Data Science for Health Workshop
Friday 17th January 2020
09:30am – 11:00am
Hack Aotearoa 2020 begins with exciting case studies delving into the machine learning and data science applications within healthcare. These case studies include:
Title: Mitara – fighting measles through digital health and social media
Presenters: Dr Canaan Aumua, Dr Sanjeev Krishna, The University of Auckland
Description: The world’s first measles chatbot Mitara is a new-age public health tool, providing concerned citizens across New Zealand and the Pacific with a trusted source of health information. This presentation will describe the bridging of the gap between technology and healthcare to provide medically sound advice in the midst of a health crisis.
Title: A Multi-Task Imputation and Classification Neural Architecture for Early Prediction of Sepsis from Multivariate Clinical Time Series
Presenter: Dr. Johnathan Rubin (PhD), Philips Research North America
Description: Sepsis is a life-threatening condition that puts 30 million lives at risk every year worldwide.
Early detection of sepsis would allow faster administration of antibiotic treatment and improve patient outcomes, as well as significantly reduce hospital expenses.
In this talk, we present our submission to the 2019 PhysioNet challenge where the goal was the early detection of sepsis using physiological data.
We also have exciting presentations by Dr. Nhung Nghiem (Department of Public Health, The University of Otago) on predicting cardiovascular disease incidence using machine learning and linked health and social administrative datasets;and
Dr. Martin Urschler, (School of Computer Science The University of Auckland) about Forensic Age Estimation of Adolescents using 3D MRI data and Machine Learning
Creating a Scientific Publication from Intensive Care Registry Data Workshop:
Friday 17th January 2020
09:00am – 11:00am
Facilitator: Professor David Pilcher – Intensive Care Specialist, The Alfred Hospital, Melbourne, Australia. The Australian and New Zealand Intensive Care Society
Number: 20-30 participants (spaces are limited – first come first served)
All levels welcome (beginner to advanced)
In this workshop, participants will get experience in constructing a scientific paper using Intensive Care registry data. Participants will work in two groups and will be given a hypothesis to test. The groups will then direct the faculty personnel in the construction of an analysis plan. Faculty will then ‘live code’ the analysis at the instruction of the participants, with immediate presentation of the results. A de-identified international dataset (GOSSIS dataset) will be used.
Participants will learn the benefits and pitfalls in using large datasets for answering clinical questions. They will be given exposure to basic comparative statistics, measurement of risk adjusted outcomes, and how to present analyses in a scientific format for a peer-review journal. No coding or statistics experience is necessary. A little clinical knowledge is desirable but not essential.
Facilitator Biography: David Pilcher is an Intensive Care Specialist at The Alfred Hospital in Melbourne. He trained in respiratory and general medicine in the UK before moving to Australia in 2002 to undertake training in Intensive Care Medicine. His interests include organ donation, lung transplantation, ECMO, severity adjustment of ICU outcomes, ICU performance monitoring and the epidemiology of Intensive Care medicine. He is the Chairman of the Australian and New Zealand Intensive Care Society (ANZICS) Centre for Outcome and Resource Evaluation (CORE) which runs the bi-National critical care registries. He is a medical advisor to DonateLife in Victoria. He is also an Adjunct Clinical Professor with the Department of Epidemiology and Preventive Medicine at Monash University.