JMIR Publications

ABSTRACT

Background:

Medication nonadherence occurs in as many as 50% of people yet studies for detecting nonadherence and delivering real-time interventions to improve adherence are lacking. mHealth technologies show promise to support medication adherence. Commercially available off the shelf technology shows promise for developing interventions using existing technologies that are readily accessible and easy to use.

Objective:

The purpose of this study was to evaluate the feasibility and acceptability of using a mHealth system using commercially available off the shelf technology to self-track medication adherence and deliver near real-time short message service (SMS) text message feedback about medication adherence patterns. The mHealth system consists of a commercially available smart button device to self-track medication-taking, a companion smart phone app, a computer algorithm used to determine adherence to a prescribed medication and deliver a standard or tailored SMS text messages based on the timing of medication taking. Standard text messages provided participant feedback indicating that the smartphone app registered the button press, whereas tailored text messages provided feedback to reinforce habit formation based on the timing that medications were taken.

Methods:

A convenience sample of 5 adults with chronic kidney disease prescribed antihypertensive medication participated in a 52-day longitudinal study. The study was conducted in three phases with a standard text message sent in phases 1 (study day 1 – 14) and 3 (study day 15 - 45) and tailored text messages sent during phase 2 (study days 46 - 52) in response to participant self-tracking of medication taking using the mHealth system. Medication adherence was measured using two different methods: (1) the smart button used to self-track medication taking as part of the mHealth system, and (2) electronic medication monitoring caps. Concordance between these two methods was evaluated using the percentage of measurements made on the same day and occurring within +/- 5 minutes of one another. Acceptability was evaluated using qualitative feedback from participants.

Results:

Five patients with chronic kidney disease stages 1-4 were enrolled in the study with the majority being men (60%), white (80%) and Hispanic/Latino (40%) of middle age (52.6 years; SD = 22.49; range 20- 70). The mHealth system was successfully initiated in the clinic setting for all enrolled participants. Of the expected 260 data points, 36.5% (n=95) were recorded with the smart button and 76.2% (n=198) with electronic monitoring. Concordant events (n=94) occurred 47.4% of the time and 58.5% occurred within +/- 5 minutes of one another. Participant comments suggested text messages were encouraging, and two participants thought messages that helped remind people who had a problem remembering to take medications would be a better approach.

Conclusions:

It was feasible to recruit participants in the clinic setting for a mHealth study and our system was successfully initiated for all enrolled participants in the clinic. The smart button is an innovative way to self-report measurement of adherence data including date and timing of medication taking not previously available from measures that rely on recall of adherence data. Although the selected smart button had poor concordance with electronic monitoring caps, participants were willing to use it to self-track medication adherence and found the mHealth system acceptable to use in most cases. The mHealth system has potential to provides real-time actionable information about medication taking, despite some noted limitations in the selected technologies.