Improving Adherence and Outcomes by Artificial Intelligence-Adapted Text Messages (Michigan)

Project Final Report (PDF, 1.68 MB) Disclaimer

This project does not have any related annual summary.

Improving Adherence and Outcomes by Artificial Intelligence-Adapted Text Messages - Final Report

Farris, K. Improving Adherence and Outcomes by Artificial Intelligence-Adapted Text Messages - Final Report. (Prepared by the University of Michigan under Grant No. R21 HS022336). Rockville, MD: Agency for Healthcare Research and Quality, 2017. (PDF, 1.68 MB)

The findings and conclusions in this document are those of the author(s), who are responsible for its content, and do not necessarily represent the views of AHRQ. No statement in this report should be construed as an official position of AHRQ or of the U.S. Department of Health and Human Services. 
Principal Investigator: 
Document Type: 
Medical Condition: 

A National Web Conference on the Role of Health IT to Improve Medication Management

Event Details

  • Date: September 13, 2018
  • Time: 1:00pm to 2:30pm
This AHRQ Web conference addressed the potential of health IT to improve medication monitoring, adherence, and medication therapy management for patients with complex conditions. Presenters discussed a text messaging system for patients with chronic conditions, the effects of a smart pillbox intervention on patient medication adherence after hospital discharge, and recommendations for clinical decision support used by community pharmacists delivering medication therapy management.
This project does not have any related resource.
This project does not have any related survey.
This project does not have any related project spotlight.
This project does not have any related survey.
This project does not have any related story.
This project does not have any related emerging lesson.

Project Details - Ended


Self-management of chronic conditions involves complex behaviors, and patients vary in their adherence to these behaviors. Patients’ failure to take their medications as prescribed is a major cause of excess morbidity and mortality and increased healthcare costs. Studies suggest that 33-50 percent of patients do not take their medications properly, contributing to nearly 100,000 premature deaths each year. Medication non-adherence is a major cause of uncontrolled hypertension, which is a major cause of stroke, coronary heart disease, heart failure, and mortality.

Text messaging, or short message system (SMS), has shown some promise in improving medication adherence for chronic conditions. However, many mobile health services lack the capacity to meet patients’ complex and changing needs. Improving medication adherence requires addressing multiple challenges because patients typically have a variety of reasons for not taking their medication as prescribed, such as beliefs about their disease and its treatment, organizational challenges, and cost barriers. In order to address these challenges, researchers used Reinforcement Learning (RL), an artificial intelligence method, to develop a model medication adherence system that automatically adapts text message communication to improve individual medication taking.

The specific aims of the project were as follows:

  • Develop RL methods for adaptive decision making in human-centered environments and demonstrate the feasibility of the resulting RL-based adaptive SMS medication adherence intervention. 
  • Demonstrate “learning” by the RL-base adaptive system using data showing adaptation of the SMS message stream according to variation across patients and over time in the reasons for non-adherence. 
  • Examine the potential efficacy of the RL-based adaptive SMS intervention with respect to improvements in medication adherence and systolic blood pressure. 

In this randomized controlled study, participants were assigned to either the Medication Event Monitoring System (MEMS) plus text messaging intervention, or to the MEMS-only control. MEMS is an electronic pill bottle cap that captures the date and time the bottle is opened. Findings indicated that self-reported medication adherence significantly improved at 3 months in the intervention group compared to the control group. The text messages and system were accepted by study participants, and almost half were interested in enrolling in a similar program post-study. The distribution of text messages themes changed over time signifying that the RL agent was learning and adapting. To further test its impact, future research should test the RL agent among a population with more variation in medication adherence.