Gold Coast Expert Leads the Evaluation of First-of-its-kind Project in Australia to Reduce Incidence of Suicide
NEWS - 26 Apr 2022
Professor Chris Stapelberg is a leading psychiatrist with Gold Coast Health and a university researcher, who has been investigating major depression since 2009. Professor Stapelberg and his team are passionate about reducing the incidence of suicide, by assessing public healthcare systems, and ensuring that vulnerable patients do not ‘fall through the cracks’ of an often-fragmented health service. Tragically, suicide was the leading cause of death for Australians aged 15-49 years in 2019, according to a recent report from the Australian Institute of Health of Wellness (AIHW).
One of Professor Stapelberg’s major projects is leading the evaluation of the Australian Zero Suicide Framework, to assess the impact of the Zero Suicide Framework implemented in 2016 to effectively support people accessing mental health services.
To effectively evaluate these large-scale mental health service implementations, Professor Stapelberg and his team have pioneered machine learning and data mining technologies. These systems can understand and work with large volumes of unstructured clinical records, to help identify trends and outcomes for specific suicide prevention programs. They have already produced the first evidence worldwide that their approach is effective.
Strong partnerships between universities and hospitals are vital to achieving success in innovation and research
Professor Stapelberg is based at Gold Coast University Hospital (GCUH), which is one of the busiest hospitals in Queensland, and co-locates with Lumina, the health and innovation community within the wider Gold Coast Health and Knowledge Precinct.
Professor Stapelberg has a clinical psychiatry role, he is the Joint Chair in Mental Health for Bond University and Gold Coast Hospital and Health Service, and in addition, is the Mental Health and Specialist Services Co-Lead for Innovation, Evaluation, and Research.
As both a clinician and researcher, Professor Stapelberg can undertake important research and demonstrate its effectiveness in a real patient setting, like GCUH. “I work as a clinician psychiatrist, on the medical and surgical wards, seeing people with mental health presentations. I’m active on the medical side of the hospital, and the university side. A lot of the work I’m doing is leveraged across both the hospital and the university,” Professor Stapelberg explained.
“It is an important partnership. The hospital and health service benefit from having academic input, and the university benefits from research and innovation in a real-world context. The outcomes of research, evaluation and innovation ultimately benefit the people who have mental health needs,” he said.
Determining the impact of suicide prevention programs involves evaluating them with real-world data
Professor Stapelberg leads the evaluation of the Gold Coast Health Suicide Prevention Strategy, one of the largest implementations of the Zero Suicide Framework in Australia. The Zero Suicide Framework is a quality improvement, practical operating framework for the mental healthcare system, based on the belief that deaths by suicide for people under the care of public health systems are preventable.
To determine if the framework is having a positive impact on people accessing mental health services, its implementation must by necessity be heavily data-driven. “If you have a good implementation, you need to actually demonstrate that it works. And to do that, you need data,” Professor Stapelberg said.
Mental health data can be difficult to analyse – but Professor Stapelberg’s team is using machine learning to help
Professor Stapelberg and his team have collected and analysed large volumes of patients’ data. “Mental health data, like most health data, is mostly unstructured. A lot of the really interesting data exists in the form of clinical notes or triage text. It is sometimes difficult to sift through that data and pull out something that we could put into a spreadsheet as structured data to be analysed,” Professor Stapelberg said.
The challenge with unstructured data is that it takes a lot of highly-skilled people, often with clinical experience, to read and understand the clinical records. From this, they can gather data such as suicidal presentations to an emergency department, or if a particular treatment has been started. It’s time-consuming, dependent on the availability of skilled people, expensive, and very difficult to do at scale.
However, Professor Stapelberg has been using machine learning to ‘read’ the clinical records and mine the data, enabling him to pull-out relevant information from massive, unstructured datasets, and turn it into structured data that the team can then analyse and use.
“For example, we successfully used one of our machine learning systems to analyse 1.4 million presentations to the GCUH emergency department over 10 years and outputted the presentation trends. The evaluation shows that numbers of people representing to hospital after attempting suicide were reduced with our clinical suicide prevention pathway. That’s the first evidence of its kind in the world, and that has arisen from that data evaluation,” he said.
Mental health issues are increasing, so ongoing assessment of mental health services is important
According to the Australian Institute of Health and Wellness, almost half of all Australian adults will face mental health issues during their lives – and this is increasing over time.
Professor Stapelberg has seen this in the data. “We established baseline data of presentations of suicidal crisis in our emergency departments, and what we discovered was a linear trend, that was growing faster than the population growth on the Gold Coast, meaning over the last few years, more people were presenting to emergency departments in crisis,” he said.
With needs increasing, we must invest in healthcare programs that work. His ongoing work is an important step – it evaluates the impact of the Zero Suicide Framework, and in particular, the clinical ‘suicide prevention pathway’. “It means that we can demonstrate what we are doing is making a true difference to people in our community.”
What’s next for AI and machine learning in mental health research?
Unlike other research projects, which generally have a start date and end date, the type of evaluation work Professor Stapelberg leads is on a rolling timeline. “Evaluation is essentially an ongoing process. We’re always collecting more data.”
But there is also the opportunity to look at the data more deeply. “We’ve also been trying to drill down and say, okay, how do certain aspects of our work apply to specific areas of mental health, like child and youth mental health, and how can it be used to support clinical decision making?”
In the next four or five years, he and his team want to develop a suite of sophisticated tools, including decision support tools. “We want to get to a point in the future where we use machines that can look at a lot of data, present a summary to a clinician who’s working in the emergency department or on a psychiatric ward, and provide a true benefit in terms of their decision making.”
Diversity within Lumina’s community of experts is important to help foster new ideas
In a health, science, and innovation community like Lumina, Professor Stapelberg believes that diversity is key to the growth of new ideas. “I think having a diverse range of people or businesses in the Precinct, that might have very little to do with mental health, might be incredibly interesting in terms of cross-fertilisation of ideas,” he said.
“For example, we work with a team of people who have extensive data analysis skills, including a brilliant biostatistician, who has become a very close collaborator over the years. We’ve also got the team at Bond Business School who is helping to develop some of our tools and make them better.”
Professor Stapelberg also emphasises the importance of nurturing key partnerships to achieve research success, and the role of education, training, and knowledge sharing within the academic and clinical community.
“At GCUH, we are a major teaching and research hospital, and that is key to establishing strong partnerships with universities. What those two types of organisations can do for each other is extensive in terms of potential partnerships and leverage that you can gain from those partnerships. One or more universities, close to a hospital, is vital.”
If you or a loved one need someone to talk to, call: