2021 Healthcare Industry Challenges - Applying Open Source DataWA Health Hackathon Week
Applying Open Source Datasets
Participants who select one of the below challenges, will explore available open source datasets to discover new and innovative solutions to real world healthcare challenges experienced by the WA healthcare sector. These participants will still have access to a Domain Expert and will need to follow appropriate intellectual property and confidentiality procedures. Participants who work on one of the these challenges will be required to sign a version of this Collaboration Agreement prior to commencing the Hackathon.
To explore a list of open source data repositories, select the link below.
Using Social Media to Understand Public Sentiment Related to COVID-19
Social media text data can be a rich source of information on current mental health challenges encountered during COVID pandemic lock down events. Data science can unlock these important topics and sentiments which can help address growing mental health concerns of ordinary Australian communities undergoing COVID lockdown or Vaccine hesitation. Causation or corellations in different variables or factors can lead to better patient outcomes in Physical and Mental Health wellbeing.
Can you apply natural language processing techniques to social media text data to gain insights into public sentiment and mental health challenges related to COVID-19 lockdowns and/or vaccine hesitancy?
Domain Expert: Dr Guanjin (Brenda) Wang, Lecturer and researcher in Artifical Intelligence and Machine Learning, Murdoch University
Automatic Emergency Department Referrals
Emergency Department (ED) clinicians use many systems and programs, limiting their efficiency. Automated referrals to specialities would reduce administrative burden on clinicians. ED is not a treatment service, rather the focus is on initial management/stabilisation. It is critical that ED visits finish with suitable referral and communication to a service that a patient is referred to. Currently referral to public health services is sometimes forgotten or not done optimally.
Can you develop an automated referral option?
Domain Expert: Dr Piers Truter, Advanced Scope Physiotherapist at Fiona Stanley Hospital, and Senior Lecturer at the University of Notre Dame.
Automation to Help Clinical Coding
Clinical coding is a very manual and personnel-intensive job. A small team of 30 processes an average of ~100,000 admissions/year for 4 hospitals in the Perth metro-area. In addition, inaccurate clinical coding costs the hospitals money and resources. Automation could significantly reduce manual processing and risks associated with inaccurate coding.
Can you build a model that can read medical records and identify which diagnoses and procedures meet the criteria for clinical coding, incorporating any relevant standards?
- Western Australian Clinical Coding Authority (WACCA) https://ww2.health.wa.gov.au/Articles/A_E/Clinical-Coding-Authority
Domain Expert: Francis Lee, Medical Administration Registrar at Sir Charles Gairdner Hospital (SCGH)
Automation in Outpatient Referral Process
WA Health processes all outpatient referrals centrally, then redistributes them to the relevant health service, and finally to the relevant specialities. These referrals are then manually reviewed and triaged by medical staff. The outpatient referrals from various sources require intense manual processing both at the receiving end and the triaging end.
Can you develop an automated process to improve outpatient referrals that reduces processing time, decreases administrative burden on clinical staff, and improves patient outcomes?
Central Referral Service Guide for Referrers https://ww2.health.wa.gov.au/Articles/A_E/Central-Referral-Service-guide-for-referrers
Domain Expert: Vicky Gee, Health Services Planner, Clinical Planning
Wearables for Remote Health Monitoring
Remote health monitoring requires wearable devices to collect health data in real-time and monitor health status of patients at the comfort of their homes and detect any abnormalities indicating medical emergencies. When a user such as elderly person or a patients becomes unconscious, by a fall or accident, and thus unable to call help, they will be left until someone noticed or found the situation. If there is a way to raise an alarm by recognising the situation from the health and activity data collected, a search and rescue team can be dispatched while the users are unconscious.
Can you develop a proof on concept ‘remote health monitoring wearable device’ to collect health data in real-time and monitor the health status of a patient in the comfort of their home? It needs to be an end to end product that can detect any abnormalities indicating a medical emergency.
Domain Expert: Dr Jumana Abu-Khalaf, Vice Chanecellor’s Research Fellow in Computing and Security at Edith Cowan University (ECU)
Door-to-Door Navigation for Patients
Many patients have difficulty navigating within hospitals to attend outpatient or in-patient appointments or to visit relatives. This adds additional stress at a time they may be preoccupied with their own health concerns or that of a relative, and this negatively impacts on the hospital user experience. Whilst navigation-assists for hospitals have been revolutionised in the last 20 years, navigation within many hospitals remains old fashioned, inefficient, and ineffective.
Can you develop a product to assist patients and visitors to navigate within a hospital?
Domain Expert: Dr Andrew Toner, Consultant Anaesthetist