pain history na (%)	Settings na (%)	Sickle cell information na (%)
Baseline to 6 weeks	54 (90%)	54 (90%)	24 (40%)	54 (90%)	47 (78%)	52 (87%)
6–12 weeks	12 (20%)	24 (40%)	2 (3%)	14 (23%)	14 (23%)	11 (18%)
12–24 weeks	15 (25%)	20 (33%)	1 (2%)	16 (27%)	15 (25%)	14 (23%)
Past 24 weeks	4 (7%)	5 (8%)	1 (2%)	4 (7%)	5 (8%)	6 (10%)

^a n = dyads.

App use was further assessed by which members of the dyad were enrolled into the app’s database and which members subsequently used the app (Table 3). In the majority of dyads, the caregiver was enrolled in the database and used the app, followed by both the caregiver and child. Children in dyads with caregiver-only enrollment and use were younger; however, during postintervention interviews, caregivers of younger children (less than 8 years) reported allowing the child to “play” with the app on the caregiver’s device.

Table 3.

Enrollment in app database and app use by dyad member

Participant by role entered into database	Participant by role app use	Number of dyads	Child age mean, median (range)
Caregiver only	Caregiver only	34	4.6, 4.5 (0–13)
Caregiver only	Child only	7	10, 10 (8–13)
Child only	Child only	7	12.9, 13 (8–17)
Caregiver and child	Child only	3	12.7, 14 (8–16)
Caregiver and child	Caregiver and child	9	12.6, 13 (8–17)

Overall quality, assessed using the MARS with caregivers, was a mean 3.9 at 6 and 12 weeks, which corresponds with a “good” response. Similarly, median and mean scores for each of the subscales were near a 4.0, with subjective quality mean scores slightly higher. Mean scores for all six app-specific items were 4.2–4.3 on a five-point scale.

In postintervention interviews, all 15 dyads stated that they were satisfied or very satisfied with the app. The I’m in Pain and the Crisis Care components, particularly the avatar for recording pain and the medication reminders and tracking, were reported as most helpful. While the messaging component was rarely used, most participants described the potential utility of this feature for communicating with providers. Themes emerged from interviews around self-management behavior development, including “pinpointing” the child’s pain, which included the caregiver report of better understanding the child’s pain characteristics, monitoring pain triggers, and determining the most helpful treatments. Teaching and learning was another theme that included teaching the child about their SCD and, for caregivers of younger children, learning what to expect in the future. A theme that emerged among older children (ages 8–17) and their caregivers was the use of the app to encourage the child to begin assuming responsibility for their SCD self-management.

Implementation

Twelve reports of problems were made to the study personnel by 12 dyads during the study, with no repeat problems reported (all were resolved after the first report). Problems involved logging into the app (n = 8). Of these, half were system issues resolved with bug fixes and updates to the app. The other half were the result of lost log-in information or passwords. Two of the remaining four problems were other system issues that were resolved with bug fixes and updates, and the final two problems were requests for assistance with downloading the app and registering on a new phone or device. Of the six system problems, five occurred within the first 4 months of the study. All but 1 of the 12 dyads reporting a problem were retained in the study through the end-of-intervention (12 weeks from baseline). Three additional participants reported problems during postintervention interviews. Problems reported during interviews included being “kicked out” of the app after logging in, the app “shutting off” after accessing Sickle Cell Information, snooze function on the medication reminder system, and occasional app freezing. All but one participant reported that the problems were resolved after either installing updates or reinstalling the app.

Early maintenance

From the end of the intervention period (at 12 weeks) to the 24 week follow-up visit, 22 dyads (37%) used the app. Eight dyads (13%) continued to use the app after the study ended. During postintervention interviews, all participants reported that they would continue to use the app, though some anticipated that they would use only one component or would use one component more than others (i.e., I’m in Pain and Crisis Care).

DISCUSSION

Findings from this study demonstrate the feasibility of a theory-based mHealth intervention (app) designed to enhance self-management of SCD in children and their caregivers. Feasibility and implementation approaches were assessed using RE-AIM.

For reach, recruitment goals to enroll 60 children/parent dyad at the rate of 2 dyads/week were exceeded; 60 dyads were enrolled in 28 weeks, supporting the feasibility of the recruitment approach. Few studies of eHealth interventions for children with SCD also enrolled caregivers; our findings support the feasibility of including caregivers. Including caregivers in mHealth and eHealth intervention studies in SCD, particularly those focused on self-management, is critical as the caregiver inexorably is part of the child’s care and understanding individual roles and dyadic interactions is important for promoting health behavior change.

The sample size in this study was larger than many other eHealth intervention studies for SCD [25]. In addition, of the 64 dyads that expressed interest in the study, 60 proceeded to enrollment (93.8%), indicating high interest in the intervention. Feasibility and pilot studies of other eHealth interventions for children with SCD reported lower enrollment rates [26–28]; differences may be attributable to variations in the complexity and expected time investment of study participants. One explanation for successful recruitment and enrollment in this study may have been the availability of the app on both iOS and Android. While loaner devices were available for this study, none was needed as all participants had personal mobile devices (tablet or smartphone). In addition, no participants were excluded for the lack of Wi-Fi access. Similar to prior studies [28], the majority of participants had devices that use the Android platform.

While the intervention period and retention to end-of-intervention (12 weeks) in this study was comparable to other studies [28–30], retention decreased by approximately 20% at the postintervention follow-up study visit (24 weeks). Compensation or incentivization has been commonly associated with an increase in survey response [31,32], with other methods, such as reminders and flexibility in scheduling and data collection methods, having a less substantial influence on survey completion [31]. Greater numbers of retention strategies result in higher rates of participation [33], and studies conducted with minority research participants require careful consideration of cultural context [34]. Ely and Coleman [35] described successful strategies to retain children with SCD and their families in a longitudinal study that included staff awareness of cultural differences, understanding of scheduling challenges with corresponding flexibility in study procedures, and building trust with participants. Multiple strategies were used to promote the completion of study surveys and retention in our study, including compensation for completing surveys, reminder phone calls or text messages, and flexibility in scheduling and method for completing surveys (days, evenings, and weekends; by phone or in person). However, future studies will entail additional strategies to improve retention to follow-up, such as ensuring that study personnel are aware of cultural differences and taking measures to build rapport through more frequent contact with participants.

Adoption was supported by high rates of app use and high satisfaction. Scores on the MARS indicated that participants found the app to be of good quality and believed it had a beneficial impact on SCD-specific awareness, knowledge, attitude change, intention to change, help-seeking behaviors, and likelihood of self-management behavior change. These findings were supported by postintervention interviews in which all participants were satisfied or very satisfied. Prior feasibility studies of eHealth interventions in SCD also reported high satisfaction or acceptability [25]; the consistently high rates of satisfaction across eHealth platforms are indicative of the need for this type of intervention to improve SCD care.

For implementation, few participants reported technical issues, and of those who did, all but one remained in the study till the end-of-intervention, suggesting that technical difficulties were not deterrents to continued participation. This finding is divergent from studies on other eHealth interventions for SCD in which technical difficulties led to participants discontinuing the use of the intervention or becoming less involved [36,37]. A possible explanation for this difference may be that the majority of early issues in this study were system problems that were resolved quickly. Additional issues not reported to study staff were discovered during postintervention interviews, which may indicate underreporting of issues to study staff, resolution of issues prior to participants reporting problems, or participants understanding that a new app was being tested and problems were likely to occur.

Despite all interview participants reporting that they would continue to use the app, actual use went from 45% from 6 to 12 weeks to 37% from 12 to 24 weeks. The decrease in app use over time was similar to findings on sustained use in other feasibility and pilot studies of eHealth interventions in SCD [27,28,30,37,38]. Despite the decreased use over time, findings from postintervention interviews support the possibility that app use created new habits and/or behavioral changes that may be lasting, thereby allowing families to sustain perceived benefits without needing to keep using the app. For example, by tracking pain and actions taken for pain management, participants may have learned new nonpharmacological strategies to successfully mitigate minor or oncoming pain that can be applied without continuing to use the app. A longer prospective study is needed to evaluate the possibility of sustained behavior change.

The application of a user-centered design has been suggested as one method for promoting sustained use of an mHealth intervention [39]. While a user-centered design was applied in this study [13], incorporating additional suggestions from participants obtained during postintervention interviews may increase sustained use. For example, while the app included some aspects of gamification, such as a customizable avatar, one suggestion made by multiple participants was the incorporation of more enhanced gamification, which may increase the intention to use the app, particularly in younger populations [40]. Other strategies for increasing sustained use include providing greater context to health data tracking and monitoring and increased incorporation of interaction with health care professionals [41]. The app included communication with a provider, but this was the least frequently used component of the app, likely because participants had multiple established methods in place for contacting providers (e.g., phone and text) prior to the study. One method for increasing the interaction with providers may be to share the data generated via the app during routine health care appointments to communicate everyday health and disease and symptoms self-management. In addition, during postintervention interviews, several participants admitted to being unaware of some features of the app. A critical revision to the app is an embedded tutorial users can access on demand to supplement the instructional session and written instructions, which may enhance fidelity to the intervention, increase rates of use, sustain use over time, and potentially influence health outcomes.

The findings of the study should be considered in the context of limitations. One potential limitation was the use of convenience sampling from children and caregivers attending clinic appointments, which may have led to a sample of participants who are more highly engaged in the child’s care. In addition, because participants were recruited from a single clinic, and all were African American and non-Hispanic residing in the Southeast, conclusions on feasibility may not be generalizable to other settings with more diverse patient populations. Future studies should include multiple sites with greater diversity in patient populations to enhance generalizability. Participants were instructed to use the app as they would in everyday life and included a wide range of child age and developmental level, leading to variation in which members of the dyad used the app. Additionally, it was unknown how many children used the app on a parent’s device. This limitation prevented full characterization of use by each member of the dyad. Another potential limitation pertains to the subsample of participants included in the postintervention interviews. Only participants who completed the end-of-intervention study visit were approached for an interview. While this strategy was designed to include participants who would have greater familiarity with the intervention, the participants who did not continue till the end-of-intervention and with a less favorable view of the intervention may have been missed. In future studies, participants who had minimal to no app use or who did not complete the study should be included in the interviews to explore all perspectives of the intervention and to better understand the barriers or challenges with sustained use of the intervention. In addition, the MARS was administered to caregiver participants only and satisfaction data from interviews were only obtained from a subset of participants, which may not adequately represent satisfaction from the full sample. Future studies should include the assessment of satisfaction with either all participants or a greater representation of the sample. Finally, the linear nature of the app led to limitations in the evaluation of the use of features within app components, such as the avatar within the I’m in Pain component.

Theory-based self-management interventions for children and adolescents with SCD and their families are needed to facilitate the development of effective self-management behaviors, leading to improved health and utilization outcomes. mHealth is a well-suited delivery mechanism for self-management interventions to be used in real-world settings. While eHealth and mHealth interventions have been developed for children with SCD and their families, few have undergone effectiveness testing in large-scale trials. Findings from this study indicate that implementation processes were feasible and may be applied with modifications to an adequately powered effectiveness trial; similarly, the self-management intervention was feasible for use by children with SCD and their families with modifications to improve sustained use.

Compliance with Ethical Standards

Conflicts of Interest: J.K. has received honoraria from Novartis Pharmaceuticals Corporation, MEDSCAPE, MD Magazine, Terumo, and Bluebird Bio, Inc. All other authors have no conflicts to disclose.

Author Contributions: S.P., J.K., M.M., K.R., M.J., T.K. were responsible for the conception or design of the work; S.P. and M.J. acquired the data; S.P., A.G., and M.M. analyzed the data; S.P., J.K., M.M., K.R., A.G., T.K. interpreted the data; S.P. drafted the work; S.P., J.K., M.M., A.G., K.R., M.J., T.K. revised the work for critically important intellectual content; S.P., J.K., M.M., A.G., K.R., M.J., T.K. approved the final version to be published.

Ethical Approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The study was approved by the Institutional Review Board at the Medical University of South Carolina, Pro00062837 and Pro00068250. This article does not contain any studies with animals performed by any of the authors.

Informed Consent: Informed consent was obtained from all individual participants included in the study.