Developing machine learning-based triage tools for ambulance dispatch

Primary Investigator: Hans Blomberg
Project manager: Douglas Spangler

This project aims to develop, validate, and spread open-source tools for use in performing risk assessments in field of pre-hospital care. The project is funded by Vinnova, and is being executed in cooperation between the Uppsala Ambulance Service, UCPR, and the Uppsala University Health Services Research group.

Original Research
Spangler D, Hermansson T, Smekal D, Blomberg H. A validation of machine learning-based risk scores in the prehospital setting. PLOS ONE. 2019;14: e0226518. doi:10.1371/journal.pone.0226518
Source code

Prehospital risk score demoApplication Source code
This web app demonstrates the behavior of a set of publicly released models based on the above research article.

White paperFull text Summary Source code
This paper documents the background, methodology, and preliminary results of the project in developing predictive models based on prehospital data. We plan to further develop this work into a full peer-reviewed validation of our approach to prehospital risk estimation.

Freetext ExplorerApplication Source code
This web app enables the exploration of the associations between free-text notes documented by nurses at the dispatch center with a number of pertinent patient outcomes. This is the first of several open-source applications we plan to release over the course of the project.