The Feasibility of a Telemedicine Platform to Monitor Adherence and Adverse Effects of ABL Kinase Inhibitors

Background: Poor adherence to tyrosine kinase inhibitor (TKI) therapy for chronic myeloid leukemia (CML) is one of the most powerful predictors of suboptimal responses. In fact, patients who adhere to ≤90% of imatinib (26.4% of patients in one series), virtually never achieve complete molecular response, and are at high risk (relative risk 11.7, p=0.001) of not achieving a major molecular response (Marin D et al. J Clin Oncol, 2010. 28(14): 2381-8). Despite the known risks of TKI non-adherence, there are no standardized interventions to improve compliance. eMedonline is a web-based, medication telemonitoring system that integrates smartphones, radio frequency identification (RFID), and barcode technology to optimize adherence and adverse effect reporting. In this pilot study, we tested the feasibility of this platform to track adherence and describe patient-reported outcomes while on TKI therapy.

Methods: Subjects were patients (≥18 years of age) with CML, on nilotinib or imatinib therapy. They were randomized to two groups: self-monitored drug administration versus technology-monitored drug administration (eMedonline). After three months, each group crossed over to the alternate group for another three months. Subjects in the telemonitoring arm were provided with devices, which sent alerts at the scheduled time of TKI administration. Subjects then registered taking their medication, using the radiofrequency identification component of their device. Subjects recorded symptoms (using the M. D. Anderson Symptom Inventory, MDASI-CML), as well as qualitative descriptions of general well-being in an eDiary. Subjects returned with unused pills to be manually counted by clinical research staff at the 90-day study visits -- at the time of crossover and end-of-study. We also compared adverse effects in the medical record to daily patient-reported outcomes using the eMedonline platform.

Results: A total of 24 chronic phase CML patients were enrolled at 3 different sites Rex/UNC Healthcare (12), UNC (9), and Northwestern (3). Twenty-one (88%) subjects were on imatinib, and 18 (75%) had greater than 6 months of experience on therapy prior to enrollment. Twenty-one (88%) subjects completed the study. TKI adherence rates were high, with median adherence of 100% (range 93-100%) during the eMedonline phase. However, only 11 (46%) subjects successfully completed the manual pill count. The average adherence time for individual subjects varied greatly (median 4.83 minutes; range 0.8-111). Patient-reported symptoms, via the MDASI-CML six core symptoms, were cataloged in the eMedonline e-diary by 14 subjects. Among the 9 subjects with complete patient-reported symptom data in eMedonline and in the medical record, 5 subjects self-reported ≥1 moderate-severe symptom on the MDASI-CML, and 30% (6/18) of these symptoms were also reported by the patient's provider in the medical record.

Conclusion: We demonstrate that it is feasible to both assess adherence to CML TKI therapy and capture MDASI-CML core symptoms using the telemedicine platform eMedonline. Adherence rates were uniformly high during the eMedonline phase of the study, and standard pill counts, as a measure of adherence, were less feasible - likely due to differing study sites, pharmacies, and/or medication refill schedules. Finally, eMedonline is a feasible tool to capture patient-reported medication adverse effects, many of which are not routinely discussed or documented in the context of clinical care. Further prospective study of eMedonline and other telemedicine platforms is needed to better define their role in the care of CML patients taking TKI therapies.