Rinaldo Bellomo

Wendy Chapman
Thank you, Vincent. That was fabulous and amazing that you could pull that off with such short notice. Our next speaker is Rinaldo Bellomo. He directs the data analytics research and evaluation Centre at Austin Health and the University of Melbourne, and his Centre goes by the short name of DARE. And any of you who don't know Rinaldo, when you hear him talk you will think that that acronym fits him and his group perfectly.

He's the Director of Intensive Care Research at the Austin Hospital, and Professor of Intensive Care Medicine at the University of Melbourne, and he's also the co-director of the Australian and New Zealand Intensive Care Research Centre, and an NHMRC Practitioner Fellow.

He's on a lot of lists. Since 2006 he's been the most published intensive care investigator in the world. In 2014, Thompson Reuters recognised him as one of the most influential scientific minds in clinical medicine, and he's the most published biomedical investigator in the history of Australian medicine.

So we're just so lucky to have him tonight and looking forward to his talk. He's going to be speaking on the evolution of our ability to use electronic medical records in acute care.

Rinaldo Bellomo
Thank you very much, Wendy. So we were lucky enough, about two years and a bit ago, to be supported by Shitij and the university to start exploring the ability to use electronic medical records in acute care to evolve our ability to understand disease and inform treatment. We were supported to form a small group of investigators, which were clinicians and data scientists, to see if we were able to evolve our ability to use the data that is produced in electronic medical records within a hospital system, such as the Austin Hospital, to inform future investigations in the field elsewhere.

We got going and we tried to focus on at least three major areas of investigation, which related to sepsis, delirium, and the facilitation of randomised control trials. We focused first of all on trying to understand if there were better ways of identifying septic patients ,that were a particular risk when presented to the emergency department, using a tool called a qSOFA. Why will we do that? I was lucky enough to be part of the third international consensus for the definition of sepsis and septic shock, or sepsis three, which identify that in people presented to hospital with infection, the assessment of the respiratory rate, the blood pressure and the conscious state, so-called qSOFA, Quick Sequential Organ Failure Assessment, might be helpful in identifying those very patients that need to have advanced care or are at risk of doing badly.

So we wanted to see if the EMR of our hospital would enable us to identify such patients. And you can see there that using the EMR you can identify all the adult ED presentations. 165,000. You can identify those suspected infection by linking the admissions to the ordering or microbiological tests as well as the administration of antibiotics, and you can identify that 20% of these patients were qSOFA positive. They can identify that these patients had almost a 13% mortality. That has to be put in the context that if you present to the hospital with an acute myocardial infarction with ST segment elevation, your mortality at the Austin is about 5 to 6%. So clearly a very high risk population.

And this now has created the ability to develop alerts for clinicians in the emergency department that can tell them that there is a patient that fulfills this criteria in the emergency department and can hopefully change the rapidity, the context, and the efficacy of care. And the currently advanced process of delivering such alerts is holding out for COVID to finish so that we can actually move to a more normal world and we can implement such alerts.

Identifying people with delirium is a very difficult thing to do, because there is no standard definition or gold standard that says if somebody is or is not delirious. We define somebody such, in language. We use words like confused, agitated, disorientated, aggressive, combative. And these words can give us a window on the presence or absence of delirium in ICU patients. We can use the EMR notes entered by nurses, residents, registrars, physiotherapists, to analyse whether patients with delirium are present and how many there are and what their characteristics might be if we can identify them by such means.

So this is reading 60 million words, thousands of EMR notes, and you can identify over a period of several years and 12,000 patients admitted to the Austin ICU that about 5,000 of them are characterized by notes written by nurses and doctors that identify that they are disorientated, confused, agitated and so on. The identification of these patients then opens the door not only to alerts, but also to important epidemiological assessment of what medications can or cannot do for these patients.

Doing trials is a really expensive business. Randomising large groups of patients is a very costly undertaking. We wanted to see if doing EMR leveraged, randomised control trials would allow us to randomise thousands of patients, a large group of patients, and achieve high quality data capture at minimal cost. This is a paper we recently published in the Journal of the American Medical Association. It's one of the biggest randomised control trials in anaesthesia, and try to test whether the large tidal volume that is typically given to patients during anaesthesia is necessary, or it is actually potentially detrimental.

The study was conducted at the Austin Hospital and it's a study that, in terms of numbers randomised, would typically cost three to four million dollars in Australia and New Zealand, and more than $10 million in the USA. But we did it and it cost $35 per patient randomised. And it's been published in one of the top journals in the world and it was able to identify that there is no advantage of using high tidal volumes, and in fact there's a trend in the other direction. Particularly for laparoscopic surgery, such large tidal volumes may be injurious.

So let me conclude by saying that the DARE experiment has delivered multiple lessons, insights, and has opened the door to interventions. We're now incredibly busy dealing with multiple trials and multiple activities, and we've got more EMR-based trials coming along. We have contributed and are affiliated to the publication of more than 30 papers, about a paper a month, since we were created. And I would like to say that it's been a wonderful experience to work with data scientists and discover that if you've got a clinician alone, you don't produce anything like what you could with a data scientist. If you're a data scientist alone, you cannot have the domain knowledge to deliver clinical science. But in DARE, by putting them together, we've had new maths of zero plus zero delivering 30 papers over two years. Thank you very much.

Wendy Chapman
Thank you, Rinaldo. So now those of you who haven't met Rinaldo before know firsthand he's a force to be reckoned with. And we hope through the Centre that we're able to learn from what they've done with DARE and build a similar kind of environment so that other hospitals and services can have similar experiences.