Ai-Based Therapy Compliance for Heart Failure Patients in Clinical Trials and In Real-Life.

Insufficient compliance to therapies can be a challenge when treating heart failure patients. Inadequate compliance can lead to increased exacerbations, reduced physical function and higher risk of hospitalization and death.

Flaskdata.io is an AI-based solution that improves compliance for patients in clinical trials especially for pill or multi pill therapies. It integrates packaging, point of care and is based on break-through technology which can be adapted for commercialization post launch.

The compliance problem is reframed as an adaptive control process for a system formulated in analytic terms. We write an equation for a control algorithm:

dx/dt = g(x,y, r(t)), x(0) = c

where x(t) is the compliance state vector of the patient and y(t) is a compliance control vector and r(t) is a stochastic vector representing certain unknown effects like patient’s belief in treatment efficacy which the system learns in the course of clinical trials. The rate of change in the compliance system is measured by the derivative of the state vector dx/dt.

The control algorithm is used in a mobile agent to reinforce compliant patient behavior. In addition to enabling the agent to personalize its behavior for each patient, the state vectors enables message selection policies learned from one patient to be generalized to other patients who share the same state vector. This enables the agent to immediately reinforce patients who are recruited into the study with similar characteristics to other patients who are already enrolled.