STAN is a Supportive Triage Assistant which uses Machine Learning to help triage emergency department patients. To date, STAN has directly supported over 60,000 patients in New Zealand.
We think so. Nurses regularly adjust their decision making due to STAN's recommendations (around 25% of the time when STAN recommends an opposing course of action).
In fact, a survey of 16 nurses using STAN at Nelson Marlborough District Health Board found that:
- 88% of nurses thought STAN improved triage consistency
- 100% of nurses thought STAN helped junior nurses triage more appropriately
- 88% of nurses thought STAN improved recognition and uptake of clinical pathways
- The majority of nurses thought STAN helped with Sepsis identification (88%), appropriate mental health triage (81%) and recognition of high risk symptoms (69%)
You can find more information at stantriage.com
We are in the midst of publishing research to further showcase how STAN has helped support triage in New Zealand. So keep an eye out for this.
We have decided not to pursue commercial use of STAN despite significant interest from around New Zealand. The simple fact is that (unfortunately) there are easier ways to make money than selling a niche decision support tool to the public health sector of a tiny island economy.
Consequently, we've made some of our code public in the hope that it may be helpful to someone. Let us know if it is, and get in touch anytime - mackdelany [at] gmail.com