IntHealth: Process Intelligence in Emergency Rooms

Description

Process improvement in Healthcare, and specially in Emergency Rooms (ER), is vital to diminish the waiting and attention times, improving the quality of the services and increasing the amount of patients attended.
To reduce the waiting times and patient queues in ER, ER boxes should be freed faster. The quicker a physician can diagnose accurately and discharge patients, the faster a box in the ER can be released, contributing to shorten the waiting times for future patients.
The tool IntHealth is aimed at analyzing the process executed in ER, through data mining and process mining techniques. IntHealth includes a series of rules to predict the discharge for hospitalization of patients based on historical information stored in event logs. It enables ER experts to improve their knowledge of the process and help them make better and faster decisions.
By including data mining rules for predicting hospitalization and process mining techniques, IntHealth can reduce waiting times by discharging patients for hospitalization faster and releasing ER boxes promptly.

 

Screenshots

Following you can see some screenshots of the application:

 

Episode View

1-episode-view

 

Episode View with events and rules matched

2-episode-view-with-events

 

Adding events to an Episode

 

3-adding-events

 

 

Generated Process Model Through ProM 6.6

4-generated-process-model

 

Screencast

To demonstrate the use of IntHealth a screencast can be seen here:

 

This video includes a walkthrough of the different components of the tool and an example of the inclusion of a new episode. It includes the login, the creation of a new episode, adding new activities and characteristics, the trigger of hospitalization rules and the generation of a process model, through process mining techniques. It also includes a view of historical episode data.

About

Rojas, E., Munoz-Gama, J., Sepúlveda, M., & Capurro, D. (2016). Process mining in healthcare: A literature review. Journal of biomedical informatics, 61, 224-236.

Acknowledgments

This project was partially funded by Fondecyt (Chile) grants 1150365 and 11130577, the Ph.D. Scholarship Program of CONICYT Chile (CONICYT-Doctorado Nacional/2014-63140180), and by Universidad de Costa Rica Professor Fellowships.