EDDI Joins the Team in Theatre

In an operating room (OR), machines use musical tones for their alarms.

This means someone must look at the equipment to see why it’s made a sound. These musical tones – initiated by threshold alarms – won’t necessarily explain any changes or their significance.

But what if the surgical team heard a voice saying, “hypovolaemia is possible”, in English, French, Mandarin or any other local language?

Meet EDDI. This New Zealand-invented Early Detection Decision Information software takes data from the anaesthesia machine in the OR, self-checks it, and compares it against patterns based on physiological principles.

“This is not straight big data; rather, it’s “micro big-data”, and it’s never been done before,” says Mark Hildesley who’s responsible for getting EDDI to market.

“EDDI is a support tool. The anaesthetist still has complete responsibility for interpreting the diagnostic output, but EDDI will alert the anaesthetist and OR team much earlier and with spoken real time information.”

Right now, EDDI is being tested in Lower Hutt, Auckland and Melbourne where the team is getting to grips with some challenges.

“EDDI has to communicate with all types of anaesthesia monitor to work globally. It has to fit in with all the other OR electronics as well. Think Mac versus PC. EDDI was originally built on a GE machine, but Melbourne uses a German Draeger, so we need to teach EDDI German, as it were.

“Every time an OR uses a different machine for anaesthesia, we’ll need to write some new code so EDDI can talk to it.”

EDDI is still pre-revenue, waiting for someone who sees the potential and will take the risk. Before A.I. jumped into the public awareness, Mark pointed out that the early responses were either “It’s not blue-sky enough and it’s too close to the market”, or “It’s too blue-sky and too far from the market.” Times have changed and interest in EDDI is circling.

Mark and team already see potential beyond the OR because EDDI offers a new method of health data comparisons in real time.

“For example, you could put EDDI to work in many of the existing digital health data streams where there’s a data process involving a health decision whether it’s ICU or monitoring a discharged patient.”

He says EDDI has the potential to help with safety, through earlier and better supported decisions. As the techniques improve, each time it’s used, it will be able to expands its own reference library with machine learning. EDDI could improve databases, right down to those extremely rare conditions that affect a dozen people in the world. The continuous improvement and early pattern spotting may support all digital health processes as the techniques are deployed world-wide.

EDDI will help in the OR first, but elsewhere soon.

By Prue Scott