Our program, titled HappyPatient, is designed for one specific purpose: to identify how hospitals can improve the patient experience. At present, hospitals rely mainly on patient exit surveys. However, there is an untapped resource waiting to be utilized.
Online reviews currently act as a complaint board for unsatisfied patients. All too often, an unhappy customer rants online, is contacted by a hospital representative, and has their complaint addressed individually. This is a labor-intensive process that can be improved by using OpenAI’s text analysis capabilities. The OpenAI API can read through hundreds of reviews quickly and efficiently and determine precisely what factors a hospital needs to improve upon.
What HappyPatient Does
Our program enables the API to do just this. Simply provide it with a list of reviews for a specific hospital – we’ve seeded the API with a few sample reviews to give it an idea of what to do – and the API will return a list of ratings, along with specific suggestions on what to improve. These ratings and suggestions are then compiled into a concise report that provides targeted feedback without the need for reading dozens of reviews manually.
There are several benefits to using HappyPatient. First, it enables a hospital to quickly gain meaningful feedback from patients that would otherwise have to be collected manually. Second, it is easily scalable and applicable to any hospital that is reviewed online. Finally, it provides specific focus points for a hospital to improve on.
There are some limitations to this program. It cannot explain how to fix issues in a hospital; it can only tell what the issues are. Additionally, it requires that reviews be fed into it in the form of a text file. Future releases will include a web crawler to automatically find reviews from the internet, but this functionality has not yet been implemented. Finally, there is the possibility that the API will become confused by poor grammar, which may obscure the true intent of a review.
In the process of creating HappyPatient, we continually ran into one issue: how do we get the API to consistently provide accurate suggestions for improvement? Through extensive testing, we managed to craft a set of reviews to demonstrate to the API precisely what it needed to do. Over time, the output got better until it reached an acceptable output.
Overall, HappyPatient will benefit hospitals by providing them with critical information on how to improve their patient experience. It will also reduce the burden placed on administrative and customer service staff by reducing the workload required of them. We hope to see widespread implementation of HappyPatient soon, and plan to expand this review analysis system to other businesses in the future.