A pilot program conducted by Kaiser Permanente Southern California has demonstrated that interactive text messaging can improve medication refill rates among Medicare patients who have one or more chronic diseases.
Kaiser Permanente Southern California partnered with mPulse Mobile, an mHealth engagement vendor, whose platform delivers text messages to patients through meaningful dialog generated by machine learning-based natural language processing that classifies patient responses into most commonly occurring categories—which, in turn, automatically triggers appropriate actions.
Patients in the study received tailored medication refill reminders through a series of messages that included prompts to authenticate by date of birth, complete a refill, ask for help, share reasons why they had not refilled, or choose to opt out of the program by numeric or textual responses on their phones.
The three-month study resulted in a 14 percentage point difference in medication refill rate between the intervention group of more than 12,000 patients who received the text messages, compared with the control group of more than 76,000 patients who didn’t receive the messages.
“The program results far exceed our expectations with 44 percent refill rate in the text message group as compared to 30 percent in the non-text group,” says Rena Brar Prayaga, lead author and behavioral data scientist at mPulse Mobile.
“In addition to the difference in refill rates, the 37 percent response rate by this older Medicare population was higher than expected, and patient feedback was very positive, with 96 percent of the patients indicating that the solution was easy to use,” she added.
Results of the study were published last week in the Journal of Medical Internet Research’s mHealth and uHealth.
In light of the success of the pilot, Kaiser Permanente is rolling out the text message intervention program to other regions besides southern California.
“As we expand the program to other Kaiser Permanente regions, we expect to rely more heavily on machine learning–based natural language processing to improve recognition accuracy,” the study indicated. “While we also rely on human intelligence to validate and handle outliers and unexpected responses, our goal is to reduce manual processing of member queries and responses to less than 5 percent in future programs.”
Register or login for access to this item and much more
All Health Data Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access