Our research collaborations are changing:
The Speed Extract Project tracks the journey of chest pain patients when they present at hospital with the objective to improve the delivery of care and outcomes for future patients.
Chest pain is one of the leading medical reasons for presentation to emergency departments. If related to the heart, it is often due to an acute coronary syndrome, a major cause of mortality, morbidity and health care costs in Australia.
Whilst data collection processes are in place for patients presenting with chest pain, historically they have only been used for administrative purposes providing high level reporting or tracking of individual patients.
Currently best practice for the treatment of chest pain patients states that an electrocardiograph (ECG) should be performed within 10 minutes of the patient presenting to a hospital or an ambulance attending, followed by a blood test. The results of these tests currently dictate the treatment pathway the patient travels along.
As part of an interdisciplinary team spanning cardiologists, digital health and data experts, Charmaine and Aldo use existing deidentified data to provide an overview of multiple groups of patients to identify patterns in how they present with symptoms of chest pain. By applying these novel data analytics, additional factors are taken into consideration when planning a patient’s treatment pathway, minimising unfavourable outcomes, but also improving their long-term care.
From their analysis, evidence is now available to support changes to the processes and procedures for the management of chest pain patients.
This evidence supports:
This work has far reaching implications and is currently being scaled across Sydney Health Partners.
This type of data analysis can be extended to more effectively manage the care of patients with many other health conditions. In the future demographic and environmental data could be linked to identify risk factors in patient presenting with different health issues. Current research into this type of data analysis is being used to better manage the care of children who present with febrile neutropenia (fever and a low white blood cell count, indicating infection).
It is only recently that data science techniques are being applied to the social sciences.
With the implementation of Sydney’s 2014 lockout laws in the CBD and Kings Cross areas, a study led by Dr Roman Marchant, PhD student Nicholas James and Professor Sally Cripps analysed the effectiveness of the policies by assessing the level of crime in the form of non-domestic assaults and violent public behaviour.
The question was “Did crime levels change when the laws were introduced?”
It is important to look at the wide-reaching ramification of this change in policy as it has an impact on the way the city performs economically and socially.
Businesses including hotels, other accommodation and venues, employment and attracting a talented workforce to Sydney, tourism, nightlife and social activities are all affected. It also impacts residents’ and visitors’ perceptions of the city, leading to the question “Is Sydney a vibrant and diverse 24-hour city?”
Roman’s group analysed over 9 million crime records from 2005 to 2017 which were then narrowed down one type of crime in a specific localised area. They employed complex analytics to estimate the change in underlying patterns of crime occurrence.
They changed the way crime data is analysed by grouping assaults more accurately to specific locations, as opposed to a postcode level, and as a result were able to isolate the effects of the laws on the CBD and Kings Cross.
This showed that Sydney’s lockout laws haven’t curbed violence in the CBD as there was a decreasing trend in non-domestic assaults prior to the laws being introduced. In Kings Cross the number of assaults reduced with the implementation of the laws, however this could be attributed to the closure of late-night venues and decreasing numbers of visitors to the area.
They produced a report, which was submitted to the NSW Parliament Committee on Sydney’s Late Night Economy, which is currently assessing the effectiveness on the laws and it is using their study as evidence.
It’s already having flow on effect to other areas of crime analysis and is transferable to other data sets such a census information and population dynamics.
For example, it could be used to analyse the growth of unemployment in an area and identify any interdependencies that emerge over time or how changes in retail trading hours may have implications for local transport decision making.
This is the first time that cutting-edge data science techniques have been used for government policy evaluation.
Its applications are far reaching and have the potential to change policies, based on factual evidence, to better reflect local society requirements.
Dr Richard Scalzo and Professor Hugh Durrant-Whyte worked with Sydney Water to develop a next generation data strategy for their entire business. As a result of their analysis they identified areas and opportunities to improve internal processes and services to consumers.
It was quickly identified that whilst the company had data relating to many different aspects of supplying water to end users, such as water input, pressure, flow, demand, energy use and technical issues, it was held in different areas by different groups with no oversight of the entire pool of information available.
Initially the CTDS team met with dozens of individuals from different areas of the business such as customers service, corporate strategy, engineers and marketing. As a result, they mapped where data sat within the organisation and how it was used to make decisions.
The map makes the company’s data more visible and allows the organisation to see bottle necks to data access and use and enables them to identify better opportunities to use the data they have.
Once implemented, these recommendations will enable them to use the instrumentation they have in place is smarter ways. For example, customer service was only alerted to technical faults when customers called to report them, however they could now use historical modelling of existing data to predict when and where maintenance may be required and prevent technical issues occurring.
Further recommendation to move to a centrally organised data server with a good interface were made. This allows improved data access so identification and interpretation of more complex challenges can take place whilst also take into account the role of uncertainty.
The framework used to map data and identify an improved data strategy is one which can be applied to many business models.
There is interest in adopting this approach by the dairy industry and other Australian water companies.