Problem

Combinaut is a small software development firm that produces high quality software for the healthcare industry. Health care systems like Atrium Health, one of the largest in the nation, strive to provide the highest quality care possible. To that end, they make use of a doctor finder (developed by Combinaut) that provides a searchable index of their health care providers. The system worked well for them, but they wanted their patients to be able to search using natural language in a way that would be similar to what users have become accustomed to when using search engines like Google.

Solution

Making use of an open-source Named Entity Recognition (NER) model a simple prototype was built. The sample data to test the prototype began to be gathered. The prototype validated the basic approach. As the range of inputs began to expand, a more sophisticated approach combining a basic framework consisting of a data-driven pipline of filters was developed that made use of the NER model, traditional pattern matching and data indexes that is accessed through a REST API.

The result is a more accessible and user-friendly service that delights their users, and more importantly, eases access to critical care.