Adults recently diagnosed with an ischemic stroke who used an artificial intelligence (AI) platform on their mobile devices had a 50 percent improvement in adherence to their anticoagulation therapy, according to a randomized trial.

The study used the AI platform developed by AiCure, which visually confirms medication ingestion on smartphones. AiCure sponsored the trial, while the National Center for Advancing Translational Sciences
 provided a grant to fund the study.

Lead researcher Daniel L. Labovitz, MD, of the Montefiore Medical Center in the Bronx, New York, and colleagues published their results online in Stroke on April 6.

“In the absence of routine laboratory monitoring, artificial intelligence has the potential to automate a critical component of care—adherence monitoring—and provide continuity of care between visits to ensure patients persist with their therapy and get full therapeutic benefit,” Labovitz said in a news release.

The researchers enrolled 28 patients who were diagnosed with ischemic stroke and received warfarin, dabigatran, rivaroxaban or apixaban. They randomized 15 patients to daily monitoring by the AI platform or to no daily monitoring for 12 weeks.

At baseline, the mean age was 57 years old, and 53.6 percent of patients were women.

The patients in the AI group received smartphones, which had software that identified the patient, medication and confirmed ingestion and provided medication reminders and dosing instructions. If patients had late doses, they were notified within the hour and before the end of the dosing window. Clinic staff members received text messages or emails if doses were missed, late or based on incorrect usage.

After 12 weeks, the mean cumulative adherence based on pill count was 97.2 percent in the AI platform group and 90.6 percent in the control group. Based on plasma samples, 100 percent of patients in the AI platform group and 50 percent of patients in the control group were adherent. The researchers marked plasma samples as adherent if the drug concentration levels were above the minimum required therapeutic range.

Of the patients who were nonadherent based on plasma samples, all were in the control group and all were prescribed direct oral anticoagulants.

For the 19 patients who received direct oral anticoagulants, the mean cumulative adherence was 90.1 percent based on visual confirmation of drug administration using the AI app. The mean cumulative adherence based on pill count was 96.4 percent in the AI platform group and 90.9 percent in the control group.

The researchers also asked patients in the AI platform group to complete questionnaires before and after the study. They found that 73.3 percent of patients before the study and 83.3 percent of patients after the study rated the AI platform as “extremely good” as a medication management tool and a means to improve the doctor/patient relationship.

“Few studies have deployed smartphone apps in middle-aged or elderly populations,” the researchers wrote. “Consistent use and general likability of the AI app over 12 weeks underscores the possibility of harnessing new technologies to optimize adherence in patients on direct oral anticoagulants for whom routine laboratory tests are not good indicators of adherence. AI platforms have the potential to accurately monitor medication ingestion and change patient behavior.”