Patient Engagement and Medication Consistency with Adolescent Heart Transplant Recipients

Abstract

This study explored the relationship between in-app messaging and patient time between doses of medication during mHealth video directly observed therapy (vDOT) intervention. The vDOT intervention sought to improve medication adherence among adolescent heart transplant recipients. The study utilized a 12-week vDOT pilot with ten adolescent patients, focusing on tacrolimus, a critical immunosuppressant medication. vDOT aimed to enhance patient engagement, provide interpersonal support, and ensure medication-taking consistency (inter-dose timeliness and intervals between doses). Results demonstrated the relationship between in-app communication and more consistent medication-taking behaviors, revealing that messages containing praise, rapport-building, and medically relevant information were positively associated with patients taking medication every 12 hours. The findings suggest that integrating direct observation and personal engagement via mHealth apps could be a promising strategy for managing medication adherence in pediatric transplant care.

I. Introduction

A. Non-adherence and Pediatric Transplant Care

Success and improvement in post-transplant care has contributed to higher survival rates and increased quality of life for pediatric patients following receipt of an organ transplant [1–3], yet concerns remain about medication adherence and consistency of medication-taking behavior [2, 3]. For pediatric patients and especially those in adolescence, poor adherence to medication regimens significantly predict poor post-transplant outcomes including acute rejection, hospitalizations, and even mortality [1, 2]. Adolescent transplant recipients may exhibit non-adherence with immunosuppressant medication as high as 40% to 60% [2, 4]. Adolescents are estimated to be about twice as likely as younger pediatric patients to exhibit medication non-adherence [5], experience more rejection episodes [6–8], and have significantly earlier incidents of the first rejection episode [8].

Post-transplant immunosuppressant medication maintenance therapy for pediatric heart recipients typically involves a calcineurin inhibitor such as tacrolimus [9]. Medication adherence includes consistency in tacrolimus dose timing such as taking the medication twice a day in 12-hour intervals. Success with a daily medication plan leads to stable tacrolimus levels and sufficient bioavailability of the medication within the blood. Inconsistent medication-taking can lead to fluctuations in the absorption rate, affect drug levels in the blood, and increase the likelihood of adverse effects such as rejection [2, 10, 11]. Tacrolimus taken in twelve-hour intervals helps to achieve recommended stable blood trough concentrations [12–14]. A single missed dose of tacrolimus may significantly impact bioavailability, such as a 49% decrease in simulations of kidney transplant patients [15].

Neglecting or inconsistency with tacrolimus doses can lead to gradual harm that often remains undetected until complications arise. Routine monitoring of medication blood levels helps assure that the drug concentration is within the safe and effective range, allowing for dosage adjustments that ensure effectiveness while reducing the risk of harmful effects. However, even with blood level monitoring, continuously tracking and even improving medication adherence presents an ongoing challenge for both pediatric patients and their transplant care teams.

B. Mobile Health and Medication Adherence

Intervention efforts to improve medication adherence and consistency in medication taking-behavior have not been successful among higher risk adolescents transplant patients [1999, 15]. These challenges have led multidisciplinary researchers and transplant care teams to integrate technology and mobile health (mHealth) approaches into many chronic disease management interventions, including those for cardiovascular disease and cystic fibrosis [16, 17] and even pediatric transplant patients[18, 19]. Studies have emphasized the need to tailor intervention efforts to pediatric and adolescent patients to improve medication adherence and bolster the effectiveness of behavioral interventions. Patient engagement, interpersonal support, and efforts to modify multiple identified barriers have been identified as key aspects of any successful intervention [20–23]. Each of these are possible within mHealth technologies and approaches which may also facilitate realtime monitoring and observation of medication-taking behavior. Patient-facing mHealth platforms and user interfaces may act as tools for improving patients’ organizational skills, planning around medication-taking, and help raise patient conscientiousness around their medication regimen [23, 24].

mHealth interventions have the ability to increase patient engagement and offer interpersonal support, yet maintaining engagement remains a significant challenge within intervention efforts [21, 22, 25]. Engagement is a multi-faceted concept encompassing both practical aspects like system use and psychological elements such increased involvement and self-efficacy to change health behaviors [26]. Short et al. [27] highlighted the importance of both microlevel engagement (i.e., use of system features and the degree of user interest and focus during activities), and macrolevel engagement (i.e., motivation for behavioral change) in enhancing the effectiveness of mHealth interventions for behavior modification. Patients’ proactive engagement facilitates their motivation to manage their care [28]. mHealth apps may increase the likelihood of patients intentionally adhering to their medication plans when interactive communication features and social opportunities are incorporated in the intervention and mHealth design [29].

