The final report of the New Zealand Health and Disability System review makes clear the vision for a ‘data-driven, digitally-enabled ecosystem’ in New Zealand and the need for ‘technologies to support decision-making and innovation’ to achieve this. While many decision-support tools are in use across the health sector, we are unaware of the specific algorithms used and who is using them, their implementation and governance, or the practical lessons learned from implementing those algorithms.
This research project, which was completed in December 2021, focused on predictive algorithms aligning with the approach in ‘Government Use of Artificial Intelligence in New Zealand‘. In healthcare, these predictive algorithms range from simple score-based tools for patient risk to advanced machine learning such as neural networks.
This work addressed questions below related to the NZ healthcare algorithm landscape, with a focus on predictive algorithms for decision support.
- How are predictive algorithms being used in practice within the New Zealand healthcare sector (primary and secondary care)?
- What factors are associated with successful implementation and use?
- What are the key challenges with the development of predictive algorithms?
- What are the high-value areas to focus on for new algorithm development, tuning/localisation and deployment?
The lessons learned through implementation and use of predictive algorithms in practice have been published in a final report.
In parallel to this research project supported by PDH, Orion Health was funded by MBIE to deliver the NZ Algorithm Hub to make models and algorithms supporting NZ’s response to COVID-19 freely available. We hope this project lays the foundation for a national algorithm management solution that will provide value beyond the current pandemic response. Commercially, this research aligns with a long-term objective to support national algorithm management by informing the development of product and services offerings.
This research aligns with the ‘Predictive Analytics‘ research theme by understanding and supporting more widespread use of predictive analytics by the health sector.
Research team
Pieta Brown, Orion Health
Rachel Owens, Orion Health
Dr Karol Czuba
A/Prof Alain Vandal, University of Auckland
Haze White