In the United States, poor adherence accounts for up to 70% of all medication-related hospital admissions [1], resulting in $100 billion in health care costs annually [2]. An estimated 50% to 80% of pediatric patients are nonadherent to medications [3-6]. Reasons for nonadherence are multifactorial. The most important determinants of nonadherence are consistently documented as complexity and duration of treatment regimens, as well as forgetfulness [7-9]. Thus, children undergoing difficult hematopoietic stem cell transplants (HSCTs) who require medication indefinitely are at high risk for medication nonadherence.

HSCT is the transplantation of hematopoietic stem cells derived from bone marrow (ie, “bone marrow transplant” or peripheral blood stem cells). HSCT became the common treatment for life-threatening malignant (eg, leukemia, lymphoma) and nonmalignant (eg, aplastic anemia, sickle cell disease, immune deficiencies) disorders in the mid-1900s. Recipients of bone marrow transplants must adhere to multifaceted outpatient regimens that require strict personal hygiene, environmental restrictions, and complex medication regimens. These include wearing masks, avoiding large crowds, frequent mouth care, and taking multiple time-sensitive medications throughout the day.

Lack of adherence to any of these regimens can be life-threatening [10]. Adherence to immunosuppressant medications during the acute phase (ie, the first 100 days) post transplant is critical to prevent graft versus host disease and avoid graft failure. Studies in children with HSCT report suboptimal adherence rates (52%-78%) that worsen over time [11-13]. Nonadherence to oral medications has been associated with a greater incidence of infections in children during the outpatient HSCT treatment phase [12]. Providers also report adherence as a major concern for outpatient pediatric HSCT recipients [14]. However, no studies have examined adherence specifically to immunosuppressant medications, which are key to engraftment and ultimately survival.

Multiple factors influence adherence, but, ultimately, the final common pathway to adherence is human behavior [15]. For children, caregivers are ultimately responsible for adherence, including refilling prescriptions, retrieving medications, and administering them correctly. In a high-risk HSCT population, caregivers are susceptible to fatigue and stress from being isolated with their child due to infection risk and independently managing complex treatment regimens at home, often with limited time and support. Unfortunately, behavioral interventions employed to address adherence are often difficult to deliver and manage in the context of these daily tasks. Specifically, these interventions can involve 8 to 12 face-to-face sessions, tracking behaviors, practicing skills, and other homework, which can be overwhelming to caregivers.

Alternatively, behavioral economics (BE) theory suggests that small “nudges” can produce and sustain behavior change [16]. A BE approach is a significant paradigm shift from the complex cognitive behavioral theories that inform most adherence interventions. Instead, BE assumes decision-making can be influenced through low-intensity interventions to lead patients to optimal choices [17]. Adherence to medication and exercise programs using BE-designed interventions in adults have some initial promising data [18-21]. Within pediatrics, for example, BE has been successful in increasing fruit consumption and improving vaccination rates [22-24].

With mobile health (mHealth) access nearly ubiquitous and an estimated 250 billion mHealth app downloads in 2022 [25], technology has great potential to improve adherence [26]. mHealth interventions based on BE principles may be ideal, as they provide sophisticated yet simple “nudges” to individuals [27]. Such reminder apps target forgetfulness, a common reason for poor adherence [28,29]. These mHealth interventions are ideally suited for complex HSCT regimens requiring multiple-dose medications that are time sensitive.

A recent systematic review in solid organ transplant reported the most successful interventions, resulting in increased adherence and improved clinical outcomes (eg, complications, hospital admissions, survival) have used mHealth reminders and a direct measure of adherence [30]. Although direct measures of adherence (eg, serum assays) can avoid bias inherent in self- or parent reports, no studies have used direct measures of adherence or mHealth interventions to promote immunosuppression adherence in the acute phase after an HSCT (ie, discharge to day 100). Furthermore, the feasibility, acceptability, and efficacy of BE-designed mHealth interventions have not been examined in the HSCT population. Thus, this study utilizes a BE approach to mHealth app development, as well as an iterative approach of cyclical evaluation and revision, to maximize the usability and adoption of the digital intervention, BMT4me.