C. Video Directly Observed Therapy Intervention

An mHealth application supporting a video directly observed therapy intervention (vDOT) was developed by Scene Health Inc. and includes features which facilitate patient engagement, rapport-building, and interpersonal support around medication adherence and their care. Through the app, patients are able to record and submit videos demonstrating dose-by-dose videos of themselves taking their immunosuppressant medication [30]. These medication adherence videos are then reviewed and approved by a healthcare professional (e.g., transplant nurse) through a centralized platform asynchronously. This vDOT mHealth app also allows patients to track their progress and report symptoms of illness or medication side-effects. Patients and transplant team members may message and communicate through the mHealth app (Figure 1). Transplant nurses approving videos and other transplant team members interacting with the patients receive training on interpersonal engagement and possible ways to interact with, build rapport, and encourage patients via in-app messaging. Altogether, these features support macrolevel engagement, can improve patient motivation for behavioral change, foster their active participation in care coordination, and increase their treatment self-efficacy.

Figure 1.

In-App Patient and Nurse Messaging Interface

For transplant care teams, the vDOT app and approach to promoting better medication adherence represents a supportive feedback system for the patient. Messaging the patient and in-app reminders serve as behavioral cues for medication-taking and may help to increase consistency in medication-taking with the aim of maintaining consistent dose intervals (e.g., every 12-hours for a twice-a-day medication). The active encouragement offered by the care team through the mHealth app can support development of the patient’s organizational skills, planning, and conscientiousness around medication taking. Realtime data from the provider-facing side of the platform can assist transplant team members in identifying emerging patterns of non-adherence and inconsistency and swiftly offer patients additional support and care.

II. Prior Work with vDOT

A. Pilot Study of vDOT

Ten adolescent heart transplant patients from a children’s hospital in the southeastern United States who were exhibiting poor medication adherence were recruited for a pilot study of the vDOT mHealth intervention [31, 32]. Adolescent patients and caregivers reported high levels of intervention acceptability over the course of the 12-week vDOT protocol, with 8 patients completing all 12 weeks. These patients submitted as evidence of adherence 1211 of a possible 1344 videos equating to 90.1% adherence rate for twice-a-day medication. When including the two patients who discontinued the intervention given their time in the intervention, the adherence rate was 82.9% for all possible adherence videos. Patient medication level variability index [1] (MLVI, a biomedical indicator of tacrolimus blood levels) improved by 21.7% reduction among these 10 patients. Scores from patient self-reports and parental proxy-reports demonstrated significant improvements from pre- to post-intervention surveys concerning perceived barriers to medication adherence and transplant-specific health-related quality of life. Most importantly, no patients who completed the 12-week vDOT protocol experienced transplant-related hospitalizations due to rejection in the 6-month period following the intervention. In comparison, 83.3% of those patients who did not use the vDOT or discontinued participation (n=6) were either hospitalized or experienced episodes of rejection. Unfortunately, two of these patients passed away secondary to complications associated with coronary allograft vasculopathy.

B. In-app Communication and Patient Engagement

We conducted a qualitative analysis of in-app messaging data from the pilot study to examine the interpersonal and supportive interactions between the adolescent patients and nursing staff during the 12-week vDOT intervention [32]. A total of 894 in-app messages exchanged by the 10 participating patients were extracted, coded, and analyzed using content analysis [33] to understand the content, purpose, and context of patient engagement and interpersonal support. Two researchers coded the patient-nurse messages independently using an inductive and iterative process and achieved high inter-rater reliability. Each portion of each message exchanged was given a separate code depending on the content (i.e., no multicoding).

Messages were categorized into three broad categories: 1) interpersonal support and rapport-building, 2) medically-related questions and patient education, or 3) supportive information about the mHealth intervention and functional aspects of the vDOT app. Patients receiving more messages with praise, gratitude from the nurses, and rapport-building tended to be more adherent with their medication videos. The results of the study and analysis demonstrated the richness and depth of the interpersonal exchange and patient engagement that occurred during the intervention and their relationship to intervention success.

C. Inter-Dose Timing and Deviation in Medication Adherence

The vDOT pilot data also provided data on whether patients consistently took their medication 12 hours from their previous dose. All submitted videos demonstrating medication adherence included time and date information used for calculation of inter-dose timing and deviation from 12-hours from the patients’ last doses [34]. Results indicated significant variability with regard to when adolescents took their medication and how much these patients deviated from a 12-hour dosing interval. Greater deviation from a 12-hour dosing interval was more prevalent among older patients and those reporting greater barriers to medication adherence. More consistent inter-dose timing was associated with improvements in MLVI and no experiences of rejection and hospitalization. Findings from these analyses demonstrated the importance of tacrolimus dose timing, the impact of greater deviation from 12-hour intervals, and on the utility of vDOT mHealth app to observe and record exact dose timing data.

D. Patient Engagement and Inter-Dose Timing

The purpose of this study and subsequent analyses are to examine the direct relationship between the in-app messaging during the vDOT and individual patient deviation from a 12-hour dosing interval. Interpersonal support and engagement can be powerful tools for behavior change interventions, thus the nature and degree of in-app messaging was hypothesized to be associated with consistency of inter-dose timing.

III. Methods

Data from the 12-week asynchronous DOT intervention pilot was analyzed in the current study. The intervention employed a single-group design involving 10 transplant recipients ranging in age from 10 and 21 years and their parents [31, 32]. Enrollment into the DOT program occurred on a rolling basis and remained open until 10 patients initiated participation. Patients and their parents completed baseline measures of quality of life and medication adherence. Feasibility and acceptability were assessed through phone interviews at the 3-week mark. Post-intervention surveys and exit interviews were completed by patients and parents at the conclusion of the 12-week period.

A. Sample

Adolescents aged 10–21 years with documented evidence of or concern for medication non-adherence were recruited from the outpatient clinic during routine visits to their pediatric heart transplant center. Eligibility criteria included recipients who had undergone a heart transplant more than 6 months prior and were considered medically stable by the transplant team. Medication non-adherence was defined as medication level variability index [1] values ≥ 2.0 or concerns reported by the patient’s transplant care team. Patients exhibiting medication nonadherence were recruited to provide a representative sample of high-risk patients. Nonadherent patients are less likely to consistently attend clinical care and be identified for study participation. Intentional identification and recruitment of nonadherent patients offered greater internal and external validity of potential results. Due to the communicative features of the intervention and the availability of only English-speaking staff, non-English speaking participants were excluded.

B. Measures

1). Inter-dose Timing, Intervals, and Deviation Measures

As evidence of medication adherence, patients submitted videos in accordance with their prescribed medication regimen, which, in the current study, consisted of twice-daily tacrolimus. These videos, along with their precise submission times, were used to calculate the time elapsed between instances of tacrolimus administration. The time lapse between observed medication taking was computed for each dose for each patient, with an ideal interval of 12 hours. For analysis purposes, deviation from the 12-hour benchmark was used. Deviation was determined as the absolute value of the observed interval between tacrolimus doses minus 12 hours. For example, doses taken at either 8 or 16 hours after the previous dose were equally regarded as a 4-hour deviation from the 12-hour dosing interval.

2). In-app Messaging and Types

Content analysis was used to examine the content, purpose, and context of messages exchanged between 10 patients and their reviewing nurses [33]. Using initial codes derived from the Pervasive Systems Design model [35] and other relevant theories and research, the first two authors developed a codebook a priori. Independently, they coded the data, with up to 3 codes selected for each text as each message could contain several sentences and sentiments. An inductive, iterative approach was used to code the data and devise additional codes for guiding a thematic analysis of the 894 in-app messages comprising the dataset. Any disagreements in coding were addressed through iterative discussion. Following the coding phase, inter-rater reliability was calculated and was found to range from 81.5%−100%.

C. Analysis

Descriptive statistics were used to characterize the sample, number and type of in-app messages exchanged, and the inter-dose timing of immunosuppressant medication. Due to the low sample size (n = 10), non-parametric Spearman’s rho (ρ) correlation coefficient was used when estimating the correlation between the patient mean deviation from their inter-dose interval of 12 hours and each the number of messages sent and the number per category and sub-type.

IV. Results

The 10 patients who initiated the 12-week DOT program (Table 1) ranged from 12 to 21 years of age. A majority of these patients were female, had private insurance, and had received a pre-transplant diagnosis of congenital heart disease or cardiomyopathy.

Table 1.

Patient, Dosing, and Message Characteristics

Patient Characteristics	M±SD, n(%)
Patient age, years, M±SD	16.60±2.95
Median	15
Range (min. — max.)	12 – 21
Gender, female, n (%)	6 (60.0%)
Race, minority, n (%)	2 (20.0%)
Parent age, years, M±SD	39.29±3.35
Type of insurance at transplant, n (%)
Medicaid	4 (40.0%)
Private	6 (60.0%)
Inter-dose interval time, M±SD	13.42±7.33
Median	12.02
Range (min. — max.)	0.97 – 167.51
Inter-dose deviation from 12-hours, M±SD	2.11±7.16
Median	0.38
Range (min. — max.)	0.00 – 155.51
Number of unique days with messages per patient,	48.20±11.63
M (SD)
Range of messages sent per patients, days	33–70
Messages sent, total, n	894
Sent by patient	227 (25.4)
Sent by nurse	613 (68.6)
Sent by Scene Health staff	54 (6.0)
Proportion of days which message was sent
Sent by patient	15.4%
Sent overall	65.6%
Timing of messages sent by patient
1:00am — 8:59am	8 (3.5)
9:00am — 11:59 am	64 (28.2)
12:00pm — 4:59pm	36 (15.9)
5:00pm — 7:59pm	11 (4.8)
8:00pm — 12:59am	108 (47.6)
Messages by Content Area
Interpersonal rapport and support
Praise	210 (23.5%)
Rapport building	318 (35.6%)
Empathy	37 (4.1%)
Gratitude	146 (16.3%)
Medically related
Medical problems	6 (0.7%)
Patient education	39 (4.4%)
Intervention related
Requesting communication	15 (1.7%)
Feedback	174 (19.5%)
Functional information and protocol	176 (19.7%)
Questions or request for information	174 (19.5%)
Information exchange or answers	218 (24.4%)

A. Patient and Nurse Messaging

Ten patients and reviewing nurses sent 894 messages during the 12-week intervention (Table 1). Patients sent 227 messages (25.4%) and 613 were sent by the nurses (68.6%). The mean number of exchanged messages per patient was 89.30±17.96 with a mean of 27.50±10.23 sent by the patients and 61.30±16.12 sent by the nurses. Patients sent almost half of their messages to the nurses during the evening hours (n = 108, 47.6% of 227). The content areas and subcategories are provided in Table 1.

B. Patient Inter-Dose Timing Intervals

The 10 patients submitted a total of 1342 videos which were counted as evidence of medication adherence. Calculated in 12-hour intervals, the mean time between doses was 13.42±7.33 hours with a median of 12.02 hours. Patients deviated from their scheduled 12-hours by a mean of 2.11±7.16 hours (22.8 minutes) with a median value of 0.38 hours.

