SLIDE 1 A Geovisual Epiphanee for the Health Industry
Simon Moncrieffa, Rob Rowellb, Narelle Mullana
aCurtin University, CRCSI bInsight GIS
The Cooperative Research Centre has supported this work for Spatial Information, whose activities were funded by the Australian Commonwealth Cooperative Research Centres Programme. This work is part of a larger body of work for which a provisional patent has been applied.
US Provisional Patent Application No. 62/074970
SLIDE 2 What is CRC•SI ?
Universities (RESEARCH) Federal & State Govt Agencies (SPATIAL INFORMATION) Australian/NZ Companies
(COMMERCIAL CAPABILITY)
CRCSI SI
SLIDE 3
Restrict tricted ed Access cess = Li Limit ited d An Analysi lysis sensitivity of health information (and mapped or geo-enable it.... OMG) RISK …. creates restrictions and limitations LOST… health + place analysis
SLIDE 4
Big Data + Many Data Sources + Numerous Queries
health industry = extremely large data sets enormous number of potential queries unfeasible to produce all results ahead of time
SLIDE 5
(Analysis by) Visualisation
SLIDE 6
(Analysis by) Visualisation
relate disease, diagnosis and service delivery to a location visualise the relationships between exposure, healthcare and outcomes issues quickly identified + actions effectively implemented
SLIDE 7 (Communication by) Visualisation “Our eyesi esight t is is a g a grea eat t so sour urce ce of
SLIDE 8
Epiphanee: Where did it begin
Research Project between WA Dept of Health, CRCSI and Curtin Uni “Enabling wider access to Health Data for efficient delivery of services while ensuring confidentiality is maintained”
SLIDE 9
- Processes data on the fly allowing
users to generate their own query specifications.
- Delivers results without the data
maintaining privacy
- Privacy module allows richer
interpretation and analysis
- Being used for mapping of summary
statistics (e.g. age standardised rate).
Query Generation & Visualisation Epiphanee Generation Data
Dataset Result1 V1 V2 V3 Result2 V4 V5 Result3
Epiphanee – Maximising Result / Minimising Risk
SLIDE 10 Epiphanee – Comparing Traditional Outputs
Results with no privacy filter (unrestricted access) Results using conventional privacy methods Results using Epiphanee developed privacy filter
SLIDE 11
– Hospitalisation (~11 million records) and mortality data (~100,000 records) – Summary statistics with a thematic map as output. – Data export: Excel, ShapeFile, PDF.
– Hospitalisation data (~50,000 events) – Point location, demographic, epidemiological and financial data – Result: Rapid generation of thematic maps for statistical areas (SA1, SA2) – Linked to ABS data for extended visualisations
- In both cases, Epiphanee dynamically linked data to normalise the
- utput
Case Study –
SLIDE 12
Example Image 1 -
SLIDE 13
Example Image 2
SLIDE 14 Outcomes
Proportion of queries returning a result compared to the number
privacy filter is employed. Propor portion ion of Total Reporti ting Defau efault lt Dynamic ic 50% 6.8% 100% 90% 0.1% 98.7%
SLIDE 15
Outcomes
minutes not days currency of data
SLIDE 16 Outcomes
time for deeper analysis “Detect the expected + Discover the unexpected”
Thomas and Cook 2005
SLIDE 17
Outcomes
easier consumption
SLIDE 18
“The flexibility and privacy protection built into Epiphanee will allow access to detailed aggregated health information that is more timely, and enable a wider range of health users to produce customised outputs from their desktop“
Dr Tarun Weeramanthri Executive Director Public Health and Clinical Services Division Dept of Health, WA
SLIDE 19
Epiphanee has demonstrated that it is possible to enable wider access to Health Data for efficient delivery of services while ensuring confidentiality is maintained
Thank You
