Triaging GP referrals is a time-consuming task for doctors in New Zealand. It is common for doctors to spend more than ten hours per week on triaging electronic referrals to various risk categories or direct to investigation tests. Presently, there are over 200 untriaged referrals at Waitematā District Health Board (WDHB) which are close to breaching key performance indicators in the hospital. Currently, very little research has been done to address this problem using machine learning (ML) techniques. This is partly due to the challenging nature of working with electronic GP referrals in New Zealand, as they contain both structured and unstructured data.

The primary goal of this project is to provide a decision support system for triaging referrals to make this task more efficient. In addition to achieving this goal, several secondary research goals are possible. These include:

• Predictive model sharing between hospitals or populations
• Explain-ability of prediction results
• Monitoring of input data concept drift detection
• Provide capability for performing data science and model training within a hospital setting

This research is ambitious because we aim to change the way data-driven health is done in New Zealand. One outcome of this research will be to move the focus away from ad hoc and hospital-specific solutions to a smart data platform that supports the sharing of population knowledge and data-derived assets, integrating local patient context and making the practice of data science a first-class citizen.

We build on multiple prior Precision Driven Health research programmes and, if successful, will potentially become a key enabler for more efficient future data-driven health research projects. We also expect that the work of this research enhances the Orion Health Smart Data Platform, which provides a reusable, scalable and flexible data environment for this and future data-driven PDH projects.