C. Message within vDOT and Dose Timing Intervals

Despite the low sample size, several message types were significantly associated with patient deviation from inter-dose interval of 12 hours (Table 2). The overall number of messages exchanged did not significantly correlate with inter-dose timing deviation (ρ = .198, p = .583). However, the number of messages patients and nurses exchanged which were characterized as rapport-building had a moderate to strong correlation (ρ = −.571, p = .084) indicating greater rapport-building efforts between nurses and patients was associated with more consistent dosing intervals. Praise (ρ = −.377, p = .283) and nurses offering gratitude (ρ = −.561, p = .092) had similar associations with dosing deviation from 12-intervals although these were not statistically significant given the small sample size. The strongest and statistically significant correlations were found between deviation from the 12-hour dosing interval and messages in which the nurse requested communication with the patient (ρ = .722, p = .018), provided functional information concerning the patient’s protocol (ρ = .742, p = .014), and asked direct questions to the patient (ρ = .706, p = .023). In these cases, patients who experienced greater dose timing consistency challenges were those requiring greater direct support from the supportive nursing staff within the vDOT intervention.

Table 2.

Corelation of Messages and Inter-Dose Deviation

Message Categories and sub-types	Spearman’s rho (ρ)
Total number of messages
Interpersonal rapport and support
Praise	ρ = −.377
Rapport building	ρ = −.571+
Empathy	ρ = .001
Gratitude	ρ = −.561+
Medically related
Medical problems	ρ = .080
Patient education	ρ = .076
Intervention related
Requesting communication	ρ = .722*
Feedback	ρ = .261
Functional information and protocol	ρ = .742*
Questions or request for information	ρ = .706*
Information exchange or answers	ρ = .012

^*p < .05,

^**p < .01,

^***p < .001,

⁺p < .10

V. Discussion and Conclusion

The findings from this current analysis and prior research within this vDOT pilot study [31, 32, 34, 36, 37] support the promise of vDOT to promote medication adherence among a high-risk sample of pediatric heart transplant recipients. With the importance of medication adherence and consistency of medication dosing, the connection between interpersonal aspects of the vDOT intervention and mHealth approach and inter-dose timeliness is an important aspect to explore. Results from the current study further support the promise of vDOT to increase patient engagement and medication adherence.

Incorporating direct observation of medication-taking behavior while ensuring patient engagement and interpersonal support is difficult both clinically and methodologically. Transplant care teams must monitor and assess patient progress during the post-transplant period challenge, yet each method of assessing medication adherence have distinct disadvantages [18, 38]. The results demonstrate the ability vDOT to allow direct observation of medication taking-behavior while facilitating patient and provider communication and social support. Twice-daily check-ins offer direct assessment of medication dosing consistency through unobtrusive methods. Inter-dose intervals and clinical monitoring of medication adherence through vDOT offers multidisciplinary transplant care teams the option to provide immediate support and action when patients experience barriers to adherence and consistency in daily routines around medication.

Several limitations to the current study must be considered. Despite over 1300 dosing submissions and nearly 900 in-app messages exchanged, data collected from the pilot study of the 12-week DOT intervention were from only 10 patients and from a singular transplant center. Of those recruited to participate, four patients did not initiate the vDOT program and an additional two patients did not complete the intervention. Attrition was associated with older adolescent patients and those with higher MLVI values [36]. Additionally, all analyses are correlational, meaning that the temporal order of factors and any causal relationship cannot be determined. Currently, a larger and more robust randomized control trial has begun with plans to recruit a larger sample of adolescent across multiple pediatric heart transplant centers to offer a more robust examination of patient engagement, inter-dose timing, and efficacy of vDOT to promote medication adherence and long-term post-transplant outcomes [37].

The findings show that vDOT interventions delivered through an mHealth app can facilitate interpersonal support and engagement and further promote positive behavior change. Interactive dialogue, direct communication, and exchange of more personalized feedback may motivate patients to internalize the need and responsibility for medication adherence and self-monitoring of health behaviors. Improvement in organizational skills and conscientiousness around their medication may support adherence and behavioral consistency – which are key aspects of post-transplant care. Many of the messages within the app included language that built rapport, complimented patient efforts, and shared medically-relevant information. Nurses frequently offered commendations, feedback, and posed helpful questions to patients, serving as cues for medication adherence. Adolescence is a challenging period for these patients [2, 39–41], and vDOT may hold some promise in promoting their long-term quality of life and medication adherence during post-transplant care.