CIO50 #26-50 Senthu Jegadheesan, Pen CS
Preventable hospitalisation costs Australia a staggering $2.3 billion annually, accounting for some 2.4 million days in hospital.
Looking at the sector as a whole, Australia’s annual health bill was a whopping $115.5 billion for 2020-21, according to Health Ministry data.
For years digital health experts attributed much of this spending to inefficiencies stemming from outdated, siloed legacy systems – often manual – and generally low levels of interoperability, which has led to poor continuity of patient care.
For Senthu Jegadheesan, chief technology officer with integrated health solutions company, Pen CS technologies for improving collection and management of health data have the potential help solve some of the biggest challenges affecting all Australians.
Providing early detection and intervention remains one of the biggest challenges, especially given the lack of visibility on how patients move across the sprawling health system, and the mounting security and compliance constraints of doing so.
“The first step to reform is linking patient information using ‘whole patient’ journey modelling,” Jegadheesan tells CIO Australia.
Over the past 12 months, he and his team built and launched a solution they call ‘privacy-preserving patient data linkage’.
Designed to be used on-demand, the solution enhances delivery of clinical services at the point of care, in real time,” Jegadheesan explains. Importantly, it also gets around the security challenges relating to health data typically being ‘identifiable’.
“The current market approach of linkage necessitates the sharing of identifiable information with the linking peer or a third party. With a large number of data variables changing hands and a largely manual process, human errors are common and costly.”
Secure linkage keys with advanced encryption protocol and PenCS’ unique data matching algorithm applied to source data can then be used to link with other data sets using the same linkage method, Jegadheesan adds.
“This linkage method is deterministic (rather than probabilistic) to minimise false positives, so that “Jill Smith” is only matched with “Jill Smith”. And it’s privacy preserving, meaning that no personally identifiable information has to leave the premises of the source data".
Real world problems
But the solution isn’t limited to health data, allowing for the input of data related to social and other determinants of health to gain a better understanding of individual patients and develop more targeted treatment programs.
“Useful aggregated insights from the merged dataset can be derived organically - with the use of AI and machine learning techniques - and then shared with participating parties, maintaining compliance with data governance policies,” Jegadheesan says.
“The real value of this linkage technology is that it can be utilised to solve real-world problems like risk stratification of patients and identification of factors leading to hospitalisations. These can potentially translate into strategies to reduce preventable hospitalisations and address mental health issues.”
PenCS’s privacy-preserving patient data linkage solution was built on Microsoft Azure using a cloud native technology stack with end-to-end automation.
Jegadheesan cites a number of benefits realised to date, including high-accuracy absolute patient data matching, with accuracy levels at 85 percent or more. Data quality has also improved at the client source.
Further demonstrating the value of the solution to the business, he and his team have assisted in expanding “commissioning opportunities” by establishing joint venture data sharing arrangements with peer agencies or other upstream clients.