Introduction: Digital medicine consists of the marriage between active pharmaceuticals and wearable/ingestible sensors combined with mobile and web-based tools in the hope of improving the management of medication adherence. However, even in this space, objective adherence levels may not provide enough information to make the most informed decision.
Methods: Ingestion data from two clinical trials (NCT02722967, NCT02219009) with a digital medicine system (DMS) in patients with serious mental illness was used. The presence, or absence, of an observed ingestion on any particular day while the patient was engaged with the DMS was used to create a binary Markov Chain, which we have previously shown as a viable way of representing this data.1 The evolution of the entropy rate of each patient’s adherence chain over time was used to visualize the concept of adherence volatility.
Results: Adherence volatility was found in the presence of (relatively) stable adherence rates while on treatment regardless of the level of adherence. Such an observation may indicate impending adherence changes.
Conclusions: The adherence volatility metric provides a way of assessing a type of underlying steady-state for the observed adherence rates used to make decisions with digital medicine systems. The initial finding that adherence volatility may be identified in patients regardless of how adherent they are overall, suggests that knowledge of whether or not adherence behaviors are shifting towards some new adherence ‘steady-state’ may be a critical component in the future of digital medicine.
This poster was presented at the 32nd annual Psych Congress, held Oct. 3-6, 2019, in San Diego, California.