BMT4me (Figure 1) is a digital health intervention designed to aid in medication management, optimize adherence, and track symptoms or side effects in real time. The app reminds users of medication doses, records the time medication is taken, and asks reasons for missed doses through pop-up notifications. Medications can be typed in the app by the user or entered using an image-to-text option, which converts the medication label into text. As symptom severity can affect adherence and inform the need for medical intervention, it is illustrated by both corresponding emojis and numbers on a 1-to-10 scale to aid understanding and for convenient entry. Additional features include a notes page for recording any details of care to communicate with the provider and the ability to upload pictures. The app can be downloaded to both iOS and Android devices. Currently, availability is limited to the English language, and no clinical data is shared with the medical team during the research study.

The overall aim of this mixed methods study is to design and test a theory-driven mHealth app that can virtually assist caregivers of children post HSCT in improving medication adherence, tracking symptoms, and recording pertinent medical information via the journal feature. This app is different from other medication adherence apps available to patients with cancer in several ways: (1) features (eg, reminders, engaging emojis) in the app are designed to support medication adherence by applying BE principles to provide reminders and tailored notifications for caregivers; (2) it is specific to children after an HSCT; and (3) it provides more than medication reminders, with the design encouraging caregivers to engage with the app via symptom tracking and a journal feature that allows uploading of pictures.

The larger mixed methods mHealth app development study consists of 3 phases: (1) prototype development, (2) feasibility and usability, and (3) usability and clinical efficacy (Figure 2). This protocol corresponds to phase 2, feasibility and usability among 15 caregivers immediately upon discharge from the inpatient unit post HSCT. Inclusion criteria include (1) caregiver of child aged 2 to 18 years undergoing allogeneic transplant, (2) access to a smartphone (Android or iOS), and (3) English speaking. Findings from phase 1 informed phase 2, and thus findings from phase 2 will lead to further refinement prior to phase 3. In phase 3, the smartphone app will be compared to usual care in a pilot randomized controlled trial.

All phases of the study have been approved by the Nationwide Children’s Hospital Institutional Review Board (approval number STUDY00000910).

This is a convergent mixed methods design focused on analysis of quantitative and qualitative data simultaneously collected to obtain feedback on the adherence mHealth app interface, usability, and flow and content of the software. Usability testing is the principal means of determining if the system meets its intended purpose [31]. Purposive sampling will be used to identify caregivers of children across the spectrum of age, sex, and diagnoses of pediatric patients treated with HSCT. Caregivers will be evaluated for engagement and understanding of the app in the following steps (Figure 3):

At the conclusion, caregivers will complete a 15-minute semistructured interview related to participation in the intervention, as well as the System Usability Scale [32]. Open-ended feedback related to suggestions for further app development will undergo qualitative analysis.

Trained research assistants will recruit and onboard eligible participants. The participants will sign written consent forms. The research assistant will then conduct one-on-one audio-recorded sessions, which will include a facilitator and a note taker. A project scientist may also participate. The measures listed below will be collected either electronically or in paper-and-pencil form from each caregiver. All participants will receive US $25 at enrollment and US $25 at the conclusion of the study after the exit interview and measures are complete.

All interviews will be transcribed verbatim. During qualitative analysis, investigators will use an iterative process of reading, summarizing, and rereading data. The qualitative team is composed of 3 individuals with formal research training in psychology, behavioral health, and nursing. The interview transcripts will be organized and coded using NVIVO software (QSR International). In-vivo coding will occur as new themes emerge from the data. The data will be triangulated by comparing and integrating quantitative survey data with qualitative findings.

Quantitative data will be summarized with descriptive statistics (frequencies, means, and SDs). During the passive use observation period, passive data modules will capture phone activity and caregivers’ application use (eg, time, date, duration of use). Descriptive statistics will be used to analyze phone activity. Correlation analysis will be used to investigate use behavior over time. Acceptability will be assessed by averaging total scores from the System Usability Scale. Consistent with the literature [36], scores >68% on the System Usability Scale will be considered acceptable. The proportion of participants that enroll and complete the study will be examined to assess study feasibility.

The nature of the study is exploratory rather than confirmatory. Thus, the objective of the analysis is effect size estimation, rather than formal hypothesis testing. Threats to power (eg, participant attrition due to patient death, early taper of immunosuppressant medication) are not a primary concern. We will compute pre-post effect sizes (eg, a standardized mean difference) to inform the future randomized controlled trial (phase 3) assessing the efficacy of the intervention.