Mobile video directly observed therapy

A mobile video directly observed therapy mHealth application developed by Scene Health Inc. offers patients direct tracking of dose-by-dose medication-taking behaviours and interpersonal communication and support from their transplant team. ^{Reference Killian, Clifford and Gupta4,Reference Killian, Clifford, Lustria, Skivington and Gupta5} Within the application, patients record and submit videos of themselves taking medication. Nurses review these videos asynchronously as evidence medication was correctly taken. Patients can report any symptoms of illness or medication side effects, and reviewing nurses may escalate administration problems or patient health and safety concerns to the patient’s transplant care provider.

This mHealth application provides instant feedback and dose tracking for the patient while providing the transplant team directly observed behaviour. Most importantly, the application is designed to foster communication and rapport-building with the patient. Interpersonal support and engagement were reported as valued by patients and used by nurses to offer support, encouragement, build rapport, answer medically related questions, offer patient education, and offer functional support about the therapy. ^{Reference Killian, Schelbe, Lustria, Watkivs and Gupta6} Importantly, this significant rapport and interpersonal support was developed with the reviewing nurses who were previously unknown to the patients. ^{Reference Killian, Clifford, Lustria, Skivington and Gupta5,Reference Killian, Schelbe, Lustria, Watkivs and Gupta6}

Recent advancements in the mHealth application build on trends and innovations within mHealth approaches promoting adherence. A newly implemented “video-back” feature gives reviewing nurses the additional option to send patients live or pre-recorded videos, a further avenue for developing rapport and interpersonal support. Pre-recorded videos may be from the patient’s own transplant team members or other supportive individuals (e.g., patient’s family, individuals identified as sources of support by the patient). Another key innovation is the centralised review and monitoring of all patients in the programme through one transplant centre. A singular nurse or small team may provide remote, asynchronous review and patient support for a geographically dispersed number of patients. The integration of a caregiver reminder option is another new feature designed to support behavioural change and recognises the process of gradual transition of adherence responsibility from caregiver to the patient. The mHealth application and new features are designed to address multiple, modifiable risk factors for non-adherence through actively encouraging development of organisational skills, planning, and conscientiousness around the behaviour.

Current study and procedures

The current study builds upon the prior pilot study ^{Reference Killian, Clifford and Gupta4,Reference Killian, Clifford, Lustria, Skivington and Gupta5} and uses similar methods to examine a 12-week asynchronous, directly observed therapy intervention with adolescent heart transplant recipients from four Southeast US transplant centres during a 2-year trial period. The current randomised control trial will involve 100 patients who are at least 6 months post-transplant and non-adherent (medication level variability index² > 2.0) and otherwise medically stable. Potential participants will be screened and recruited via phone calls and outpatient clinic appointments. Consenting patients aged 10–21 years will receive pre-intervention questionnaires and subsequently randomised into either the mHealth intervention or enhanced goal-setting group intervention using a randomisation procedure stratified by transplant centre (1:1 assignment ratio).

Patients receiving directly observed therapy via submission of mobile videos on a mHealth application will be given orientation materials and assistance with download and setup. Each patient’s care team will enter all medications into the provider-facing portion of the Scene platform. Following setup, the 12-week directly observed therapy intervention will begin and medication-taking monitored by centralised review within the application. The enhanced goal-setting condition will be a standard-of-care intervention including routine care and monitoring. Patients in this condition will receive standard patient education, materials on the importance of medication adherence, and information about standard medication reminder applications available for smart phones or similar devices. To avoid participant compensation as a motivator for adherence, all participants will receive the same compensation regardless of intervention condition (gift cards valued at 84USD, equal to 1USD per day of participation). Also, all patients will be provided with a phone and data plan to avoid selective sampling. A flow chart detailing study procedures and timeline is included in Figure 1. The study is registered with ClinicalTrials.gov.

Figure 1.

Participant flow diagram.

Measurement

Patient demographic and family characteristics will be collected from medical records and surveys (Table 1). Primary outcomes include number of successfully submitted videos reviewed as evidence of medication taking-behaviour and medication level variability index² scores across the pre- and post-intervention periods. Medication adherence during the directly observed therapy intervention will be considered as the percentage of submitted and accepted patient videos demonstrating appropriate medication-taking behaviour. Long-term pre- and post-intervention adherence will be measured by medication level variability index scores calculated and from at least three patient tacrolimus (i.e., Prograf or FK506) or sirolimus blood levels from each 6 months before, during, and 6 months after participation. These scores are patient-level standard deviation values across medication blood levels with higher scores indicating greater variation in blood levels and less consistency in medication-taking. Pre- and post-intervention episodes of organ rejection and hospitalisation will also be collected as indicators of post-transplant health outcomes.

Table 1.

Study measures

Patients and parents will receive invitations to complete online surveys for secondary outcomes measures before beginning medication level variability index and after completion of the 12-week programme. Patient self-reported and parental proxy-reported outcome measures completed pre- and post-intervention include measures of health-related quality of life, ^{Reference Weissberg-Benchell, Zielinski and Rodgers7} barriers to medication adherence, ^{Reference Simons and Blount8} engagement and relationship with their transplant care team, ^{Reference Hibbard, Mahoney, Stockard and Tusler9} and perceptions of quality of healthcare received. ^{Reference Rick, Rowe and Hann10} Patient engagement with the application will be measured using system usage data (e.g., amount, frequency, duration, and depth of use) and self-reports of user satisfaction.

Assuming 100 patients complete the 12-week intervention, the study is sufficiently powered (.80) to detect a 26.9% between-groups difference in post-intervention medication level variability index values as statistically significant. Evenly distributed between groups, and a rejection level of α = 0.05, the study will be able to detect a small effect size of Cohen’s f = 0.245 as statistically significant (power = .80) for secondary outcomes. We will use an intent-to-treat analytic approach to protect against bias introduced by participant attrition when estimating the effect of treatment group. We will examine pre-intervention differences between the directly observed therapy and control groups to determine the success of randomisation to generate comparable groups on observed characteristics. If necessary, sensitivity analyses will be conducted to determine the impact of baseline group imbalance by treating these factors as covariates in the modelling to adjust for their influence on post-intervention outcomes.