Atidia AI Risk
Prediction
Overview

BACKGROUND

An astounding 10% of patient interactions result in complications and half of those are preventable (Grattan Institute Report). A major opportunity exists to predict the risks of complications early, enabling clinicians to intervene and optimise patients, preventing complications: cancellations, re-admission, other morbidities and mortality.

Risk Prediction can be implemented at different levels of sophistication. We define the categories in the following way:

 

  • Manual (current practice)
  • Standard risk calculators (standardising care)
  • AI: Expert system (encoding Anaesthetic expertise into a set of rules)
  • AI: Machine Learning

The current practice is to manually assess risk based on a patient’s history. In some cases, surgical risk calculators are used. This can range from simple calculations such as ASA and DASI scores, providing a measure of general health, and cardiovascular health respectively, to linear statistical models.

These calculators and models are designed to be used manually. Therefore, they are designed to exploit a minimal number of input fields, to be practical to use. Recent adoption of medical records means that a much larger set of data fields are available. In addition, large datasets are available for training more sophisticated models, and more sophisticated ML techniques are available. In addition, calculation of risk is usually not automated, so is often not done and is error prone. It’s an ideal time to develop advanced AI based Risk Prediction, the focus of Atidia.

Simply presenting relevant clinical information, these standard risk calculators and automating the process, will standardise care, which, as estimated by the Grattan Report, may save Australia up to $1.5 billion (comprising just 1% surgery worldwide). Atidia’s current Patient Optimizer product is ready to address this market and is currently gaining commercial traction.

We estimate that it can prevent up to 55% of preventable complications. However, we are not content to standardise care. We aim to radically enhance the standard of care.