Event

Abstract

Background and Aims: Type 1 diabetes and its related complications significant impact on individuals and society across a wide spectrum. Our objective was to utilise machine learning techniques to predict diabetic ketoacidosis (DKA) and HbA1c>7%.

Methods and Results: Nine different models were implemented and model performance evaluated via the Area under the Curve (AUC). These models were applied to a large multi-centre dataset of  13761  type 1 diabetes individuals prospectively recruited from Australia and New Zealand. Predictive features included a number of clinical demographic and socio-economic measures collected at previous visits.

In our study, 6.7% reported at least one episode of DKA since their last clinic visit. A number of features were significantly associated with DKA. Our results showed that Fast Large Margin (FLM) performed well in predicting DKA with an AUC of 0.859. The FLM also provided the lowest classification error rate of 1.4%, highest sensitivity of 100%, F-measure of 99.3% and recall value of 100%. As for HbA1c >7%, the optimal Deep Learning model provided a good AUC of 0.863.  

Conclusion: Machine learning models can be effectively implemented on real-life large clinical datasets and they perform well in terms of identifying individuals with type 1 diabetes at risk of adverse outcomes.

Prof. Arul Earnest is the deputy head of the registries program and senior biostatistician at the School of Public Health& Preventive Medicine at Monash University. He has authored a textbook, three book chapters, and over 245 publications in peer-reviewed international journals. He has garnered 19 scientific awards and scholarships, with a H-index of 52 and I10-index of 163, with 10176 citations. Dr. Earnest's research is globally recognized, with engagements like the International Society of Bayesian Analysis Conference and the World Congress of Epidemiology. He collaborates internationally, including in the United States, Singapore, Canada, and the Netherlands. He leads a data science team of 4 data analysts and 7 PhD students, with a research focus on Bayesian spatio-temporal models and machine learning techniques, showcasing his commitment to advancing healthcare methodologies in clinical registries.

This seminar will be held in person and online via Zoom. It will not be recorded.

For external visitors wishing to attend in person, please email .
There are security measures in place at the venue and access is restricted.

Location:
Ground floor, Conference room 3
Monash University School of Public Health and Preventive Medicine
553 St Kilda Road,
Melbourne VIC 3004

Zoom meeting:

Please click this URL to start or join. https://monash.zoom.us/j/89690912410?pwd=DvDbE5OVYG3DOpDbC6DWgIOhijSX2R.1

Or, go to https://monash.zoom.us/join and enter

meeting ID: 896 9091 2410 and

passcode: 798127