JMIR Publications

ABSTRACT

Background:

The emergence of Chat-Bots in healthcare is fast-approaching. Data on the feasibility of Chat-Bots for chronic disease management is scarce.

Objective:

We explore the feasibility of creating a Chat-Bot for patients with Inflammatory Bowel Diseases by categorizing electronic dialogue data from patients to their healthcare providers using Natural Language Processing.

Methods:

Electronic dialogue data collected between 2013 and 2018 from a care management platform (UCLA eIBD) at a tertiary referral center for IBD at the University of California, Los Angeles was used. Part of the data was manually reviewed and an algorithm for categorization was created. The algorithm categorized all relevant dialogues into a set number of categories using Natural Language Processing. Additionally, three independent physicians evaluated the appropriateness of the categorization.

Results:

16,453 lines of dialogue were collected and analyzed. We categorized 8,324 messages from 424 patients into seven categories, since there was overlap in these categories, their frequencies was measured independently into: symptoms (32.8%), medications (38.7%), appointments (24.5%), labs (34.0%), finance/insurance (7.2%), communications (34.9%), procedures (10.0%), and miscellaneous (10.1%). Furthermore, the algorithm showed 95% similarity in categorization compared to the 3 independent physicians.

Conclusions:

With increased adaptation of electronic health (e-health) technologies, Chat-Bots could have great potential in interacting with patients, collecting data, and increasing efficiency. Our categorization showcases the feasibility of using large amounts of electronic dialogue for the development of a Chat-Bot algorithm. Chat-Bots could allow for the monitoring of patients beyond consultations and potentially empower and educate patients and improve clinical outcomes.