Patient portal helps predict medication discontinuation

Register now

Researchers say electronic communication between patients and providers via an online portal can indicate their potential for stopping therapy.

A team of researchers at Vanderbilt University Medical Center analyzed messages sent over a 12-year period by more than 1,100 breast cancer patients prescribed hormone therapy at VUMC. The source of the de-identified electronic health record data for their study was the online patient portal My Health at Vanderbilt.

The study—published in the Journal of the American Medical Informatics Association—examined patients’ patterns of messaging with providers, the topics they communicated, and the extent to which this messaging indicated whether a patient will discontinue a prescribed five-year regimen of therapy.

“This investigation suggests that patient-generated content can inform the study of health-related behaviors,” state the study’s authors. “Given that approximately 50 percent of breast cancer patients do not complete a course of hormonal therapy as described, the identification of factors associated with medication discontinuation can facilitate real-time interventions to prevent early discontinuation.”

Also See: FDA approves digital pill that tracks patient compliance

Through their analysis using machine learning, researchers identified 10 topics in the patient messages that were determined to be associated with increased risk of discontinuing hormone therapy, while another 13 topics were found to be associated with decreased risk.

Specifically, patients who mentioned surgery-related topics or side effects caused by hormone therapy were associated with increased risk of early medication discontinuation, while patients who sought professional suggestions, expressed gratitude to the healthcare team, or mentioned drugs to cope with side effects or symptoms were associated with decreased risk of medication discontinuation.

“To our knowledge, there are no previous studies linking the content of these patient messages to health outcomes or patient behaviors,” says Zhijun Yin, the first author and assistant professor of biomedical informatics at VUMC. “This study is notable because it suggests a method for automatically calculating this risk for each patient as a matter of day-to-day clinical practice. We believe methods like these will further streamline and personalize our system of care.”

For reprint and licensing requests for this article, click here.